2026 Abstracts
Shivakalyani Adepu, PhD
Fellow (Clinical or Postdoctoral Researcher)
3D printed radiopaque bioresorbable citrate-based cardiovascular scaffolds
Introduction: Radiopaque bioresorbable stents are designed to provide temporary mechanical support to blood vessels while allowing real-time imaging during and after implantation. Unlike traditional metal stents, which remain permanently and can lead to chronic inflammation or complications, bioresorbable stents degrade over time, supporting natural vessel healing and reducing long-term risks. However, most biodegradable polymers are radiolucent and not visible under X-ray or fluoroscopy, making it difficult for clinicians to accurately place and monitor the stents. To address radiopacity, metal markers were added to the stent ends, but issues with delivery and clinical outcomes persisted. Magnesium alloy alternatives degraded too quickly, and PLA coatings did not resolve early scaffold failures. Thinner struts, shown to reduce restenosis, became a development focus. Radiopaque bioresorbable scaffolds (BVSs) are especially valuable in coronary, peripheral, and pediatric applications, where precise placement and post-operative monitoring are critical.
Methods: In the current work, A new radiopaque, bioresorbable cardiovascular scaffold (BVS) with a strut thickness of 80μm was fabricated by μCLIP stereolithography based 3D printing using iodine-bound polydiol citrate to improve strength, flexibility, and visibility under X-ray. Iodinated polymer was synthesized by covalently attaching the Triiodo benzoic acid (TIBA) to poly-1,12-dodecamethylene citrate (PDC) followed by methacrylation to yield methacrylated(poly 1,12-dodecamethylene citrate)-co-(3-hydroxy-2,4,6-triiodo benzoic acid) (mPDC-I) scaffold with native radiopacity and mechanical properties that are comparable to metallic stent. The developed stent was tested for its in vitro cytotoxicity as per ISO10993 on HUVECs by MTT and live-dead assays. The CT visibility fluoroscopic visibility was tested in a tissue mimic and euthanized pig's heart. bioresorbable and visualized in CT as well as fluoroscopy when placed in euthanized pig's heart.
Results: The mPDC was successfully functionalized with TIBA by ester linkage and confirmed by NMR and FTIR. In SEM-EDS analysis, presence of ~10 wt% elemental iodine positively correlates covalent attachment of TIBA. The mPDC-I stents displayed a greater radial force of ~13.5 N comparable to the metal stent (~15N). The BVSs were visualized in both microCT and Fluoroscopy when placed in euthanized pig heart comparable to the metallic stent. The developed BVSs were found to be biocompatible as demonstrated in the MTT Cytotoxicity Assay with a 80-90 % cell viability.
Conclusions: By combining degradability with visibility, the developed iodine-bound citrate-based stents may enhance procedural safety, effectiveness, and long-term patient outcomes. In vivo evaluation of mPDC-30% INPs stents in a porcine model for 28 and 90 days is in process.
Methods: In the current work, A new radiopaque, bioresorbable cardiovascular scaffold (BVS) with a strut thickness of 80μm was fabricated by μCLIP stereolithography based 3D printing using iodine-bound polydiol citrate to improve strength, flexibility, and visibility under X-ray. Iodinated polymer was synthesized by covalently attaching the Triiodo benzoic acid (TIBA) to poly-1,12-dodecamethylene citrate (PDC) followed by methacrylation to yield methacrylated(poly 1,12-dodecamethylene citrate)-co-(3-hydroxy-2,4,6-triiodo benzoic acid) (mPDC-I) scaffold with native radiopacity and mechanical properties that are comparable to metallic stent. The developed stent was tested for its in vitro cytotoxicity as per ISO10993 on HUVECs by MTT and live-dead assays. The CT visibility fluoroscopic visibility was tested in a tissue mimic and euthanized pig's heart. bioresorbable and visualized in CT as well as fluoroscopy when placed in euthanized pig's heart.
Results: The mPDC was successfully functionalized with TIBA by ester linkage and confirmed by NMR and FTIR. In SEM-EDS analysis, presence of ~10 wt% elemental iodine positively correlates covalent attachment of TIBA. The mPDC-I stents displayed a greater radial force of ~13.5 N comparable to the metal stent (~15N). The BVSs were visualized in both microCT and Fluoroscopy when placed in euthanized pig heart comparable to the metallic stent. The developed BVSs were found to be biocompatible as demonstrated in the MTT Cytotoxicity Assay with a 80-90 % cell viability.
Conclusions: By combining degradability with visibility, the developed iodine-bound citrate-based stents may enhance procedural safety, effectiveness, and long-term patient outcomes. In vivo evaluation of mPDC-30% INPs stents in a porcine model for 28 and 90 days is in process.
Competition Category: Basic Science or Translational
Mentor: Guillermo Ameer, D Sc
Megan Alagna, BS
Student
Longitudinal analysis of patient- and caregiver-reported outcomes after leg amputation secondary to peripheral artery disease
Introduction: Severe peripheral artery disease (PAD) can lead to major leg amputation, which is a life-altering event for both patients and their caregivers. The longitudinal changes in patient- and caregiver-reported outcomes after amputation are not well understood. We hypothesized that these outcomes are dynamic in the first year after amputation.
Methods: This is a prospective single-center cohort study. Patients undergoing leg amputation for PAD in the Division of Vascular Surgery at NMH beginning in September 2024 and their primary caregivers were eligible to participate. All participants provided written informed consent. The validated survey instruments for patients were the WHO Quality of Life Brief Questionnaire (WHOQOL-BREF) and the PROMIS Short Form Physical Function (v. 11a) and Depression (v. 8a) measures. Caregivers completed the Oberst Caregiving Scale and Zarit Caregiver Burden assessment. Surveys were distributed to participants at five discrete timepoints in the first year after amputation.
Results: To date, 34 patients and 27 caregivers have completed at least one survey. Among patients, mean age is 65.5 ± 2.3 years, 52.9% are female, 52.9% are Black, and 32.3% reported an annual household income ≤ $50,000. Among caregivers, mean age is 60.1 ± 2.5 years, 66.7% are female, and 88.9% are family members. Most (74.1%) have been a caregiver for over 1 year, 22.2% reported prior medical training, and 44.4% were also a primary caregiver to others. WHOQOL overall mean scores decreased after the 3-month timepoint, suggesting worsened overall QOL in subsequent timepoints. PROMIS Physical Function scores were 2 standard deviations below the mean population score and did not significantly change across time points, indicating decreased physical function throughout the first year after amputation. Zarit Caregiver Burden and Oberst Caregiving scores increased after 6 months and throughout the first year, respectively, indicating a significant increase in caregiver burden after amputation.
Discussion: Our pilot analysis findings of a lack of significant improvement in patient- and caregiver- reported outcomes over 1 year suggest the poor experience of patients and caregivers who experience leg amputation. Future work will focus on completing enrollment, understanding these findings in qualitative and quantitative detail, and identifying interventions to better support patients and caregivers experiencing this major life-changing event.
Methods: This is a prospective single-center cohort study. Patients undergoing leg amputation for PAD in the Division of Vascular Surgery at NMH beginning in September 2024 and their primary caregivers were eligible to participate. All participants provided written informed consent. The validated survey instruments for patients were the WHO Quality of Life Brief Questionnaire (WHOQOL-BREF) and the PROMIS Short Form Physical Function (v. 11a) and Depression (v. 8a) measures. Caregivers completed the Oberst Caregiving Scale and Zarit Caregiver Burden assessment. Surveys were distributed to participants at five discrete timepoints in the first year after amputation.
Results: To date, 34 patients and 27 caregivers have completed at least one survey. Among patients, mean age is 65.5 ± 2.3 years, 52.9% are female, 52.9% are Black, and 32.3% reported an annual household income ≤ $50,000. Among caregivers, mean age is 60.1 ± 2.5 years, 66.7% are female, and 88.9% are family members. Most (74.1%) have been a caregiver for over 1 year, 22.2% reported prior medical training, and 44.4% were also a primary caregiver to others. WHOQOL overall mean scores decreased after the 3-month timepoint, suggesting worsened overall QOL in subsequent timepoints. PROMIS Physical Function scores were 2 standard deviations below the mean population score and did not significantly change across time points, indicating decreased physical function throughout the first year after amputation. Zarit Caregiver Burden and Oberst Caregiving scores increased after 6 months and throughout the first year, respectively, indicating a significant increase in caregiver burden after amputation.
Discussion: Our pilot analysis findings of a lack of significant improvement in patient- and caregiver- reported outcomes over 1 year suggest the poor experience of patients and caregivers who experience leg amputation. Future work will focus on completing enrollment, understanding these findings in qualitative and quantitative detail, and identifying interventions to better support patients and caregivers experiencing this major life-changing event.
Competition Category: Health Services Outcomes or Clinical
Mentor: Karen Ho, MD
Giselle Alrachid, BS
Student
Does fat injection alter scar tissue? A clinical review
Introduction: Autologous fat injection has been increasingly applied in the management of scar tissue, including postsurgical, burn-related, radiation associated, fibrotic, and atrophic acne scars. Multiple fat processing techniques have been described, including conventional microfat grafting, nanofat, and stromal vascular fraction (SVF) based preparations. However, variability in scar type, processing methods, and outcome assessment complicates interpretation of clinical outcomes and benefits. The purpose of this review was to evaluate whether autologous fat injection alters scar tissue characteristics and whether specific processing techniques confer clinical advantages.
Methods: A PubMed review was performed to identify human clinical studies evaluating fat injections for treatment of scar tissue. The initial search yielded 285 articles. After abstract and full text screening, 32 studies met the inclusion criteria. Studies were then organized by scar type and processing method. Outcomes were categorized as patient reported, observer reported, functional, objective biomechanical, or histological. Comparative studies were examined to evaluate whether fat processing techniques had an effect on reported outcomes.
Results: Fat injection has been studied across postsurgical, acne/atrophic, burn related, adherent/fibrotic, radiation associated, and inflammatory/sclerotic scars. Patient reported outcomes (e.g., pain, stiffness, POSAS patient scores) and observer reported scar scales (e.g., POSAS, VSS) demonstrated the most consistent improvement across scar types. Functional improvement, including range of motion and severity of contractures, was notably reported in radiation associated and fibrotic scars. Biomechanical assessments were variably reported, and histologic findings were not consistently correlated with clinical outcomes. Comparative studies evaluating processing techniques demonstrated early or scar-specific advantages with nanofat enrichment or SVF-based preparations; however, overall superiority over conventional fat grafting was not consistently shown.
Conclusions: Current clinical literature supports a beneficial role for fat injection in scar management, particularly in improving patient reported symptoms and functional outcomes. Evidence suggesting consistent advantage of specific fat processing techniques remains limited. The available data supports a scar type dependent application of fat-based therapies rather than a universally superior fat processing technique.
Methods: A PubMed review was performed to identify human clinical studies evaluating fat injections for treatment of scar tissue. The initial search yielded 285 articles. After abstract and full text screening, 32 studies met the inclusion criteria. Studies were then organized by scar type and processing method. Outcomes were categorized as patient reported, observer reported, functional, objective biomechanical, or histological. Comparative studies were examined to evaluate whether fat processing techniques had an effect on reported outcomes.
Results: Fat injection has been studied across postsurgical, acne/atrophic, burn related, adherent/fibrotic, radiation associated, and inflammatory/sclerotic scars. Patient reported outcomes (e.g., pain, stiffness, POSAS patient scores) and observer reported scar scales (e.g., POSAS, VSS) demonstrated the most consistent improvement across scar types. Functional improvement, including range of motion and severity of contractures, was notably reported in radiation associated and fibrotic scars. Biomechanical assessments were variably reported, and histologic findings were not consistently correlated with clinical outcomes. Comparative studies evaluating processing techniques demonstrated early or scar-specific advantages with nanofat enrichment or SVF-based preparations; however, overall superiority over conventional fat grafting was not consistently shown.
Conclusions: Current clinical literature supports a beneficial role for fat injection in scar management, particularly in improving patient reported symptoms and functional outcomes. Evidence suggesting consistent advantage of specific fat processing techniques remains limited. The available data supports a scar type dependent application of fat-based therapies rather than a universally superior fat processing technique.
Competition Category: Health Services Outcomes or Clinical
Mentor: Arun Gosain, MD
Christian Arcelona, BS
Fellow (Clinical or Postdoctoral Researcher)
Expanding access to surgical simulation: Apple integration of the SkinFlaps soft-tissue surgical simulator
Introduction: Simulation-based rehearsal of facial soft-tissue procedures, including local flap design, defect closure, and cleft lip repair, is an increasingly important tool for both pre-operative planning and resident education. As surgical simulation technology advances toward patient-individualized virtual surgical planning, the need for consistent and reproducible tissue behavior across different computing platforms becomes critical. SkinFlaps is an open-source surgical simulator that employs projective dynamics and finite element modeling to support a range of facial soft-tissue procedures. Since its creation, SkinFlaps has only been available for Windows, limiting access for many plastic surgery programs and trainees who use Apple laptops. In this project, we developed a native macOS version of SkinFlaps to ensure broad accessibility and identical tissue behavior, enabling surgical rehearsals and teaching cases to yield consistent results on both platforms.
Methods: To create a native macOS version for Apple silicon, SkinFlaps was adapted without altering tissue parameters, simulation settings, or the core projective dynamics engine. Windows-specific software components were replaced with Mac-compatible alternatives, while the underlying physics model and material constants were preserved. Cross-platform consistency was verified through automated, side-by-side comparisons using standardized scenarios. At each simulation step, tissue displacement, solution accuracy, and volume preservation were measured and compared to predefined agreement thresholds.
Results: The macOS version of SkinFlaps operated successfully without any changes to tissue or simulation parameters. In all tested scenarios, tissue displacement, solution accuracy, and volume preservation were consistent between the macOS and Windows versions. The simulator performed interactively on Apple silicon, offering the complete set of surgical tools while maintaining the same tissue behavior as the Windows build.
Conclusions: We have developed a native macOS version of the open-source SkinFlaps surgical simulator that preserves identical tissue physics through projective dynamics and finite element modeling. This update enables plastic surgeons and trainees to use SkinFlaps on Apple hardware, supporting reproducible rehearsal of facial procedures. Achieving this cross-platform consistency is a key step toward our future enhancements, including integration of human-derived tissue properties, incorporation of patient-specific CT imaging for digital twinning and individualized planning, and educational validation studies to assess the simulator's value as a training tool. Following presentation, the Apple (macOS) version of SkinFlaps will be freely available as open-source software via GitHub.
Methods: To create a native macOS version for Apple silicon, SkinFlaps was adapted without altering tissue parameters, simulation settings, or the core projective dynamics engine. Windows-specific software components were replaced with Mac-compatible alternatives, while the underlying physics model and material constants were preserved. Cross-platform consistency was verified through automated, side-by-side comparisons using standardized scenarios. At each simulation step, tissue displacement, solution accuracy, and volume preservation were measured and compared to predefined agreement thresholds.
Results: The macOS version of SkinFlaps operated successfully without any changes to tissue or simulation parameters. In all tested scenarios, tissue displacement, solution accuracy, and volume preservation were consistent between the macOS and Windows versions. The simulator performed interactively on Apple silicon, offering the complete set of surgical tools while maintaining the same tissue behavior as the Windows build.
Conclusions: We have developed a native macOS version of the open-source SkinFlaps surgical simulator that preserves identical tissue physics through projective dynamics and finite element modeling. This update enables plastic surgeons and trainees to use SkinFlaps on Apple hardware, supporting reproducible rehearsal of facial procedures. Achieving this cross-platform consistency is a key step toward our future enhancements, including integration of human-derived tissue properties, incorporation of patient-specific CT imaging for digital twinning and individualized planning, and educational validation studies to assess the simulator's value as a training tool. Following presentation, the Apple (macOS) version of SkinFlaps will be freely available as open-source software via GitHub.
Competition Category: Education
Mentor: Arun Gosain, MD
Jane Frances Aruma, MD
Senior Resident (Clinical PGY3-5)
Evaluating the association between baseline food insecurity and financial distress in patients undergoing cancer treatment in Nigeria: Results from the COST-FIN Trial (NCT06630962)
Introduction: Cancer care imposes a significant financial burden on patients, particularly in low- and middle-income countries (LMICs), like Nigeria, where healthcare financing is largely out-of-pocket. Concurrently, over 40% of the population experiences food insecurity. With over 100,000 new cancer cases annually in Nigeria, the dual burden of limited food security and high treatment costs may worsen financial distress (FD). While the link between food insecurity and FD is well-documented in high-income settings, this relationship remains understudied in LMICs. This study explores the relationship between baseline food insecurity and FD among Nigerian patients undergoing cancer treatment.
Methods: In this prospective randomized controlled trial, adult patients (≥18 years) with a new diagnosis of breast, colorectal, or prostate cancer were enrolled within six weeks of diagnosis at a public and a private cancer center in Nigeria. A structured questionnaire was used to collect demographic, clinical, and socioeconomic data. Food insecurity was measured using the 2-item Hunger Vital Sign™, categorizing patients as food-secure (score 0), marginal/at-risk (1), or food-insecure (2). Financial distress was assessed with the 12-item FACIT-COST tool (score range: 0-44; lower = greater distress). Statistical analysis included the Kruskal-Wallis test, χ² or Fisher's exact test, Spearman rank correlation, and a univariable linear trend model (FACIT-COST ~ food-security score), with significance set at p < 0.05.
Results: Among 135 patients (median age 56; 65% women; cancer types: 64% breast, 21% prostate, 15% colorectal), 58.5% were food-insecure, 9.6% marginally insecure, and 31.9% food-secure at baseline. Financial distress worsened with increasing food insecurity (p < 0.001). Spearman correlation confirmed an inverse association (ρ = -0.41, p = 6.5 × 10⁻⁷). The linear trend model showed that worsening by one food-security level corresponded to a 4.3-point decrease in FACIT-COST score (β = -4.3 ± 0.8, p = 7.4 × 10⁻⁷). A corresponding gradient was seen in economic indicators: median personal income dropped from ₦79,500 ($52) (food-secure) to ₦50,000 ($33) and ₦48,000 ($31) (p = 0.018), while median household income declined from ₦185,000 ($121) to ₦90,000 ($59) to ₦50,000 ($33) (p = 0.002). Employment, cancer stage, age, and head-of-household status did not differ significantly across food-security groups (p ≥ 0.10).
Conclusions: Food insecurity was highly prevalent at diagnosis and strongly associated with financial distress in Nigerian cancer patients. Food insecurity is a clear early indicator of financial vulnerability during treatment. Routine screening for food insecurity could help identify high-risk patients and guide support services.
Methods: In this prospective randomized controlled trial, adult patients (≥18 years) with a new diagnosis of breast, colorectal, or prostate cancer were enrolled within six weeks of diagnosis at a public and a private cancer center in Nigeria. A structured questionnaire was used to collect demographic, clinical, and socioeconomic data. Food insecurity was measured using the 2-item Hunger Vital Sign™, categorizing patients as food-secure (score 0), marginal/at-risk (1), or food-insecure (2). Financial distress was assessed with the 12-item FACIT-COST tool (score range: 0-44; lower = greater distress). Statistical analysis included the Kruskal-Wallis test, χ² or Fisher's exact test, Spearman rank correlation, and a univariable linear trend model (FACIT-COST ~ food-security score), with significance set at p < 0.05.
Results: Among 135 patients (median age 56; 65% women; cancer types: 64% breast, 21% prostate, 15% colorectal), 58.5% were food-insecure, 9.6% marginally insecure, and 31.9% food-secure at baseline. Financial distress worsened with increasing food insecurity (p < 0.001). Spearman correlation confirmed an inverse association (ρ = -0.41, p = 6.5 × 10⁻⁷). The linear trend model showed that worsening by one food-security level corresponded to a 4.3-point decrease in FACIT-COST score (β = -4.3 ± 0.8, p = 7.4 × 10⁻⁷). A corresponding gradient was seen in economic indicators: median personal income dropped from ₦79,500 ($52) (food-secure) to ₦50,000 ($33) and ₦48,000 ($31) (p = 0.018), while median household income declined from ₦185,000 ($121) to ₦90,000 ($59) to ₦50,000 ($33) (p = 0.002). Employment, cancer stage, age, and head-of-household status did not differ significantly across food-security groups (p ≥ 0.10).
Conclusions: Food insecurity was highly prevalent at diagnosis and strongly associated with financial distress in Nigerian cancer patients. Food insecurity is a clear early indicator of financial vulnerability during treatment. Routine screening for food insecurity could help identify high-risk patients and guide support services.
Competition Category: Basic Science or Translational
Mentor: Juliet Lumati, MD MPH
Ahad Azimuddin, MD MBA
Junior Resident (Clinical PGY1-2)
Improving HPV vaccination in Ukraine: impact of targeted physician education on knowledge, attitudes, and clinical practice
Introduction: Cervical cancer remains a leading cause of cancer-related mortality among women in low- and middle-income countries and is driven by human papillomavirus (HPV) infection. Although HPV vaccination is effective and available in Ukraine, uptake remains low. This study aimed to identify barriers to HPV vaccination and evaluate the impact of targeted physician education
Methods: This prospective study conducted among primary care physicians and patients in Ivano-Frankivsk, Ukraine. Pre-intervention surveys were conducted testing knowledge and understanding of HPV vaccines. Based on baseline findings, a structured educational program covering areas of misconceptions was developed using European and US recommendations and delivered through continuing medical education events, supplemented by email and social media outreach.
Results: A total of 40 physicians and 34 patients/parents completed baseline surveys. Initial assessment revealed knowledge gaps among patients regarding asymptomatic transmission of HPV (29.4%) and male eligibility for vaccination (29.4%). Among physicians, gaps were noted in understanding of male vaccination (27.5%), age of eligibility (53%), dosing and route of administration (45%), and appropriate follow-up (27.5%).
Following the educational campaign, physicians showed knowledge improvement in dosing schedule (45% to 0%), age of eligibility (53% to 8%), PAP smear follow-up (from 28% to 18%), and male vaccination eligibility (from 28% to 7%). However, major misconceptions persisted, particularly regarding HPV risk perception, with 81% believing HPV does not pose cancer risk in individuals with limited sexual exposure (increased by 29%).
Perceived barriers shifted modestly: perceived cost role decreased (85% to 73%), while concerns about vaccine availability increased (80% to 95%), last one is likely related to ongoing war. Safety concerns declined (30% to 21%), though doubts about vaccine efficacy slightly increased (13% to 18%). There was change in clinical practice: omission of HPV discussions decreased 50% to 12%, routine sexual health discussions increased from 10% to 53%. Nearly all physicians reported recommending vaccination (95%), with high willingness to vaccinate their own children (98%). Uncertainty in recommending vaccination even if free decreased from 28% to 2%.
Conclusions: Targeted physician education significantly improved knowledge and clinical communication regarding HPV vaccination. Persistent misconceptions and structural barriers highlight the need for continued education and system-level interventions to improve vaccination uptake in Ukraine.
Methods: This prospective study conducted among primary care physicians and patients in Ivano-Frankivsk, Ukraine. Pre-intervention surveys were conducted testing knowledge and understanding of HPV vaccines. Based on baseline findings, a structured educational program covering areas of misconceptions was developed using European and US recommendations and delivered through continuing medical education events, supplemented by email and social media outreach.
Results: A total of 40 physicians and 34 patients/parents completed baseline surveys. Initial assessment revealed knowledge gaps among patients regarding asymptomatic transmission of HPV (29.4%) and male eligibility for vaccination (29.4%). Among physicians, gaps were noted in understanding of male vaccination (27.5%), age of eligibility (53%), dosing and route of administration (45%), and appropriate follow-up (27.5%).
Following the educational campaign, physicians showed knowledge improvement in dosing schedule (45% to 0%), age of eligibility (53% to 8%), PAP smear follow-up (from 28% to 18%), and male vaccination eligibility (from 28% to 7%). However, major misconceptions persisted, particularly regarding HPV risk perception, with 81% believing HPV does not pose cancer risk in individuals with limited sexual exposure (increased by 29%).
Perceived barriers shifted modestly: perceived cost role decreased (85% to 73%), while concerns about vaccine availability increased (80% to 95%), last one is likely related to ongoing war. Safety concerns declined (30% to 21%), though doubts about vaccine efficacy slightly increased (13% to 18%). There was change in clinical practice: omission of HPV discussions decreased 50% to 12%, routine sexual health discussions increased from 10% to 53%. Nearly all physicians reported recommending vaccination (95%), with high willingness to vaccinate their own children (98%). Uncertainty in recommending vaccination even if free decreased from 28% to 2%.
Conclusions: Targeted physician education significantly improved knowledge and clinical communication regarding HPV vaccination. Persistent misconceptions and structural barriers highlight the need for continued education and system-level interventions to improve vaccination uptake in Ukraine.
Competition Category: Quality Improvement
Mentor: Vitaliy Poylin, MD FACS FASCRS
Paola Barrios, MD MSAI
Senior Resident (Clinical PGY3-5)
Remodeling of protein networks across cirrhosis stages reveals molecular signatures of disease severity and therapeutic targets
Introduction: While 5-7% of compensated patients decompensate annually, we lack both diagnostic tools to predict this critical transition and therapies to halt progression. Here we analyze changes to protein network architecture through various stages of cirrhosis and identify key proteins in these stages.
Methods: Plasma samples from 30 cirrhosis patients at three stages: compensated, portal hypertension (pHTN), and decompensated were analyzed. Protein interaction networks were constructed using top-down proteomics expression data and Pearson correlations (r < 0.7). Interactions were parsed through STRING protein-protein interaction database and filtered to retain only connections found to have biological evidence. Rich-Club (RC) analysis, a method based on social network interactions, was applied to identify groups of highly connected proteins (“hubs”) that form core clusters integral for network maintenance. RC assesses whether these connections are stronger than expected by chance, identifying key regulatory and therapeutic proteins across disease stages.
Results: Network breakdown correlates with clinical deterioration: We analyzed the 44 proteins that had differentially expressed proteoforms across compensated, pHTN, decompensated cirrhosis. With disease progression, connections per protein decreased from compensated (12.9) to pHTN (11.4) to decompensated (6.5), as did network density (0.61, 0.52, 0.48), revealing decreasing network complexity (Fig). Normalized network coefficients at ρ>1 confirmed meaningful, non-random, biologically relevant networks. Identifying significant proteins: Compensated cirrhosis exhibited RC at k = 9 (ρ = 1.46; 24 proteins), portal hypertension showed a tighter RC at k = 9 (ρ = 1.49; 23 proteins), and decompensated cirrhosis displayed RC at k = 5 (ρ = 1.54; 18 proteins). Identifying unique proteins: Among RC proteins, Apolipoprotein A-II (apoA-II) emerged as a unique hub protein in decompensation when compared to compensated and pHTN and no unique hub proteins in pHTN.
Conclusions: Cirrhosis progression involves measurable collapse of plasma protein networks, with ApoA-II emerging as a molecular signature of decompensation. Increased connectivity in pHTN potentially signals final compensatory state before network deterioration in decompensation. These results offer an objective biomarker to investigate prediction of this critical transition, and possible therapeutic targets, addressing a major unmet need in cirrhosis care.
Methods: Plasma samples from 30 cirrhosis patients at three stages: compensated, portal hypertension (pHTN), and decompensated were analyzed. Protein interaction networks were constructed using top-down proteomics expression data and Pearson correlations (r < 0.7). Interactions were parsed through STRING protein-protein interaction database and filtered to retain only connections found to have biological evidence. Rich-Club (RC) analysis, a method based on social network interactions, was applied to identify groups of highly connected proteins (“hubs”) that form core clusters integral for network maintenance. RC assesses whether these connections are stronger than expected by chance, identifying key regulatory and therapeutic proteins across disease stages.
Results: Network breakdown correlates with clinical deterioration: We analyzed the 44 proteins that had differentially expressed proteoforms across compensated, pHTN, decompensated cirrhosis. With disease progression, connections per protein decreased from compensated (12.9) to pHTN (11.4) to decompensated (6.5), as did network density (0.61, 0.52, 0.48), revealing decreasing network complexity (Fig). Normalized network coefficients at ρ>1 confirmed meaningful, non-random, biologically relevant networks. Identifying significant proteins: Compensated cirrhosis exhibited RC at k = 9 (ρ = 1.46; 24 proteins), portal hypertension showed a tighter RC at k = 9 (ρ = 1.49; 23 proteins), and decompensated cirrhosis displayed RC at k = 5 (ρ = 1.54; 18 proteins). Identifying unique proteins: Among RC proteins, Apolipoprotein A-II (apoA-II) emerged as a unique hub protein in decompensation when compared to compensated and pHTN and no unique hub proteins in pHTN.
Conclusions: Cirrhosis progression involves measurable collapse of plasma protein networks, with ApoA-II emerging as a molecular signature of decompensation. Increased connectivity in pHTN potentially signals final compensatory state before network deterioration in decompensation. These results offer an objective biomarker to investigate prediction of this critical transition, and possible therapeutic targets, addressing a major unmet need in cirrhosis care.
Competition Category: Basic Science or Translational
Mentor: Daniela Ladner, MD MPH
Raheem Bell, MD MS MScL
Senior Resident (Clinical PGY3-5)
The association between county-level air pollution and advanced stage lung cancer at initial diagnosis in the United States
Introduction: Air pollution is a well-established carcinogen, yet the extent to which it is associated with advanced stage lung cancer at primary diagnosis remains unclear. This analysis examines the association between county-level air pollution and advanced-stage lung cancer presentation while controlling for socioeconomic and lifestyle confounders.
Methods: We conducted a retrospective ecological analysis of incident lung cancer cases with known clinical stage from the National Cancer Database (2010-2023), linked to annual satellite-derived county-level estimates of fine particulate matter (PM2.5) and nitrogen dioxide (NO2) across contiguous U.S. counties. The primary outcome was the rate of advanced-stage (stage III-IV) diagnoses per county-year. We fit generalized estimating equation Poisson regression models using standardized 3-year moving averages of PM2.5 and NO2 as the primary exposures. Models were adjusted for county-level confounders including smoking prevalence, Social Vulnerability Index, median household income, educational attainment, and healthcare access.
Results: Among 2,622,528 lung cancer cases (mean age 68.6, 51.0% male), NO2 exposure was associated with a higher likelihood of advanced-stage diagnosis. Each standard deviation increase in NO2 was associated with a 0.4% higher rate of advanced-stage presentation (RR=1.004, 95% CI: 1.001-1.007). In contrast, PM2.5 was not associated with advanced-stage diagnosis in any model (RR=1.001, 95% CI: 0.997-1.006). County-level smoking prevalence was the strongest predictor, with each standard deviation increase associated with a 1.4-1.7% higher rate of advanced-stage presentation (p<0.001). In histology-specific analyses, NO2 was associated with squamous cell carcinoma (RR=1.009, 95% CI: 1.005-1.012), while PM2.5 was associated with large cell carcinoma (RR=1.021, 95% CI: 1.001-1.042); no associations were observed for adenocarcinoma or small cell carcinoma. Counties with the highest social vulnerability had advanced-stage rates of 66.5%, representing a 14.2 percentage-point difference compared to the lowest-risk strata defined by vulnerability, pollution, and smoking.
Conclusions: County-level NO2 exposure was associated with a consistent increase in the likelihood of advanced-stage lung cancer at diagnosis, particularly for squamous cell carcinoma, whereas PM2.5 showed no independent association. Structural and lifestyle factors like smoking prevalence and social vulnerability remain the dominant drivers of late-stage presentation. Efforts to improve early detection should prioritize reducing these underlying disparities alongside environmental risk mitigation.
Methods: We conducted a retrospective ecological analysis of incident lung cancer cases with known clinical stage from the National Cancer Database (2010-2023), linked to annual satellite-derived county-level estimates of fine particulate matter (PM2.5) and nitrogen dioxide (NO2) across contiguous U.S. counties. The primary outcome was the rate of advanced-stage (stage III-IV) diagnoses per county-year. We fit generalized estimating equation Poisson regression models using standardized 3-year moving averages of PM2.5 and NO2 as the primary exposures. Models were adjusted for county-level confounders including smoking prevalence, Social Vulnerability Index, median household income, educational attainment, and healthcare access.
Results: Among 2,622,528 lung cancer cases (mean age 68.6, 51.0% male), NO2 exposure was associated with a higher likelihood of advanced-stage diagnosis. Each standard deviation increase in NO2 was associated with a 0.4% higher rate of advanced-stage presentation (RR=1.004, 95% CI: 1.001-1.007). In contrast, PM2.5 was not associated with advanced-stage diagnosis in any model (RR=1.001, 95% CI: 0.997-1.006). County-level smoking prevalence was the strongest predictor, with each standard deviation increase associated with a 1.4-1.7% higher rate of advanced-stage presentation (p<0.001). In histology-specific analyses, NO2 was associated with squamous cell carcinoma (RR=1.009, 95% CI: 1.005-1.012), while PM2.5 was associated with large cell carcinoma (RR=1.021, 95% CI: 1.001-1.042); no associations were observed for adenocarcinoma or small cell carcinoma. Counties with the highest social vulnerability had advanced-stage rates of 66.5%, representing a 14.2 percentage-point difference compared to the lowest-risk strata defined by vulnerability, pollution, and smoking.
Conclusions: County-level NO2 exposure was associated with a consistent increase in the likelihood of advanced-stage lung cancer at diagnosis, particularly for squamous cell carcinoma, whereas PM2.5 showed no independent association. Structural and lifestyle factors like smoking prevalence and social vulnerability remain the dominant drivers of late-stage presentation. Efforts to improve early detection should prioritize reducing these underlying disparities alongside environmental risk mitigation.
Competition Category: Health Services Outcomes or Clinical
Mentor: David Odell, MD FACS
Johanna Borst, MD
Senior Resident (Clinical PGY3-5)
No place like home: discharge timing and costs following pediatric gastrostomy tube placement in the United States
Introduction: Same-day discharge (SDD) after elective pediatric gastrostomy tube placement is safe in selected patients, yet contemporary national adoption and its impact on costs and practice variation remain poorly defined. We evaluated the association between SDD and adjusted hospital costs, as well as temporal trends and hospital-level variation in SDD use.
Methods: We performed a retrospective cohort study using the Pediatric Health Information System, including children ≤18 years undergoing elective gastrostomy tube placement from 2016 to 2025. Encounters with concurrent procedures and postoperative length of stay >2 days were excluded. Discharge was categorized as SDD versus postoperative day (POD) 1-2. Costs were modeled using generalized linear models with a gamma distribution and log link, adjusting for demographics, medical complexity, illness severity, hospital volume, and year, with clustering at the hospital level. Temporal trends were assessed using logistic regression, and hospital-level variation was quantified using mixed-effects models with median odds ratios (MORs). Thirty-day revisits were evaluated as a balancing measure.
Results: Among 11,818 children across 46 hospitals, SDD occurred in 1,155 encounters (9.8%). Median costs were substantially lower for SDD compared with POD 1-2 discharge ($34,018 [IQR $17,959-$46,770] vs $49,491 [IQR $33,284-$71,159]). After adjustment, SDD was associated with 31% lower costs (cost ratio 0.69, 95% CI 0.58-0.82), corresponding to approximately $17,000 lower costs per encounter. Each additional postoperative day was associated with a 29% increase in adjusted costs. SDD increased from 4.7% in 2016 to 12.0% in 2025, although this trend was not significant after accounting for clustering within hospitals. Adoption was highly concentrated: the five highest-volume hospitals accounted for 59% of all SDD cases. Hospital-level variation was substantial (median SDD rate 3.1%, IQR 1.3%-9.5%), with a MOR of 4.91, indicating nearly five-fold differences in odds of SDD for otherwise similar patients across hospitals. Revisits were uncommon (1.45%) and were not significantly associated with SDD (adjusted OR 1.60, 95% CI 0.93-2.76).
Conclusions: SDD after pediatric gastrostomy placement is associated with markedly lower costs without an increase in short-term revisits, yet remains underutilized and highly variable across hospitals. These findings suggest that discharge timing is a major, modifiable driver of cost variation and that broader implementation of standardized perioperative pathways represents a high-yield opportunity to improve value in pediatric surgical care.
Methods: We performed a retrospective cohort study using the Pediatric Health Information System, including children ≤18 years undergoing elective gastrostomy tube placement from 2016 to 2025. Encounters with concurrent procedures and postoperative length of stay >2 days were excluded. Discharge was categorized as SDD versus postoperative day (POD) 1-2. Costs were modeled using generalized linear models with a gamma distribution and log link, adjusting for demographics, medical complexity, illness severity, hospital volume, and year, with clustering at the hospital level. Temporal trends were assessed using logistic regression, and hospital-level variation was quantified using mixed-effects models with median odds ratios (MORs). Thirty-day revisits were evaluated as a balancing measure.
Results: Among 11,818 children across 46 hospitals, SDD occurred in 1,155 encounters (9.8%). Median costs were substantially lower for SDD compared with POD 1-2 discharge ($34,018 [IQR $17,959-$46,770] vs $49,491 [IQR $33,284-$71,159]). After adjustment, SDD was associated with 31% lower costs (cost ratio 0.69, 95% CI 0.58-0.82), corresponding to approximately $17,000 lower costs per encounter. Each additional postoperative day was associated with a 29% increase in adjusted costs. SDD increased from 4.7% in 2016 to 12.0% in 2025, although this trend was not significant after accounting for clustering within hospitals. Adoption was highly concentrated: the five highest-volume hospitals accounted for 59% of all SDD cases. Hospital-level variation was substantial (median SDD rate 3.1%, IQR 1.3%-9.5%), with a MOR of 4.91, indicating nearly five-fold differences in odds of SDD for otherwise similar patients across hospitals. Revisits were uncommon (1.45%) and were not significantly associated with SDD (adjusted OR 1.60, 95% CI 0.93-2.76).
Conclusions: SDD after pediatric gastrostomy placement is associated with markedly lower costs without an increase in short-term revisits, yet remains underutilized and highly variable across hospitals. These findings suggest that discharge timing is a major, modifiable driver of cost variation and that broader implementation of standardized perioperative pathways represents a high-yield opportunity to improve value in pediatric surgical care.
Competition Category: Health Services Outcomes or Clinical
Mentor: Timothy Lautz, MD
Clara Bosco, MD
Senior Resident (Clinical PGY3-5)
Secondary interventions after total arch replacement with frozen elephant trunk
Introduction: Total aortic arch replacement with frozen elephant trunk (FET) is a hybrid procedure that combines aortic arch replacement with stent grafting of the descending thoracic aorta. Insight into patient-related factors, indications, and timing of distal re-intervention after FET is limited. This study aims to evaluate outcomes following FET implantation and identify factors associated with and timing of distal aortic reintervention.
Methods: Single-center retrospective review of adult patients undergoing total arch reconstruction with FET from 2013-2024. Data were drawn from a prospectively maintained aortic database, with additional variables collected through retrospective review.
Results: Included were 146 patients with a mean age of 57.9 ± 12.3 years. 21.9% (32/146) were female. 21.9% (32/146) underwent aortic reintervention, primarily endovascular (28/32, 87.5%). Median time to reintervention was 228 days (IQR 130.5-410.75). The leading indication for reintervention was late aortic diameter progression (23/32, 71.9%) with residual dissection (20/32, 71.9%). Hyperlipidemia (53.1% [17/32] vs 29.8% [34/114], p = 0.021) or prior aortic intervention (59.4% [19/32] vs 30.7% [35/114], p = 0.003) were associated with reintervention. Patients with a reintervention were less likely to have experienced a stroke following the index FET procedure (3.1% [1/32] vs 22.8% [26/114], p = 0.009). Overall survival (OS) did not differ between patients with and without reintervention (p = 0.12; Figure 1). The cumulative incidence of reintervention at 12 and 36 months was 19.2% (95% CI 12.2-27.4) and 33.0% (95% CI 23.1-43.1), respectively (Figure 2).
Conclusions: Aortic reintervention after FET was necessary in 33% of patients within 36 months after the index procedure. Most reinterventions were endovascular and did not affect overall mortality. Patients with greater comorbidity burden and prior aortic interventions had increased reintervention risk. These findings highlight FET as a platform for staged aortic repair and emphasize the importance of longitudinal surveillance and multidisciplinary longitudinal decision-making.
Methods: Single-center retrospective review of adult patients undergoing total arch reconstruction with FET from 2013-2024. Data were drawn from a prospectively maintained aortic database, with additional variables collected through retrospective review.
Results: Included were 146 patients with a mean age of 57.9 ± 12.3 years. 21.9% (32/146) were female. 21.9% (32/146) underwent aortic reintervention, primarily endovascular (28/32, 87.5%). Median time to reintervention was 228 days (IQR 130.5-410.75). The leading indication for reintervention was late aortic diameter progression (23/32, 71.9%) with residual dissection (20/32, 71.9%). Hyperlipidemia (53.1% [17/32] vs 29.8% [34/114], p = 0.021) or prior aortic intervention (59.4% [19/32] vs 30.7% [35/114], p = 0.003) were associated with reintervention. Patients with a reintervention were less likely to have experienced a stroke following the index FET procedure (3.1% [1/32] vs 22.8% [26/114], p = 0.009). Overall survival (OS) did not differ between patients with and without reintervention (p = 0.12; Figure 1). The cumulative incidence of reintervention at 12 and 36 months was 19.2% (95% CI 12.2-27.4) and 33.0% (95% CI 23.1-43.1), respectively (Figure 2).
Conclusions: Aortic reintervention after FET was necessary in 33% of patients within 36 months after the index procedure. Most reinterventions were endovascular and did not affect overall mortality. Patients with greater comorbidity burden and prior aortic interventions had increased reintervention risk. These findings highlight FET as a platform for staged aortic repair and emphasize the importance of longitudinal surveillance and multidisciplinary longitudinal decision-making.
Competition Category: Health Services Outcomes or Clinical
Mentor: Neel Mansukhani, MD
Mariana Bustamante Eduardo, PhD
Fellow (Clinical or Postdoctoral Researcher)
Fatty acid exposure promotes age-related mammary tissue alterations with pro-tumorigenic potential
Introduction: Exposure of non-transformed breast cells and breast microstructures to the medium-chain (MC) fatty acid (FA) octanoic acid (OA) induces a metabolic shift toward the serine, one-carbon, glycine and methionine pathways (SOG/methionine), enhancing epigenetic plasticity, increasing reactive oxygen species (ROS), promoting cell survival and disrupting cell-cell communication. Similarly, the aged mammary gland is characterized by disrupted cell-cell communication, epigenetic plasticity and increased ROS. We hypothesize that FA-induced metabolic reprogramming leads to biological aging of the mammary gland, contributing to pro-tumorigenic alterations observed during chronologic aging.
Methods: MCF-10A cells were exposed to ± OA for proteomics. Breast microstructures exposed to ± OA were analyzed by scRNAseq. Breast microstructures and 3D mammary spheres derived from primary cells were embedded in Matrigel, exposed to ± OA for 7 days, stained for luminal and basal markers, F-actin, and nuclei, and imaged by confocal microscopy to assess migration/invasion. Migratory cell populations enriched in OA-containing media were identified with scRNAseq. Raman spectroscopy (RS) was used to characterize the lipid content in normal breast tissue.
Results: OA treatment induced changes previously reported in aging and tumorigenic contexts, including: (1) upregulation (p < 0.01) of aging-related genes (GDF15, MDK, PLIN2), and downregulation (p < 0.01) of lineage markers and MMP7, a gene whose downregulation promotes mammary epithelial aging; (2) upregulation (p < 0.01) of Senescence-Associated Secretory Phenotype (SASP) genes, including AREG (reprogramming) and ANGPTL4 (migration); (3) increased secreted signaling via AREG, GDF15, and MDK; and (4) reduced extracelular matrix (ECM)-receptor and cell-cell interactions. Ex vivo, OA altered tissue architecture disrupting the basal barrier and promoting cellular migration. BMYO1, LASP1, and LHS1 epithelial subtypes were among the migratory cells in OA media and expressed SASP, cancer (MYC, EGFR, SREBF1), migration (S100A4, NCAM1), aging and SOG/methionine genes. FB1 fibroblasts dominated in vehicle media, but OA favored ECM-disassembling FB2 cells. RS analysis demostrates the presence of both saturated and unsaturated FAs and revealed the presence of MCFAs, such as OA, with higher intensities observed in the postmenopausal tissue supporting the in vivo plausibility of our in vitro/ex vivo findings.
Conclusions: Our data supports a model suggesting that chronological and biological aging processes increase the release of free FAs, due to elevated GDF15-induced lipolysis. The rise in FAs drives mammary gland remodeling and accelerates aging of the gland. Chronological and biological aging increase vulnerability to breast cancer. This model suggests potential preventive strategies such as targeting GDF15 and SOG/methionine.
Methods: MCF-10A cells were exposed to ± OA for proteomics. Breast microstructures exposed to ± OA were analyzed by scRNAseq. Breast microstructures and 3D mammary spheres derived from primary cells were embedded in Matrigel, exposed to ± OA for 7 days, stained for luminal and basal markers, F-actin, and nuclei, and imaged by confocal microscopy to assess migration/invasion. Migratory cell populations enriched in OA-containing media were identified with scRNAseq. Raman spectroscopy (RS) was used to characterize the lipid content in normal breast tissue.
Results: OA treatment induced changes previously reported in aging and tumorigenic contexts, including: (1) upregulation (p < 0.01) of aging-related genes (GDF15, MDK, PLIN2), and downregulation (p < 0.01) of lineage markers and MMP7, a gene whose downregulation promotes mammary epithelial aging; (2) upregulation (p < 0.01) of Senescence-Associated Secretory Phenotype (SASP) genes, including AREG (reprogramming) and ANGPTL4 (migration); (3) increased secreted signaling via AREG, GDF15, and MDK; and (4) reduced extracelular matrix (ECM)-receptor and cell-cell interactions. Ex vivo, OA altered tissue architecture disrupting the basal barrier and promoting cellular migration. BMYO1, LASP1, and LHS1 epithelial subtypes were among the migratory cells in OA media and expressed SASP, cancer (MYC, EGFR, SREBF1), migration (S100A4, NCAM1), aging and SOG/methionine genes. FB1 fibroblasts dominated in vehicle media, but OA favored ECM-disassembling FB2 cells. RS analysis demostrates the presence of both saturated and unsaturated FAs and revealed the presence of MCFAs, such as OA, with higher intensities observed in the postmenopausal tissue supporting the in vivo plausibility of our in vitro/ex vivo findings.
Conclusions: Our data supports a model suggesting that chronological and biological aging processes increase the release of free FAs, due to elevated GDF15-induced lipolysis. The rise in FAs drives mammary gland remodeling and accelerates aging of the gland. Chronological and biological aging increase vulnerability to breast cancer. This model suggests potential preventive strategies such as targeting GDF15 and SOG/methionine.
Competition Category: Basic Science or Translational
Mentor: Susan Clare, MD PhD
Austin Chang, BA
Student
Recovery-focused robotic transhiatal esophagectomy enables early discharge without compromising oncologic outcomes
Introduction: Esophagectomy is a crucial component of esophageal cancer treatment. However, it is associated with high morbidity and prolonged hospitalization. We evaluated whether a recovery-focused robotic transhiatal esophagectomy (RTHE) approach could enable earlier discharge without compromising safety or oncologic outcomes.
Methods: We conducted a retrospective cohort analysis using single institutional data for esophageal cancer patients who underwent RTHE or transthoracic esophagectomy (TTE) (Feb 2020-Dec 2025). Multivariate Cox proportional hazards regression and Fine and Gray subdistribution hazard models were used to determine factors associated with overall survival and recurrence-free survival, respectively. A multivariate regression model identified factors associated with hospital length of stay (LOS). RTHE cases were divided into chronological tertiles to assess learning-curve effects.
Results: Of 206 patients undergoing esophagectomy, 79 patients (38.3%) underwent RTHE and 127 patients (61.7%) received TTE. Patient demographics and lymph node harvest were comparable. RTHE had earlier oral intake (1 vs 7 days, p<0.001), reduced feeding jejunostomy-tube placement (17.7% vs 89.8%, p<0.001), fewer ICU admissions (67.1% vs 94.5%, p<0.001), shorter ICU stay (1 vs 2 days, p<0.001), and shorter hospital LOS (4 vs 9 days, p<0.001). More RTHE patients were discharged home directly (98.7% vs 85.8%, p=0.016), with no difference in all-cause 30-day readmission rates. RTHE had more laryngeal nerve paresis (10.1% vs 2.4%; p=0.024) but comparable anastomotic leak rate (7.6% vs 17.3%; p=0.060) and less acute respiratory distress syndrome (0% vs 8.7%, p=0.008). Adjusted overall and recurrence-free survival were comparable. RTHE was independently associated with shorter median LOS [-6 days (-6.85 to -5.00, 95% CI), p<0.001]. Across chronological tertiles, LOS decreased to 3 days and ICU admission decreased to 14.8% (p<0.001).
Conclusions: A recovery-focused robotic transhiatal esophagectomy pathway enabled earlier oral intake, reduced ICU utilization, and hospital discharge as early as 3 days without compromise in oncologic adequacy or survival.
Methods: We conducted a retrospective cohort analysis using single institutional data for esophageal cancer patients who underwent RTHE or transthoracic esophagectomy (TTE) (Feb 2020-Dec 2025). Multivariate Cox proportional hazards regression and Fine and Gray subdistribution hazard models were used to determine factors associated with overall survival and recurrence-free survival, respectively. A multivariate regression model identified factors associated with hospital length of stay (LOS). RTHE cases were divided into chronological tertiles to assess learning-curve effects.
Results: Of 206 patients undergoing esophagectomy, 79 patients (38.3%) underwent RTHE and 127 patients (61.7%) received TTE. Patient demographics and lymph node harvest were comparable. RTHE had earlier oral intake (1 vs 7 days, p<0.001), reduced feeding jejunostomy-tube placement (17.7% vs 89.8%, p<0.001), fewer ICU admissions (67.1% vs 94.5%, p<0.001), shorter ICU stay (1 vs 2 days, p<0.001), and shorter hospital LOS (4 vs 9 days, p<0.001). More RTHE patients were discharged home directly (98.7% vs 85.8%, p=0.016), with no difference in all-cause 30-day readmission rates. RTHE had more laryngeal nerve paresis (10.1% vs 2.4%; p=0.024) but comparable anastomotic leak rate (7.6% vs 17.3%; p=0.060) and less acute respiratory distress syndrome (0% vs 8.7%, p=0.008). Adjusted overall and recurrence-free survival were comparable. RTHE was independently associated with shorter median LOS [-6 days (-6.85 to -5.00, 95% CI), p<0.001]. Across chronological tertiles, LOS decreased to 3 days and ICU admission decreased to 14.8% (p<0.001).
Conclusions: A recovery-focused robotic transhiatal esophagectomy pathway enabled earlier oral intake, reduced ICU utilization, and hospital discharge as early as 3 days without compromise in oncologic adequacy or survival.
Competition Category: Health Services Outcomes or Clinical
Mentor: Samuel Kim, MD
Calvin Chao, MD
Senior Resident (Clinical PGY3-5)
Association of vascular surgery interest groups with applicant volume and match outcomes: A multi-year analysis of National Residency Match Program data
Introduction: Vascular Surgery Interest Groups (VSIGs) have been established at numerous medical schools to promote early specialty exposure. However, their influence on applicant volume and match outcomes remains incompletely defined. This study evaluates the impact of VSIG presence in an applicant's medical school on national application volume to integrated vascular surgery residency and match performance.
Methods: U.S. allopathic medical schools with active VSIGs were identified via Society for Vascular Surgery registration. Aggregate, de-identified data were obtained from the National Resident Matching Program for 2020-2025 match cycles and compared to U.S. allopathic non-VSIG institutions. Data were normalized by number of institutions in each group and used for univariate analysis.
Results: 428 applicants were identified from VSIG institutions (n=88) versus 142 applicants from non-VSIG institutions (n=76). VSIG institutions demonstrated significantly higher normalized applicant volume per cycle (0.81 vs 0.31, p<0.0001) and higher mean annual match rate (79.8% vs 64.9%, p=0.006). Applicant volume per school increased at VSIG institutions over time (2020: 0.63; 2025: 0.97), while non-VSIG institutions showed a slight decline (2020: 0.36; 2025: 0.32). No significant differences were found between groups in research, work, or volunteer experiences, or abstract/presentation/publication count. Nationally, research experiences remained srows, work and volunteer experiences declined, and abstract/presentation/publication count increased across the study period.
Conclusions: The presence of a VSIG at a medical school is associated with increased integrated vascular surgery residency applicant volume and improved match outcomes. Early engagement through VSIGs may be an effective strategy to enhance recruitment into vascular surgery.
Methods: U.S. allopathic medical schools with active VSIGs were identified via Society for Vascular Surgery registration. Aggregate, de-identified data were obtained from the National Resident Matching Program for 2020-2025 match cycles and compared to U.S. allopathic non-VSIG institutions. Data were normalized by number of institutions in each group and used for univariate analysis.
Results: 428 applicants were identified from VSIG institutions (n=88) versus 142 applicants from non-VSIG institutions (n=76). VSIG institutions demonstrated significantly higher normalized applicant volume per cycle (0.81 vs 0.31, p<0.0001) and higher mean annual match rate (79.8% vs 64.9%, p=0.006). Applicant volume per school increased at VSIG institutions over time (2020: 0.63; 2025: 0.97), while non-VSIG institutions showed a slight decline (2020: 0.36; 2025: 0.32). No significant differences were found between groups in research, work, or volunteer experiences, or abstract/presentation/publication count. Nationally, research experiences remained srows, work and volunteer experiences declined, and abstract/presentation/publication count increased across the study period.
Conclusions: The presence of a VSIG at a medical school is associated with increased integrated vascular surgery residency applicant volume and improved match outcomes. Early engagement through VSIGs may be an effective strategy to enhance recruitment into vascular surgery.
Competition Category: Education
Mentor: Tadaki Tomita, MD
Aboubacar Cherif, BA
Student
A comparison of a PAD-specific customized generative AI chatbot (VERA) versus general-purpose AI chatbots for peripheral artery disease patient education
Introduction: Patients with peripheral artery disease (PAD) are known to have poor awareness and understanding of the diagnosis. The role of generative AI chatbots in improving PAD patient education is unknown. Our goal is to compare a generative AI chatbot customized for PAD patient education to publicly available AI chatbots.
Methods: This is a cross-sectional comparative evaluation of the responses of four AI chatbots to ten prompts that are commonly asked questions about PAD. The three publicly available AI chatbots were ChatGPT-5, Gemini 2.5 Flash, and Claude Sonnet 4.5. We created a customized, voice AI chatbot for PAD education grounded on curated and prompt-injected guidance called Vascular Education and Resources using Artificial Intelligence, or “VERA.” De-identified chatbot-generated responses to inputs were assessed for readability (Flesch-Kincaid Grade Level, Flesch Reading Ease, Gunning Fog Index, Simple Measure of Gobbledygook Index, and Average Reading Level Consensus Score), accuracy, comprehensiveness, and patient education quality (Patient Education Materials Assessment Tool; PEMAT) using validated instruments and expert scoring rubrics. Nonparametric statistical testing was used to compare chatbot performance across all evaluation domains.
Results: VERA generated the most accessible text compared to the other chatbots and produced responses at a median grade level of 6.6, which was lower than responses from the other chatbots. PAD expert-rated accuracy scores were high across all the chatbots without significant differences between them. Comprehensiveness scores were more varied and demonstrated that VERA was less comprehensive than the other chatbots. PEMAT understandability scores were uniformly high. PEMAT actionability scores were low overall but did not differ significantly across chatbots on post hoc analysis.
Conclusions: A generative AI chatbot research tool customized for PAD patient education generates textual information about PAD that is more accessible (mean grade level 6.6) than publicly available AI chatbots without loss of accuracy, albeit with modestly reduced comprehensiveness that reflects intentional simplification for patient-centered communication. Future research will assess the acceptability and feasibility of this research tool to be adopted as part of PAD patient education.
Methods: This is a cross-sectional comparative evaluation of the responses of four AI chatbots to ten prompts that are commonly asked questions about PAD. The three publicly available AI chatbots were ChatGPT-5, Gemini 2.5 Flash, and Claude Sonnet 4.5. We created a customized, voice AI chatbot for PAD education grounded on curated and prompt-injected guidance called Vascular Education and Resources using Artificial Intelligence, or “VERA.” De-identified chatbot-generated responses to inputs were assessed for readability (Flesch-Kincaid Grade Level, Flesch Reading Ease, Gunning Fog Index, Simple Measure of Gobbledygook Index, and Average Reading Level Consensus Score), accuracy, comprehensiveness, and patient education quality (Patient Education Materials Assessment Tool; PEMAT) using validated instruments and expert scoring rubrics. Nonparametric statistical testing was used to compare chatbot performance across all evaluation domains.
Results: VERA generated the most accessible text compared to the other chatbots and produced responses at a median grade level of 6.6, which was lower than responses from the other chatbots. PAD expert-rated accuracy scores were high across all the chatbots without significant differences between them. Comprehensiveness scores were more varied and demonstrated that VERA was less comprehensive than the other chatbots. PEMAT understandability scores were uniformly high. PEMAT actionability scores were low overall but did not differ significantly across chatbots on post hoc analysis.
Conclusions: A generative AI chatbot research tool customized for PAD patient education generates textual information about PAD that is more accessible (mean grade level 6.6) than publicly available AI chatbots without loss of accuracy, albeit with modestly reduced comprehensiveness that reflects intentional simplification for patient-centered communication. Future research will assess the acceptability and feasibility of this research tool to be adopted as part of PAD patient education.
Competition Category: Health Services Outcomes or Clinical
Mentor: Karen Ho, MD
Anastasia Chibucos, BS
Student
Surgical outcomes in pancreatic cancer: Implications for perioperative risk stratification
Introduction: Pancreatic ductal adenocarcinoma (PDAC) is the third leading cause of cancer related mortality in the United States. While systemic therapy and surgical resection improve survival, perioperative morbidity remains substantial. This analysis aims to look at potential impacts of the Area Deprivation Index (ADI), a social determinant of health, and the Prognostic Nutritional Index (PNI) on perioperative outcomes.
Methods: Patients with localized PDAC enrolled in a prospective phase II trial (NCT04616131) between October 2020 and December 2025 were analyzed. ADI (national percentile) was grouped into tertiles (Low/Medium/High). Readmission was defined as hospital encounters within 30 days or 31-90 days following discharge. Comparisons between ADI and readmission were assessed using chi-square tests and independent samples of t-tests. Associations between PNI and readmission were analyzed using t-tests. The relationship between PNI and length of stay (LOS) was evaluated using Pearson correlation analysis, with LOS treated as a continuous variable. Additional analyses evaluated circulating tumor DNA (ctDNA), operative variables, and comorbidities using univariate statistical methods.
Results: In a cohort of 76 patients undergoing pancreatic resection, 28 (37%) were admitted within 90 days. In multivariable Cox regression, LOS was independently associated with increased hazard of readmission (HR 1.12, 95% CI 1.04-1.20, p=0.004), while PNI was not (HR 0.95, 95% CI 0.89-1.02, p=0.18).
However, lower PNI was significantly associated with 30-day readmission (44.91 vs 47.91; p = 0.039), but not with 90-day readmission (47.61 vs 47.18; p = 0.798) suggesting impaired nutritional status in early postoperative readmissions. ADI was not associated with 30-day readmission (mean 35.46 vs 39.18; p = 0.461), but was significantly higher among patients readmitted at 31-90 days (46.68 vs 34.46; p = 0.012) suggesting a potential association between socioeconomic disadvantage and delayed readmission.
No significant associations were observed between readmission and ctDNA detection at diagnosis, post-adjuvant therapy, or postoperatively.
Conclusions: Lower PNI associated with 30-day readmission, while ADI contributed to delayed (31-90 day) readmission, suggesting distinct biologic and socioeconomic influences on postoperative outcomes. In contrast, tumor-derived biomarkers and intraoperative factors were not associated with readmission in this cohort. These findings demonstrate the importance of patient level physiologic and social determinants in perioperative risk for PDAC.
Methods: Patients with localized PDAC enrolled in a prospective phase II trial (NCT04616131) between October 2020 and December 2025 were analyzed. ADI (national percentile) was grouped into tertiles (Low/Medium/High). Readmission was defined as hospital encounters within 30 days or 31-90 days following discharge. Comparisons between ADI and readmission were assessed using chi-square tests and independent samples of t-tests. Associations between PNI and readmission were analyzed using t-tests. The relationship between PNI and length of stay (LOS) was evaluated using Pearson correlation analysis, with LOS treated as a continuous variable. Additional analyses evaluated circulating tumor DNA (ctDNA), operative variables, and comorbidities using univariate statistical methods.
Results: In a cohort of 76 patients undergoing pancreatic resection, 28 (37%) were admitted within 90 days. In multivariable Cox regression, LOS was independently associated with increased hazard of readmission (HR 1.12, 95% CI 1.04-1.20, p=0.004), while PNI was not (HR 0.95, 95% CI 0.89-1.02, p=0.18).
However, lower PNI was significantly associated with 30-day readmission (44.91 vs 47.91; p = 0.039), but not with 90-day readmission (47.61 vs 47.18; p = 0.798) suggesting impaired nutritional status in early postoperative readmissions. ADI was not associated with 30-day readmission (mean 35.46 vs 39.18; p = 0.461), but was significantly higher among patients readmitted at 31-90 days (46.68 vs 34.46; p = 0.012) suggesting a potential association between socioeconomic disadvantage and delayed readmission.
No significant associations were observed between readmission and ctDNA detection at diagnosis, post-adjuvant therapy, or postoperatively.
Conclusions: Lower PNI associated with 30-day readmission, while ADI contributed to delayed (31-90 day) readmission, suggesting distinct biologic and socioeconomic influences on postoperative outcomes. In contrast, tumor-derived biomarkers and intraoperative factors were not associated with readmission in this cohort. These findings demonstrate the importance of patient level physiologic and social determinants in perioperative risk for PDAC.
Competition Category: Health Services Outcomes or Clinical
Mentor: Akhil Chawla, MD
Elizabeth Christian, MPPA
Student
Navigating financial toxicity in global cancer Care: A systematic review of financial navigation interventions
Introduction: The rising cost of cancer care is increasing the financial burden on patients and health systems globally, driving recognition of financial toxicity (FT) as a critical outcome in oncology. FT is associated with treatment nonadherence, poorer quality of life, and increased mortality, with a disproportionate impact in low- and middle-income countries (LMICs). Financial navigation programs (FNPs) have emerged to mitigate FT by addressing cost-related barriers through insurance optimization, financial assistance, and cost management. However, their structure, implementation, and effectiveness remain heterogeneous, and evidence has not been systematically synthesized and is largely concentrated in high-income settings. The objective of this study was to systematically review and synthesize the existing literature on financial navigation programs in cancer care, with a focus on their impact on financial and clinical outcomes across diverse health system settings.
Methods: We conducted a systematic review following Cochrane Handbook methodology and PRISMA 2020 guidelines (PROSPERO ID: 1239362). MEDLINE, Embase, Cochrane Library, Web of Science, and other databases were searched from inception to November 26, 2025. Eligible studies included original research evaluating oncology FNPs among patients with cancer and/or caregivers. Multiple study designs were included. Two independent authors (. & .) reviewed the articles for inclusion, extracted the data, and conducted a quality assessment of the studies. Data were extracted on study characteristics, intervention components, and outcomes. Risk of bias was assessed using RoB 2, ROBINS-I, CASP, and MMAT tools. Findings were synthesized narratively.
Results: Fifteen studies (2018-2025), all from the United States, were included; three were randomized trials. Interventions commonly involved financial screening, resource coordination, insurance navigation, and longitudinal follow-up. Eleven studies reported participant financial outcomes, with eight reporting cost savings, including reductions in out-of-pocket costs, with mean per-patient savings ranging from $350 to $16,633. Evidence on clinical outcomes was limited, with only two studies assessing adherence. Four studies reported positive program-level return on investment. Overall, methodological quality was variable, with moderate to high risk of bias.
Conclusions: Financial navigation programs (FNPs) demonstrate promise in addressing financial burden; however, evidence remains limited and heterogeneous. This burden is disproportionately high in low- and middle-income countries (LMICs), where it is a key driver of treatment abandonment and adverse outcomes. Future research should prioritize standardized frameworks, rigorous evaluation, and expanded implementation in LMIC settings.
Methods: We conducted a systematic review following Cochrane Handbook methodology and PRISMA 2020 guidelines (PROSPERO ID: 1239362). MEDLINE, Embase, Cochrane Library, Web of Science, and other databases were searched from inception to November 26, 2025. Eligible studies included original research evaluating oncology FNPs among patients with cancer and/or caregivers. Multiple study designs were included. Two independent authors (. & .) reviewed the articles for inclusion, extracted the data, and conducted a quality assessment of the studies. Data were extracted on study characteristics, intervention components, and outcomes. Risk of bias was assessed using RoB 2, ROBINS-I, CASP, and MMAT tools. Findings were synthesized narratively.
Results: Fifteen studies (2018-2025), all from the United States, were included; three were randomized trials. Interventions commonly involved financial screening, resource coordination, insurance navigation, and longitudinal follow-up. Eleven studies reported participant financial outcomes, with eight reporting cost savings, including reductions in out-of-pocket costs, with mean per-patient savings ranging from $350 to $16,633. Evidence on clinical outcomes was limited, with only two studies assessing adherence. Four studies reported positive program-level return on investment. Overall, methodological quality was variable, with moderate to high risk of bias.
Conclusions: Financial navigation programs (FNPs) demonstrate promise in addressing financial burden; however, evidence remains limited and heterogeneous. This burden is disproportionately high in low- and middle-income countries (LMICs), where it is a key driver of treatment abandonment and adverse outcomes. Future research should prioritize standardized frameworks, rigorous evaluation, and expanded implementation in LMIC settings.
Competition Category: Health Services Outcomes or Clinical
Mentor: Juliet Lumati, MD MPH
Jessica Colin Escobar, MD MBA
Senior Resident (Clinical PGY3-5)
Improving belonging in the house of surgery: The need for a nationwide peer network for residents
Introduction: Minoritized general surgery trainees have higher burnout and attrition rates. We surveyed residents prior to enrollment in a pilot national peer mentorship network to assess needs.
Methods: General surgery residents were recruited through affinity-based societies and social media. A survey collected demographics, validated belonging and burnout measures, thoughts of leaving the program, and satisfaction with pursuing surgery.
Results: 83 residents enrolled. 60% were female and 20% LGBTQ+. 40% identified as Asian, 37% Black, and 8% Hispanic/Latino, of whom 30% reported 0-1 residents with a shared identity in their program. 64% reported a guardian/parent with a graduate/professional degree, 22% a college degree, and 14% a high school diploma or less. Median belonging score was 41 (range 15 - 55). 77% were satisfied with pursuing surgery, but 36% were burned out and 24% had thoughts of leaving their program. Among those considering leaving, lack of program support was common (68%). Desired discussion topics were professional development, fellowship, burnout, navigating research, identity-related issues, and family planning.
Conclusions: Residents recruited for participation in a national mentorship network through affinity societies reported a lower belonging scores and higher attrition intentions than have been previously established for general surgery residents nationally. Limited opportunities for identity-concordant connections and lack of support within programs are common drivers, suggesting a potential role for a national peer network. Future work will evaluate the impact of identity-concordant support groups, consisting of 3-4 PGY1-2s and 1-2 senior residents (peer support- and crisis response-trained), prompted monthly using evidence-based discussion questions.
Methods: General surgery residents were recruited through affinity-based societies and social media. A survey collected demographics, validated belonging and burnout measures, thoughts of leaving the program, and satisfaction with pursuing surgery.
Results: 83 residents enrolled. 60% were female and 20% LGBTQ+. 40% identified as Asian, 37% Black, and 8% Hispanic/Latino, of whom 30% reported 0-1 residents with a shared identity in their program. 64% reported a guardian/parent with a graduate/professional degree, 22% a college degree, and 14% a high school diploma or less. Median belonging score was 41 (range 15 - 55). 77% were satisfied with pursuing surgery, but 36% were burned out and 24% had thoughts of leaving their program. Among those considering leaving, lack of program support was common (68%). Desired discussion topics were professional development, fellowship, burnout, navigating research, identity-related issues, and family planning.
Conclusions: Residents recruited for participation in a national mentorship network through affinity societies reported a lower belonging scores and higher attrition intentions than have been previously established for general surgery residents nationally. Limited opportunities for identity-concordant connections and lack of support within programs are common drivers, suggesting a potential role for a national peer network. Future work will evaluate the impact of identity-concordant support groups, consisting of 3-4 PGY1-2s and 1-2 senior residents (peer support- and crisis response-trained), prompted monthly using evidence-based discussion questions.
Competition Category: Education
Mentor: Charity Glass, MD MPP
Teja Devarakonda, MD PhD
Senior Resident (Clinical PGY3-5)
Pulmonary microvasculature is as an effective filter against tumor re-metastasis
Introduction: The pulmonary capillary bed is a critical gateway for systemic metastasis, yet its role as a potential mechanical filter remains underappreciated. We utilized a mouse parabiosis model to definitively test whether Circulating Tumor Cells (CTCs) trapped in the lung can undergo secondary recirculation to seed distant hosts. This model uniquely isolates the recipient from the initial tumor inoculum, ensuring that any metastatic seeding results exclusively from CTCs capable of spontaneous recirculation and capillary traversal.
Methods: To assess metastatic transfer, we established parabiotic pairs using Syngeneic BalbC, C57B6, and immunodeficient SCID mice. Host mice were inoculated with luciferase-tagged Lewis Lung adenocarcinoma (LL2-Luc2) cells via tail vein injection three days prior to surgical union. We monitored metastatic burden weekly using longitudinal luciferase-based imaging (LAGO). Crucially, the establishment of shared physiological circulation was validated via flow cytometry in a separate cohort of C57B6 45.1/45.2 mice to confirm chimerism.
Results: Flow cytometry confirmed the rapid establishment of shared circulation and chimerism between parabiotic partners within just 7 days. Despite this verified blood exchange, LAGO imaging revealed a striking absence of metastatic transfer to the recipient mice across all pairs over the 4 week monitoring period (n = 7 pairs; p < 0.01).
Conclusions: Our data indicates that the pulmonary microcirculation acts as a potent blockade that traps CTCs and prevents secondary recirculation, even after the establishment of shared systemic blood flow. These findings suggest that mechanical arrest in the lung capillaries is a significant limiting factor in the metastatic cascade that requires further investigation regarding how human cancers eventually bypass this physiological filter.
Methods: To assess metastatic transfer, we established parabiotic pairs using Syngeneic BalbC, C57B6, and immunodeficient SCID mice. Host mice were inoculated with luciferase-tagged Lewis Lung adenocarcinoma (LL2-Luc2) cells via tail vein injection three days prior to surgical union. We monitored metastatic burden weekly using longitudinal luciferase-based imaging (LAGO). Crucially, the establishment of shared physiological circulation was validated via flow cytometry in a separate cohort of C57B6 45.1/45.2 mice to confirm chimerism.
Results: Flow cytometry confirmed the rapid establishment of shared circulation and chimerism between parabiotic partners within just 7 days. Despite this verified blood exchange, LAGO imaging revealed a striking absence of metastatic transfer to the recipient mice across all pairs over the 4 week monitoring period (n = 7 pairs; p < 0.01).
Conclusions: Our data indicates that the pulmonary microcirculation acts as a potent blockade that traps CTCs and prevents secondary recirculation, even after the establishment of shared systemic blood flow. These findings suggest that mechanical arrest in the lung capillaries is a significant limiting factor in the metastatic cascade that requires further investigation regarding how human cancers eventually bypass this physiological filter.
Competition Category: Basic Science or Translational
Mentor: Ankit Bharat, MBBS
Joshua DeYoung, BS
Student
A mechanical analysis of suture anchor strength using braided compared to mesh suture
Introduction: Suture pull through and tendon rupture present ongoing challenges after tendon repair, limiting early range of motion and functional rehabilitation after surgery. Previous work has demonstrated increased tissue ingrowth and lower pull through when mesh suture is used in abdominal fascial closure. The purpose of this study was to examine the behavior of mesh compared to braided suture in an ex vivo mechanical model with different suture anchor constructs to determine the performance of each.
Methods: Sutures tested were FiberWire 2-0, Duramesh 1, and Duramesh 2. Anchors tested included the Arthrex 3.9 mm BioComposite SwiveLock, the Mitek 4.75 mm Healix Advance Knotless PEEK (polyether ether ketone), and the Stryker 3.5 mm Citrefix Biocomposite body/PEEK eyelet Anchor systems. Each anchor was tested in a synthetic sawbones model under perpendicular cyclic and load to failure conditions using a single column Instron Universal Testing System.
Results: In the Arthrex anchor system, the mean load at failure for Duramesh 1 was 38.2 N, Duramesh 2 was 93.7 N, and FiberWire 2-0 was 249.9 N. Welch ANOVA demonstrated significant differences in mean load at failure between suture types (p < 0.0001), primarily driven by differences between FiberWire and the mesh suture constructs. In contrast, in the Mitek anchor system Duramesh 1 had a mean load at failure of 92.6 N, Duramesh 2 had a mean load at failure of 129.6 N, and FiberWire 2-0 had a mean load at failure of 100.1 N, with no significant differences detected between groups (p = 0.134). Similarly, the Stryker anchor system did not significantly differ in load at failure between suture types (p = 0.093). Duramesh 1 had a mean load at failure of 131.7 N, Duramesh 2 had a mean load at failure of 166.4 N, and FiberWire 2-0 had a mean load at failure of 179.1 N. Failure mode patterns varied across anchor systems and suture types.
Conclusions: Mesh suture constructs demonstrated biomechanical performance comparable to braided suture constructs in select suture anchor systems while providing load-to-failure values within or above physiologic loading ranges.
Methods: Sutures tested were FiberWire 2-0, Duramesh 1, and Duramesh 2. Anchors tested included the Arthrex 3.9 mm BioComposite SwiveLock, the Mitek 4.75 mm Healix Advance Knotless PEEK (polyether ether ketone), and the Stryker 3.5 mm Citrefix Biocomposite body/PEEK eyelet Anchor systems. Each anchor was tested in a synthetic sawbones model under perpendicular cyclic and load to failure conditions using a single column Instron Universal Testing System.
Results: In the Arthrex anchor system, the mean load at failure for Duramesh 1 was 38.2 N, Duramesh 2 was 93.7 N, and FiberWire 2-0 was 249.9 N. Welch ANOVA demonstrated significant differences in mean load at failure between suture types (p < 0.0001), primarily driven by differences between FiberWire and the mesh suture constructs. In contrast, in the Mitek anchor system Duramesh 1 had a mean load at failure of 92.6 N, Duramesh 2 had a mean load at failure of 129.6 N, and FiberWire 2-0 had a mean load at failure of 100.1 N, with no significant differences detected between groups (p = 0.134). Similarly, the Stryker anchor system did not significantly differ in load at failure between suture types (p = 0.093). Duramesh 1 had a mean load at failure of 131.7 N, Duramesh 2 had a mean load at failure of 166.4 N, and FiberWire 2-0 had a mean load at failure of 179.1 N. Failure mode patterns varied across anchor systems and suture types.
Conclusions: Mesh suture constructs demonstrated biomechanical performance comparable to braided suture constructs in select suture anchor systems while providing load-to-failure values within or above physiologic loading ranges.
Competition Category: Basic Science or Translational
Mentor: Lindsay Janes, MD
Ashley Dodd, MD
Senior Resident (Clinical PGY3-5)
Quantitative perfusion assessment during fluorescence angiography
Introduction: Fluorescence angiography is increasingly used by pediatric surgeons to assess organ perfusion intraoperatively; however, interpretation remains largely subjective. A newly available commercial quantification platform offers objective assessment, but its performance characteristics have not been well described.
Methods: Intraoperatively during intestinal resections and testicular detorsion, a series of indocyanine green (ICG) angiograms were performed using a Stryker® 1788 imaging tower incorporating the Quantitative Perfusion (QP) function. QP-derived perfusion metrics were compared to surgeon judgement as well as full computer-automated segmented quantitative fluorescence analysis.
Results: The QP function demonstrated a measurable decrement in plateau intensity tissue perfusion that was concordant with expert surgeon intraoperative judgment. Quantitative trends were consistent across analyzed regions. Computer-automated analysis confirmed similar findings as well as demonstrated correlation of numerous time-based parameters, including time to maximum intensity and maximum inflow slope.
Conclusions: Objective fluorescence quantification shows promise as an adjunct for pediatric surgeons in assessing tissue viability. While reliance on magnitude-based metrics from currently available commercial platforms has clinical utility, future approaches incorporating dynamic parameters such as the rate of fluorescence ingress are likely to provide additional value.
Methods: Intraoperatively during intestinal resections and testicular detorsion, a series of indocyanine green (ICG) angiograms were performed using a Stryker® 1788 imaging tower incorporating the Quantitative Perfusion (QP) function. QP-derived perfusion metrics were compared to surgeon judgement as well as full computer-automated segmented quantitative fluorescence analysis.
Results: The QP function demonstrated a measurable decrement in plateau intensity tissue perfusion that was concordant with expert surgeon intraoperative judgment. Quantitative trends were consistent across analyzed regions. Computer-automated analysis confirmed similar findings as well as demonstrated correlation of numerous time-based parameters, including time to maximum intensity and maximum inflow slope.
Conclusions: Objective fluorescence quantification shows promise as an adjunct for pediatric surgeons in assessing tissue viability. While reliance on magnitude-based metrics from currently available commercial platforms has clinical utility, future approaches incorporating dynamic parameters such as the rate of fluorescence ingress are likely to provide additional value.
Competition Category: Health Services Outcomes or Clinical
Mentor: Seth Goldstein, MD MPhil
Kacie Ford, BS
Student
Fecal microbiota transfer (FMT) from ApoE4 targeted replacement mice fails to rescue post-TBI cognitive impairment
Introduction: Apolipoprotein E (ApoE) is a lipid-transport protein that plays an important role in lipid metabolism and redistribution. The major human ApoE isoforms are encoded by distinct alleles (2, 3, and 4). Compared with ApoE 2 and 3, ApoE4 increases the risk of cognitive impairment. In fact, ApoE4 is the strongest genetic risk factor for sporadic Alzheimer's disease (AD), whereas the ApoE2 allele has been identified as a protective genetic variant. Additionally, the ApoE variants have genotype-dependent differences in their gut microbiome. Whether the baseline differences in gut microbial community structure between ApoE allelic variants contribute to this difference has yet to be studied. Therefore, we hypothesized that fecal microbiota transplantation (FMT) from ApoE2 donor mice would attenuate neurocognitive decline following traumatic brain injury (TBI), whereas FMT from ApoE4 donors would exacerbate neurocognitive impairment.
Methods: Male and female C57BL/6 (B6) mice (14 weeks old; n = 80) underwent TBI via controlled cortical impact or sham injury. Two hours post-TBI, mice received FMT via oral gavage using donor stool from healthy B6, ApoE2, or ApoE4 targeted replacement mice. FMT was administered three times per week for four weeks. One-month post-injury, mice underwent neurocognitive testing with Open Field testing, Novel Object Location (NOL), and Novel Object Recognition (NOR) to evaluate anxiety, learning, and memory. Behavioral data was analyzed using Limelight software.
Results: At one-month post-TBI, no difference in anxiety levels was noted between groups. NOR, which assesses non-spatial learning involving multiple brain regions, revealed significant differences between TBI and sham injury indicating a significant TBI effect (p ≤ 0.0015). This difference was mitigated in mice receiving FMT from both B6 and ApoE2 donor mice (p ≤.0018 and p ≤0.0001 respectively). However, no protection was seen in TBI mice receiving FMT from ApoE4 donor mice (p ≥ 0.4177). In contrast, no significant effects were detected in the NOL test, which assesses hippocampal-dependent spatial learning.
Conclusions: These findings support our previously published data showing that FMT from healthy donors attenuates neurocognitive deficits after TBI. They also support our hypothesis that the genotype-dependent differences in the ApoE4 gut microbiome are nonprotective. These data suggest an additional mechanism underlying the neurocognitive outcome differences between the ApoE allelic variants. Ongoing studies will further investigate transcriptional changes in the brain to identify molecular correlations underlying these microbiome-driven effects on neurocognitive recovery.
Methods: Male and female C57BL/6 (B6) mice (14 weeks old; n = 80) underwent TBI via controlled cortical impact or sham injury. Two hours post-TBI, mice received FMT via oral gavage using donor stool from healthy B6, ApoE2, or ApoE4 targeted replacement mice. FMT was administered three times per week for four weeks. One-month post-injury, mice underwent neurocognitive testing with Open Field testing, Novel Object Location (NOL), and Novel Object Recognition (NOR) to evaluate anxiety, learning, and memory. Behavioral data was analyzed using Limelight software.
Results: At one-month post-TBI, no difference in anxiety levels was noted between groups. NOR, which assesses non-spatial learning involving multiple brain regions, revealed significant differences between TBI and sham injury indicating a significant TBI effect (p ≤ 0.0015). This difference was mitigated in mice receiving FMT from both B6 and ApoE2 donor mice (p ≤.0018 and p ≤0.0001 respectively). However, no protection was seen in TBI mice receiving FMT from ApoE4 donor mice (p ≥ 0.4177). In contrast, no significant effects were detected in the NOL test, which assesses hippocampal-dependent spatial learning.
Conclusions: These findings support our previously published data showing that FMT from healthy donors attenuates neurocognitive deficits after TBI. They also support our hypothesis that the genotype-dependent differences in the ApoE4 gut microbiome are nonprotective. These data suggest an additional mechanism underlying the neurocognitive outcome differences between the ApoE allelic variants. Ongoing studies will further investigate transcriptional changes in the brain to identify molecular correlations underlying these microbiome-driven effects on neurocognitive recovery.
Competition Category: Basic Science or Translational
Mentor: Steven Schwulst, MD
Meryl Pearl Franco, BS
Student
A volumetric and angular MRI analysis of the thoracic outlet for predicting surgical success
Introduction: Neurogenic thoracic outlet syndrome (nTOS) remains a challenging clinical diagnosis due to the lack of objective, universally accepted radiographic criteria. While surgical decompression through scalenectomy and rib resection is often successful, predicting which patients will achieve significant symptomatic relief remains difficult. This study aims to investigate the utility of specific MRI measurements of thoracic outlet in predicting resolution of nTOS symptoms following decompression surgery.
Methods: A retrospective cohort study was conducted on patients undergoing nTOS decompression surgery between 2017 and 2024. Pre-operative MRI scans were utilized to measure anatomical parameters of the interscalene triangle (IT), including IT volume, IT angle and linear dimensions of anterior scalene (AS) height/width, middle scalene (MS) height, and IT base length. To account for variations in patient habitus, all measurements were normalized using residual analysis against clavicle length, established as a superior skeletal proxy (R2 = 0.10 for volume) compared to T1 vertebral width. Clinical outcomes were categorized into significant improvement, partial improvement or worsened. Binary outcomes for post-operative sensory, motor and pain deficits were also recorded. Statistical analysis included Spearman's correlation for ranked outcomes, McNemar's test for symptom resolution and logistic regression.
Results: The primary finding was a statistically significant correlation between the size-corrected interscalene angle and overall clinical outcome (Spearman's ? = -0.346, p =0.038). A narrower corrected angle was predictive of a “significant improvement” outcome, whereas wider angles were associated with partial or worsened states. Corrected IT volume showed a weak trend toward motor recovery (p = 0.26) but did not reach global significance (p = 0.29). Individual muscle dimensions, such as AS width (p = 0.74) were not predictive of surgical success, suggesting that the geometric relationship (angle) is more clinically relevant than isolated muscle hypertrophy.
Conclusions: Size-normalized MRI analysis provides a robust framework for identifying objective markers of surgical success in nTOS. The interscalene angle, when corrected for patient habitus via clavicle-length residuals, serves as a significant predictor of post-operative improvement. Patients with a more acute/narrow angle relative to their frame appear to be the ideal candidates for surgical decompression. Incorporating normalized angular and volumetric measurements into pre-operative planning can enhance patient selection and provide more accurate prognostic counseling for patients undergoing nTOS decompression.
Methods: A retrospective cohort study was conducted on patients undergoing nTOS decompression surgery between 2017 and 2024. Pre-operative MRI scans were utilized to measure anatomical parameters of the interscalene triangle (IT), including IT volume, IT angle and linear dimensions of anterior scalene (AS) height/width, middle scalene (MS) height, and IT base length. To account for variations in patient habitus, all measurements were normalized using residual analysis against clavicle length, established as a superior skeletal proxy (R2 = 0.10 for volume) compared to T1 vertebral width. Clinical outcomes were categorized into significant improvement, partial improvement or worsened. Binary outcomes for post-operative sensory, motor and pain deficits were also recorded. Statistical analysis included Spearman's correlation for ranked outcomes, McNemar's test for symptom resolution and logistic regression.
Results: The primary finding was a statistically significant correlation between the size-corrected interscalene angle and overall clinical outcome (Spearman's ? = -0.346, p =0.038). A narrower corrected angle was predictive of a “significant improvement” outcome, whereas wider angles were associated with partial or worsened states. Corrected IT volume showed a weak trend toward motor recovery (p = 0.26) but did not reach global significance (p = 0.29). Individual muscle dimensions, such as AS width (p = 0.74) were not predictive of surgical success, suggesting that the geometric relationship (angle) is more clinically relevant than isolated muscle hypertrophy.
Conclusions: Size-normalized MRI analysis provides a robust framework for identifying objective markers of surgical success in nTOS. The interscalene angle, when corrected for patient habitus via clavicle-length residuals, serves as a significant predictor of post-operative improvement. Patients with a more acute/narrow angle relative to their frame appear to be the ideal candidates for surgical decompression. Incorporating normalized angular and volumetric measurements into pre-operative planning can enhance patient selection and provide more accurate prognostic counseling for patients undergoing nTOS decompression.
Competition Category: Health Services Outcomes or Clinical
Mentor: Lindsay Janes, MD
Isabel Gippo, MPH
Student
Impact of direct-to-operating room transfer protocol for ruptured abdominal aortic aneurysm
Introduction: Ruptured abdominal aortic aneurysm (rAAA) carries high mortality, with many patients presenting to facilities unable to provide definitive repair. Multidisciplinary protocols have improved outcomes, but their sustained impact on surgical volume and regionalization remains unclear. The goal of this study is to evaluate the impact of a direct-to-operating room transfer protocol for rAAA on surgical volume, interhospital transfer patterns, and surgical outcomes at a tertiary referral center.
Methods: This retrospective, single-institution cohort study used institutional databases and the Vascular Quality Initiative registry to identify all patients undergoing rAAA repair between 2014 and 2024. In January 2020, a multidisciplinary rapid-response transfer protocol was implemented to facilitate direct-to-operating room transfers. Patients were stratified into pre- and post-implementation cohorts. Differences in case volume, transfer patterns, transfer distance, time to treatment, and surgical outcomes between groups were assessed using univariable analysis, with statistical significance set at p < 0.05.
Results: A total of 96 patients underwent rAAA repair during the study period, with 28.1% treated before and 71.9% treated after implementation of the transfer protocol. Baseline characteristics were similar between cohorts, except for a higher prevalence of diabetes (40.7% vs 14.5%, p=0.005) and former smokers (59.3% vs 34.8%, p=0.029) in the pre-implementation group. The annual rAAA repair volume increased threefold after implementation, from a mean of 4.5 to 13.8 cases per year. Overall, 86.5% of patients were transferred from outside hospitals, with a significant increase in the proportion of external transfers after implementation (70.4% vs 92.8%, p = 0.007). Median transfer distance increased from 23.7 to 49.1 miles (p=0.017) and time from transfer acceptance to operating room arrival decreased from 173 to 124 minutes (p=0.145). Distance-adjusted transfer time decreased significantly from 9.5 to 4.1 minutes per mile (p=0.017). Endovascular repair was performed in 68.8% of patients, with no difference between cohorts. In-hospital stroke (6.2%), myocardial infarction (7.3%), and mortality (22.9%) rates were similar between groups. Median length of stay was 6.6 days. Thirty-day outcomes did not differ between cohorts, with readmission in 13.5%, reintervention in 36.5%, and mortality in 25.0% of patients.
Conclusions: Implementation of a multidisciplinary direct-to-operating room protocol for rAAA was associated with a sustained increase in surgical volume, higher rates of interhospital transfer, expansion of the regional referral network, and lower distance-adjusted transfer times, without compromising perioperative outcomes.
Methods: This retrospective, single-institution cohort study used institutional databases and the Vascular Quality Initiative registry to identify all patients undergoing rAAA repair between 2014 and 2024. In January 2020, a multidisciplinary rapid-response transfer protocol was implemented to facilitate direct-to-operating room transfers. Patients were stratified into pre- and post-implementation cohorts. Differences in case volume, transfer patterns, transfer distance, time to treatment, and surgical outcomes between groups were assessed using univariable analysis, with statistical significance set at p < 0.05.
Results: A total of 96 patients underwent rAAA repair during the study period, with 28.1% treated before and 71.9% treated after implementation of the transfer protocol. Baseline characteristics were similar between cohorts, except for a higher prevalence of diabetes (40.7% vs 14.5%, p=0.005) and former smokers (59.3% vs 34.8%, p=0.029) in the pre-implementation group. The annual rAAA repair volume increased threefold after implementation, from a mean of 4.5 to 13.8 cases per year. Overall, 86.5% of patients were transferred from outside hospitals, with a significant increase in the proportion of external transfers after implementation (70.4% vs 92.8%, p = 0.007). Median transfer distance increased from 23.7 to 49.1 miles (p=0.017) and time from transfer acceptance to operating room arrival decreased from 173 to 124 minutes (p=0.145). Distance-adjusted transfer time decreased significantly from 9.5 to 4.1 minutes per mile (p=0.017). Endovascular repair was performed in 68.8% of patients, with no difference between cohorts. In-hospital stroke (6.2%), myocardial infarction (7.3%), and mortality (22.9%) rates were similar between groups. Median length of stay was 6.6 days. Thirty-day outcomes did not differ between cohorts, with readmission in 13.5%, reintervention in 36.5%, and mortality in 25.0% of patients.
Conclusions: Implementation of a multidisciplinary direct-to-operating room protocol for rAAA was associated with a sustained increase in surgical volume, higher rates of interhospital transfer, expansion of the regional referral network, and lower distance-adjusted transfer times, without compromising perioperative outcomes.
Competition Category: Quality Improvement
Mentor: Neel Mansukhani, MD
Hyebin Han, MS
Student
Traumatic brain injury (TBI) increases gut lymphoid aggregates in aged compared to young mice
Introduction: Traumatic brain injury (TBI) affects nearly 3 million people annually in the United States, with older adults representing a growing proportion of cases. TBI induces a neuroinflammatory response driven by persistent microglial activation. Our prior work shows that aged mice (>18 months) exhibit constitutively activated microglia and increased recruitment of peripheral T cells after TBI, resulting in greater neurocognitive impairment than in young mice (12-14 weeks). We have also demonstrated age- and TBI-associated gut dysbiosis. Although TBI disrupts gut structure and systemic homeostasis in adult animals, the impact of aging on TBI-induced gut changes remains unclear. We hypothesized that the aged gut is more susceptible to structural and immune alterations following TBI than the young gut.
Methods: Young (12-week-old) and aged (84-week-old) C57BL/6 mice underwent severe TBI via open-head controlled cortical impact. Colons were harvested at 4, 24, and 48 hours post-injury (n = 3-4/time point; no data were collected for young mice at the 48-hour timepoint). Colons were paraffin-embedded and sectioned (5 µm; four serial sections per tissue) then stained with hematoxylin and eosin (H&E). Slides were manually quantified then analyzed using GraphPad Prism v10. Histology was evaluated in a blinded manner by a board-certified pathologist.
Results: Unlike previous findings, H&E analysis revealed no gross structural damage, including changes in crypt morphology or muscularis propria in either age group following TBI. However, aged mice exhibited a significant increase in gut lymphoid aggregates at 24 and 48 hours (adjusted p-values of 0.0004 and <0.0001, respectively) post-TBI compared to naïve aged controls. In contrast, young mice showed only a modest increase in lymphoid aggregates at 24 hours post-injury (adj p-value = 0.0032).
Conclusions: Our findings support the hypothesis that the aged gut is more susceptible to TBI-induced alterations. Although overt structural damage was not observed, aged mice exhibited significantly greater lymphocyte infiltration than young mice, consistent with our previous findings that aging is associated with enhanced T cell recruitment after TBI. Together, these data suggest that aging amplifies gut immune activation, which may, in turn, fuel neuroinflammation via the brain-gut axis. Building on our prior work showing that fecal microbiota transplant (FMT) improves neurocognitive outcomes in young mice, ongoing studies will determine whether age-specific FMT can modulate immune responses in the brain and gut after TBI.
Methods: Young (12-week-old) and aged (84-week-old) C57BL/6 mice underwent severe TBI via open-head controlled cortical impact. Colons were harvested at 4, 24, and 48 hours post-injury (n = 3-4/time point; no data were collected for young mice at the 48-hour timepoint). Colons were paraffin-embedded and sectioned (5 µm; four serial sections per tissue) then stained with hematoxylin and eosin (H&E). Slides were manually quantified then analyzed using GraphPad Prism v10. Histology was evaluated in a blinded manner by a board-certified pathologist.
Results: Unlike previous findings, H&E analysis revealed no gross structural damage, including changes in crypt morphology or muscularis propria in either age group following TBI. However, aged mice exhibited a significant increase in gut lymphoid aggregates at 24 and 48 hours (adjusted p-values of 0.0004 and <0.0001, respectively) post-TBI compared to naïve aged controls. In contrast, young mice showed only a modest increase in lymphoid aggregates at 24 hours post-injury (adj p-value = 0.0032).
Conclusions: Our findings support the hypothesis that the aged gut is more susceptible to TBI-induced alterations. Although overt structural damage was not observed, aged mice exhibited significantly greater lymphocyte infiltration than young mice, consistent with our previous findings that aging is associated with enhanced T cell recruitment after TBI. Together, these data suggest that aging amplifies gut immune activation, which may, in turn, fuel neuroinflammation via the brain-gut axis. Building on our prior work showing that fecal microbiota transplant (FMT) improves neurocognitive outcomes in young mice, ongoing studies will determine whether age-specific FMT can modulate immune responses in the brain and gut after TBI.
Competition Category: Basic Science or Translational
Mentor: Steven Schwulst, MD
Maurissa Harris, BA
Student
Streamlining patient access in plastic and reconstructive surgery through a standardized scheduling and referral decision guide
Introduction: Plastic and reconstructive surgery practices encompass a wide range of procedures and expertise, creating complexity during initial patient intake. Non-clinical scheduling staff frequently receive patient inquiries regarding conditions and procedures, leading to uncertainty about service availability and appropriate provider assignment. In the absence of standardized guidance, staff rely on internal messaging and clinical escalation, resulting in scheduling delays, workflow inefficiency, and other system-based barriers to access to care. This study aims to evaluate the impact of newly developed standardized scheduling and referral decision tool on patient access and workflow efficiency in a plastic and reconstructive surgery practice. We hypothesize that implementation of this tool improves scheduling efficiency, reduces reliance on internal escalation and/or informal communication tools and increases staff confidence in intake-related decision-making.
Methods: A comprehensive intake reference guide was developed to outline conditions and procedures commonly sought by patients seeking plastic and reconstructive surgery, identify appropriate internal provider(s) for each, and specify cases requiring external referral. The tool was implemented for use by non-clinical scheduling staff. Evaluation was accomplished via a post-utilization survey to evaluate several domains including staff confidence, scheduling efficiency, and guide usability. The questionnaire incorporates both Likert-scale items and assesses overall acceptability of the guide via a Net Promoter Score (NPS) style recommendation question. Data collection remains ongoing.
Results: Preliminary survey responses suggest improved staff confidence in determining appropriate provider assignments, reduced reliance on informal messaging and internal clinical escalation, and increased efficiency in scheduling decisions. Preliminary survey responses (N = 8) demonstrate positive average pre-post changes across all Likert scale domains, with the largest improvement observed in staff confidence determining scheduling (mean increase +1.7 on a 5-point scale). Preliminary analysis of usability of the guide demonstrates an NPS of 88 indicating a high acceptability and willingness of staff to recommend and continue to use the guide.
Conclusions: Implementation of a standardized scheduling and referral decision guide for non-clinical scheduling staff may improve workflow efficiency and facilitate timely patient access in a plastic and reconstructive practice. This low-cost, scalable intervention highlights the values of structured intake resources in complex surgical practices and may be generalizable across specialties. Final results will further define the impact of this intervention on scheduling workflow and how it improves access to care.
Methods: A comprehensive intake reference guide was developed to outline conditions and procedures commonly sought by patients seeking plastic and reconstructive surgery, identify appropriate internal provider(s) for each, and specify cases requiring external referral. The tool was implemented for use by non-clinical scheduling staff. Evaluation was accomplished via a post-utilization survey to evaluate several domains including staff confidence, scheduling efficiency, and guide usability. The questionnaire incorporates both Likert-scale items and assesses overall acceptability of the guide via a Net Promoter Score (NPS) style recommendation question. Data collection remains ongoing.
Results: Preliminary survey responses suggest improved staff confidence in determining appropriate provider assignments, reduced reliance on informal messaging and internal clinical escalation, and increased efficiency in scheduling decisions. Preliminary survey responses (N = 8) demonstrate positive average pre-post changes across all Likert scale domains, with the largest improvement observed in staff confidence determining scheduling (mean increase +1.7 on a 5-point scale). Preliminary analysis of usability of the guide demonstrates an NPS of 88 indicating a high acceptability and willingness of staff to recommend and continue to use the guide.
Conclusions: Implementation of a standardized scheduling and referral decision guide for non-clinical scheduling staff may improve workflow efficiency and facilitate timely patient access in a plastic and reconstructive practice. This low-cost, scalable intervention highlights the values of structured intake resources in complex surgical practices and may be generalizable across specialties. Final results will further define the impact of this intervention on scheduling workflow and how it improves access to care.
Competition Category: Quality Improvement
Mentor: Sumanas Jordan, MD PhD
Ricky Hill, PhD
Fellow (Clinical or Postdoctoral Researcher)
Financial navigation for cancer care in Nigeria: interim findings from the COST-FIN trial
Introduction: Cancer outcomes in low- and middle-income countries (LMICs) are shaped by financial barriers that delay, interrupt, or prevent care. In Nigeria, most patients pay out-of-pocket (OOP) for cancer treatment, placing households at risk for financial catastrophe (FC), financial distress (FD), and treatment interruption. Financial navigation (FN) helps patients anticipate costs, develop financial plans, and connect to resources. While FN has shown benefit in high-income settings, COST-FIN is, to our knowledge, the first randomized trial in sub-Saharan Africa evaluating FN among patients with breast, colorectal, and prostate cancer.
Methods: COST-FIN is a pragmatic randomized trial conducted within the African Research Group for Oncology network at Lakeshore Cancer Center and Obafemi Awolowo University Teaching Hospitals Complex. Adults with newly diagnosed breast, colorectal, or prostate cancer were randomized 1:1 to structured FN or usual care and followed at baseline, 3, 6, and 12 months. FN included financial literacy assessment, individualized financial planning, resource linkage, and risk-stratified follow-up. FC was primarily defined as OOP cancer spending exceeding 40% of non-subsistence household expenditure; 10% household income and 25% total household expenditure thresholds were assessed in sensitivity analyses. FD was measured using FACIT-COST, where higher scores indicate better financial well-being. As of the January 23, 2026 interim data lock, 159 participants had been randomized; 156 were eligible for analysis.
Results: Using the 40% non-subsistence expenditure definition, FC was highly prevalent at baseline: 62% in the control arm and 56% in the FN arm. At 3 months, FC was lower among FN participants than controls (44% vs. 64%; p = 0.070), with a similar directional difference at 6 months (39% vs. 58%). Using the 25% household expenditure threshold, 3-month FC was significantly lower in the FN arm (29% vs. 51%; p = 0.049). Longitudinal modeling suggested a protective association between FN and FC, though not statistically significant (OR = 0.43; p = 0.11). Predictors of FC included non-breast cancer diagnosis, female sex, lower household income, and lower discretionary expenditure. Overall, 81% of analyzable participants lacked health insurance. FACIT-COST scores improved directionally in the FN arm, increasing from 13 at baseline to 17 at 6 months, compared with 16 to 15 in controls. Following interim review, the DSMB recommended crossover; beginning October 2025, control participants were offered FN.
Conclusions: FC was highly prevalent and remained common during early follow-up. Interim findings suggest structured FN may reduce FC during early cancer treatment, with the strongest signal at 3 months. Ongoing follow-up will assess durability, FD, treatment interruption, and adherence.
Methods: COST-FIN is a pragmatic randomized trial conducted within the African Research Group for Oncology network at Lakeshore Cancer Center and Obafemi Awolowo University Teaching Hospitals Complex. Adults with newly diagnosed breast, colorectal, or prostate cancer were randomized 1:1 to structured FN or usual care and followed at baseline, 3, 6, and 12 months. FN included financial literacy assessment, individualized financial planning, resource linkage, and risk-stratified follow-up. FC was primarily defined as OOP cancer spending exceeding 40% of non-subsistence household expenditure; 10% household income and 25% total household expenditure thresholds were assessed in sensitivity analyses. FD was measured using FACIT-COST, where higher scores indicate better financial well-being. As of the January 23, 2026 interim data lock, 159 participants had been randomized; 156 were eligible for analysis.
Results: Using the 40% non-subsistence expenditure definition, FC was highly prevalent at baseline: 62% in the control arm and 56% in the FN arm. At 3 months, FC was lower among FN participants than controls (44% vs. 64%; p = 0.070), with a similar directional difference at 6 months (39% vs. 58%). Using the 25% household expenditure threshold, 3-month FC was significantly lower in the FN arm (29% vs. 51%; p = 0.049). Longitudinal modeling suggested a protective association between FN and FC, though not statistically significant (OR = 0.43; p = 0.11). Predictors of FC included non-breast cancer diagnosis, female sex, lower household income, and lower discretionary expenditure. Overall, 81% of analyzable participants lacked health insurance. FACIT-COST scores improved directionally in the FN arm, increasing from 13 at baseline to 17 at 6 months, compared with 16 to 15 in controls. Following interim review, the DSMB recommended crossover; beginning October 2025, control participants were offered FN.
Conclusions: FC was highly prevalent and remained common during early follow-up. Interim findings suggest structured FN may reduce FC during early cancer treatment, with the strongest signal at 3 months. Ongoing follow-up will assess durability, FD, treatment interruption, and adherence.
Competition Category: Health Services Outcomes or Clinical
Mentor: Juliet Lumati, MD MPH
Jessie Ho, MD
Fellow (Clinical or Postdoctoral Researcher)
Design and implementation of a time management and task prioritization curriculum
Introduction: Time management has become an essential skill for physicians to improve efficiency and continue to manage the competing draws on physician time. Though time management is imperative to physician success and well-being, there remains a gap in training at all levels. The objective of this study was to develop, implement, and assess a time management and task prioritization workshop (TMTP) for medical students transitioning to residency.
Methods: A TMTP curriculum was administered to all fourth-year medical (M4) students considering procedural or anatomy focused specialties. The learners were comprised of two distinct cohorts: rising M4 students as a part of the “Introduction to Phase 3” (IP3) course and students completing their M4 in a “Capstone” curriculum. A one-hour TMTP was administered to groups of (IP3 n= 57, Capstone n=43). A voluntary pre- and post- curriculum survey was administered. A Likert scale was used to assess confidence and short answers for qualitative responses. Responses were dichotomized and analysis completed using chi-squared test.
Results: Eighty-three students (83% response rate) completed the pre-course survey and 92 students completed the post-course survey (92% response rate). All respondents felt neutral to more confident that task prioritization and time management were important for their success in residency. Most respondents (86%) had not previously participated in a TMTP curriculum. For both time management and task prioritization, there was a significant increase in confidence following completion of the workshop in both groups.
Conclusions: M4 students are not developing confidence in TMTP skill through their traditional clinical rotations. The TMTP curriculum was effectively implemented with promising initial formative feedback. Future work will focus on continued sustainability and evaluating the effects of the workshop following clinical rotations.
Methods: A TMTP curriculum was administered to all fourth-year medical (M4) students considering procedural or anatomy focused specialties. The learners were comprised of two distinct cohorts: rising M4 students as a part of the “Introduction to Phase 3” (IP3) course and students completing their M4 in a “Capstone” curriculum. A one-hour TMTP was administered to groups of (IP3 n= 57, Capstone n=43). A voluntary pre- and post- curriculum survey was administered. A Likert scale was used to assess confidence and short answers for qualitative responses. Responses were dichotomized and analysis completed using chi-squared test.
Results: Eighty-three students (83% response rate) completed the pre-course survey and 92 students completed the post-course survey (92% response rate). All respondents felt neutral to more confident that task prioritization and time management were important for their success in residency. Most respondents (86%) had not previously participated in a TMTP curriculum. For both time management and task prioritization, there was a significant increase in confidence following completion of the workshop in both groups.
Conclusions: M4 students are not developing confidence in TMTP skill through their traditional clinical rotations. The TMTP curriculum was effectively implemented with promising initial formative feedback. Future work will focus on continued sustainability and evaluating the effects of the workshop following clinical rotations.
Competition Category: Education
Mentor: Tadaki Tomita, MD
Alex Horowitz, MD
Student
Feasibility of resilience training and its impact on quality of life in new ostomates and their partners
Introduction: Approximately 100,000 people in the United States undergo colostomy or ileostomy every year. Stoma has been associated with drastic effects on patient's quality of life (QoL) in multiple studies, however, effects on partner's life have never been fully evaluated. Resiliency has been shown to be beneficial in improving perioperative outcomes. The purpose of this study is to evaluate effects of new stoma on partners' QoL as well as benefits of resiliency training on patients and their significant other.
Methods: This is a single institution randomized controlled trial. All new ostomates and their partners were randomly assigned to either standard care (SC) or resilience training group (RT). SC included pre-operative visit with wound ostomy nurse and two inpatient post-operative visits. RT group watched resiliency training videos in addition to standard care. Participants were asked to fill out pre-and post-operative surveys that included Connor-Davidson Resilience Scale (CD-RISK-25), Resilience, Information, Support and Empowerment (RISE-IBD), and the City of Hope Quality of Life questionnaire for a Patient with an Ostomy (COH-QOL-Ostomy).
Results: Overall, 53 individuals completed surveys including 35 patients and 18 significant others including 23 men and 30 women with average age was 54 years (range 21-77 years). In patients, according to CD-RISC-25, overall QoL declined significantly postoperatively by Month 1 (71.2 ± 12.4 to 62.8 ± 15.4), recovering by Month 3 (66.1 ± 16.5). At the subscale level, significant declines were observed in disease coping (14.0 ± 3.0 to 12.2 ± 2.9), adaptability/flexibility (9.1 ± 1.6 to 8.0 ± 2.1), and regulation of emotion and cognition (5.8 ± 1.3 to 4.8 ± 1.9) at Month 1, with return to baseline levels by Month 3. COH-QOL-Ostomy total score remained unchanged; however, psychological well-being subscale declined significantly by Month 1 (6.7 ± 1.4 to 5.9 ± 1.1). Social well-being declined significantly by Month 3 (6.7 ± 2.0 to 5.8 ± 1.3).), while physical well-being improved (5.1 ± 1.5 to 5.8 ± 1.0).
In partners, according to CD-RISC-25, RISE-IBD total scores, overall QoL remained srows. However, the RISE IBD disease acceptance domain declined significantly by Month 1 (5.6 ± 1.6 to 4.2 ± 0.8), and self-reliance scores decreased significantly by Month 3 (8.5 ± 2.3 to 5.9 ± 1.5).
Conclusions: Creation of stoma imposes substantial effect on quality of life on both patients and their significant others. Peri-operative resiliency training may lead to better outcomes and faster adjustments for patients and their partners.
Methods: This is a single institution randomized controlled trial. All new ostomates and their partners were randomly assigned to either standard care (SC) or resilience training group (RT). SC included pre-operative visit with wound ostomy nurse and two inpatient post-operative visits. RT group watched resiliency training videos in addition to standard care. Participants were asked to fill out pre-and post-operative surveys that included Connor-Davidson Resilience Scale (CD-RISK-25), Resilience, Information, Support and Empowerment (RISE-IBD), and the City of Hope Quality of Life questionnaire for a Patient with an Ostomy (COH-QOL-Ostomy).
Results: Overall, 53 individuals completed surveys including 35 patients and 18 significant others including 23 men and 30 women with average age was 54 years (range 21-77 years). In patients, according to CD-RISC-25, overall QoL declined significantly postoperatively by Month 1 (71.2 ± 12.4 to 62.8 ± 15.4), recovering by Month 3 (66.1 ± 16.5). At the subscale level, significant declines were observed in disease coping (14.0 ± 3.0 to 12.2 ± 2.9), adaptability/flexibility (9.1 ± 1.6 to 8.0 ± 2.1), and regulation of emotion and cognition (5.8 ± 1.3 to 4.8 ± 1.9) at Month 1, with return to baseline levels by Month 3. COH-QOL-Ostomy total score remained unchanged; however, psychological well-being subscale declined significantly by Month 1 (6.7 ± 1.4 to 5.9 ± 1.1). Social well-being declined significantly by Month 3 (6.7 ± 2.0 to 5.8 ± 1.3).), while physical well-being improved (5.1 ± 1.5 to 5.8 ± 1.0).
In partners, according to CD-RISC-25, RISE-IBD total scores, overall QoL remained srows. However, the RISE IBD disease acceptance domain declined significantly by Month 1 (5.6 ± 1.6 to 4.2 ± 0.8), and self-reliance scores decreased significantly by Month 3 (8.5 ± 2.3 to 5.9 ± 1.5).
Conclusions: Creation of stoma imposes substantial effect on quality of life on both patients and their significant others. Peri-operative resiliency training may lead to better outcomes and faster adjustments for patients and their partners.
Competition Category: Health Services Outcomes or Clinical
Mentor: Vitaliy Poylin, MD FACS FASCRS
Saad Hussain, MD
Fellow (Clinical or Postdoctoral Researcher)
High mortality, low transplant: real-world trajectories of compensated and decompensated cirrhosis in a multicenter US cohort
Introduction: Cirrhosis carries substantial mortality, and liver transplant is typically reserved for advanced disease. We examined stage-specific progression, mortality and transplant rates in a real-world cirrhosis cohort.
Methods: A retrospective cohort (CAPriCORN) examined EHR data from patients with cirrhosis (2016-2021), across 7 health systems in a major US metropolitan area. Data were linked with transplant (SRTR) and death data (Veritas®). Patients were classified into 5 stages according to D'Amico et al: 1: compensated, 2: compensated w/ portal hypertension, 3a: variceal bleeding, 3b: 1st non-variceal bleed decomp, 4: ≥2 decompensating events. Patients with HCC at any time were excluded. Primary outcomes were 1 year stage progression, mortality and transplant. Secondary outcomes extended these analyses to 2 years. Adjusted competing risk analysis for death/transplant was performed (demographics, etiology, insurance, frailty, CCI). P-value significant <0.05.
Results: Among 17,855 patients with cirrhosis, mean age was 59.5 years, 45.5% female. 48.6% were White, 18.1% Black, 17.3% Hispanic. MASLD (40.5%) and MetALD (17.7%) were the most common. Insurance included Medicare (31.9%), private insurance (30.8%), and Medicaid (20.3%). Mean CCI was 2.4, mean MELD was 13.4. At inclusion, 54.7% of patients were Stage 1, 6.7% Stage 2, 4.1% Stage 3a, 19.8% Stage 3b and 14.6% Stage 4. At 1 yr, percent of patients remaining in the same stage were: Stage 1, 74%; Stage 2, 72%; Stage 3a, 67%; Stage 3b, 56%; and Stage 4, 67%. Mortality/transplant rates increased progressively at 1 year (Stage 1: 6%/1%; Stage 2: 7%/2%; Stage 3a: 9%/0.3%; Stage 3b: 15%/4%; Stage 4: 26%/7%) and 2 years (Stage 1: 9%/1%; Stage 2: 11%/2%; Stage 3a: 13%/1%; Stage 3b: 21%/5%; Stage 4: 30%/7%). On adjusted analysis, mortality risk rose significantly with advancing stage (sHR: Stage 3b 1.96, Stage 4 3.02), while transplant risk increased significantly only at Stages 2, 3b, and 4 (sHR: 1.99, 3.44, 5.47).
Conclusions: The likelihood of patients with ≤1 decompensating events receiving a transplant remains low, despite a 1-year mortality ~6%. More aggressive pursuit of living donor liver transplantation or use of marginal organs will likely save many lives for those patients.
Methods: A retrospective cohort (CAPriCORN) examined EHR data from patients with cirrhosis (2016-2021), across 7 health systems in a major US metropolitan area. Data were linked with transplant (SRTR) and death data (Veritas®). Patients were classified into 5 stages according to D'Amico et al: 1: compensated, 2: compensated w/ portal hypertension, 3a: variceal bleeding, 3b: 1st non-variceal bleed decomp, 4: ≥2 decompensating events. Patients with HCC at any time were excluded. Primary outcomes were 1 year stage progression, mortality and transplant. Secondary outcomes extended these analyses to 2 years. Adjusted competing risk analysis for death/transplant was performed (demographics, etiology, insurance, frailty, CCI). P-value significant <0.05.
Results: Among 17,855 patients with cirrhosis, mean age was 59.5 years, 45.5% female. 48.6% were White, 18.1% Black, 17.3% Hispanic. MASLD (40.5%) and MetALD (17.7%) were the most common. Insurance included Medicare (31.9%), private insurance (30.8%), and Medicaid (20.3%). Mean CCI was 2.4, mean MELD was 13.4. At inclusion, 54.7% of patients were Stage 1, 6.7% Stage 2, 4.1% Stage 3a, 19.8% Stage 3b and 14.6% Stage 4. At 1 yr, percent of patients remaining in the same stage were: Stage 1, 74%; Stage 2, 72%; Stage 3a, 67%; Stage 3b, 56%; and Stage 4, 67%. Mortality/transplant rates increased progressively at 1 year (Stage 1: 6%/1%; Stage 2: 7%/2%; Stage 3a: 9%/0.3%; Stage 3b: 15%/4%; Stage 4: 26%/7%) and 2 years (Stage 1: 9%/1%; Stage 2: 11%/2%; Stage 3a: 13%/1%; Stage 3b: 21%/5%; Stage 4: 30%/7%). On adjusted analysis, mortality risk rose significantly with advancing stage (sHR: Stage 3b 1.96, Stage 4 3.02), while transplant risk increased significantly only at Stages 2, 3b, and 4 (sHR: 1.99, 3.44, 5.47).
Conclusions: The likelihood of patients with ≤1 decompensating events receiving a transplant remains low, despite a 1-year mortality ~6%. More aggressive pursuit of living donor liver transplantation or use of marginal organs will likely save many lives for those patients.
Competition Category: Health Services Outcomes or Clinical
Mentor: Daniela Ladner, MD MPH
Shareni Jeyamogan, PhD
Fellow (Clinical or Postdoctoral Researcher)
Augmented delivery of regulatory T cells for the prevention of transplant rejection by using fucosylation
Introduction: Organ transplantation requires life-long administration of non-specific immunosuppressive agents, all of which have substantial side-effects. Newer, more targeted therapies are needed that will allow for minimization of such agents. Regulatory T cells (Tregs) are important mediators of immune homeostasis, promote transplantation tolerance in animals, and are elevated in tolerant human transplant recipients. Donor-specific Tregs (Ds-Tregs) are more targeted and potent in their effects given they recognize donor-derived MHC:peptide complexes. Fucosylation of Tregs has been shown to be more effective at graft homing and in preventing graft-versus-host disease in animal models and may improve the efficacy of Ds-Tregs when administered as an adoptive cell therapy.
Methods: Ds-Tregs were expanded for 21-28 days using IL-2, TGF-beta and Everolimus, with stimulation from activated donor B cells. These expanded cells were fucosylated using fucosyltransferase and 1mM GDP-fucose for 30 minutes. Phenotypic stability of Ds-Tregs were assessed using flow cytometry, and the functionality of these cells were confirmed using Mixed Lymphocyte Reaction (MLR) assays. For in vivo testing, NSG skin graft models were used. Mice received donor skin at Day 0 and were treated with various doses of fucosylated or non-fucosylated Ds-Tregs between Days 42-49.
Results: Generated Ds-Tregs maintained their phenotype with >90% CD4+CD25+CD127-Foxp3+. MLR assays showed donor-specific suppression of alloimmune responses, with >50% suppression at a 1:250 Treg: Effector ratio. In the NSG model, high-dose fucosylated Ds-Tregs (2M FUC-DsTregs) significantly prolonged skin allograft survival compared to the rejection control. While non-fucosylated and low-dose fucosylated groups showed intermediate protection, the high-dose fucosylated group provided better graft survival.
Conclusions: Ds-Tregs can be expanded and fucosylated with upregulation of cutaneous lymphocyte antigen (CLA) while maintaining adequate viability (>90%) for adoptive cell transfer experiments. Fucosylation enhances the ability of Ds-Tregs to prolong allograft survival by improving “homing” to the targeted area.
Methods: Ds-Tregs were expanded for 21-28 days using IL-2, TGF-beta and Everolimus, with stimulation from activated donor B cells. These expanded cells were fucosylated using fucosyltransferase and 1mM GDP-fucose for 30 minutes. Phenotypic stability of Ds-Tregs were assessed using flow cytometry, and the functionality of these cells were confirmed using Mixed Lymphocyte Reaction (MLR) assays. For in vivo testing, NSG skin graft models were used. Mice received donor skin at Day 0 and were treated with various doses of fucosylated or non-fucosylated Ds-Tregs between Days 42-49.
Results: Generated Ds-Tregs maintained their phenotype with >90% CD4+CD25+CD127-Foxp3+. MLR assays showed donor-specific suppression of alloimmune responses, with >50% suppression at a 1:250 Treg: Effector ratio. In the NSG model, high-dose fucosylated Ds-Tregs (2M FUC-DsTregs) significantly prolonged skin allograft survival compared to the rejection control. While non-fucosylated and low-dose fucosylated groups showed intermediate protection, the high-dose fucosylated group provided better graft survival.
Conclusions: Ds-Tregs can be expanded and fucosylated with upregulation of cutaneous lymphocyte antigen (CLA) while maintaining adequate viability (>90%) for adoptive cell transfer experiments. Fucosylation enhances the ability of Ds-Tregs to prolong allograft survival by improving “homing” to the targeted area.
Competition Category: Basic Science or Translational
Mentor: James Mathew, PhD
Kevin Jin, BS
Student
Risk factors for recurrent tricuspid regurgitation after pulmonary thromboendarterectomy for chronic thromboembolic pulmonary hypertension
Introduction: Concomitant tricuspid valve repair during pulmonary thromboendarterectomy (PTE) is recommended for severe tricuspid regurgitation (TR) or annular diameter exceeding 4.0 cm, criteria extrapolated from left-sided valve surgery. Whether these thresholds appropriately identify patients at risk for late TR recurrence in chronic thromboembolic pulmonary hypertension (CTEPH) remains unclear. We characterized long-term TR trajectories after PTE, identified independent predictors of recurrent moderate or greater TR, and assessed the impact of recurrent TR on mortality.
Methods: Adults undergoing PTE for CTEPH at Northwestern Memorial Hospital between May 2016 and September 2025 were retrospectively studied. Echocardiograms were analyzed at baseline, pre-discharge, and long-term follow-up (≥6 months post-discharge). Right atrial reservoir strain (RASr) and right ventricular strain were quantified by speckle-tracking echocardiography. Recurrent ≥moderate TR was analyzed using competing risk methods with death as a competing event. All-cause mortality was assessed using Cox proportional hazards models with TR severity as a time-varying covariate. Models were adjusted for age and sex.
Results: Among 170 patients, 36 (21%) developed recurrent ≥moderate TR at long-term follow-up (median 2.8 years, IQR 1.6-4.6). Although TR improved at pre-discharge, late recurrence was common, particularly among patients with preoperative moderate or severe TR. In multivariable analysis, independent predictors of recurrence included preoperative moderate-severe TR (aHR 2.22, p=.027), lower preoperative RASr (aHR 1.03 per 1% decrease, p=.017), pre-discharge TR severity (mild: aHR 3.05, p=.004; moderate-severe: aHR 5.48, p<.001), and higher pre-discharge pulmonary vascular resistance (aHR 1.01, p=.007). Critically, neither absolute nor indexed tricuspid annular diameter predicted recurrence. Severe recurrent TR trended toward higher mortality risk in time-varying analyses (aHR 13.5, p=.055), though limited events (n=8) precluded definitive conclusions. Three-year cumulative incidence of recurrence rose stepwise from 20% in patients with none/trivial/mild preoperative TR to 50% in those with severe preoperative TR.
Conclusions: Recurrent TR after PTE is driven by preoperative TR severity, residual post-discharge TR, persistent pulmonary vascular burden, and impaired right atrial functional reserve, not annular diameter. These findings challenge the current TAD-based threshold for concomitant repair and support routine tricuspid valve repair for severe preoperative TR regardless of annulus size. RASr may further refine repair decisions in moderate TR and guide postoperative surveillance.
Methods: Adults undergoing PTE for CTEPH at Northwestern Memorial Hospital between May 2016 and September 2025 were retrospectively studied. Echocardiograms were analyzed at baseline, pre-discharge, and long-term follow-up (≥6 months post-discharge). Right atrial reservoir strain (RASr) and right ventricular strain were quantified by speckle-tracking echocardiography. Recurrent ≥moderate TR was analyzed using competing risk methods with death as a competing event. All-cause mortality was assessed using Cox proportional hazards models with TR severity as a time-varying covariate. Models were adjusted for age and sex.
Results: Among 170 patients, 36 (21%) developed recurrent ≥moderate TR at long-term follow-up (median 2.8 years, IQR 1.6-4.6). Although TR improved at pre-discharge, late recurrence was common, particularly among patients with preoperative moderate or severe TR. In multivariable analysis, independent predictors of recurrence included preoperative moderate-severe TR (aHR 2.22, p=.027), lower preoperative RASr (aHR 1.03 per 1% decrease, p=.017), pre-discharge TR severity (mild: aHR 3.05, p=.004; moderate-severe: aHR 5.48, p<.001), and higher pre-discharge pulmonary vascular resistance (aHR 1.01, p=.007). Critically, neither absolute nor indexed tricuspid annular diameter predicted recurrence. Severe recurrent TR trended toward higher mortality risk in time-varying analyses (aHR 13.5, p=.055), though limited events (n=8) precluded definitive conclusions. Three-year cumulative incidence of recurrence rose stepwise from 20% in patients with none/trivial/mild preoperative TR to 50% in those with severe preoperative TR.
Conclusions: Recurrent TR after PTE is driven by preoperative TR severity, residual post-discharge TR, persistent pulmonary vascular burden, and impaired right atrial functional reserve, not annular diameter. These findings challenge the current TAD-based threshold for concomitant repair and support routine tricuspid valve repair for severe preoperative TR regardless of annulus size. RASr may further refine repair decisions in moderate TR and guide postoperative surveillance.
Competition Category: Health Services Outcomes or Clinical
Mentor: Stephen Chiu, MD
William Kacey, BS
Student
Cut the crap: mechanical bowel preparation does not improve engraftment of fecal microbiota transplant
Introduction: The gut microbiome heavily influences host physiology, modulating the immune system and inflammation both locally and systemically. Restoration of a healthy gut microbiome through fecal microbiota transplantation (FMT) is an effective therapy for chronic Clostridioides difficile colitis and has demonstrated therapeutic potential in neuroinflammatory conditions such as stroke and traumatic brain injury. Optimizing donor microbiota engraftment after FMT requires consideration of various factors including recipient gut preparation. Mechanical bowel preparation has a cleansing effect on the microbiome, but it is unsettled whether this improves the efficiency of FMT. We hypothesized that mechanical bowel preparation of the recipient with polyethylene glycol (PEG) prior to FMT would improve engraftment of transplanted gut microbiota.
Methods: C57BL/6 mice (12-14 week) underwent mechanical bowel preparation with PEG (425 g/L) or phosphate-buffered saline (PBS) via four 200 µL oral gavages administered at 20-minute intervals. Four hours after bowel preparation, mice underwent FMT via 200 µL oral gavage with resuspended stool (100 g/L) from human patients with an active inflammatory bowel disease (IBD) flare, healthy donor patients (HD), or PBS vehicle control. FMT was repeated on day 2 and day 4. Baseline stool was collected from each mouse one day prior to PEG treatment, and final stool was collected on day 7. 16S ribosomal RNA sequencing (16S rRNAseq) of stool samples was performed to characterize the gut microbiome before and after engraftment.
Results: PEG did not improve engraftment of transplanted gut microbiota. At the genus level, 7 days post FMT, the gut microbial community structure of FMT recipient mice did not resemble the composition of the donor stool. While FMT significantly altered the gut microbiota of IBD recipient mice regardless of pretreatment with PEG (p < 0.05) or saline (p < 0.05), FMT did not significantly alter the gut microbiota of HD recipient mice regardless of pretreatment with PEG (p = 0.69) or saline (p = 0.35). PEG and non-PEG post-engraftment microbiota were not significantly different in either IBD (p = 0.81) or HD (p = 0.61) recipients.
Conclusions: We hypothesized that PEG pretreatment prior to FMT would improve the engraftment of the donor microbiota. Our data demonstrates no superiority in microbiota transfer with PEG bowel preparation when compared to superimposition of donor microbiota onto untreated native recipient microbiota. The difference we found in engraftment between IBD and HD donor microbiota suggests a role for nonmicrobial structural components of stool and inflammatory proteins in mucosal colonization. Understanding factors that augment or inhibit donor microbiota colonization are critical to better understanding the gut microbiome, improving animal models of human gut dysbiosis, and refining FMT-based therapeutic approaches.
Methods: C57BL/6 mice (12-14 week) underwent mechanical bowel preparation with PEG (425 g/L) or phosphate-buffered saline (PBS) via four 200 µL oral gavages administered at 20-minute intervals. Four hours after bowel preparation, mice underwent FMT via 200 µL oral gavage with resuspended stool (100 g/L) from human patients with an active inflammatory bowel disease (IBD) flare, healthy donor patients (HD), or PBS vehicle control. FMT was repeated on day 2 and day 4. Baseline stool was collected from each mouse one day prior to PEG treatment, and final stool was collected on day 7. 16S ribosomal RNA sequencing (16S rRNAseq) of stool samples was performed to characterize the gut microbiome before and after engraftment.
Results: PEG did not improve engraftment of transplanted gut microbiota. At the genus level, 7 days post FMT, the gut microbial community structure of FMT recipient mice did not resemble the composition of the donor stool. While FMT significantly altered the gut microbiota of IBD recipient mice regardless of pretreatment with PEG (p < 0.05) or saline (p < 0.05), FMT did not significantly alter the gut microbiota of HD recipient mice regardless of pretreatment with PEG (p = 0.69) or saline (p = 0.35). PEG and non-PEG post-engraftment microbiota were not significantly different in either IBD (p = 0.81) or HD (p = 0.61) recipients.
Conclusions: We hypothesized that PEG pretreatment prior to FMT would improve the engraftment of the donor microbiota. Our data demonstrates no superiority in microbiota transfer with PEG bowel preparation when compared to superimposition of donor microbiota onto untreated native recipient microbiota. The difference we found in engraftment between IBD and HD donor microbiota suggests a role for nonmicrobial structural components of stool and inflammatory proteins in mucosal colonization. Understanding factors that augment or inhibit donor microbiota colonization are critical to better understanding the gut microbiome, improving animal models of human gut dysbiosis, and refining FMT-based therapeutic approaches.
Competition Category: Basic Science or Translational
Mentor: Steven Schwulst, MD
Ihor Kohut, MD PhD
Junior Resident (Clinical PGY1-2)
Feasibility and clinical yield of in-clinic germline testing in a colorectal surgery practice
Introduction: Most colorectal cancer (CRC) cases are sporadic. Hereditary cancer syndrome (HCS) risk has historically been quoted at 3-5% in patients >50 years and 10-20% in patients <50 years, with guidelines limiting testing by phenotype and family history. When all solid-tumor patients are tested, HCS incidence rises to 10-30%, and restrictive criteria miss 7.8% of CRC cases, including Lynch syndrome in older adults. Standard referral to genetic counseling causes treatment delays and reduces uptake. Multigene panel testing (MGPT) offers expedited turnaround that may inform immediate care.
The purpose of this study is to compare point of care MGPT timeliness and compliance when compared to standard genetic counseling referrals.
Methods: This is a retrospective review of MGPT for all patients with a new diagnosis of CRC presenting to the colorectal surgery clinic through the rapid testing service with Myriad Genetics. While at the office the patient would undergo pretest education and testing consent with a Myriad genetic counselor over the phone after meeting with the surgeon. Blood would be drawn in the lab while patients are getting their tumor markers and baseline labs drawn. Testing results were communicated to the surgeon and the genetic counselor, who then notified the patient.
Results: We had a total of 166 patients present with a new diagnosis of CRC between December 2024 and December 2025. Of those, 23 patients had genetic testing done prior to the visit, 9 patients declined testing, 84 patients underwent Myriad point-of-care testing (study group), while the rest 50 patients were managed according to the traditional pathway for referral to genetic counseling if they met criteria (control group). Pathogenic or likely pathogenic variants were identified in 12 patients (positive rate 14.1%). In the control group, 14 (21%) were referred to genetic counseling, and 36 did not undergo testing.
Average return of results in the study group was 9.4 calendar days compared to 75 days in the control group. Average additional clinic wait time for pretest education was 14.5 minutes in the study group. All patients who underwent point of care testing had their tests covered by insurance except in one case which was covered by the study.
Conclusions: Point of care MGPT is feasible and reproducible in all patients with newly diagnosed CRC. It allowed for fast result turnaround time, higher compliance without interrupting normal specialty clinic workflow.
The purpose of this study is to compare point of care MGPT timeliness and compliance when compared to standard genetic counseling referrals.
Methods: This is a retrospective review of MGPT for all patients with a new diagnosis of CRC presenting to the colorectal surgery clinic through the rapid testing service with Myriad Genetics. While at the office the patient would undergo pretest education and testing consent with a Myriad genetic counselor over the phone after meeting with the surgeon. Blood would be drawn in the lab while patients are getting their tumor markers and baseline labs drawn. Testing results were communicated to the surgeon and the genetic counselor, who then notified the patient.
Results: We had a total of 166 patients present with a new diagnosis of CRC between December 2024 and December 2025. Of those, 23 patients had genetic testing done prior to the visit, 9 patients declined testing, 84 patients underwent Myriad point-of-care testing (study group), while the rest 50 patients were managed according to the traditional pathway for referral to genetic counseling if they met criteria (control group). Pathogenic or likely pathogenic variants were identified in 12 patients (positive rate 14.1%). In the control group, 14 (21%) were referred to genetic counseling, and 36 did not undergo testing.
Average return of results in the study group was 9.4 calendar days compared to 75 days in the control group. Average additional clinic wait time for pretest education was 14.5 minutes in the study group. All patients who underwent point of care testing had their tests covered by insurance except in one case which was covered by the study.
Conclusions: Point of care MGPT is feasible and reproducible in all patients with newly diagnosed CRC. It allowed for fast result turnaround time, higher compliance without interrupting normal specialty clinic workflow.
Competition Category: Health Services Outcomes or Clinical
Mentor: Vitaliy Poylin, MD FACS FASCRS
Junho Lee, MD PhD
Fellow (Clinical or Postdoctoral Researcher)
Oral antibiotics alone versus mechanical bowel preparation plus oral antibiotics in elective right colectomy: A NSQIP analysis
Introduction: Mechanical bowel preparation plus oral antibiotics (MBP+OA) has been the standard for colorectal resection. However, whether MBP provides additional benefit beyond OA alone remains controversial, particularly in right colectomy where the necessity of bowel cleansing is less clear. This study aimed to compare surgical site infections (SSI) rates between OA alone and MBP+OA in patients undergoing elective right colectomy using the ACS-NSQIP database.
Methods: We conducted a retrospective cohort study using the ACS-NSQIP and Colectomy targeted database (2019 - 2023). Patients were categorized as OA alone or MBP+OA. The primary outcomes were 30-day SSI and anastomotic leak. Inverse probability of treatment weighting (IPTW) based on a propensity score including key preoperative covariates was applied to adjust for baseline differences between groups.
Results: A total of 26,462 patients undergoing elective right colectomy were included (OA alone; n = 2,648; MBP+OA: n = 23,814). Baseline characteristics differed between groups and were addressed using IPTW. After IPTW, baseline characteristics were well balanced between groups. SSI occurred in 172 patients (6.5%) receiving OA alone versus 1089 patients (4.6%) receiving MBP+OA. Anastomosis site leakage occurred in 55 patients (2.1%) receiving OA alone versus 390 patients (1.6%) receiving MBP+OA. In IPTW-weighted logistic regression analyses, the odds of SSI were significantly lower in the MBP+OA group (OR 0.77, 95% CI 0.65-0.91; p=0.002), whereas the odds of anastomotic leak were not significantly different (OR 0.86, 95% CI 0.64-1.15; p=0.309).
Conclusions: MBP+OA was associated with a lower risk of SSI compared with OA alone, without a difference in anastomotic leak. These findings suggest that the addition of MBP may provide benefit in SSI prevention in right colectomy.
Methods: We conducted a retrospective cohort study using the ACS-NSQIP and Colectomy targeted database (2019 - 2023). Patients were categorized as OA alone or MBP+OA. The primary outcomes were 30-day SSI and anastomotic leak. Inverse probability of treatment weighting (IPTW) based on a propensity score including key preoperative covariates was applied to adjust for baseline differences between groups.
Results: A total of 26,462 patients undergoing elective right colectomy were included (OA alone; n = 2,648; MBP+OA: n = 23,814). Baseline characteristics differed between groups and were addressed using IPTW. After IPTW, baseline characteristics were well balanced between groups. SSI occurred in 172 patients (6.5%) receiving OA alone versus 1089 patients (4.6%) receiving MBP+OA. Anastomosis site leakage occurred in 55 patients (2.1%) receiving OA alone versus 390 patients (1.6%) receiving MBP+OA. In IPTW-weighted logistic regression analyses, the odds of SSI were significantly lower in the MBP+OA group (OR 0.77, 95% CI 0.65-0.91; p=0.002), whereas the odds of anastomotic leak were not significantly different (OR 0.86, 95% CI 0.64-1.15; p=0.309).
Conclusions: MBP+OA was associated with a lower risk of SSI compared with OA alone, without a difference in anastomotic leak. These findings suggest that the addition of MBP may provide benefit in SSI prevention in right colectomy.
Competition Category: Quality Improvement
Mentor: Vitaliy Poylin, MD FACS FASCRS
Sonia Lele, MD
Junior Resident (Clinical PGY1-2)
Aberrant venous anatomy in anterior lumbar spinal exposures
Introduction: Vascular surgeons are often called upon to provide safe access to the anterior lumbar spine for orthopedic and neurosurgeons. The most common levels that require exposure are the 4th and 5th lumbar and the proximal sacrum. While there are many possible complications of anterior lumbar spinal exposures, one of the more feared and potential devastating is venous injury leading to hemorrhage. A good understanding of typical venous anatomy as well as possible variations is critical for a safe surgery.
Methods: Retrospective chart review was conducted of spinal surgeries performed via anterior approach for which the vascular surgery service assisted with exposure and closure at a single academic institution. Patients with aberrant venous anatomy present on pre-operative imaging were identified and operative reports were reviewed to determine alternative approaches to exposure and retraction of venous vasculature.
Results: Seven total patients with aberrant venous anatomy were identified on anterior exposures for spinal surgeries between November 2024 and February 2026. Venous abnormalities were grouped into four categories: internal iliac vein originating off the contralateral common iliac vein, internal iliac originating directly off the inferior vena cava, bilateral internal iliac veins originating off a common trunk, and an aberrant iliolumbar vein originating off the common iliac vein. Adjustments to operative approach included isolation and retraction of aberrant iliac veins, ligation of draining branches, and ligation of aberrant internal iliac veins when necessary to achieve exposure of the required disc spaces. All exposures were completed without injury to the major venous vasculature.
Conclusions: Aberrant venous anatomy should be routinely identified on pre-operative imaging prior to anterior spinal exposures to prevent accidental venous injury. Further studies are necessary to determine the overall incidence of aberrant venous anatomy in the general population.
Methods: Retrospective chart review was conducted of spinal surgeries performed via anterior approach for which the vascular surgery service assisted with exposure and closure at a single academic institution. Patients with aberrant venous anatomy present on pre-operative imaging were identified and operative reports were reviewed to determine alternative approaches to exposure and retraction of venous vasculature.
Results: Seven total patients with aberrant venous anatomy were identified on anterior exposures for spinal surgeries between November 2024 and February 2026. Venous abnormalities were grouped into four categories: internal iliac vein originating off the contralateral common iliac vein, internal iliac originating directly off the inferior vena cava, bilateral internal iliac veins originating off a common trunk, and an aberrant iliolumbar vein originating off the common iliac vein. Adjustments to operative approach included isolation and retraction of aberrant iliac veins, ligation of draining branches, and ligation of aberrant internal iliac veins when necessary to achieve exposure of the required disc spaces. All exposures were completed without injury to the major venous vasculature.
Conclusions: Aberrant venous anatomy should be routinely identified on pre-operative imaging prior to anterior spinal exposures to prevent accidental venous injury. Further studies are necessary to determine the overall incidence of aberrant venous anatomy in the general population.
Competition Category: Health Services Outcomes or Clinical
Mentor: Heron Rodriguez, MD
Daniel Liesman, MD
Senior Resident (Clinical PGY3-5)
A multiparameter probe for premature infant core physiologic monitoring (PreMo)
Introduction: Conventional premature infant monitoring devices fail in hemodynamic extremis. They largely rely on peripheral perfusion which does not adequately capture the unique transitional physiology of the neonate. To bridge this gap, we designed a multiparameter monitoring probe that can be affixed to a standard nasogastric tube and then inserted in the rectum or the esophagus. The probe, entitled PreMo, is fitted with an IR and red light spectrometer and thermistor to provide continuous measurements of heart rate, oxygen saturation, blood pressure, and temperature.
Methods: Preterm lambs were delivered in an ex-utero intra-partum procedure at 125 days' gestation (term=145). They were intubated and an umbilical artery catheter was placed. PreMo probes were inserted into the rectum and esophagus. Hypoxia from ventilator cessation with subsequent bradycardia was used to assess PreMo's ability to track heart rate and oxygen saturation. Vasopressor administration and controlled bleeds were used for the blood pressure monitoring capabilities. Hypothermia was induced with cooled saline to assess the thermistor. Gold standards included echocardiography, arterial blood gases, umbilical arterial line pressures, and an esophageal temperature probe. A conventional peripheral pulse oximeter was also affixed to the ear or tongue for an additional comparison.
Results: PreMo demonstrated excellent agreement with echocardiography for heart rate assessment achieving a Lin's concordance correlation coefficient (CCC) of 0.99 with a mean bias of -0.51 BPM and 95% limit-of-agreement (LOA) of -11.8-10.8 BPM. Oxygen saturation showed the greatest variability overall (CCC 0.74, mean bias 0.39%, LOA -45.2-41.1%). Blood pressure also showed strong concordance (CCC 0.96, mean bias 0.71 mm Hg, LOA -7.9-8.5 mm Hg). Lastly, temperature tracked consistently with the reference probe although with a negative bias (CCC 0.75, mean bias -0.3 °C, LOA -2.0-1.4 °C). PreMo demonstrated superior performance in comparison to the standard peripheral pulse oximeter during periods of hemodynamic instability correctly identifying severe hypoxia (SaO2 < 70% on blood gas analysis) in 80% of measured events in comparison to 7.7% of measured events for the standard monitor.
Conclusions: PreMo provides accurate and continuous monitoring of heart rate, oxygen saturation, temperature, and blood pressure. The probe demonstrated agreement with gold standards and maintained signal fidelity during hemodynamic derangement, suggesting its potential to reduce the need for invasive monitoring. Ultimately, PreMo addresses a critical requirement in NICU care: continuous, real-time assessment of core rather than peripheral perfusion, offering the potential for a more sensitive indicator of early hemodynamic instability in vulnerable preterm infants.
Methods: Preterm lambs were delivered in an ex-utero intra-partum procedure at 125 days' gestation (term=145). They were intubated and an umbilical artery catheter was placed. PreMo probes were inserted into the rectum and esophagus. Hypoxia from ventilator cessation with subsequent bradycardia was used to assess PreMo's ability to track heart rate and oxygen saturation. Vasopressor administration and controlled bleeds were used for the blood pressure monitoring capabilities. Hypothermia was induced with cooled saline to assess the thermistor. Gold standards included echocardiography, arterial blood gases, umbilical arterial line pressures, and an esophageal temperature probe. A conventional peripheral pulse oximeter was also affixed to the ear or tongue for an additional comparison.
Results: PreMo demonstrated excellent agreement with echocardiography for heart rate assessment achieving a Lin's concordance correlation coefficient (CCC) of 0.99 with a mean bias of -0.51 BPM and 95% limit-of-agreement (LOA) of -11.8-10.8 BPM. Oxygen saturation showed the greatest variability overall (CCC 0.74, mean bias 0.39%, LOA -45.2-41.1%). Blood pressure also showed strong concordance (CCC 0.96, mean bias 0.71 mm Hg, LOA -7.9-8.5 mm Hg). Lastly, temperature tracked consistently with the reference probe although with a negative bias (CCC 0.75, mean bias -0.3 °C, LOA -2.0-1.4 °C). PreMo demonstrated superior performance in comparison to the standard peripheral pulse oximeter during periods of hemodynamic instability correctly identifying severe hypoxia (SaO2 < 70% on blood gas analysis) in 80% of measured events in comparison to 7.7% of measured events for the standard monitor.
Conclusions: PreMo provides accurate and continuous monitoring of heart rate, oxygen saturation, temperature, and blood pressure. The probe demonstrated agreement with gold standards and maintained signal fidelity during hemodynamic derangement, suggesting its potential to reduce the need for invasive monitoring. Ultimately, PreMo addresses a critical requirement in NICU care: continuous, real-time assessment of core rather than peripheral perfusion, offering the potential for a more sensitive indicator of early hemodynamic instability in vulnerable preterm infants.
Competition Category: Basic Science or Translational
Mentor: Aimen Shaaban, MD
Lara Lopes, MD MS
Senior Resident (Clinical PGY3-5)
Left renal vein division during open abdominal aortic aneurysm repair
Introduction: Proximal aortic exposure is a critical step in open abdominal aortic aneurysm (AAA) repair, and left renal vein (LRV) division is commonly used to facilitate access. Although widely performed, concerns persist regarding its potential impact on renal function and perioperative outcomes. Existing literature reports conflicting findings, and the overall safety of this maneuver remains uncertain. The purpose of this study is to evaluate the impact of LRV division in patients undergoing open AAA repair.
Methods: This is a systematic review and meta-analysis registered in PROSPERO register of systematic reviews (CRD42025640222) and conducted in accordance to Cochrane's guidelines for systematic review and meta-analysis. PubMed, EMBASE, and Cochrane databases were systematically searched for studies comparing outcomes of patients who underwent open AAA repair with and without LRV division. Two authors screened the search results and collected data of interest independently, according to the PRIMA protocol. The primary outcome was 30-day mortality and secondary outcomes were short and long-term renal function. Risk ratio (RR) and mean difference (MD) with corresponding 95% confidence interval (CI) were estimated using a random-effects model. Significance was defined as a p-value <0.05.
Results: A total of 190 studies were screened for inclusion, of which nine studies (8 cohort studies and 1 case-control study) met the inclusion criteria. These studies included a total of 1,324 patients, 350 of whom underwent LRV division, while 974 patients did not. Meta-analysis revealed no significant difference in 30-day mortality (28/205 versus 123/512, RR= 0.90, 95% CI 0.63-1.29; p=0.42, I2=0%), AKI (RR=1.74, 95% CI 0.39-7.83; p=0.33, I2=69%), need for dialysis (4/178 versus 10/473, RR=1.61, 95% CI 0.54-4.84), discharge eGFR (MD=-0.32, 95% CI -2.57-1.92, p=0.60, I2=0%), and discharge serum creatinine (sCr) (MD=-0.01, 95% CI -0.02-0.01, p = 0.29, I2 = 0%). LRV division was associated with an increase postoperative sCr (MD=0.08, 95% CI 0.04-0.11, p=0.009, I2=0%) compared to patients with the LRV left intact.
Conclusions: Among patients undergoing AAA open surgical repair, our results demonstrate that LRV division is not associated with an increase in 30-day mortality or worse renal function.
Methods: This is a systematic review and meta-analysis registered in PROSPERO register of systematic reviews (CRD42025640222) and conducted in accordance to Cochrane's guidelines for systematic review and meta-analysis. PubMed, EMBASE, and Cochrane databases were systematically searched for studies comparing outcomes of patients who underwent open AAA repair with and without LRV division. Two authors screened the search results and collected data of interest independently, according to the PRIMA protocol. The primary outcome was 30-day mortality and secondary outcomes were short and long-term renal function. Risk ratio (RR) and mean difference (MD) with corresponding 95% confidence interval (CI) were estimated using a random-effects model. Significance was defined as a p-value <0.05.
Results: A total of 190 studies were screened for inclusion, of which nine studies (8 cohort studies and 1 case-control study) met the inclusion criteria. These studies included a total of 1,324 patients, 350 of whom underwent LRV division, while 974 patients did not. Meta-analysis revealed no significant difference in 30-day mortality (28/205 versus 123/512, RR= 0.90, 95% CI 0.63-1.29; p=0.42, I2=0%), AKI (RR=1.74, 95% CI 0.39-7.83; p=0.33, I2=69%), need for dialysis (4/178 versus 10/473, RR=1.61, 95% CI 0.54-4.84), discharge eGFR (MD=-0.32, 95% CI -2.57-1.92, p=0.60, I2=0%), and discharge serum creatinine (sCr) (MD=-0.01, 95% CI -0.02-0.01, p = 0.29, I2 = 0%). LRV division was associated with an increase postoperative sCr (MD=0.08, 95% CI 0.04-0.11, p=0.009, I2=0%) compared to patients with the LRV left intact.
Conclusions: Among patients undergoing AAA open surgical repair, our results demonstrate that LRV division is not associated with an increase in 30-day mortality or worse renal function.
Competition Category: Health Services Outcomes or Clinical
Mentor: Ashley Vavra, MD MS
Kirsten Lung, MD
Fellow (Clinical or Postdoctoral Researcher)
ADH upstage rates in the era of contemporary imaging and biopsy techniques
Introduction: Surgical excision is considered standard of care for atypical ductal hyperplasia (ADH) given the reported 15-30% upstage rate in the literature to ductal carcinoma in situ or invasive breast cancer. With advancements in breast imaging, including digital breast tomosynthesis (DBT) and vacuum-assisted core needle biopsy (CNB), the recommendation for surveillance of benign, high-risk lesions has gained traction. We aimed to evaluate whether ADH upstage rate lowers with improved radiographic and sampling techniques and to identify a subset within this cohort with biopsy-proven ADH for whom surgical excision may be omitted based on clinical, mammographic, and procedural findings.
Methods: Retrospective cohort study using institutional data from the Enterprise Data Warehouse across Northwestern Medicine Health System. Women > 18 years old who underwent screening DBT between 2016-2024 followed by surgical excision for ADH diagnosed via stereotactic or ultrasound-guided CNB were included. Patients with ADH detected on MRI or sampled via MRI-guided biopsy, prior/concurrent history of breast cancer or known pathogenic mutation were excluded. Statistical analyses were completed with chi-square testing using p-value < 0.05 as statistical significance.
Results: 677 women with median age at biopsy of 52 years (range 25-85 years) were identified. A total of 720 distinct cases of biopsy-proven ADH that underwent surgical excision were identified. The overall upstage rate was 11.7% (n = 84), including 16 cases of invasive breast cancer (19%) and 68 cases of ductal carcinoma in situ (81%). Age at biopsy (OR: 1.05 [95% CI: 1.03, 1.07], p < 0.001), post-menopausal status (OR: 3.31 [95% CI: 1.95, 5.86], p <0.001), and smoking history (OR: 1.95 [95% CI: 1.19, 3.16], p = 0.007) showed higher odds of upstaging. 85% of patients underwent stereotactic, vacuum-assisted biopsy, with 6 cores and a 9-gauge needle size being most frequent. On multivariable models, imaging findings associated with a clinically significant rate of upstage include mammographic asymmetry (p = 0.039), worrisome features on imaging (p = 0.03), multiple imaging findings for a lesion (p < 0.001), larger lesion size (p = 0.013), and fine and pleomorphic calcifications (both p < 0.001). ADH described as “focal” on CNB had lower odds of upstaging (p = 0.021).
Conclusions: In our population of ADH detected by DBT and diagnosed with vacuum-assisted biopsy, the overall upstage rate is lower than reported. Certain clinical and imaging findings are suggestive of higher risk of upstaging. This highlights the importance of shared treatment decision-making between physician and patient.
Methods: Retrospective cohort study using institutional data from the Enterprise Data Warehouse across Northwestern Medicine Health System. Women > 18 years old who underwent screening DBT between 2016-2024 followed by surgical excision for ADH diagnosed via stereotactic or ultrasound-guided CNB were included. Patients with ADH detected on MRI or sampled via MRI-guided biopsy, prior/concurrent history of breast cancer or known pathogenic mutation were excluded. Statistical analyses were completed with chi-square testing using p-value < 0.05 as statistical significance.
Results: 677 women with median age at biopsy of 52 years (range 25-85 years) were identified. A total of 720 distinct cases of biopsy-proven ADH that underwent surgical excision were identified. The overall upstage rate was 11.7% (n = 84), including 16 cases of invasive breast cancer (19%) and 68 cases of ductal carcinoma in situ (81%). Age at biopsy (OR: 1.05 [95% CI: 1.03, 1.07], p < 0.001), post-menopausal status (OR: 3.31 [95% CI: 1.95, 5.86], p <0.001), and smoking history (OR: 1.95 [95% CI: 1.19, 3.16], p = 0.007) showed higher odds of upstaging. 85% of patients underwent stereotactic, vacuum-assisted biopsy, with 6 cores and a 9-gauge needle size being most frequent. On multivariable models, imaging findings associated with a clinically significant rate of upstage include mammographic asymmetry (p = 0.039), worrisome features on imaging (p = 0.03), multiple imaging findings for a lesion (p < 0.001), larger lesion size (p = 0.013), and fine and pleomorphic calcifications (both p < 0.001). ADH described as “focal” on CNB had lower odds of upstaging (p = 0.021).
Conclusions: In our population of ADH detected by DBT and diagnosed with vacuum-assisted biopsy, the overall upstage rate is lower than reported. Certain clinical and imaging findings are suggestive of higher risk of upstaging. This highlights the importance of shared treatment decision-making between physician and patient.
Competition Category: Health Services Outcomes or Clinical
Mentor: Olga Kantor, MD
Nancy Ly, MD MS
Senior Resident (Clinical PGY3-5)
Evaluating the effects of surgical site infections on patients with different socioeconomic statuses: a mixed-methods study
Introduction: Preoperative optimization is standard practice in ventral hernia repair to reduce postoperative complications. However, optimization requirements such as smoking cessation, weight loss, and glycemic control may disproportionately burden patients of lower socioeconomic status (SES), potentially contributing to disparities in surgical access and outcomes. We sought to evaluate how SES influences postoperative outcomes and patient experience following ventral hernia repair.
Methods: We performed a mixed-methods study consisting of quantitative and qualitative components. For the quantitative arm, we conducted a retrospective cohort study using the Abdominal Core Health Quality Collaborative (ACHQC) registry (2016-2025), including adults undergoing open ventral hernia repair with mesh. SES was measured using the Distressed Communities Index (DCI), comparing patients in low- and high-SES strata. Primary outcomes included 30-day surgical site infection (SSI), surgical site occurrence (SSO), healthcare utilization (length of stay, readmission, reoperation), and patient-reported quality of life (HerQLes, PROMIS 3a). Multivariable regression identified predictors of SSI. The qualitative arm included semi-structured interviews exploring barriers to care, optimization experiences, and decision-making around surgery.
Results: A total of 16,176 patients met inclusion criteria (6,176 low SES; 10,000 high SES). Low-SES patients had higher rates of smoking (11% vs 7%, p<0.001), diabetes (23% vs 17%, p<0.001), obesity (65% vs 56%, p<0.001), contaminated wounds (9% vs 5%, p<0.001), and larger hernia defects. They also reported worse baseline quality of life (p<0.001). Unadjusted SSI rates were higher in low-SES patients (6% vs 4%); however, SES was not independently associated with SSI after adjustment (OR 1.13, 95% CI 0.97-1.31, p=0.121). Smoking, diabetes, and wound contamination remained significant predictors. Early qualitative interviews revealed profound quality-of-life burden, barriers to surgical access, and high willingness to accept operative risk across SES groups.
Conclusions: Low-SES patients present with greater modifiable risk burdens and worse baseline quality of life, but similar adjusted SSI risk after repair. Optimization practices may unintentionally contribute to disparities in surgical access by delaying care for patients with the greatest symptom burden. Patient-centered quality-of-life considerations should be incorporated into surgical decision-making.
Methods: We performed a mixed-methods study consisting of quantitative and qualitative components. For the quantitative arm, we conducted a retrospective cohort study using the Abdominal Core Health Quality Collaborative (ACHQC) registry (2016-2025), including adults undergoing open ventral hernia repair with mesh. SES was measured using the Distressed Communities Index (DCI), comparing patients in low- and high-SES strata. Primary outcomes included 30-day surgical site infection (SSI), surgical site occurrence (SSO), healthcare utilization (length of stay, readmission, reoperation), and patient-reported quality of life (HerQLes, PROMIS 3a). Multivariable regression identified predictors of SSI. The qualitative arm included semi-structured interviews exploring barriers to care, optimization experiences, and decision-making around surgery.
Results: A total of 16,176 patients met inclusion criteria (6,176 low SES; 10,000 high SES). Low-SES patients had higher rates of smoking (11% vs 7%, p<0.001), diabetes (23% vs 17%, p<0.001), obesity (65% vs 56%, p<0.001), contaminated wounds (9% vs 5%, p<0.001), and larger hernia defects. They also reported worse baseline quality of life (p<0.001). Unadjusted SSI rates were higher in low-SES patients (6% vs 4%); however, SES was not independently associated with SSI after adjustment (OR 1.13, 95% CI 0.97-1.31, p=0.121). Smoking, diabetes, and wound contamination remained significant predictors. Early qualitative interviews revealed profound quality-of-life burden, barriers to surgical access, and high willingness to accept operative risk across SES groups.
Conclusions: Low-SES patients present with greater modifiable risk burdens and worse baseline quality of life, but similar adjusted SSI risk after repair. Optimization practices may unintentionally contribute to disparities in surgical access by delaying care for patients with the greatest symptom burden. Patient-centered quality-of-life considerations should be incorporated into surgical decision-making.
Competition Category: Health Services Outcomes or Clinical
Mentor: Megan Melland-Smith, MD
Sahita Manda, BS
Student
Fast-track fasting: exploring clinical factors and outcomes of limited perioperative fasting for enhanced recovery in pediatric surgical patients
Introduction: Evidence for limited perioperative fasting mainly comes from adult studies, where it is associated with improved hydration and shortened length of stays (LOS). However, pediatric data remains limited, and uptake within enhanced recovery protocols (ERPs) is variable. This study evaluates the association between limited perioperative fasting compliance and clinical outcomes in children undergoing elective gastrointestinal (GI) surgery, as well as its relationship with adherence to other ERP elements and variation across procedure type and hospital.
Methods: We performed a secondary as-treated analysis of a prospective trial across 18 children's hospitals (July 2020-July 2024), including patients aged 10 to18 undergoing elective GI surgery. Limited fasting included two components: 1) avoiding prolonged fasting preoperatively (i.e., clear liquids within 2 hours before surgery) and 2) initiating early postoperative oral nutrition (i.e., clear liquids within 24 hours after surgery or regular diet by postoperative day 1). Patients were categorized as full compliance (both components) or no compliance (0 or 1 component). We compared groups on patient-and procedure-level characteristics and evaluated outcomes, including LOS and surgical and infectious complications. We also assessed adherence to other ERP elements. Bivariate analyses, multivariable logistic, and quantile regression were used to evaluate associations.
Results: Among 567 patients, 175 (30.9%) received both limited fasting components, 212 (37.4%) received one, and 180 (31.7%) received neither. Compared with patients receiving 0 or 1 component, full compliance was associated with shorter median LOS (β = -0.98, 95% CI: [-1.71, -0.44]), corresponding to approximately one day less in the hospital. There were no differences in surgical or infectious complication rates between groups. Full compliance was associated with adherence to other ERP elements, including naso/orogastric tube removal (OR = 3.20, 95% CI: [1.71, 5.98]), gut stimulation, (OR = 7.46, 95% CI: [3.18, 17.49]), intraperitoneal/perianastomotic drain avoidance (OR = 3.14, 95% CI: [1.06, 09.27]), and limited postoperative intravenous fluids (OR = 4.95, 95% CI: [2.71, 9.04]). Compliance rates were similar across procedure type (20.2-37.3%, p=0.07) but varied significantly by hospital (0-85.7%, p<0.001).
Conclusions: Compliance with limited fasting for children undergoing GI surgery was associated with shorter LOS without increased surgical or infectious complications. Adherence to limited fasting was associated with greater adherence to other ERP elements, suggesting hospital-level commitment to enhanced recovery and the importance of bundled interventions. Similar compliance across procedure types indicates that uptake is independent of complexity. Variation by hospital calls for additional research into implementation barriers and facilitators.
Methods: We performed a secondary as-treated analysis of a prospective trial across 18 children's hospitals (July 2020-July 2024), including patients aged 10 to18 undergoing elective GI surgery. Limited fasting included two components: 1) avoiding prolonged fasting preoperatively (i.e., clear liquids within 2 hours before surgery) and 2) initiating early postoperative oral nutrition (i.e., clear liquids within 24 hours after surgery or regular diet by postoperative day 1). Patients were categorized as full compliance (both components) or no compliance (0 or 1 component). We compared groups on patient-and procedure-level characteristics and evaluated outcomes, including LOS and surgical and infectious complications. We also assessed adherence to other ERP elements. Bivariate analyses, multivariable logistic, and quantile regression were used to evaluate associations.
Results: Among 567 patients, 175 (30.9%) received both limited fasting components, 212 (37.4%) received one, and 180 (31.7%) received neither. Compared with patients receiving 0 or 1 component, full compliance was associated with shorter median LOS (β = -0.98, 95% CI: [-1.71, -0.44]), corresponding to approximately one day less in the hospital. There were no differences in surgical or infectious complication rates between groups. Full compliance was associated with adherence to other ERP elements, including naso/orogastric tube removal (OR = 3.20, 95% CI: [1.71, 5.98]), gut stimulation, (OR = 7.46, 95% CI: [3.18, 17.49]), intraperitoneal/perianastomotic drain avoidance (OR = 3.14, 95% CI: [1.06, 09.27]), and limited postoperative intravenous fluids (OR = 4.95, 95% CI: [2.71, 9.04]). Compliance rates were similar across procedure type (20.2-37.3%, p=0.07) but varied significantly by hospital (0-85.7%, p<0.001).
Conclusions: Compliance with limited fasting for children undergoing GI surgery was associated with shorter LOS without increased surgical or infectious complications. Adherence to limited fasting was associated with greater adherence to other ERP elements, suggesting hospital-level commitment to enhanced recovery and the importance of bundled interventions. Similar compliance across procedure types indicates that uptake is independent of complexity. Variation by hospital calls for additional research into implementation barriers and facilitators.
Competition Category: Health Services Outcomes or Clinical
Mentor: Mehul Raval, MD MS MBA
Adwaiy Manerikar, MD
Senior Resident (Clinical PGY3-5)
Concomitant atrial fibrillation procedures during septal myectomy: a Society of Thoracic Surgeons database analysis
Introduction: Surgical ablation of atrial fibrillation (AF) during septal myectomy (SM) for hypertrophic obstructive cardiomyopathy has limited evaluation. We examined national outcomes and practice patterns for concomitant AF ablation during SM.
Methods: The Society of Thoracic Surgeons Adult Cardiac Surgery Database was queried for patients undergoing septal myectomy ± mitral valve surgery from 2014 to 2023. Patients undergoing SM with and without AF were compared. Patients with AF were stratified by surgical ablation vs. no ablation. 1:1 propensity score matching was used to account for baseline differences resulting in balanced groups. The primary outcome was operative mortality. Secondary outcomes were incidence of surgical ablation and left atrial appendage occlusion, mitral valve surgery, cross-clamp time, postoperative permanent pacemaker implantation, stroke, renal failure, and readmission within 30 days.
Results:Patients with AF (n=690) had more comorbidities, and higher rates of operative mortality, new-onset renal failure, and 30-day readmission than those without AF (n=2889). 380 (55%) patients with AF underwent surgical ablation: pulmonary vein isolation (30.3%), left atrial only (54.3%), and biatrial ablation (15.5%). Paroxysmal AF was most common (n=546, 81.3%), Ablated patients had higher cross-clamp times, mitral valve surgery, and left atrial appendage occlusion (97%) versus non-ablated patients (21.4%). After matching, there was no difference in operative mortality, stroke, pacemaker implantation, new-onset renal failure, or readmission between ablated and non-ablated groups.
Conclusions: With no increased risk of operative mortality, 30-day readmission, pacemaker, new-onset renal failure or stroke, concomitant AF ablation with left atrial appendage occlusion should be considered at the time of septal myectomy.
Methods: The Society of Thoracic Surgeons Adult Cardiac Surgery Database was queried for patients undergoing septal myectomy ± mitral valve surgery from 2014 to 2023. Patients undergoing SM with and without AF were compared. Patients with AF were stratified by surgical ablation vs. no ablation. 1:1 propensity score matching was used to account for baseline differences resulting in balanced groups. The primary outcome was operative mortality. Secondary outcomes were incidence of surgical ablation and left atrial appendage occlusion, mitral valve surgery, cross-clamp time, postoperative permanent pacemaker implantation, stroke, renal failure, and readmission within 30 days.
Results:Patients with AF (n=690) had more comorbidities, and higher rates of operative mortality, new-onset renal failure, and 30-day readmission than those without AF (n=2889). 380 (55%) patients with AF underwent surgical ablation: pulmonary vein isolation (30.3%), left atrial only (54.3%), and biatrial ablation (15.5%). Paroxysmal AF was most common (n=546, 81.3%), Ablated patients had higher cross-clamp times, mitral valve surgery, and left atrial appendage occlusion (97%) versus non-ablated patients (21.4%). After matching, there was no difference in operative mortality, stroke, pacemaker implantation, new-onset renal failure, or readmission between ablated and non-ablated groups.
Conclusions: With no increased risk of operative mortality, 30-day readmission, pacemaker, new-onset renal failure or stroke, concomitant AF ablation with left atrial appendage occlusion should be considered at the time of septal myectomy.
Competition Category: Health Services Outcomes or Clinical
Mentor: Christopher Mehta, MD
Ben Mazurek, BS
Student
Low-density neutrophils and monocyte remodeling are associated with distinct immune trajectories of PGD and AKI after heart transplantation
Introduction: Ischemia-reperfusion injury is a principal driver of early inflammation after heart transplantation, yet the temporal coordination between neutrophil activation, monocyte remodeling, and downstream complications such as primary graft dysfunction (PGD) and acute kidney injury (AKI) remains incompletely defined. We hypothesized that reperfusion induces a surge of circulating low-density neutrophils (LDNs) that orchestrate early monocyte reprogramming and delineate divergent immune trajectories associated with PGD and AKI.
Methods: Twelve adult heart transplant recipients were prospectively enrolled [mean age 57.4 years; BMI 26.4 kg/m2; 7 male, 5 female; 11 donation-after-brain-death and 1 donation-after-circulatory-death]. PGD occurred in 3 patients and AKI in 9; AKI was defined by KDIGO criteria and PGD by ISHLT consensus definitions. Blood was serially collected at pre-cross-clamp (Pre-Cx), post-reperfusion (Post-Cx), postoperative day 1 (POD-1), and postoperative day 2 (POD-2). Peripheral blood mononuclear cells were analyzed using multiparametric flow cytometry (CD45, CD15, CD16, CD14, HLA-DR) to quantify LDNs and monocyte subsets: classical (CM), intermediate (INT-M), and non-classical (NCM). Temporal dynamics were assessed using two-way repeated-measures ANOVA with Geisser-Greenhouse correction.
Results: LDNs were minimal prior to reperfusion but increased sharply following cross-clamp release, peaking at POD-1 before partially declining by POD-2 (p<0.0001). This neutrophil surge coincided with expansion of CM and INT-M and a marked reduction in NCM. In patients with AKI, monocyte responses demonstrated pronounced temporal remodeling, with significant effects of time (p=0.0007), subset (p<0.0001), and a strong time-subset interaction (p=0.0009), driven by progressive INT-M expansion and sustained depletion of NCM. In contrast, PGD demonstrated differences across monocyte subsets (p=0.0003) without clear temporal evolution (p=0.150). LDN levels correlated positively with INT-M expansion (r=0.72, p=0.004) and inversely with NCM (r=−0.65, p=0.009).
Conclusions: Heart transplantation is associated with a rapid LDN surge that coincides with INT-M expansion and NCM depletion. Distinct immune trajectories were observed across clinical phenotypes: AKI was characterized by dynamic, evolving immune remodeling, whereas PGD demonstrated a more static profile, potentially reflecting an early-dominant inflammatory state. The association between LDNs and monocyte subset shifts suggests a potential neutrophil-monocyte axis in early post-transplant immune responses and highlights LDNs and monocyte subsets as candidate biomarkers for early risk stratification and future investigation.
Methods: Twelve adult heart transplant recipients were prospectively enrolled [mean age 57.4 years; BMI 26.4 kg/m2; 7 male, 5 female; 11 donation-after-brain-death and 1 donation-after-circulatory-death]. PGD occurred in 3 patients and AKI in 9; AKI was defined by KDIGO criteria and PGD by ISHLT consensus definitions. Blood was serially collected at pre-cross-clamp (Pre-Cx), post-reperfusion (Post-Cx), postoperative day 1 (POD-1), and postoperative day 2 (POD-2). Peripheral blood mononuclear cells were analyzed using multiparametric flow cytometry (CD45, CD15, CD16, CD14, HLA-DR) to quantify LDNs and monocyte subsets: classical (CM), intermediate (INT-M), and non-classical (NCM). Temporal dynamics were assessed using two-way repeated-measures ANOVA with Geisser-Greenhouse correction.
Results: LDNs were minimal prior to reperfusion but increased sharply following cross-clamp release, peaking at POD-1 before partially declining by POD-2 (p<0.0001). This neutrophil surge coincided with expansion of CM and INT-M and a marked reduction in NCM. In patients with AKI, monocyte responses demonstrated pronounced temporal remodeling, with significant effects of time (p=0.0007), subset (p<0.0001), and a strong time-subset interaction (p=0.0009), driven by progressive INT-M expansion and sustained depletion of NCM. In contrast, PGD demonstrated differences across monocyte subsets (p=0.0003) without clear temporal evolution (p=0.150). LDN levels correlated positively with INT-M expansion (r=0.72, p=0.004) and inversely with NCM (r=−0.65, p=0.009).
Conclusions: Heart transplantation is associated with a rapid LDN surge that coincides with INT-M expansion and NCM depletion. Distinct immune trajectories were observed across clinical phenotypes: AKI was characterized by dynamic, evolving immune remodeling, whereas PGD demonstrated a more static profile, potentially reflecting an early-dominant inflammatory state. The association between LDNs and monocyte subset shifts suggests a potential neutrophil-monocyte axis in early post-transplant immune responses and highlights LDNs and monocyte subsets as candidate biomarkers for early risk stratification and future investigation.
Competition Category: Basic Science or Translational
Mentor: Stephen Chiu, MD
Bilal Khan Mohammed, MD
Fellow (Clinical or Postdoctoral Researcher)
A compartmentalized monocyte trafficking program defines early vascular immune remodeling after pulmonary thromboendarterectomy
Introduction: Pulmonary thromboendarterectomy (PTE) is the definitive therapy for chronic thromboembolic pulmonary hypertension (CTEPH), yet immune programs governing vascular recovery after removal of chronic thrombo-inflammatory fibrotic obstruction remain undefined. We hypothesized that reperfusion induces a compartmentalized monocyte response within the pulmonary vascular bed that encodes disease chronicity and drives early vascular remodeling.
Methods: Twenty-three patients undergoing PTE were prospectively enrolled. Paired pulmonary arterial (PA) and systemic arterial (BA) blood samples were collected at pre-cross-clamp, immediately post-reperfusion, postoperative day-1 and 2(POD-1 and 2). Flow cytometry quantified classical (CM; CD14⁺⁺CD16⁻), intermediate (INT-M; CD14⁺⁺CD16⁺), and non-classical monocytes (NCM; CD14⁺CD16⁺⁺). Thrombi were classified as organizing (n = 2), organized/fibrotic (n = 13), or mixed acute-on-chronic (n = 8) by histopathology. Time from diagnosis to surgery was correlated with clot phenotype and flow analysis. Pre- and post-PTE hemodynamic parameters were recorded.
Results: Reperfusion induced systemic and pulmonary monocytosis, increasing from 1.8 ± 0.7 × 10³ to 3.0 ± 0.9 × 10³ cells/μL by POD-2(p < 0.05). This was driven by coordinated expansion of CM and INT-M, rising from 71 ± 8%(1.3 ± 0.5 × 10³ cells/μL) to 88 ± 6%(2.6 ± 0.8 × 10³ cells/μL)(p = 0.01) and from 12 ± 5% (0.22 ± 0.09 × 10³ cells/μL) to 22 ± 7% (0.57 ± 0.18 × 10³ cells/μL)(p < 0.01), respectively, dominating the POD-1/2 reperfusion phase. In contrast, NCM declined sharply(14 ± 4% [0.26 ± 0.11 × 10³ cells/μL] to 3 ± 2% [0.09 ± 0.04 × 10³ cells/μL], p < 0.001).Immediately following reperfusion, PA samples showed a transient excess of monocytes relative to BA, indicating pulmonary vascular sequestration. By POD-2, CM levels equilibrated between compartments(Δ = 0.3%, p = 0.46; Δ = -1.35 × 10³ cells/μL, p = 0.38), consistent with redistribution into the systemic circulation. Hemodynamics improved significantly after PTE (PVR 431 ± 270 → 117 ± 49 dyn·s·cm⁻⁵, p < 0.001).Patients with organized fibrotic thrombi exhibited longer disease duration(median 945 vs 420 days, p = 0.06) and greater INT-M expansion(median 0.58 × 10³ vs 0.24 × 10³ cells/μL, p = 0.04),linking intermediate monocyte responses to thrombus chronicity and disease severity.
Conclusions: PTE unmasks a compartmentalized monocyte trafficking axis during reperfusion, characterized by early pulmonary sequestration of classical monocytes followed by systemic expansion dominated by intermediate monocytes. Selective INT-M amplification in chronic fibrotic disease identifies a severity-linked immune program that may govern vascular remodeling. Monocyte subset dynamics represent candidate biomarkers of vascular recovery and potential targets for immune modulation in CTEPH.
Methods: Twenty-three patients undergoing PTE were prospectively enrolled. Paired pulmonary arterial (PA) and systemic arterial (BA) blood samples were collected at pre-cross-clamp, immediately post-reperfusion, postoperative day-1 and 2(POD-1 and 2). Flow cytometry quantified classical (CM; CD14⁺⁺CD16⁻), intermediate (INT-M; CD14⁺⁺CD16⁺), and non-classical monocytes (NCM; CD14⁺CD16⁺⁺). Thrombi were classified as organizing (n = 2), organized/fibrotic (n = 13), or mixed acute-on-chronic (n = 8) by histopathology. Time from diagnosis to surgery was correlated with clot phenotype and flow analysis. Pre- and post-PTE hemodynamic parameters were recorded.
Results: Reperfusion induced systemic and pulmonary monocytosis, increasing from 1.8 ± 0.7 × 10³ to 3.0 ± 0.9 × 10³ cells/μL by POD-2(p < 0.05). This was driven by coordinated expansion of CM and INT-M, rising from 71 ± 8%(1.3 ± 0.5 × 10³ cells/μL) to 88 ± 6%(2.6 ± 0.8 × 10³ cells/μL)(p = 0.01) and from 12 ± 5% (0.22 ± 0.09 × 10³ cells/μL) to 22 ± 7% (0.57 ± 0.18 × 10³ cells/μL)(p < 0.01), respectively, dominating the POD-1/2 reperfusion phase. In contrast, NCM declined sharply(14 ± 4% [0.26 ± 0.11 × 10³ cells/μL] to 3 ± 2% [0.09 ± 0.04 × 10³ cells/μL], p < 0.001).Immediately following reperfusion, PA samples showed a transient excess of monocytes relative to BA, indicating pulmonary vascular sequestration. By POD-2, CM levels equilibrated between compartments(Δ = 0.3%, p = 0.46; Δ = -1.35 × 10³ cells/μL, p = 0.38), consistent with redistribution into the systemic circulation. Hemodynamics improved significantly after PTE (PVR 431 ± 270 → 117 ± 49 dyn·s·cm⁻⁵, p < 0.001).Patients with organized fibrotic thrombi exhibited longer disease duration(median 945 vs 420 days, p = 0.06) and greater INT-M expansion(median 0.58 × 10³ vs 0.24 × 10³ cells/μL, p = 0.04),linking intermediate monocyte responses to thrombus chronicity and disease severity.
Conclusions: PTE unmasks a compartmentalized monocyte trafficking axis during reperfusion, characterized by early pulmonary sequestration of classical monocytes followed by systemic expansion dominated by intermediate monocytes. Selective INT-M amplification in chronic fibrotic disease identifies a severity-linked immune program that may govern vascular remodeling. Monocyte subset dynamics represent candidate biomarkers of vascular recovery and potential targets for immune modulation in CTEPH.
Competition Category: Basic Science or Translational
Mentor: Stephen Chiu, MD
Sarbjeet Niraula, PhD
Fellow (Clinical or Postdoctoral Researcher)
Gut microbial co-occurrence network in patients with peripheral artery disease
Introduction: The human gut microbiome comprises a complex network of microorganisms and microbially-derived metabolites. The association of the gut microbiome with peripheral artery disease (PAD) is poorly understood. The goal of this study is to identify correlations of gut microbial markers with clinical features of PAD.
Methods: Study participants aged ≥ 40 years with documented PAD and non-PAD controls each provided fecal samples. Microbial DNA extracted from fecal samples was analyzed using shotgun metagenomics. Untargeted fecal metabolite quantification was performed using nuclear magnetic resonance. Raw sequence data were quality filtered using the KneadData pipeline followed by taxonomic annotation using Kraken2 against the UHGG database (v2.0.2) and species abundance estimation using Bracken. A species network was constructed using SparCC. Modules of positively correlated species were constructed and analyzed using the Weighted Gene Co-expression Network Analysis package in R. The first principal component of each module (ME) and metabolite concentration and sample traits were analyzed by Spearman correlations. Network visualization was done in Cytoscape.
Results: Among 78 patients with PAD, 40 (51.3%) had chronic limb threatening ischemia (CLTI) and 38 (48.7%) had symptoms of claudication. A network consisting of 1644 species (nodes) and 23,108 connections (edges) was clustered into 17 modules that were further grouped into 4 meta-modules. Node degree and clustering coefficients were reduced with increased PAD severity (p<0.001). The majority of modules correlated with claudication were also positively correlated with concentrations of specific microbe-derived metabolites, including butyrate, nicotinate, and glutamate. The total relative abundance of species belonging to these modules was higher in the claudication group than in the CLTI group. A module that was negatively associated with CLTI was enriched with known butyrate producers such as Faecalibacterium spp., Agathobaculum butyriciproducens and Ruminococcus bromii, whereas a module that was positively associated with CLTI was enriched with oral microbes such as Granulicatella spp., Streptococcus oralis, Actinomyces. oris, and Oribacterium. sinus.
Conclusions: In a species correlation network analysis, we identified specific microbial and metabolomic signatures linked to the clinical severity of PAD, underscoring the importance of interaction patterns beyond abundance-based changes.
Methods: Study participants aged ≥ 40 years with documented PAD and non-PAD controls each provided fecal samples. Microbial DNA extracted from fecal samples was analyzed using shotgun metagenomics. Untargeted fecal metabolite quantification was performed using nuclear magnetic resonance. Raw sequence data were quality filtered using the KneadData pipeline followed by taxonomic annotation using Kraken2 against the UHGG database (v2.0.2) and species abundance estimation using Bracken. A species network was constructed using SparCC. Modules of positively correlated species were constructed and analyzed using the Weighted Gene Co-expression Network Analysis package in R. The first principal component of each module (ME) and metabolite concentration and sample traits were analyzed by Spearman correlations. Network visualization was done in Cytoscape.
Results: Among 78 patients with PAD, 40 (51.3%) had chronic limb threatening ischemia (CLTI) and 38 (48.7%) had symptoms of claudication. A network consisting of 1644 species (nodes) and 23,108 connections (edges) was clustered into 17 modules that were further grouped into 4 meta-modules. Node degree and clustering coefficients were reduced with increased PAD severity (p<0.001). The majority of modules correlated with claudication were also positively correlated with concentrations of specific microbe-derived metabolites, including butyrate, nicotinate, and glutamate. The total relative abundance of species belonging to these modules was higher in the claudication group than in the CLTI group. A module that was negatively associated with CLTI was enriched with known butyrate producers such as Faecalibacterium spp., Agathobaculum butyriciproducens and Ruminococcus bromii, whereas a module that was positively associated with CLTI was enriched with oral microbes such as Granulicatella spp., Streptococcus oralis, Actinomyces. oris, and Oribacterium. sinus.
Conclusions: In a species correlation network analysis, we identified specific microbial and metabolomic signatures linked to the clinical severity of PAD, underscoring the importance of interaction patterns beyond abundance-based changes.
Competition Category: Basic Science or Translational
Mentor: Karen Ho, MD
Adaure Nwaba, MD MS
Senior Resident (Clinical PGY3-5)
From treatment to survivorship: A mixed-methods study of a breast cancer survivorship program in Lagos, Nigeria
Introduction: Breast cancer care is a multidisciplinary effort requiring collaboration amongst various specialties. With increased breast cancer awareness and treatment options in Sub-Saharan Africa (SSA), there is now a larger incidence of women living with a breast cancer diagnosis. However, there is a gap in literature in relation to social support systems and survivorship programs utilized by patients in SSA. This includes understanding the presence and effects of possible underdiagnosed physical and/or psychosocial side effects of having breast cancer and undergoing breast cancer treatment. Thus, this project aims to evaluate the current breast cancer survivorship program (Dew Drops) at Lakeshore Cancer Center in Lagos, Nigeria by assessing patient utilization of the program and patient-perceived needs that may be further met through the program.
Methods: Patients greater than 18 years of age, currently or previously diagnosed with breast cancer, and being treated at Lakeshore Cancer Center in Lagos, Nigeria are eligible for inclusion. This is a mixed methods study utilizing both a survey questionnaire and qualitative interviews. Study tools include investigator-defined questions related to program utilization and awareness, and the ULL-27 validated questionnaire that evaluates quality of life and upper extremity lymphedema risk in breast cancer patients. This assesses domains of physical and psychosocial well-being. Recruitment fliers are shared at the Lakeshore Cancer Center and through the Dew Drops breast cancer support program. This project is supported and funded by the University of Lagos, Northwestern University, and Havey Global Health Institute.
Results: IRB approval has been obtained by the University of Lagos and is currently pending from Northwestern University. We expect about 10-30 patients to complete both survey questionnaire and interviews. It has also provided an opportunity to understand the basic function of the Dew Drops Breast Cancer support program, which was necessary for baseline understanding of the programmatic structure during project development.
Conclusions: This study is designed to inform future survivorship programming at Lakeshore Cancer Center by identifying patterns of program utilization and unmet physical and psychosocial needs among breast cancer survivors.
Methods: Patients greater than 18 years of age, currently or previously diagnosed with breast cancer, and being treated at Lakeshore Cancer Center in Lagos, Nigeria are eligible for inclusion. This is a mixed methods study utilizing both a survey questionnaire and qualitative interviews. Study tools include investigator-defined questions related to program utilization and awareness, and the ULL-27 validated questionnaire that evaluates quality of life and upper extremity lymphedema risk in breast cancer patients. This assesses domains of physical and psychosocial well-being. Recruitment fliers are shared at the Lakeshore Cancer Center and through the Dew Drops breast cancer support program. This project is supported and funded by the University of Lagos, Northwestern University, and Havey Global Health Institute.
Results: IRB approval has been obtained by the University of Lagos and is currently pending from Northwestern University. We expect about 10-30 patients to complete both survey questionnaire and interviews. It has also provided an opportunity to understand the basic function of the Dew Drops Breast Cancer support program, which was necessary for baseline understanding of the programmatic structure during project development.
Conclusions: This study is designed to inform future survivorship programming at Lakeshore Cancer Center by identifying patterns of program utilization and unmet physical and psychosocial needs among breast cancer survivors.
Competition Category: Health Services Outcomes or Clinical
Mentor: Juliet Lumati, MD MPH
Nicole Ontiveros, MD MS
Senior Resident (Clinical PGY3-5)
Surveillance or signal? Increased diagnostic burden without increased oncologic risk following postmastectomy breast reconstruction
Introduction: Mastectomy is a cornerstone in the management of non-metastatic breast cancer, with nearly half of patients electing breast reconstruction. Reconstruction confers well-established benefits in quality of life and psychological wellbeing, and prior studies have largely supported its oncologic safety. However, the reconstructed breast presents unique challenges: altered tissue architecture, implant or flap-related imaging changes, and increased postoperative surveillance may influence downstream diagnostic and procedural rates. Additionally, concerns persist that reconstructive considerations may impact mastectomy technique and residual breast tissue, potentially affecting recurrence risk. This study sought to characterize contemporary oncologic outcomes following mastectomy with and without breast reconstruction.
Methods: A retrospective cohort analysis was conducted utilizing the TriNetX database. Patients with Stage I-III breast cancer undergoing mastectomy (simple, skin-sparing, nipple-sparing, or radical) were compared to those undergoing mastectomy with breast reconstruction (implant-based or autologous). Propensity score matching was performed based on age, cancer stage, and comorbidities. Mastectomy served as the index event. The primary outcome was breast cancer recurrence, with abnormal diagnostic mammography, breast biopsy, chest wall/breast excision, and nodal metastasis >6 months following mastectomy serving as proxies. Secondary outcomes included all-cause mortality >6 months after mastectomy.
Results: A total of 7,398 patients were included in the final matched analysis. The mean age at mastectomy was 49.9 years, and the largest proportion of patients had Stage I disease. Mean follow-up was 1,706.6 days in the mastectomy-only cohort and 2,268.1 days in the reconstruction cohort. Compared to mastectomy alone, mastectomy with reconstruction was associated with an increased risk of abnormal or inconclusive diagnostic mammography (RR 1.55, 95% CI 1.37-1.79, p < 0.0001), breast biopsy (RR 2.19, 95% CI 1.30-3.69, p < 0.001), and chest wall/breast excision (RR 1.60, 95% CI 1.28-2.01, p < 0.0001) occurring >6 months after mastectomy. There was no statistically significant difference in nodal metastatic progression between groups (HR 1.23, 95% CI 0.91-1.66, p = 0.49). Mastectomy alone was associated with increased mortality >6 months after surgery compared to mastectomy with reconstruction (HR 1.91, 95% CI 1.54-2.36, p < 0.01).
Conclusions: Breast reconstruction following mastectomy is associated with increased downstream diagnostic and procedural burden without increased nodal metastasis or mortality. These findings suggest that increased diagnostic interventions reflect surveillance burden rather than true oncologic risk and have important implications for patient counseling and post-reconstruction surveillance strategies. Further studies are needed to explore the influence of surgical technique on oncologic outcomes following breast reconstruction.
Methods: A retrospective cohort analysis was conducted utilizing the TriNetX database. Patients with Stage I-III breast cancer undergoing mastectomy (simple, skin-sparing, nipple-sparing, or radical) were compared to those undergoing mastectomy with breast reconstruction (implant-based or autologous). Propensity score matching was performed based on age, cancer stage, and comorbidities. Mastectomy served as the index event. The primary outcome was breast cancer recurrence, with abnormal diagnostic mammography, breast biopsy, chest wall/breast excision, and nodal metastasis >6 months following mastectomy serving as proxies. Secondary outcomes included all-cause mortality >6 months after mastectomy.
Results: A total of 7,398 patients were included in the final matched analysis. The mean age at mastectomy was 49.9 years, and the largest proportion of patients had Stage I disease. Mean follow-up was 1,706.6 days in the mastectomy-only cohort and 2,268.1 days in the reconstruction cohort. Compared to mastectomy alone, mastectomy with reconstruction was associated with an increased risk of abnormal or inconclusive diagnostic mammography (RR 1.55, 95% CI 1.37-1.79, p < 0.0001), breast biopsy (RR 2.19, 95% CI 1.30-3.69, p < 0.001), and chest wall/breast excision (RR 1.60, 95% CI 1.28-2.01, p < 0.0001) occurring >6 months after mastectomy. There was no statistically significant difference in nodal metastatic progression between groups (HR 1.23, 95% CI 0.91-1.66, p = 0.49). Mastectomy alone was associated with increased mortality >6 months after surgery compared to mastectomy with reconstruction (HR 1.91, 95% CI 1.54-2.36, p < 0.01).
Conclusions: Breast reconstruction following mastectomy is associated with increased downstream diagnostic and procedural burden without increased nodal metastasis or mortality. These findings suggest that increased diagnostic interventions reflect surveillance burden rather than true oncologic risk and have important implications for patient counseling and post-reconstruction surveillance strategies. Further studies are needed to explore the influence of surgical technique on oncologic outcomes following breast reconstruction.
Competition Category: Health Services Outcomes or Clinical
Mentor: Robert Galiano, MD
Bianka Progri, MS
Fellow (Clinical or Postdoctoral Researcher)
Evaluating the safety of transdermal deferoxamine patch in oncological settings
Introduction: This study investigates the impact of Deferoxamine Intradermal Delivery Patch (DIDP) applied topically as a skin radioprotectant on radiation therapy efficacy. While DIDP's radioprotective properties on the skin have been established in previous studies, its potential protective effects on cancer cells during radiotherapy remain largely unexplored. This research aims to elucidate whether DIDP applied to skin overlying tumors interferes with apoptosis induction in cancer cells following radiation therapy. Our study provides insights into the interplay between DIDP and radiation therapy effectiveness in cancer treatment.
Methods: The study used a murine model of human breast cancer, injecting 1x10^6 MCF7-Luc2 cancer cells into the mammary gland of 8-week-old female immunodeficient NSG mice. Once tumors reached an intensity of ~1x108 photons/s, mice were divided into two cohorts: control and DFO-treated. DIDP was applied daily for 30 days until the study endpoint. Radiation therapy (5 x 2 Gy daily; total 10 Gy) was administered on the last 5 days of DIDP treatment using the RS-2000 irradiator. Pre- and post-radiation tumor size was assessed using luciferin injection and LAGO bioluminescence imaging. Skin and tumor samples were preserved in formalin and embedded in paraffin for analyses. TUNEL staining and ImageJ software were used to quantify apoptotic cells. Statistical analysis was performed using ANOVA and GraphPad Prism, with p < 0.05 considered significant.
Results: The analysis of TUNEL staining revealed a significant increase in apoptotic cells following radiotherapy, in both control and DFO-treated groups (FC=4.70, p<0.0001), compared to non-irradiated control, indicating irradiation was effective. Importantly, no significant differences were found in apoptotic cells when comparing irradiated control to irradiated DFO-treated mice, suggesting DIDP does not interfere with radiation therapy efficacy. Moreover, bioluminescence assays showed no significant differences in tumor size after radiation therapy between groups (FC=0.8, p=0.337). These findings confirm that DIDP does not negatively influence radiotherapy effectiveness in breast cancer cells.
Conclusions: This study highlights the importance of evaluating radioprotective agents' in vivo effects on cancer treatment outcomes for oncological safety. Specifically, we address the interaction between DIDP and radiation therapy. The findings indicate that topical DFO treatment poses minimal risk of interfering with radiation therapy, as shown by no significant difference in tumor growth and cell apoptosis following radiation. DIDP does not compromise radiation therapy efficacy, likely due to its inability to penetrate deeper tissues. In conclusion, our study suggests that DIDP may be safely used as a radioprotective treatment in oncological settings.
Methods: The study used a murine model of human breast cancer, injecting 1x10^6 MCF7-Luc2 cancer cells into the mammary gland of 8-week-old female immunodeficient NSG mice. Once tumors reached an intensity of ~1x108 photons/s, mice were divided into two cohorts: control and DFO-treated. DIDP was applied daily for 30 days until the study endpoint. Radiation therapy (5 x 2 Gy daily; total 10 Gy) was administered on the last 5 days of DIDP treatment using the RS-2000 irradiator. Pre- and post-radiation tumor size was assessed using luciferin injection and LAGO bioluminescence imaging. Skin and tumor samples were preserved in formalin and embedded in paraffin for analyses. TUNEL staining and ImageJ software were used to quantify apoptotic cells. Statistical analysis was performed using ANOVA and GraphPad Prism, with p < 0.05 considered significant.
Results: The analysis of TUNEL staining revealed a significant increase in apoptotic cells following radiotherapy, in both control and DFO-treated groups (FC=4.70, p<0.0001), compared to non-irradiated control, indicating irradiation was effective. Importantly, no significant differences were found in apoptotic cells when comparing irradiated control to irradiated DFO-treated mice, suggesting DIDP does not interfere with radiation therapy efficacy. Moreover, bioluminescence assays showed no significant differences in tumor size after radiation therapy between groups (FC=0.8, p=0.337). These findings confirm that DIDP does not negatively influence radiotherapy effectiveness in breast cancer cells.
Conclusions: This study highlights the importance of evaluating radioprotective agents' in vivo effects on cancer treatment outcomes for oncological safety. Specifically, we address the interaction between DIDP and radiation therapy. The findings indicate that topical DFO treatment poses minimal risk of interfering with radiation therapy, as shown by no significant difference in tumor growth and cell apoptosis following radiation. DIDP does not compromise radiation therapy efficacy, likely due to its inability to penetrate deeper tissues. In conclusion, our study suggests that DIDP may be safely used as a radioprotective treatment in oncological settings.
Competition Category: Basic Science or Translational
Mentor: Arun Gosain, MD
Rayan Rahmani, BS
Student
Standardizing perioperative care for pediatric gastrostomy tube placement: impact on length of stay and hospital costs
Introduction: Perioperative management of pediatric gastrostomy tube placement varies widely across institutions, particularly with respect to time to first feed, length of stay (LOS), and resource utilization. We evaluated the impact of implementing a standardized perioperative care pathway on postoperative recovery and hospital resource utilization.
Methods: We conducted a single-institution retrospective cohort study of children undergoing planned gastrostomy tube placement from March 2019 to June 2025. Patients were grouped into pre-implementation (3/2019-11/2022), implementation (12/2022-4/2024), and post- implementation (5/2024-6/2025) periods corresponding to pathway rollout. Primary outcomes were time to first feed, postoperative LOS, and standardized hospital cost. Outcomes were compared using Kruskal-Wallis tests and multivariable generalized linear models adjusted for age, sex, race/ethnicity, and medical complexity.
Results: Among 343 patients, 179 underwent surgery pre-implementation, 96 during implementation, and 68 post-implementation. From pre- to post-implementation, median time to first feed decreased from 6.9 to 3.6 hours (p < 0.001), LOS decreased from 29.7 to 24.8 hours (p < 0.001), same-day discharge increased from 2.8% to 29.4% (p < 0.001), and median standardized costs decreased from $9,554 to $7,517 (p < 0.001). In adjusted analyses, standardized hospital costs were 28% lower post-implementation compared with pre- implementation (adjusted cost ratio 0.72, 95% CI 0.66-0.79).
Conclusions: Implementation of a standardized gastrostomy tube care pathway was associated with shorter time to first feed, shorter postoperative LOS, and lower standardized hospital costs, supporting the role of structured perioperative pathways in improving efficiency and resource utilization in pediatric surgical care.
Methods: We conducted a single-institution retrospective cohort study of children undergoing planned gastrostomy tube placement from March 2019 to June 2025. Patients were grouped into pre-implementation (3/2019-11/2022), implementation (12/2022-4/2024), and post- implementation (5/2024-6/2025) periods corresponding to pathway rollout. Primary outcomes were time to first feed, postoperative LOS, and standardized hospital cost. Outcomes were compared using Kruskal-Wallis tests and multivariable generalized linear models adjusted for age, sex, race/ethnicity, and medical complexity.
Results: Among 343 patients, 179 underwent surgery pre-implementation, 96 during implementation, and 68 post-implementation. From pre- to post-implementation, median time to first feed decreased from 6.9 to 3.6 hours (p < 0.001), LOS decreased from 29.7 to 24.8 hours (p < 0.001), same-day discharge increased from 2.8% to 29.4% (p < 0.001), and median standardized costs decreased from $9,554 to $7,517 (p < 0.001). In adjusted analyses, standardized hospital costs were 28% lower post-implementation compared with pre- implementation (adjusted cost ratio 0.72, 95% CI 0.66-0.79).
Conclusions: Implementation of a standardized gastrostomy tube care pathway was associated with shorter time to first feed, shorter postoperative LOS, and lower standardized hospital costs, supporting the role of structured perioperative pathways in improving efficiency and resource utilization in pediatric surgical care.
Competition Category: Quality Improvement
Mentor: Mehul Raval, MD MS MBA
Matthew Ramsey, MD
Senior Resident (Clinical PGY3-5)
Treatment of chronic phrenic neuropathy with a functional rectus abdominis muscle transfer and implanrows synchronized diaphragm pacing in a preclinical rat model
Introduction: Phrenic neuropathy (PhN) is a complex, debilitating condition that confers significant morbidity to both patients and the healthcare system at large. The impaired diaphragm contraction creates chronic complications from sustained mechanical ventilation. Patients with chronic PhN are outside the critical window for phrenic nerve interventions to restore diaphragm contraction and ultimately, volitional respiration. We therefore seek to treat this population by establishing a new functional muscle transfer using the rectus abdominis (RA) as a tendon transfer to power the denervated, paralyzed diaphragm in chronic PhN. The RA represents an ideal donor muscle given its proximity for use as a pedicled muscle flap, adequate excursion equal to tidal volume diaphragm excursion, ample contractile strength, optimal vector of contracture and reliable motor nerve anatomy to allow for pacer coupling.
Methods: Previous work from our lab has established a rat model of unilateral phrenic neuropathy with diaphragm pacing using a custom device from Rogers Research Group. We extend this model to establish a new surgical technique and rehabilitation pathway. Sprague-Dawley rats will be used to describe the RA surgical procedure with transdiaphragmatic pressures (Pdi) being the primary outcomes.
Results: Initial surgeries have established the reliable neurovascular anatomy of the rat RA muscle and technique for fascial harvest and inset across the unilateral muscular diaphragm. Previous work from our lab has established diaphragm excursion to be 3mm with tidal volumes on ultrasound (dUS) under anesthesia. The RA demonstrated an intraoperative excursion of 4mm with maximal, single intercostal nerve stimulation and 8mm of excursion with synchronized stimulation of two adjacent intercostal motor nerves. Ongoing work will employ the novel, custom nerve pacer currently being developed to quantify contractile force of the in vivo RA muscle-tendon construct prior to inset, establish the rehabilitation timeline for the RA tendon transfer, trace weekly Pdi as a direct assessment of airway pressures, and evaluate the potential for long term plasticity to natively synchronize RA contraction without pacing.
Conclusions: We seek to establish a new surgical technique for chronic PhN, describe the rehabilitation timeline, and validate functional recovery for ultimate bedside translation. This work will ultimately help separate patients with chronic PhN from mechanical ventilation to reduce their morbidity and mortality from pulmonary complications.
Methods: Previous work from our lab has established a rat model of unilateral phrenic neuropathy with diaphragm pacing using a custom device from Rogers Research Group. We extend this model to establish a new surgical technique and rehabilitation pathway. Sprague-Dawley rats will be used to describe the RA surgical procedure with transdiaphragmatic pressures (Pdi) being the primary outcomes.
Results: Initial surgeries have established the reliable neurovascular anatomy of the rat RA muscle and technique for fascial harvest and inset across the unilateral muscular diaphragm. Previous work from our lab has established diaphragm excursion to be 3mm with tidal volumes on ultrasound (dUS) under anesthesia. The RA demonstrated an intraoperative excursion of 4mm with maximal, single intercostal nerve stimulation and 8mm of excursion with synchronized stimulation of two adjacent intercostal motor nerves. Ongoing work will employ the novel, custom nerve pacer currently being developed to quantify contractile force of the in vivo RA muscle-tendon construct prior to inset, establish the rehabilitation timeline for the RA tendon transfer, trace weekly Pdi as a direct assessment of airway pressures, and evaluate the potential for long term plasticity to natively synchronize RA contraction without pacing.
Conclusions: We seek to establish a new surgical technique for chronic PhN, describe the rehabilitation timeline, and validate functional recovery for ultimate bedside translation. This work will ultimately help separate patients with chronic PhN from mechanical ventilation to reduce their morbidity and mortality from pulmonary complications.
Competition Category: Basic Science or Translational
Mentor: Sumanas Jordan, MD PhD
Saieesh Rao, MD MS
Senior Resident (Clinical PGY3-5)
Post-acute care disposition and complication-free survival after spinal cord injury
Introduction: Inpatient rehabilitation after spinal cord injury is known to improve injured patients' mobility and function. This is important because spinal cord injury patients are prone to immobility-associated complications. However, many spinal cord injury patients are discharged to skilled nursing facilities instead of inpatient rehabilitation, and the clinical consequences of forgoing rehabilitation are unclear. We aimed to determine whether discharge to inpatient rehabilitation is associated with lower long-term rates of immobility-associated complications following spinal cord injury compared with discharge to a skilled nursing facility.
Methods: This retrospective cohort study leveraged a quasi-experimental design to examine adult spinal cord injury patients in the Healthcare Cost and Utilization Project State Inpatient and State Emergency Department databases. Patients from six states (AR, FL, MA, MD, NY, WI) injured in years 2016-2021were included if they were discharged to either inpatient rehabilitation or a skilled nursing facility following index admission for spinal cord injury as determined using ICD10 codes. Immobility-associated complications of interest were sepsis, type 2 diabetes, venous thromboembolism, lower respiratory tract infection, pressure ulcers, and urinary tract infection. Multivariable, inverse probability of treatment-weighted Fine-Gray competing risk models adjusting for demographics, injury characteristics, and comorbidities estimated adjusted sub-distribution hazard ratios associated with inpatient rehabilitation compared to skilled nursing facility discharge for each immobility-associated complication. Model time horizon was three years with exploratory analysis up to five-year follow-up.
Results: A total of 8,542 patients were identified, of whom 5,223 were discharged to inpatient rehabilitation and 3,319 to skilled nursing facilities. Median age was significantly higher among patients discharged to skilled nursing facilities compared to inpatient rehabilitation (72 vs 58 years, p<0.001). Median and mean follow up time among patients who had at least one emergency department or inpatient encounter were 553 and 703 days, respectively. Patients discharged to inpatient rehabilitation had significantly reduced adjusted hazard ratios for sepsis (HR 0.72, [95%CI 0.62-0.84], p<0.001), type 2 diabetes (HR 0.69, [95%CI 0.49-0.97], p=0.0346), venous thromboembolism (HR 0.72, [95%CI 0.56-0.83], p=0.0123), and pressure ulcers (HR 0.69, [95%CI 0.58-0.83], P<0.001) in the first 3 years after injury. There was no difference in lower respiratory tract infections or urinary tract infections. These results were maintained under exploratory analysis with five-year follow-up.
Conclusions: Inpatient rehabilitation after spinal cord injury is associated with reduced adjusted hazard for sepsis, type 2 diabetes, venous thromboembolism, and pressure ulcers among patients three years after injury, compared to convalescence at a skilled nursing facility.
Methods: This retrospective cohort study leveraged a quasi-experimental design to examine adult spinal cord injury patients in the Healthcare Cost and Utilization Project State Inpatient and State Emergency Department databases. Patients from six states (AR, FL, MA, MD, NY, WI) injured in years 2016-2021were included if they were discharged to either inpatient rehabilitation or a skilled nursing facility following index admission for spinal cord injury as determined using ICD10 codes. Immobility-associated complications of interest were sepsis, type 2 diabetes, venous thromboembolism, lower respiratory tract infection, pressure ulcers, and urinary tract infection. Multivariable, inverse probability of treatment-weighted Fine-Gray competing risk models adjusting for demographics, injury characteristics, and comorbidities estimated adjusted sub-distribution hazard ratios associated with inpatient rehabilitation compared to skilled nursing facility discharge for each immobility-associated complication. Model time horizon was three years with exploratory analysis up to five-year follow-up.
Results: A total of 8,542 patients were identified, of whom 5,223 were discharged to inpatient rehabilitation and 3,319 to skilled nursing facilities. Median age was significantly higher among patients discharged to skilled nursing facilities compared to inpatient rehabilitation (72 vs 58 years, p<0.001). Median and mean follow up time among patients who had at least one emergency department or inpatient encounter were 553 and 703 days, respectively. Patients discharged to inpatient rehabilitation had significantly reduced adjusted hazard ratios for sepsis (HR 0.72, [95%CI 0.62-0.84], p<0.001), type 2 diabetes (HR 0.69, [95%CI 0.49-0.97], p=0.0346), venous thromboembolism (HR 0.72, [95%CI 0.56-0.83], p=0.0123), and pressure ulcers (HR 0.69, [95%CI 0.58-0.83], P<0.001) in the first 3 years after injury. There was no difference in lower respiratory tract infections or urinary tract infections. These results were maintained under exploratory analysis with five-year follow-up.
Conclusions: Inpatient rehabilitation after spinal cord injury is associated with reduced adjusted hazard for sepsis, type 2 diabetes, venous thromboembolism, and pressure ulcers among patients three years after injury, compared to convalescence at a skilled nursing facility.
Competition Category: Health Services Outcomes or Clinical
Mentor: Anne Stey, MD MSc
Margaret Reilly, MD
Senior Resident (Clinical PGY3-5)
Patient characteristics associated with readiness for behavior change in patients with peripheral artery disease
Introduction: Peripheral artery disease (PAD) is a chronic and incurable disease that is impacted by adoption of health-promoting behaviors. This study assesses readiness for behavior change (RBCh) and health beliefs in patients with PAD.
Methods: Patients with PAD were recruited from vascular surgery clinic and completed a survey on demographics, PAD-specific knowledge, and RBCh stage (pre-contemplation, contemplation, preparation, action, and maintenance) in domains of physical exercise and diet. Health beliefs including perceived severity and susceptibility to PAD and perceived barriers to management were also assessed. RBCh scores were compiled into a composite score for all domains. Multivariable logistic and linear regression was used to identify predictors of RBCh, respectively.
Results: Of 105 patients with PAD who completed the survey (mean age 68.5 years, 47.7% female, 30.2% Black, 13.4% active smokers), 77% had undergone leg revascularization and 71.6% scored as activated. Mean composite RBCh score was 17.4/20, indicating a behavior change stage of action. On bivariate analysis, lower PAD knowledge score, non-White race, lower education level, and lower income were associated with poor activation. Lower PAD knowledge score, non-White race, current smoking, and lower education level were associated with lower RBCh stage. In multivariable models adjusting for race, income, health literacy, and education level, lower knowledge score was significantly associated with poor activation (OR=1.1, p=.04), while lower knowledge score and history of revascularization were significantly associated with lower RBCh score (b-coefficient 0.13, p=.04 and b-coefficient 1.5, p=.009, respectively).
Conclusions: Patients with PAD demonstrated overall high activation and RBCh. After adjusting for demographic variables, lower PAD knowledge was associated with poor activation and low RBCh while history of revascularization was significantly associated with lower RBCh. Behavior interventions in PAD should be tailored to level of PAD knowledge and PAD severity.
Methods: Patients with PAD were recruited from vascular surgery clinic and completed a survey on demographics, PAD-specific knowledge, and RBCh stage (pre-contemplation, contemplation, preparation, action, and maintenance) in domains of physical exercise and diet. Health beliefs including perceived severity and susceptibility to PAD and perceived barriers to management were also assessed. RBCh scores were compiled into a composite score for all domains. Multivariable logistic and linear regression was used to identify predictors of RBCh, respectively.
Results: Of 105 patients with PAD who completed the survey (mean age 68.5 years, 47.7% female, 30.2% Black, 13.4% active smokers), 77% had undergone leg revascularization and 71.6% scored as activated. Mean composite RBCh score was 17.4/20, indicating a behavior change stage of action. On bivariate analysis, lower PAD knowledge score, non-White race, lower education level, and lower income were associated with poor activation. Lower PAD knowledge score, non-White race, current smoking, and lower education level were associated with lower RBCh stage. In multivariable models adjusting for race, income, health literacy, and education level, lower knowledge score was significantly associated with poor activation (OR=1.1, p=.04), while lower knowledge score and history of revascularization were significantly associated with lower RBCh score (b-coefficient 0.13, p=.04 and b-coefficient 1.5, p=.009, respectively).
Conclusions: Patients with PAD demonstrated overall high activation and RBCh. After adjusting for demographic variables, lower PAD knowledge was associated with poor activation and low RBCh while history of revascularization was significantly associated with lower RBCh. Behavior interventions in PAD should be tailored to level of PAD knowledge and PAD severity.
Competition Category: Health Services Outcomes or Clinical
Mentor: Karen Ho, MD
John Rode, MD MS
Senior Resident (Clinical PGY3-5)
Age-based disparities in evaluation and utilization of metabolic and bariatric surgery: A real-world cohort study
Introduction: Despite broadly accepted guidelines for early metabolic and bariatric surgery (MBS) referral, small studies suggest disparities in access to surgical care among adolescents and adults. We leveraged large data sources to determine age-specific differences in access to MBS among eligible patients.
Methods: A retrospective analysis was performed using Epic Cosmos, an electronic health record (EHR) database spanning >1,800 US healthcare systems, including 1,266,626 adolescents (12-<20 years) and 34,431,950 adults (20-65 years) with severe obesity from 2018-2022. MBS consultation and surgery rates were calculated by age group, and predictors of MBS were assessed using odds ratios (ORs; 95% CIs).
Results: From 2018-2022, 1.2% of eligible adolescents and 2.7% of adults received MBS consultation. Among those evaluated, 15.9% and 27.8%, respectively, underwent MBS within 24 months (2-fold higher odds for adults). Lower odds of surgery were associated with age <16 years, male sex, or rural residence in adolescents, and age >50 years, Medicare or uninsured status, or lower Social Vulnerability Index in adults. Across age groups, female sex, Hispanic ethnicity, and cardiometabolic comorbidities were more common among those who underwent MBS.
Conclusions: Using large-scale national EHR data, adolescents with severe obesity were substantially less likely than adults to receive MBS consultation or surgery, highlighting the need to improve access to adolescent MBS.
Methods: A retrospective analysis was performed using Epic Cosmos, an electronic health record (EHR) database spanning >1,800 US healthcare systems, including 1,266,626 adolescents (12-<20 years) and 34,431,950 adults (20-65 years) with severe obesity from 2018-2022. MBS consultation and surgery rates were calculated by age group, and predictors of MBS were assessed using odds ratios (ORs; 95% CIs).
Results: From 2018-2022, 1.2% of eligible adolescents and 2.7% of adults received MBS consultation. Among those evaluated, 15.9% and 27.8%, respectively, underwent MBS within 24 months (2-fold higher odds for adults). Lower odds of surgery were associated with age <16 years, male sex, or rural residence in adolescents, and age >50 years, Medicare or uninsured status, or lower Social Vulnerability Index in adults. Across age groups, female sex, Hispanic ethnicity, and cardiometabolic comorbidities were more common among those who underwent MBS.
Conclusions: Using large-scale national EHR data, adolescents with severe obesity were substantially less likely than adults to receive MBS consultation or surgery, highlighting the need to improve access to adolescent MBS.
Competition Category: Health Services Outcomes or Clinical
Mentor: Justin Ryder, PhD
Daniel Romary, MD MSBME
Junior Resident (Clinical PGY1-2)
Using pulmonary angiography disease burden to predict pulmonary vascular resistance and guide surgical candidacy for pulmonary thromboendarterectomy
Introduction: Pulmonary thromboendarterectomy (PTE) is the surgical treatment for chronic thromboembolic pulmonary hypertension (CTEPH). Successful PTE reduces pulmonary vascular resistance (PVR) and subsequently improves functional outcomes. Estimating expected PVR reduction after PTE remains difficult and increases subjectivity of PTE candidacy. We sought to gain insight from our institution's relatively unique practice of obtaining pre- and post-PTE pulmonary angiography (PAg).
Methods: We retrospectively reviewed institutional PTE patients from 2016-2025. Those with both pre- and post- (<14 months) operative PAg with simultaneous right heart catheterization were included. Each segment was graded as totally occluded, partially occluded (affected by web, stenosis, or subtotal occlusion) or patent. Segmental comparison was performed to calculate improvement with PTE. Simple and multivariable linear and logistic regression were used to estimate the relationships between angiography, hemodynamics, and functional outcomes. Maximization of Youden index or r2 was used to identify optimal cut points for the number of residually diseased (totally or partially occluded) segments associated with improved outcomes.
Results: The study included 118 cases. After adjusting for pre-PTE PVR, achieving post-PTE PVR <300 dynes was associated with improved outcomes, as was greater reduction of disease on PAg. Specifically, restricting disease to seven or fewer segments was associated with: lower post-PTE PVR (median 183, IQR [152-240] vs. 279 [187-388] dynes, p<0.001), a lower incidence of RV failure (9% vs. 36%, p=0.001) and RV dilation (17% vs. 35%, p=0.047) at 3-6 months, lower Borg scores (0 [0-0.5] vs. 1.5 [1-3], p<0.001) and less likelihood of requiring BPA (11% vs. 31%, p=0.009). Change in PVR can also be estimated by inputting the pre-PTE PVR, number of improved segments, and whether the patient was on a pulmonary vasodilator pre-PTE into our regression model.
Conclusions: Overall, these results indicate that CTEPH care teams can use PAg findings to guide their decisions on surgical candidacy for PTE. In this single-center experience, reducing disease burden to seven or fewer segments resulted in reduced right heart failure, improved Borg scores, and a decreased need for follow-up BPA. These superior outcomes were driven by a reduction in PVR, the extent of which can be estimated by applying a mathematical model with a small number of inputs.
Methods: We retrospectively reviewed institutional PTE patients from 2016-2025. Those with both pre- and post- (<14 months) operative PAg with simultaneous right heart catheterization were included. Each segment was graded as totally occluded, partially occluded (affected by web, stenosis, or subtotal occlusion) or patent. Segmental comparison was performed to calculate improvement with PTE. Simple and multivariable linear and logistic regression were used to estimate the relationships between angiography, hemodynamics, and functional outcomes. Maximization of Youden index or r2 was used to identify optimal cut points for the number of residually diseased (totally or partially occluded) segments associated with improved outcomes.
Results: The study included 118 cases. After adjusting for pre-PTE PVR, achieving post-PTE PVR <300 dynes was associated with improved outcomes, as was greater reduction of disease on PAg. Specifically, restricting disease to seven or fewer segments was associated with: lower post-PTE PVR (median 183, IQR [152-240] vs. 279 [187-388] dynes, p<0.001), a lower incidence of RV failure (9% vs. 36%, p=0.001) and RV dilation (17% vs. 35%, p=0.047) at 3-6 months, lower Borg scores (0 [0-0.5] vs. 1.5 [1-3], p<0.001) and less likelihood of requiring BPA (11% vs. 31%, p=0.009). Change in PVR can also be estimated by inputting the pre-PTE PVR, number of improved segments, and whether the patient was on a pulmonary vasodilator pre-PTE into our regression model.
Conclusions: Overall, these results indicate that CTEPH care teams can use PAg findings to guide their decisions on surgical candidacy for PTE. In this single-center experience, reducing disease burden to seven or fewer segments resulted in reduced right heart failure, improved Borg scores, and a decreased need for follow-up BPA. These superior outcomes were driven by a reduction in PVR, the extent of which can be estimated by applying a mathematical model with a small number of inputs.
Competition Category: Health Services Outcomes or Clinical
Mentor: Stephen Chiu, MD
Ankita Roy-Adhia, MD
Fellow (Clinical or Postdoctoral Researcher)
Variant lobular carcinoma in-situ of the breast: A single institution review of management & outcomes
Introduction: Lobular carcinoma in situ (LCIS) is a risk marker for the development of invasive carcinoma (IC) in either breast. LCIS includes classic (c-LCIS) and variant LCIS (v-LCIS, i.e. pleomorphic + florid LCIS). Data to guide the clinical management of v-LCIS are scant and outcomes are uncertain; we conducted a retrospective review of experience at Northwestern Medicine (NM).
Methods: The NM Enterprise Data Warehouse was queried for adult patients diagnosed with v-LCIS on core needle biopsy (CNB) from January 2000 - December 2023. Patients diagnosed with IC or DCIS <4cm from the v-LCIS within the previous year were excluded, as were those with concurrent lesions who underwent mastectomy. Clinical, radiological, and pathological variables and ipsilateral recurrences were studied. Positive margins were defined as v-LCIS, IC, or DCIS <2mm of ink. Ipsilateral recurrences were defined as v-LCIS, DCIS, or IC during the follow-up period.
Results: There were 101 patients with pure v-LCIS on CNB that met inclusion criteria. There were 75 pleomorphic, 25 florid, and 6 mixed cases. The median age was 56 years.
The most common mammographic finding was suspicious calcifications in about 80% of the patients. 41 patients underwent ultrasound, and of those, 16 patients presented with a mass. Importantly, there was no significant associations between radiological presentation and upgrade. Of 101 patients who underwent surgical excision, 29% upgraded to DCIS or invasive cancer, similar to current literature with upgrade rates ranging from 25-35%. The remaining 72 patients had no upgrade on final pathology. Only 3 of 29 upgraded patients—10%—experienced local recurrence, compared to 13 of 72 non-upgraded patients—18%. This difference was not statistically significant, but the trend favoring the upgraded group suggests that the more frequent use of radiotherapy and endocrine therapy in the upgraded group was meaningful. 6-year recurrence free survival was 81% for the no-upgrade group and 93% for the upgrade group. The paradoxically lower recurrence in upgraded patients likely reflects the protective effect of adjuvant treatments they received.
Conclusions: In summary, we could not identify any clinical or radiological features that predicted lower risk of upgrade. Thus, excision remains the standard of care for v-LCIS. Our data suggest that variant LCIS may function as a precursor lesion, similar to DCIS (rather than a risk marker) and that its management should be similar to DCIS management. Larger scale studies are needed to verify these results.
Methods: The NM Enterprise Data Warehouse was queried for adult patients diagnosed with v-LCIS on core needle biopsy (CNB) from January 2000 - December 2023. Patients diagnosed with IC or DCIS <4cm from the v-LCIS within the previous year were excluded, as were those with concurrent lesions who underwent mastectomy. Clinical, radiological, and pathological variables and ipsilateral recurrences were studied. Positive margins were defined as v-LCIS, IC, or DCIS <2mm of ink. Ipsilateral recurrences were defined as v-LCIS, DCIS, or IC during the follow-up period.
Results: There were 101 patients with pure v-LCIS on CNB that met inclusion criteria. There were 75 pleomorphic, 25 florid, and 6 mixed cases. The median age was 56 years.
The most common mammographic finding was suspicious calcifications in about 80% of the patients. 41 patients underwent ultrasound, and of those, 16 patients presented with a mass. Importantly, there was no significant associations between radiological presentation and upgrade. Of 101 patients who underwent surgical excision, 29% upgraded to DCIS or invasive cancer, similar to current literature with upgrade rates ranging from 25-35%. The remaining 72 patients had no upgrade on final pathology. Only 3 of 29 upgraded patients—10%—experienced local recurrence, compared to 13 of 72 non-upgraded patients—18%. This difference was not statistically significant, but the trend favoring the upgraded group suggests that the more frequent use of radiotherapy and endocrine therapy in the upgraded group was meaningful. 6-year recurrence free survival was 81% for the no-upgrade group and 93% for the upgrade group. The paradoxically lower recurrence in upgraded patients likely reflects the protective effect of adjuvant treatments they received.
Conclusions: In summary, we could not identify any clinical or radiological features that predicted lower risk of upgrade. Thus, excision remains the standard of care for v-LCIS. Our data suggest that variant LCIS may function as a precursor lesion, similar to DCIS (rather than a risk marker) and that its management should be similar to DCIS management. Larger scale studies are needed to verify these results.
Competition Category: Health Services Outcomes or Clinical
Mentor: Seema Khan, MD
Haroon Salahuddin, BA
Student
Outcomes of pediatric patients after vascular ring repair: Do current repair techniques effectively relieve respiratory and digestive symptoms?
Introduction: No existing studies document outcome differences including symptom persistence and reoperation among vascular ring repair techniques. This study aims to expand upon existing documentation of symptom persistence and reoperation after vascular ring repair while also assessing the impact of different repair techniques.
Methods: This was a single-center retrospective cohort study of patients who underwent vascular ring repair between 2008 and 2025. Respiratory and digestive symptoms were compared between pre-operative, first post-operative, and last post-operative visits. Symptoms and reoperation were also compared between different vascular ring repair techniques.
Results: Among all patients who underwent vascular ring repair and 202 met the inclusion criteria. Of these patients, 67 (33.2%) underwent resection of double aortic arch, 72 (35.6%) division with translocation, 33 (16.3%) ligamentum/PDA division with vascular pexy, 28 (13.9%) ligamentum/PDA division without vascular pexy , and 2 (1.0%) underwent innominate arteriopexy. Overall, symptoms decreased dramatically between pre-operative and first follow up visit with limited incidence of symptom persistence or recurrence at last follow up. No difference in symptom abatement was noted between repair types, however, patients with RAA ALSA (Right Aortic Arch with Aberrant Left Subclavian Artery) that underwent ligamentum/PDA division with or without pexy demonstrated higher re-intervention compared with those that underwent translocation. Of the five RAA ALSA patients that underwent reintervention due to persistent or recurrent symptoms, all had undergone ligation/PDA division with or without pexy initially. Additionally, Kaplan Meier curve analysis demonstrated that the translocation p p reintervention of 100% compared to 55% for the ligamentum/PDA division +/- pexy group in terms of symptom related reinterventions (log-rank p=0.012).
Conclusions: While symptom abatement is similarly successful among vascular ring repair types, reoperation may be avoidable through performing division with translocation in patients with RAA ALS rather than division with pexy alone. In addition, taking a conservative approach in thoracic dissection may reduce incidence long-term reintervention rates in patients undergoing RAA ABLSA repair.
Methods: This was a single-center retrospective cohort study of patients who underwent vascular ring repair between 2008 and 2025. Respiratory and digestive symptoms were compared between pre-operative, first post-operative, and last post-operative visits. Symptoms and reoperation were also compared between different vascular ring repair techniques.
Results: Among all patients who underwent vascular ring repair and 202 met the inclusion criteria. Of these patients, 67 (33.2%) underwent resection of double aortic arch, 72 (35.6%) division with translocation, 33 (16.3%) ligamentum/PDA division with vascular pexy, 28 (13.9%) ligamentum/PDA division without vascular pexy , and 2 (1.0%) underwent innominate arteriopexy. Overall, symptoms decreased dramatically between pre-operative and first follow up visit with limited incidence of symptom persistence or recurrence at last follow up. No difference in symptom abatement was noted between repair types, however, patients with RAA ALSA (Right Aortic Arch with Aberrant Left Subclavian Artery) that underwent ligamentum/PDA division with or without pexy demonstrated higher re-intervention compared with those that underwent translocation. Of the five RAA ALSA patients that underwent reintervention due to persistent or recurrent symptoms, all had undergone ligation/PDA division with or without pexy initially. Additionally, Kaplan Meier curve analysis demonstrated that the translocation p p reintervention of 100% compared to 55% for the ligamentum/PDA division +/- pexy group in terms of symptom related reinterventions (log-rank p=0.012).
Conclusions: While symptom abatement is similarly successful among vascular ring repair types, reoperation may be avoidable through performing division with translocation in patients with RAA ALS rather than division with pexy alone. In addition, taking a conservative approach in thoracic dissection may reduce incidence long-term reintervention rates in patients undergoing RAA ABLSA repair.
Competition Category: Health Services Outcomes or Clinical
Mentor: Allison Davila, MD
Jes Sanders, MD
Senior Resident (Clinical PGY3-5)
Pre-transplant T-cell receptor network analysis can risk stratify and may predict kidney allograft rejection
Introduction: Despite pre-transplant serological screening and HLA matching, 10-15% of kidney allografts experience acute rejection within the first year. Currently, risk stratification for transplantation relies primarily on antibody reactivity to HLA molecules, with no assessment of the T cell compartment before or after transplantation. We have previously used bulk T-cell receptor (TCR) sequencing to identify alloreactive clones pre-transplant and to monitor for these clones over time in a cohort of kidney transplant recipients. The results identified a subset of CD8+ donor reactive clones that expanded in patients with rejection. Using this same cohort, we applied network analysis to TCR repertoires in order to establish a prediction model for rejection.
Methods: Subjects were enrolled as part of a previously completed observational study. Blood, allograft, and urine samples were obtained from subjects at varying time intervals. TCR repertoire analyses were performed using Adaptive Immunosequencing (Adaptive Biologics). Sequence-similarity network analysis was then applied to TCR sequences. Networks were constructed based upon how similar a TCR sequence from one clone was compared to other T-cell clones. Graph metrics were computed across networks with size normalization to control for repertoire size, a known confounder of such analyses. Elastic net, random forest, and gradient boosted models were constructed using a leave one patient out cross validation methodology. The best performing algorithm was selected to predict rejection using a single pre-transplant blood draw. Model performance was also assessed by implementing changes in TCR sequences from pre-transplant to 3 months post-transplant. This study was approved by the Northwestern IRB.
Results: In total, 297 networks were constructed. After adjusting for repertoire size, graft status was the strongest signal for the underlying differences in network metrics. Individuals who rejected the kidney graft generally exhibited more fragmented and less connected networks at baseline. Notably, pre-transplant peripheral blood mononuclear cell network topology alone predicted non-srows outcomes with an AUC of 0.81, sensitivity of 76%, and specificity of 76%. The performance of this prediction model was independent of HLA mismatch, while changes in network topology at three months post-transplantation further improved prediction to an AUC of 0.88 (p = 0.009).
Conclusions: Collectively, TCR sequencing and network analysis represent a potential novel, non-invasive approach for pre-transplant risk stratification and immune monitoring, capturing functional immunological risk that may not be accessible through HLA genotyping or serology.
Methods: Subjects were enrolled as part of a previously completed observational study. Blood, allograft, and urine samples were obtained from subjects at varying time intervals. TCR repertoire analyses were performed using Adaptive Immunosequencing (Adaptive Biologics). Sequence-similarity network analysis was then applied to TCR sequences. Networks were constructed based upon how similar a TCR sequence from one clone was compared to other T-cell clones. Graph metrics were computed across networks with size normalization to control for repertoire size, a known confounder of such analyses. Elastic net, random forest, and gradient boosted models were constructed using a leave one patient out cross validation methodology. The best performing algorithm was selected to predict rejection using a single pre-transplant blood draw. Model performance was also assessed by implementing changes in TCR sequences from pre-transplant to 3 months post-transplant. This study was approved by the Northwestern IRB.
Results: In total, 297 networks were constructed. After adjusting for repertoire size, graft status was the strongest signal for the underlying differences in network metrics. Individuals who rejected the kidney graft generally exhibited more fragmented and less connected networks at baseline. Notably, pre-transplant peripheral blood mononuclear cell network topology alone predicted non-srows outcomes with an AUC of 0.81, sensitivity of 76%, and specificity of 76%. The performance of this prediction model was independent of HLA mismatch, while changes in network topology at three months post-transplantation further improved prediction to an AUC of 0.88 (p = 0.009).
Conclusions: Collectively, TCR sequencing and network analysis represent a potential novel, non-invasive approach for pre-transplant risk stratification and immune monitoring, capturing functional immunological risk that may not be accessible through HLA genotyping or serology.
Competition Category: Basic Science or Translational
Mentor: James Mathew, PhD
Jack Scaife, MD
Junior Resident (Clinical PGY1-2)
Discordance in diagnostic imaging interpretation between pediatric and community radiologist reads in suspected appendicitis
Introduction: For pediatric patients with suspected appendicitis, imaging interpretations from transferring institutions may differ from those of pediatric radiologists. At our freestanding children's hospital, it is standard practice to repeat ultrasounds (US) and re-interpret CT scans by pediatric radiologists. This study aimed to determine the discordance rate between outside hospital (OSH) interpretation and overread CT or repeated US for suspected appendicitis and to assess whether secondary review remains necessary.
Methods: We retrospectively compared OSH and pediatric radiologist interpretations of CT scans from patients with suspected appendicitis evaluated through our center's telehealth service between 2022 and 2025. Inclusion criteria were: (1) outside interpretation available and (2) pediatric overread completed. We also compared OSH US interpretations to repeat US performed at our institution. Discordance was defined as either downgrading from appendicitis or upgrading to appendicitis from an equivocal or normal read/interpretation.
Results: A total of 268 patients met the inclusion criteria for the CT cohort. Discordance between OSH and pediatric interpretations occurred in 35 cases (13%). Demographics were similar between discordant and concordant groups. Patients with discordant CT reads had lower rates of appendectomy (23% vs 91%; p<0.001) and no perforations (0% vs 34%; p<0.001). Discordant cases had significantly lower white blood cell count, absolute neutrophil count, and immature white blood cell percentage. The negative appendectomy rate was higher in discordant reads (25% vs 3.7%; p<0.001). For the US cohort, 110 patients met the inclusion criteria; 42 (38%) had discordant reads between OSH and repeat US. Patients with discordant US reads were less likely to undergo surgery (9.5% vs 85%; p<0.001) and had lower perforation rates (2% vs 24%; p<0.001). Laboratory values again differed significantly between concordant and discordant groups.
Conclusion: Clinically significant discordance exists between outside and pediatric interpretations of both CT and US imaging for suspected pediatric appendicitis. Patients with discordant studies were less likely to have surgery, have complicated appendicitis, or have laboratory findings consistent with appendicitis. These results support the continued value of secondary review by pediatric radiologists at freestanding children's hospitals to ensure accurate diagnosis and avoid unnecessary surgery.
Methods: We retrospectively compared OSH and pediatric radiologist interpretations of CT scans from patients with suspected appendicitis evaluated through our center's telehealth service between 2022 and 2025. Inclusion criteria were: (1) outside interpretation available and (2) pediatric overread completed. We also compared OSH US interpretations to repeat US performed at our institution. Discordance was defined as either downgrading from appendicitis or upgrading to appendicitis from an equivocal or normal read/interpretation.
Results: A total of 268 patients met the inclusion criteria for the CT cohort. Discordance between OSH and pediatric interpretations occurred in 35 cases (13%). Demographics were similar between discordant and concordant groups. Patients with discordant CT reads had lower rates of appendectomy (23% vs 91%; p<0.001) and no perforations (0% vs 34%; p<0.001). Discordant cases had significantly lower white blood cell count, absolute neutrophil count, and immature white blood cell percentage. The negative appendectomy rate was higher in discordant reads (25% vs 3.7%; p<0.001). For the US cohort, 110 patients met the inclusion criteria; 42 (38%) had discordant reads between OSH and repeat US. Patients with discordant US reads were less likely to undergo surgery (9.5% vs 85%; p<0.001) and had lower perforation rates (2% vs 24%; p<0.001). Laboratory values again differed significantly between concordant and discordant groups.
Conclusion: Clinically significant discordance exists between outside and pediatric interpretations of both CT and US imaging for suspected pediatric appendicitis. Patients with discordant studies were less likely to have surgery, have complicated appendicitis, or have laboratory findings consistent with appendicitis. These results support the continued value of secondary review by pediatric radiologists at freestanding children's hospitals to ensure accurate diagnosis and avoid unnecessary surgery.
Competition Category: Health Services Outcomes or Clinical
Mentor: Robert Swendiman, MD
Claire Shen, Student
Introduction: Early recurrence remains prevalent in pancreatic ductal adenocarcinoma (PDAC) despite multimodality treatment. Circulating tumor DNA (ctDNA) is an emerging prognostic biomarker showing promise in the sensitive detection of minimal residual disease (MRD). We evaluate KRAS-mutant ctDNA using a high sensitivity assay for identifying persistent MRD in patients receiving neoadjuvant chemotherapy (NAC), surgery, and adjuvant chemotherapy (AC).Methods: Patients with localized PDAC were enrolled in a phase II observational study (NCT04616131) between October 2020-December 2024 and treated with NAC, surgery, and AC. [AC1.1]Digital droplet PCR probed KRAS G12D, G12V, and G12R mutations and quantified maximum variant allele frequency (mVAF) to assess early recurrence (<12 months). The primary outcome of this study was recurrence-free survival (RFS). Multivariable Cox regressions assessed risk of recurrence or death by KRAS ctDNA detection and by ctDNA clearance, adjusting for age, sex, and clinical stage. Differences in mVAF were assessed by Wilcoxon rank-sum test. Restricted mean survival time (RMST) quantified absolute differences in RFS by ctDNA status in patients receiving AC.Results: Among 55 patients receiving NAC and resection[AC2.1][AC2.2][AC2.3], 197 blood samples were collected. Median NAC duration was 2.6 months, and 96.4% achieved R0 resection. ctDNA was detected a median 7.6 months before imaging recurrence. Detecrows ctDNA after surgery (n=31) independently predicted worse RFS (adjusted hazard ratio [aHR] 3.38 95% CI 1.37-8.37, p=0.009). Failure to clear ctDNA despite NAC and surgery was also prognostic (aHR 4.06, 95% CI 1.00-6.43, p=0.035). Patients with early recurrence had higher mutational burden postoperatively (median 0.09% vs 0.00% mVAF[KS3.1][cs3.2], p=0.006[KS4.1]) and lower mVAF reduction following treatment (-40% vs -97.75%, p=0.046). In 43 patients completing AC, treatment failed tomitigate the poor prognosis of those who were ctDNA-positive after NAC and surgery (12-month RMST 9.7 vs 11.3; restricted mean time lost 1.6 months, p=0.025). Clearance of ctDNA following AC was limited in patients with detecrows ctDNA following NAC and surgery; only 34.8% of patients who were ctDNA-positive after NAC and surgery became ctDNA-negative after AC, whereas ctDNA-negative patients after NAC and surgery largely maintained their negative status (81.8%).Conclusions: Persistent KRAS-mutant ctDNA after NAC and surgery reflects MRD and identifies patients with chemotherapy-resistant disease susceptible to early recurrence despite AC. These findings highlight that AC fails to overcome poor biology of ctDNA-positive disease. Evaluation of ctDNA mutational burden and clearance support quantitative ctDNA monitoring for risk-adapted postoperative management. This high-risk group may benefit most from novel KRAS-directed therapies.
Basic Science or Translational
Competition Category: Basic Science or Translational
Mentor: Chawla MD, https://www.feinberg.northwestern.edu/faculty-profiles/az/profile.html?xid=46916
Joseph Shilati, BA
Student
Shifts in cirrhosis mortality across the COVID-19 pandemic
Introduction: The COVID-19 pandemic altered mortality patterns in the United States, but its impact on cirrhosis-related deaths remains unknown. We examined death trends in patients with cirrhosis before during and after the pandemic.
Methods: Death certificate data were obtained from the WONDER Multiple Cause of Death files, 2016-2024. Alcohol-related cirrhosis was defined as ICD-10 K70.3; non-alcohol-related cirrhosis comprised K74.3-K74.6. We calculated average annual deaths for three periods: pre-pandemic (2016-2019), peri-pandemic (2020-2022), and post-pandemic (2023-2024). Percent changes were computed relative to pre-pandemic means.
Results: Average annual all-cause deaths rose from 2,812,948 pre-pandemic to 3,375,939 during the pandemic (+20.0%) and 3,081,815 post-pandemic (+9.6%). Alcohol-related cirrhosis deaths increased from 22,218 to 30,942 (+39.3%) and 29,536 (+32.9%) across the same intervals, while non-alcohol-related cirrhosis deaths rose from 42,558 to 53,187 (+25.0%) and 54,378 (+27.8%). Although all-cause mortality declined modestly after 2022, cirrhosis-related deaths, especially alcohol-related, remained markedly above pre-pandemic baselines.
Conclusions: Cirrhosis mortality surged disproportionately during the COVID-19 pandemic and did not return to pre-pandemic levels by 2024. Alcohol-related deaths showed the largest relative increase and minimal post-pandemic recovery, consistent with evidence of intensified alcohol consumption, delayed care, and persistent hepatic decompensation in affected individuals. Non-alcohol-related cirrhosis deaths also remained elevated, suggesting enduring behavioral and system-wide disruptions in chronic-disease management. These findings highlight that excess liver-disease mortality is not merely a transient pandemic artifact but a sustained public-health consequence requiring focused prevention, screening, and access strategies.
Methods: Death certificate data were obtained from the WONDER Multiple Cause of Death files, 2016-2024. Alcohol-related cirrhosis was defined as ICD-10 K70.3; non-alcohol-related cirrhosis comprised K74.3-K74.6. We calculated average annual deaths for three periods: pre-pandemic (2016-2019), peri-pandemic (2020-2022), and post-pandemic (2023-2024). Percent changes were computed relative to pre-pandemic means.
Results: Average annual all-cause deaths rose from 2,812,948 pre-pandemic to 3,375,939 during the pandemic (+20.0%) and 3,081,815 post-pandemic (+9.6%). Alcohol-related cirrhosis deaths increased from 22,218 to 30,942 (+39.3%) and 29,536 (+32.9%) across the same intervals, while non-alcohol-related cirrhosis deaths rose from 42,558 to 53,187 (+25.0%) and 54,378 (+27.8%). Although all-cause mortality declined modestly after 2022, cirrhosis-related deaths, especially alcohol-related, remained markedly above pre-pandemic baselines.
Conclusions: Cirrhosis mortality surged disproportionately during the COVID-19 pandemic and did not return to pre-pandemic levels by 2024. Alcohol-related deaths showed the largest relative increase and minimal post-pandemic recovery, consistent with evidence of intensified alcohol consumption, delayed care, and persistent hepatic decompensation in affected individuals. Non-alcohol-related cirrhosis deaths also remained elevated, suggesting enduring behavioral and system-wide disruptions in chronic-disease management. These findings highlight that excess liver-disease mortality is not merely a transient pandemic artifact but a sustained public-health consequence requiring focused prevention, screening, and access strategies.
Competition Category: Health Services Outcomes or Clinical
Mentor: Daniela Ladner, MD MPH
Mariia Shipovskaia, MD
Junior Resident (Clinical PGY1-2)
Impact of timing between stages in three-stage ileal pouch-anal anastomosis on functional and clinical outcomes
Introduction: Three-stage restorative proctocolectomy (RPC) with ileal pouch-anal anastomosis (IPAA) is favored for high-risk patients with ulcerative colitis (UC), yet the optimal timing of completion proctectomy after subtotal colectomy remains undefined.
Methods: This retrospective cohort study conducted in a tertiary academic medical center and analyzed outcomes of patients undergoing three-stage restorative proctocolectomy to evaluate the impact of delays in IPAA creation. Primary outcomes were complications associated with pouch creation within or after 6 months of the Stage 1 procedure. The Stage 1-2 interval was analyzed as a continuous variable (per 3-month increase) and dichotomized at 6 months. Multivariable logistic regression and Firth penalized regression were used. Patient-reported outcomes were assessed using the Ileal Pouch Syndrome Severity Index (IPSSI) and the Cleveland Global Quality of Life (CGQL) questionnaire.
Results: A total of 110 patients who underwent three-stage RPC with IPAA for UC between 2012 and 2025 were included. The median stage 1-2 interval was 5.7 months (IQR 4.2-7.7), and 59 patients (53.6%) were assigned to the ≤6-month group vs 51 patients (46.4%) to the >6-month group. Among patients undergoing stage 2 within 6 months, the median interval was 4.5 months (IQR 3.7-4.9), whereas in those with stage 2 after 6 months, the median interval was 8.1 months (IQR 6.7-14.0). In continuous analysis, the interval was not associated with complications (OR 1.10, 95% CI 0.93-1.31, p = 0.278), pouchitis (OR 0.84, 95% CI 0.62-1.15, p = 0.277), or restoma (OR 0.38, 95% CI 0.13-1.16, p = 0.090). The rate of 90-day Clavien-Dindo grade ≥II complications after stage 2 was similar between the ≤6-month and >6-month groups (27.6% vs 25.5%, p=0.831). In dichotomized analysis, an interval >6 months was associated with reduced odds of restoma formation (OR 0.16, 95% CI 0.03-0.89, p = 0.036) but not with complications or pouchitis. Mean IPSSI scores (42.7 ± 22.5 vs 45.7 ± 28.4, p = 0.683) and median CGQL scores (7.67 vs 7.33, p = 0.802) were comparable between groups. When comparing IPAA creation before and after 9 and 12 months, no significant differences in clinical or functional outcomes were observed.
Conclusions: A longer stage 1-2 interval was not associated with worse outcomes. Although an interval > 6 months was associated with lower odds of restoma formation, this finding should be interpreted with caution. These results support a flexible approach to stage 2 timing.
Methods: This retrospective cohort study conducted in a tertiary academic medical center and analyzed outcomes of patients undergoing three-stage restorative proctocolectomy to evaluate the impact of delays in IPAA creation. Primary outcomes were complications associated with pouch creation within or after 6 months of the Stage 1 procedure. The Stage 1-2 interval was analyzed as a continuous variable (per 3-month increase) and dichotomized at 6 months. Multivariable logistic regression and Firth penalized regression were used. Patient-reported outcomes were assessed using the Ileal Pouch Syndrome Severity Index (IPSSI) and the Cleveland Global Quality of Life (CGQL) questionnaire.
Results: A total of 110 patients who underwent three-stage RPC with IPAA for UC between 2012 and 2025 were included. The median stage 1-2 interval was 5.7 months (IQR 4.2-7.7), and 59 patients (53.6%) were assigned to the ≤6-month group vs 51 patients (46.4%) to the >6-month group. Among patients undergoing stage 2 within 6 months, the median interval was 4.5 months (IQR 3.7-4.9), whereas in those with stage 2 after 6 months, the median interval was 8.1 months (IQR 6.7-14.0). In continuous analysis, the interval was not associated with complications (OR 1.10, 95% CI 0.93-1.31, p = 0.278), pouchitis (OR 0.84, 95% CI 0.62-1.15, p = 0.277), or restoma (OR 0.38, 95% CI 0.13-1.16, p = 0.090). The rate of 90-day Clavien-Dindo grade ≥II complications after stage 2 was similar between the ≤6-month and >6-month groups (27.6% vs 25.5%, p=0.831). In dichotomized analysis, an interval >6 months was associated with reduced odds of restoma formation (OR 0.16, 95% CI 0.03-0.89, p = 0.036) but not with complications or pouchitis. Mean IPSSI scores (42.7 ± 22.5 vs 45.7 ± 28.4, p = 0.683) and median CGQL scores (7.67 vs 7.33, p = 0.802) were comparable between groups. When comparing IPAA creation before and after 9 and 12 months, no significant differences in clinical or functional outcomes were observed.
Conclusions: A longer stage 1-2 interval was not associated with worse outcomes. Although an interval > 6 months was associated with lower odds of restoma formation, this finding should be interpreted with caution. These results support a flexible approach to stage 2 timing.
Competition Category: Health Services Outcomes or Clinical
Mentor: Vitaliy Poylin, MD FACS FASCRS
Mark John Siringan, BA
Student
Phenotyping immune sensitization reveals pattern-based association with post-lung transplant complications
Introduction: Lung transplantation remains the definitive treatment for end-stage lung disease; however, long-term outcomes remain limited, with a median post-transplant survival of approximately six years. Immune sensitization is an important contributor to adverse outcomes after lung transplantation, including impaired pulmonary function, increased risk of chronic lung allograft dysfunction (CLAD), and reduced survival. Prior studies have examined individual immunologic features, such as panel reactive antibodies (PRA), donor-specific antibodies (DSA), acute cellular rejection (ACR), and antibody-mediated rejection (AMR). However, existing literature has evaluated each of these in isolation, creating a gap in understanding regarding the potential for synergistic and overlapping effects.
Methods: In this retrospective cohort study, we developed five immunological sensitization phenotypes based on established alloimmune mechanisms of graft injury: non-sensitized, pre-transplant humoral sensitization, post-transplant humoral sensitization, cellular-mediated sensitization, and mixed sensitization. A total of 502 adult lung transplant recipients were classified into these phenotypes and compared with respect to initial and longitudinal post-transplant FEV1, CLAD risk, and overall survival. Kaplan-Meier, Cox regression models, and restricted mean survival time analyses were conducted to compare post-transplant outcomes. A 6-month landmark sensitivity analysis was also performed to validate our findings against time bias. Patients alive at 6 months post-transplant were re-assigned phenotypes based on immune exposures only occurring before the 6-month landmark and re-evaluated for transplant outcomes.
Results: Across outcomes, the mixed sensitization phenotype consistently demonstrated worse lung function, increased risk of CLAD, and higher mortality compared with other phenotypes. Within the mixed sensitization group, the AMR-subtype was found as the primary contributor to adverse lung outcomes compared to the ACR-subtype. The non-sensitized group did not show a significant difference in post-transplant outcomes to several phenotypes, including pre- and post-transplant humoral or cellular-mediated groups. Additionally, a 6-month landmark sensitivity analysis demonstrated consistent results with the main analysis, particularly showing a significant reduction in mortality of the mixed sensitization group compared to other phenotypes.
Conclusions: These findings indicate that immune sensitization should not be treated as a binary exposure. Rather, sensitization patterns and their composite immunologic features allow more meaningful risk stratification and may better inform post-transplant surveillance and tailored immunomodulatory strategies. Additionally, our findings collectively challenge the practice of prioritizing lung transplant candidates based on sensitization status. Rather than serving as a blanket contraindication, patterns of sensitization may be more appropriately used to identify patients who could benefit from enhanced surveillance following transplantation.
Methods: In this retrospective cohort study, we developed five immunological sensitization phenotypes based on established alloimmune mechanisms of graft injury: non-sensitized, pre-transplant humoral sensitization, post-transplant humoral sensitization, cellular-mediated sensitization, and mixed sensitization. A total of 502 adult lung transplant recipients were classified into these phenotypes and compared with respect to initial and longitudinal post-transplant FEV1, CLAD risk, and overall survival. Kaplan-Meier, Cox regression models, and restricted mean survival time analyses were conducted to compare post-transplant outcomes. A 6-month landmark sensitivity analysis was also performed to validate our findings against time bias. Patients alive at 6 months post-transplant were re-assigned phenotypes based on immune exposures only occurring before the 6-month landmark and re-evaluated for transplant outcomes.
Results: Across outcomes, the mixed sensitization phenotype consistently demonstrated worse lung function, increased risk of CLAD, and higher mortality compared with other phenotypes. Within the mixed sensitization group, the AMR-subtype was found as the primary contributor to adverse lung outcomes compared to the ACR-subtype. The non-sensitized group did not show a significant difference in post-transplant outcomes to several phenotypes, including pre- and post-transplant humoral or cellular-mediated groups. Additionally, a 6-month landmark sensitivity analysis demonstrated consistent results with the main analysis, particularly showing a significant reduction in mortality of the mixed sensitization group compared to other phenotypes.
Conclusions: These findings indicate that immune sensitization should not be treated as a binary exposure. Rather, sensitization patterns and their composite immunologic features allow more meaningful risk stratification and may better inform post-transplant surveillance and tailored immunomodulatory strategies. Additionally, our findings collectively challenge the practice of prioritizing lung transplant candidates based on sensitization status. Rather than serving as a blanket contraindication, patterns of sensitization may be more appropriately used to identify patients who could benefit from enhanced surveillance following transplantation.
Competition Category: Health Services Outcomes or Clinical
Mentor: Chitaru Kurihara, MD
Krishay Sridalla, BA
Student
High-sensitivity ctDNA analysis uncovers relevant signals missed by NGS in pancreatic cancer
Introduction: Pancreatic ductal adenocarcinoma (PDAC) carries high mortality despite multimodal therapy, and improved biomarkers are needed to guide perioperative care. This study evaluated the prognostic significance of KRAS-mutant circulating tumor DNA (ctDNA) detected by next-generation sequencing (NGS) and digital droplet PCR (ddPCR) in localized PDAC.
Methods: In this prospective cohort study (October 2020-October 2024), patients with localized PDAC undergoing neoadjuvant chemotherapy (NAC) were enrolled across multiple sites within Northwestern Medicine (NCT04616131). Blood samples for ctDNA were assessed at diagnosis, post-NAC, and post-resection using tumor-agnostic NGS and ddPCR targeting KRAS G12D/V/R mutations. Log-rank tests and a Cox proportional hazards model assessed factors associated with overall survival (OS).
Results: The cohort included 106 patients. At diagnosis, KRAS ctDNA was detected in 17.2% (17/99) by NGS and 64.9% (63/97) by ddPCR , with higher concordance to matched tumor tissue for ddPCR than NGS (60.6% [20/33] vs 26.5% [9/34]). Using maximum variant allele frequency (mVAF) as a proxy for mutational burden, NGS-derived mVAF remained srows across treatment, whereas ddPCR showed decreased detection and mVAF: 64.9% (63/97) detecrows at diagnosis (median mVAF 0.5%) to 56.4% (31/55) detecrows post-resection (median mVAF 0.2%; p = 0.019). ctDNA detection by either platform was associated with shorter OS, with the higher-sensitivity ddPCR assay providing greater prognostic discrimination by identifying additional patients with poor outcomes not captured by NGS (NGS median OS [mOS] 11.2 vs 30.5 months, p<0.001; ddPCR mOS 24.7 vs 70.9 months, p=0.004). Stratified by detection method, median OS was shortest in patients with ctDNA detected by both NGS and ddPCR (10.9 months), longest in those not detected by either platform (40.7 months), and intermediate in patients detected only by ddPCR (26.9 months; p<0.001). In multivariable analysis, ctDNA detection by NGS (aHR 4.17, 95% CI: 2.00-8.71, p<0.001) and ddPCR (aHR 3.46, 95% CI: 1.49-8.07, p=0.004)
remained independently associated with worse survival, whereas clinical stage and CA 19‑9 were not.
Conclusions: This study showed that integrating higher-sensitivity ctDNA assays with standard NGS-based approaches may provide additional prognostic information in patients with localized PDAC. Although both NGS- and ddPCR-detected KRAS-mutant ctDNA were prognostic at diagnosis, the use of ddPCR uncovered an intermediate-risk subset of patients who were not detected by NGS alone. These findings suggest that extending ctDNA analysis beyond tumor-agnostic NGS and using higher-sensitivity technologies such as ddPCR may better refine risk stratification and inform personalized management strategies, including KRAS-targeted therapies, in patients with PDAC.
Methods: In this prospective cohort study (October 2020-October 2024), patients with localized PDAC undergoing neoadjuvant chemotherapy (NAC) were enrolled across multiple sites within Northwestern Medicine (NCT04616131). Blood samples for ctDNA were assessed at diagnosis, post-NAC, and post-resection using tumor-agnostic NGS and ddPCR targeting KRAS G12D/V/R mutations. Log-rank tests and a Cox proportional hazards model assessed factors associated with overall survival (OS).
Results: The cohort included 106 patients. At diagnosis, KRAS ctDNA was detected in 17.2% (17/99) by NGS and 64.9% (63/97) by ddPCR , with higher concordance to matched tumor tissue for ddPCR than NGS (60.6% [20/33] vs 26.5% [9/34]). Using maximum variant allele frequency (mVAF) as a proxy for mutational burden, NGS-derived mVAF remained srows across treatment, whereas ddPCR showed decreased detection and mVAF: 64.9% (63/97) detecrows at diagnosis (median mVAF 0.5%) to 56.4% (31/55) detecrows post-resection (median mVAF 0.2%; p = 0.019). ctDNA detection by either platform was associated with shorter OS, with the higher-sensitivity ddPCR assay providing greater prognostic discrimination by identifying additional patients with poor outcomes not captured by NGS (NGS median OS [mOS] 11.2 vs 30.5 months, p<0.001; ddPCR mOS 24.7 vs 70.9 months, p=0.004). Stratified by detection method, median OS was shortest in patients with ctDNA detected by both NGS and ddPCR (10.9 months), longest in those not detected by either platform (40.7 months), and intermediate in patients detected only by ddPCR (26.9 months; p<0.001). In multivariable analysis, ctDNA detection by NGS (aHR 4.17, 95% CI: 2.00-8.71, p<0.001) and ddPCR (aHR 3.46, 95% CI: 1.49-8.07, p=0.004)
remained independently associated with worse survival, whereas clinical stage and CA 19‑9 were not.
Conclusions: This study showed that integrating higher-sensitivity ctDNA assays with standard NGS-based approaches may provide additional prognostic information in patients with localized PDAC. Although both NGS- and ddPCR-detected KRAS-mutant ctDNA were prognostic at diagnosis, the use of ddPCR uncovered an intermediate-risk subset of patients who were not detected by NGS alone. These findings suggest that extending ctDNA analysis beyond tumor-agnostic NGS and using higher-sensitivity technologies such as ddPCR may better refine risk stratification and inform personalized management strategies, including KRAS-targeted therapies, in patients with PDAC.
Competition Category: Basic Science or Translational
Mentor: Akhil Chawla, MD
Nikhil Sriram, BA
Student
Financial and structural drivers of cancer care delays in Nigeria: evidence from the COST-FIN trial
Introduction: In low- and middle-income countries (LMICs), delays in cancer diagnosis and treatment initiation are common, contributing to late-stage presentation and poor survival outcomes. While financial navigation (FN) programs have reduced delays in high-income countries, evidence from LMICs is limited. This study examines the baseline prevalence of care delays and associated risk factors among patients enrolled in COST-FIN, the first randomized controlled trial of FN in sub-Saharan Africa.
Methods: Patients with newly diagnosed breast, colorectal, or prostate cancer were randomized to receive FN or usual care. Baseline surveys captured timelines for symptom onset, diagnosis, and treatment initiation, as well as perceived delays and reasons. Multivariable logistic regression assessed patient-, disease-, and system-level factors associated with delays.
Results: Among 169 patients, the median time from symptom onset to diagnosis was 23.9 weeks, and from diagnosis to treatment initiation was 2.6 weeks. Diagnostic delays were reported by 52% of patients (n=88), commonly due to lack of cancer awareness (69%). Colorectal cancer was associated with lower odds of diagnostic delay (OR 0.19). Treatment delays were reported by 24% of patients (n=39), primarily due to treatment-related costs (69%). Greater financial well-being (OR 0.95) and colorectal cancer diagnosis (OR 0.13) were independently associated with lower odds of treatment delay.
Conclusions: Delays in cancer care are common among COST-FIN participants, driven by financial barriers and limited cancer awareness. Financial well-being and cancer type predict delayed care. FN may reduce financial barriers driving care delays. Complementary strategies to increase public cancer awareness in LMICs remains essential. NCT06630962.
Methods: Patients with newly diagnosed breast, colorectal, or prostate cancer were randomized to receive FN or usual care. Baseline surveys captured timelines for symptom onset, diagnosis, and treatment initiation, as well as perceived delays and reasons. Multivariable logistic regression assessed patient-, disease-, and system-level factors associated with delays.
Results: Among 169 patients, the median time from symptom onset to diagnosis was 23.9 weeks, and from diagnosis to treatment initiation was 2.6 weeks. Diagnostic delays were reported by 52% of patients (n=88), commonly due to lack of cancer awareness (69%). Colorectal cancer was associated with lower odds of diagnostic delay (OR 0.19). Treatment delays were reported by 24% of patients (n=39), primarily due to treatment-related costs (69%). Greater financial well-being (OR 0.95) and colorectal cancer diagnosis (OR 0.13) were independently associated with lower odds of treatment delay.
Conclusions: Delays in cancer care are common among COST-FIN participants, driven by financial barriers and limited cancer awareness. Financial well-being and cancer type predict delayed care. FN may reduce financial barriers driving care delays. Complementary strategies to increase public cancer awareness in LMICs remains essential. NCT06630962.
Competition Category: Health Services Outcomes or Clinical
Mentor: Juliet Lumati, MD MPH
Valentina Velasco, BA
Student
Economic hardships and financial distress among cancer patients in Nigeria: Baseline findings from the COST-FIN trial
Introduction: Low and middle-income countries (LMICs) bear a growing share of the global cancer burden, with patients facing substantial financial barriers to care. COST-FIN is a
randomized controlled trial evaluating the impact of financial navigation on financial distress and cancer outcomes among newly diagnosed cancer patients in Nigeria. This study examines economic hardship prevalence and its association with financial distress, prior to intervention.
Methods: Baseline data from COST-FIN were analyzed cross-sectionally. Adults within six weeks of a breast, colorectal, or prostate cancer diagnosis at two tertiary centers completed an 11-item economic-hardship checklist and 11-item COST-FACIT (range 0-44, higher scores indicated less financial distress). Economic hardship was assessed as any hardship (yes/no) and hardship count (0-11). COST-FACIT differences by hardship status were evaluated using Welch t-tests. Multivariable linear regression modeled COST-FACIT as a function of hardship count.
Results: Among 169 participants, 59% reported at least one unmet expense in the prior 12 months; most frequently the inability to afford meals (31%), transportation (29%), and medicines (29%). The mean baseline COST-FACIT score was 14.4 (SD 10.03). Participants reporting any hardship had significantly lower COST-FACIT scores than those without, indicating greater financial distress (mean 10.8 vs 20.8). Each additional hardship was independently associated with a 1.78-point decrease in COST-FACIT (β=-1.78, p <0.001), explaining 23% of variance. Public hospital patients experienced significantly greater financial distress than those at the private facility.
Conclusions: Economic hardship was highly prevalent among Nigerian cancer patients at time of diagnosis and strongly associated with greater financial distress. These findings underscore the importance of addressing basic unmet needs through interventions like financial navigation to mitigate financial toxicity and promote equirows cancer care in LMIC settings.
randomized controlled trial evaluating the impact of financial navigation on financial distress and cancer outcomes among newly diagnosed cancer patients in Nigeria. This study examines economic hardship prevalence and its association with financial distress, prior to intervention.
Methods: Baseline data from COST-FIN were analyzed cross-sectionally. Adults within six weeks of a breast, colorectal, or prostate cancer diagnosis at two tertiary centers completed an 11-item economic-hardship checklist and 11-item COST-FACIT (range 0-44, higher scores indicated less financial distress). Economic hardship was assessed as any hardship (yes/no) and hardship count (0-11). COST-FACIT differences by hardship status were evaluated using Welch t-tests. Multivariable linear regression modeled COST-FACIT as a function of hardship count.
Results: Among 169 participants, 59% reported at least one unmet expense in the prior 12 months; most frequently the inability to afford meals (31%), transportation (29%), and medicines (29%). The mean baseline COST-FACIT score was 14.4 (SD 10.03). Participants reporting any hardship had significantly lower COST-FACIT scores than those without, indicating greater financial distress (mean 10.8 vs 20.8). Each additional hardship was independently associated with a 1.78-point decrease in COST-FACIT (β=-1.78, p <0.001), explaining 23% of variance. Public hospital patients experienced significantly greater financial distress than those at the private facility.
Conclusions: Economic hardship was highly prevalent among Nigerian cancer patients at time of diagnosis and strongly associated with greater financial distress. These findings underscore the importance of addressing basic unmet needs through interventions like financial navigation to mitigate financial toxicity and promote equirows cancer care in LMIC settings.
Competition Category: Health Services Outcomes or Clinical
Mentor: Juliet Lumati, MD MPH
Veronica Villanueva, PhD
Fellow (Clinical or Postdoctoral Researcher)
How old is old? An assessment of age in traumatic brain injury
Introduction: Traumatic brain injury (TBI) affects over 3 million Americans every year leading to subsequent long-term neurocognitive morbidity. Patients over 65 years of age experience increased mortality and greater long-term neurocognitive morbidity compared to young adults after TBI. Our previously published data shows an age-specific influx of CD8+ effector T-cells into the aged brain after TBI. However, when the transition from a “young” outcome to an “aged” outcome after TBI occurs remains unknown. Therefore, we hypothesized that “middle-aged” mice would have increased infiltration of effector T-cells into the brain after TBI as compared to young adult mice.
Methods: C57BL/6 mice were grouped into young adult (12-weeks-old), middle-aged (55-weeks-old), and aged (80-weeks-old) groups (N = 60). Each group underwent TBI via controlled cortical impact or sham injury. Brains were collected one month after injury for analysis via flow cytometry to assess infiltrating immune cells. Data was analyzed using two-way ANOVA and Tukey's multiple comparison posttest; p<0.05 was considered statistically significant.
Results: Flow cytometry showed a significant increase in CD8+ T-cells within the injured brain in the aged mice as compared to young mice (6567 CD8+ T-cells vs. 1400 CD8+ T-cells, p<0.01). The middle-aged group had an increase in CD8+ T-cells similar to the aged mice and significantly greater than the young mice (10650 CD8+ T-cells vs. 1400 CD8+ T-cells, p<0.01). This increase in T-cell infiltration in the middle-aged and aged cohorts was consistent across most investigated T-cell subsets including activated cytotoxic T-cells, effector memory T-cells, and resident memory T-cells. In addition to T-cells, we observed significant increases across all leukocytes in the aged and middle-aged groups as compared to the young group (16598, 10883, and 2400 leukocytes respectively p<0.01).
Conclusions: We hypothesized that middle-aged mice would have increased T-cell infiltration into the brain after TBI as compared to young adult mice. Consistent with our hypothesis, we observed a marked increase in T-cell infiltration into the injured brain of middle-aged and aged mice as compared to young mice. Furthermore, there were no significant differences noted between middle-aged and aged groups suggesting that immune and transcriptional changes associated with aging may start at an earlier age than current research suggests. These data provide novel insight into aging and injury, highlighting the importance of taking age into consideration in clinical trial design.
Methods: C57BL/6 mice were grouped into young adult (12-weeks-old), middle-aged (55-weeks-old), and aged (80-weeks-old) groups (N = 60). Each group underwent TBI via controlled cortical impact or sham injury. Brains were collected one month after injury for analysis via flow cytometry to assess infiltrating immune cells. Data was analyzed using two-way ANOVA and Tukey's multiple comparison posttest; p<0.05 was considered statistically significant.
Results: Flow cytometry showed a significant increase in CD8+ T-cells within the injured brain in the aged mice as compared to young mice (6567 CD8+ T-cells vs. 1400 CD8+ T-cells, p<0.01). The middle-aged group had an increase in CD8+ T-cells similar to the aged mice and significantly greater than the young mice (10650 CD8+ T-cells vs. 1400 CD8+ T-cells, p<0.01). This increase in T-cell infiltration in the middle-aged and aged cohorts was consistent across most investigated T-cell subsets including activated cytotoxic T-cells, effector memory T-cells, and resident memory T-cells. In addition to T-cells, we observed significant increases across all leukocytes in the aged and middle-aged groups as compared to the young group (16598, 10883, and 2400 leukocytes respectively p<0.01).
Conclusions: We hypothesized that middle-aged mice would have increased T-cell infiltration into the brain after TBI as compared to young adult mice. Consistent with our hypothesis, we observed a marked increase in T-cell infiltration into the injured brain of middle-aged and aged mice as compared to young mice. Furthermore, there were no significant differences noted between middle-aged and aged groups suggesting that immune and transcriptional changes associated with aging may start at an earlier age than current research suggests. These data provide novel insight into aging and injury, highlighting the importance of taking age into consideration in clinical trial design.
Competition Category: Basic Science or Translational
Mentor: Steven Schwulst, MD
Jayalekshmi VS, PhD
Fellow (Clinical or Postdoctoral Researcher)
The metaorganismal trimethylamine-flavin containing monooxygenase 3-trimethylamine N-oxide axis regulates endothelial denudation after arterial injury and promotes vascular smooth muscle cell phenotype switching
Introduction: Neointimal hyperplasia is a major cause for failure of interventions such as balloon angioplasty, stenting, and surgical bypass. Gut microbe-derived metabolites from dietary components may contribute to the host arterial remodeling response after injury. Flavin containing monooxygenase 3 (FMO3) catalyzes the oxidation of trimethylamine (TMA), which is produced from the catabolism of dietary choline by gut microbiota, to trimethylamine-N-oxide (TMAO), which is associated with cardiovascular disease. We hypothesize that the TMA/FMO3/TMAO axis regulates endothelial denudation and other cell mechanisms leading to neointimal hyperplasia.
Methods: Fourteen-week-old female transgenic mice overexpressing human FMO3 (hFMO3-Tg+) and sex- and age-matched wild-type (WT) mice were fed a diet supplemented with 1% choline for 3 weeks prior to undergoing left femoral artery wire injury. After 5 days, Evans blue dye was injected intravenously. Bilateral femoral arteries were collected for quantification of endothelial denudation at the site of arterial injury. Plasma was analysed by metabolomics. Cultured human aortic smooth muscle cells (HAoSMC) were treated with 100 µM TMA or 200 µM TMAO. MTT and transwell migration assays were used to assess cell viability and migration. Immunofluorescence for alpha-smooth muscle actin (αSMA) was performed to assess for changes in cell morphology at 3, 5, and 7 days.
Results: Metabolomic analysis demonstrated that plasma TMA was significantly lower (4.3 μM ± 0.5 vs. 360.9 μM ± 49.7; p<0.0001) and TMAO was significantly higher in choline-treated hFMO3-Tg+ (501.4 μM ± 68.92 vs. 368.7 μM ± 14.81; p=0.02) compared to WT mice. Femoral arteries of choline-treated hFMO3-Tg+ mice also had more Evans blue staining (8.6% ± 0.8 vs. 368.7% ± 0.7; p=0.01), indicating more vascular leakage and endothelial denudation. In vitro, HAoSMC treated with TMAO had more viability compared to controls at 24, 48 and 72 hours (p=0.0008) and greater migration (p=0.03) at 6 hours. HAoSMCs treated with TMA and TMAO also exhibited changes in morphology from spindle to hypertrophic appearance with differential expression of αSMA.
Conclusions: The TMA-FMO3-TMAO axis plays an important role in neointimal hyperplasia development by modulating the endothelial and vascular smooth muscle cell responses to injury.ses to injury.
Methods: Fourteen-week-old female transgenic mice overexpressing human FMO3 (hFMO3-Tg+) and sex- and age-matched wild-type (WT) mice were fed a diet supplemented with 1% choline for 3 weeks prior to undergoing left femoral artery wire injury. After 5 days, Evans blue dye was injected intravenously. Bilateral femoral arteries were collected for quantification of endothelial denudation at the site of arterial injury. Plasma was analysed by metabolomics. Cultured human aortic smooth muscle cells (HAoSMC) were treated with 100 µM TMA or 200 µM TMAO. MTT and transwell migration assays were used to assess cell viability and migration. Immunofluorescence for alpha-smooth muscle actin (αSMA) was performed to assess for changes in cell morphology at 3, 5, and 7 days.
Results: Metabolomic analysis demonstrated that plasma TMA was significantly lower (4.3 μM ± 0.5 vs. 360.9 μM ± 49.7; p<0.0001) and TMAO was significantly higher in choline-treated hFMO3-Tg+ (501.4 μM ± 68.92 vs. 368.7 μM ± 14.81; p=0.02) compared to WT mice. Femoral arteries of choline-treated hFMO3-Tg+ mice also had more Evans blue staining (8.6% ± 0.8 vs. 368.7% ± 0.7; p=0.01), indicating more vascular leakage and endothelial denudation. In vitro, HAoSMC treated with TMAO had more viability compared to controls at 24, 48 and 72 hours (p=0.0008) and greater migration (p=0.03) at 6 hours. HAoSMCs treated with TMA and TMAO also exhibited changes in morphology from spindle to hypertrophic appearance with differential expression of αSMA.
Conclusions: The TMA-FMO3-TMAO axis plays an important role in neointimal hyperplasia development by modulating the endothelial and vascular smooth muscle cell responses to injury.ses to injury.
Competition Category: Basic Science or Translational
Mentor: Karen Ho, MD
Grant Wiarda, BA
Student
Visceral artery mapping in patients with aortic aneurysms treated with branched/fenestrated endografts
Introduction: Branched/Fenestrated endografts (B/FEVAR) offer a valuable treatment modality for patients with aortic aneurysms involving the visceral segment aorta. In this study, we investigate the positions of the relevant visceral arteries to determine anatomic patterns and measure variability in arterial positioning. We hypothesize that the anatomic relationships of visceral arteries follow a predicrows pattern and can be used for planning urgent/emergent procedures when commercially available devices are not available.
Methods: An institutional data set of 346 patients who have received B/FEVAR at Northwestern Memorial Hospital was reviewed. Measurements included the longitudinal and arc distances of the celiac and renal arteries relative to the SMA. Descriptive statistics were utilized, and the distributions of each measurement were visualized using histograms.
Results: All visceral artery measurements demonstrated unimodal distributions, suggesting the presence of a recognizable anatomic pattern relative to the SMA. Despite this central tendency, however, patient-to-patient variability in visceral artery location was observed. The mean SMA to celiac longitudinal distance was 19.4mm (st. dev. 5.8mm), with mean longitudinal distances of SMA to left and to right renal arteries of 14.9mm (8.1mm) and 17.1 (10.0mm), respectively. Corresponding arc distances also varied, with an average distance of 3.1mm (2.5mm) for the CA, 15.9mm (4.6mm) for the right renal artery, and 18.2mm (5.4mm) for the left renal artery.
Conclusions: At the population level, a recognizable anatomic framework can be built upon during patient-specific graft design. The variability observed in renal artery anatomy highlights meaningful patient-level variability that cannot be captured by population averages alone. Although moderate variability was present, the position of the visceral arteries remained within a relatively limited range, suggesting they follow a generally predicrows anatomic pattern. Thus, visceral artery positions relative to the SMA demonstrate predicrows relative position and variability.
Methods: An institutional data set of 346 patients who have received B/FEVAR at Northwestern Memorial Hospital was reviewed. Measurements included the longitudinal and arc distances of the celiac and renal arteries relative to the SMA. Descriptive statistics were utilized, and the distributions of each measurement were visualized using histograms.
Results: All visceral artery measurements demonstrated unimodal distributions, suggesting the presence of a recognizable anatomic pattern relative to the SMA. Despite this central tendency, however, patient-to-patient variability in visceral artery location was observed. The mean SMA to celiac longitudinal distance was 19.4mm (st. dev. 5.8mm), with mean longitudinal distances of SMA to left and to right renal arteries of 14.9mm (8.1mm) and 17.1 (10.0mm), respectively. Corresponding arc distances also varied, with an average distance of 3.1mm (2.5mm) for the CA, 15.9mm (4.6mm) for the right renal artery, and 18.2mm (5.4mm) for the left renal artery.
Conclusions: At the population level, a recognizable anatomic framework can be built upon during patient-specific graft design. The variability observed in renal artery anatomy highlights meaningful patient-level variability that cannot be captured by population averages alone. Although moderate variability was present, the position of the visceral arteries remained within a relatively limited range, suggesting they follow a generally predicrows anatomic pattern. Thus, visceral artery positions relative to the SMA demonstrate predicrows relative position and variability.
Competition Category: Health Services Outcomes or Clinical
Mentor: Neel Mansukhani, MD
Andrew Yoon, BA
Student
Linguistic analysis of endocrine surgery fellowship graduates' online biographies
Introduction: A physician's online biography is becoming increasingly more important as they shape patients' perceptions and influence provider selections. The linguistic characteristics of endocrine surgeons, one of the most gender-diverse surgical subspecialties, remain largely unexplored. This study aims to evaluate linguistic patterns among recent endocrine surgery fellowship graduates to examine how communication styles vary by demographic and practice characteristics.
Methods: A retrospective review of American Association of Endocrine Surgeons (AAES) graduates (2014-2024) was conducted. The online biographies were identified by a standardized Google search and uploaded into the Linguistic Inquiry and Word Count (LIWC-22) software, generating scores for “Word Count”, “Analytic”, “Clout”, “Authentic”, and “Tone”. Demographic data included the physician's gender, region of medical school, residency, fellowship, and current practice, institution type, and graduation year.
Results: 256 graduates were identified, and 187 (74%) had publicly available biographies. Gender and medical school region were not associated with differences in linguistic scores. Residency and fellowship regions were minimally associated with different linguistic scores. Current practice region was associated with variation in Clout and Tone: surgeons in the Midwest exhibited lower Clout scores, while surgeons practicing in the South had lower Tone scores. Academic surgeons demonstrated significantly higher Analytic scores and lower Tone scores compared with community surgeons.
Conclusions: Endocrine surgery biographies display meaningful linguistic variation based on practice setting and region, but not gender or training location. Institutional environment and regional communications may influence physicians' online and professional presentation. Understanding these patterns may inform the development of patient-centered, transparent, and effective physician biographies.
Methods: A retrospective review of American Association of Endocrine Surgeons (AAES) graduates (2014-2024) was conducted. The online biographies were identified by a standardized Google search and uploaded into the Linguistic Inquiry and Word Count (LIWC-22) software, generating scores for “Word Count”, “Analytic”, “Clout”, “Authentic”, and “Tone”. Demographic data included the physician's gender, region of medical school, residency, fellowship, and current practice, institution type, and graduation year.
Results: 256 graduates were identified, and 187 (74%) had publicly available biographies. Gender and medical school region were not associated with differences in linguistic scores. Residency and fellowship regions were minimally associated with different linguistic scores. Current practice region was associated with variation in Clout and Tone: surgeons in the Midwest exhibited lower Clout scores, while surgeons practicing in the South had lower Tone scores. Academic surgeons demonstrated significantly higher Analytic scores and lower Tone scores compared with community surgeons.
Conclusions: Endocrine surgery biographies display meaningful linguistic variation based on practice setting and region, but not gender or training location. Institutional environment and regional communications may influence physicians' online and professional presentation. Understanding these patterns may inform the development of patient-centered, transparent, and effective physician biographies.
Competition Category: Education