Citations

Abstract Generative artificial intelligence (AI) is rapidly populating medical records with synthetic or partially AI-generated content, creating a feedback loop where future models are increasingly at risk of training on uncurated AI-generated data. However, the clinical consequences of this AI-generated data contamination remain unexplored. Here, we show that in the absence of mandatory human verification, this self-referential cycle drives a rapid erosion of pathological variability and diagnostic reliability of medical data at population scale. By analysing more than 800,000 synthetic data points across clinical text generation, vision–language reporting, and medical image synthesis, we find that models progressively converge toward generic phenotypes regardless of the model architectures. Specifically, rare but critical findings, including pneumothorax and effusions, vanish from the synthetic content generated by AI models, while demographic representations skew heavily toward middle-aged male phenotypes. Crucially, this degradation is masked by false diagnostic confidence. Models continue to issue reassuring reports while failing to detect life-threatening pathology, with false reassurance rates tripling to 40%. Blinded physician evaluation confirms that this decoupling of confidence and accuracy renders AI-generated documentation clinically useless after just two generations. We systematically evaluate three mitigation strategies that can be easily integrated into existing clinical workflows, finding that while synthetic volume scaling fails to prevent collapse, mixing real data with quality-aware filtering effectively preserves diversity. Ultimately, our results suggest that without policy-mandated human oversight, the deployment of generative AI threatens to degrade the very healthcare data ecosystems it relies upon.

Abstract Background Emerging evidence suggests that premature ventricular contractions (PVCs) may influence atrial electrophysiology and remodeling, potentially leading to higher rates of atrial fibrillation (AF) recurrence. However, existing data is limited to small-scale studies. Therefore, we conducted a retrospective study to evaluate the potential association between PVCs and AF ablation outcomes, which may suggest the need for more aggressive management of ventricular ectopy in this population. Objective This study aims to assess the 5-year outcomes of catheter ablation for AF in patients with and without a diagnosis of PVCs. Methods Using the TriNetX Analytics Research Network, we included patients aged ≥ 18 years who underwent AF catheter between January 1, 2014, and January 1, 2020. Patients were further categorized into two groups based on the presence of a diagnosis of PVCs. Propensity score matching (PSM) was performed using patient demographics, left ventricular ejection fraction (LVEF), cardiac comorbidities, and medications. Study outcomes included all-cause mortality & hospitalization, acute heart failure (HF), and rhythm control interventions during a 5-year follow-up. Results A total of 15,987 patients with PVCs and 40,605 patients without PVCs were identified. After PSM, 15,152 patients from each group were included for further analysis. Compared to control group, patients with PVCs had significantly higher odds of all-cause mortality (odds ratio [OR], 1.138; 95% confidence interval [CI], 1.034–1.253), all-cause hospitalization (OR, 1.300; 95% CI, 1.243–1.360), acute heart failure (HF) (OR, 1.223; 95% CI, 1.162–1.287), and cardiogenic shock (OR, 2.059; 95% CI, 1.751–2.420) following catheter ablation for AF at 5-year follow-up. Additionally, higher odds of AF requiring cardioversion (OR, 1.135; 95% CI, 1.074–1.199) or repeated ablation (OR, 1.164; 95% CI, 1.085–1.249) were observed in the PVCs group. Conclusion Premature ventricular contractions were associated with poorer clinical outcomes and a higher odds of requiring additional rhythm control intervention following catheter ablation for AF. These findings suggest that PVCs may be a marker for worse outcomes after AF ablation.

Introduction Mitral regurgitation (MR) is a common valvular disease associated with complications such as pulmonary hypertension, atrial fibrillation, and heart failure. However, its full impact on the cardiovascular system, especially on right heart function, is not yet fully understood. Understanding this relationship is important because the right ventricle (RV) is critical for maintaining cardiovascular function. Dysfunction of the RV, which may be contributed by conditions like MR, is strongly associated with poor clinical outcomes. Despite its importance, comprehensively studying MR's effect on the RV has been challenging due to the complex, interdependent nature of cardiovascular dynamics, limited patient data, and the difficulty in synthesizing disparate information to clarify the left heart-right heart connection. Methods The primary goal of this study is to investigate the effects of MR on cardiovascular hemodynamics and RV function by integrating 3D models of the left heart with a closed-loop 0D models of the entire cardiovascular system. We further conduct detailed analyses using patient-specific models to explore how various system modifications impact the RV, providing insights into the nuanced effects of MR on the right heart. Results and Discussion This analysis provides several clinically relevant insights. First, progressive MR markedly increases RV afterload and predisposes the RV to dysfunction, even when intrinsic RV contractility is preserved or enhanced. Second, MR-specific severity indices and left-heart metrics alone fail to capture the true burden on the right heart; RV impairment can progress despite stable or only modestly changing MR descriptors. Finally, these findings highlight the need to incorporate direct assessment of RV structure and function into the evaluation of MR, as RV vulnerability plays a critical role in determining patient risk and guiding management decisions.

Abstract Generative artificial intelligence (AI) is rapidly populating medical records with synthetic or partially AI-generated content, creating a feedback loop where future models are increasingly at risk of training on uncurated AI-generated data. However, the clinical consequences of this AI-generated data contamination remain unexplored. Here, we show that in the absence of mandatory human verification, this self-referential cycle drives a rapid erosion of pathological variability and diagnostic reliability of medical data at population scale. By analysing more than 800,000 synthetic data points across clinical text generation, vision–language reporting, and medical image synthesis, we find that models progressively converge toward generic phenotypes regardless of the model architectures. Specifically, rare but critical findings, including pneumothorax and effusions, vanish from the synthetic content generated by AI models, while demographic representations skew heavily toward middle-aged male phenotypes. Crucially, this degradation is masked by false diagnostic confidence. Models continue to issue reassuring reports while failing to detect life-threatening pathology, with false reassurance rates tripling to 40%. Blinded physician evaluation confirms that this decoupling of confidence and accuracy renders AI-generated documentation clinically useless after just two generations. We systematically evaluate three mitigation strategies that can be easily integrated into existing clinical workflows, finding that while synthetic volume scaling fails to prevent collapse, mixing real data with quality-aware filtering effectively preserves diversity. Ultimately, our results suggest that without policy-mandated human oversight, the deployment of generative AI threatens to degrade the very healthcare data ecosystems it relies upon.

Purpose This study assessed the association between holding an additional graduate degree beyond a Doctor of Medicine (MD) or Doctor of Osteopathic Medicine (DO) and residency match outcomes. Specifically, the authors examined graduate degree attainment trends among residency applicants and evaluated whether graduate degree holders experience lower match rates than non-holders. Methods A retrospective, observational analysis was conducted using publicly available data from the National Resident Matching Program's Charting Outcomes in the Match reports (2016-2024). The 10 specialties with the highest applicant volumes were analyzed, encompassing 80 136 matched and 6820 unmatched U.S. MD and DO senior applicants. Trends in graduate degree attainment were evaluated using linear regression models. Match rates were stratified by graduate degree status and compared using two-proportion z-tests based on aggregated matched and unmatched counts for each specialty and overall. Results From 2016 to 2024, the proportion of graduate degree holders entering the match increased significantly for MD (15.77% to 18.29%, P = .0008) and DO (19.93% to 23.31%, P = .0151) applicants. Across all 10 specialties, graduate degree holders matched at lower rates than non-holders. Among MD applicants, graduate degree holders matched at 92.57% compared with 94.12% for non-graduates (P < .0001), with specialty-specific differences ranging from 0.22% in Psychiatry to 4.61% in Anesthesiology; five specialties demonstrated statistically significant disparities. Among DO applicants, graduate degree holders matched at 84.54% compared with 88.64% for non-graduates (P < .0001), with specialty-specific differences ranging from 1.44% in Family Medicine to 10.48% in Obstetrics and Gynecology; five specialties also reached statistical significance. Conclusions The proportion of graduate degree holders in the residency match is increasing, and these applicants experience lower match rates than non-holders. These findings challenge the assumption that additional graduate education provides a competitive advantage and underscore the need for further research into the applicant-level factors contributing to these disparities.

Immunotherapy has been effective in many cancer types but has failed in multiple clinical trials in prostate cancers, with the underlying mechanisms remaining largely unclear. Here, we demonstrate that androgen receptor pathway inhibitor (ARPI) plus irradiation (IR) triggered robust anticancer immunity in prostate cancers in both patients and mice. We show that androgen-activated AR suppressed innate immune signaling by inducing inhibitor of nuclear factor kappa-B kinase subunit epsilon ( IKBKE ) gene repression through HDAC2 interaction with an IKBKE enhancer RNA ( IKBKE eRNA, or IKBKE-e ). ARPI treatment caused IKBKE derepression and enhanced an IR-induced innate immune response via action of RIG-I and MDA5 dsRNA sensors. IKBKE-e ablation largely enhanced innate immunity in prostate cancer cells in culture and anticancer immunity in mice. Our results revealed AR, HDAC2, and IKBKE eRNA as critical intrinsic immune suppressors in prostate cancer cells, suggesting that rejuvenating inhibitor of nuclear factor kappa-B kinase subunit epsilon (IKKε) signaling by targeting IKBKE-e is an actionable strategy to elicit synthetic anticancer immunity in immunologically “cold” cancers such as prostate cancer.

837 Background: A recent randomized trial suggested that CR may influence ICI activity in patients with lung cancer. However, the same has not been established in GI cancers, where reports have shown mixed results with variable definitions of 'early' vs 'late' IT. We evaluated multiple approaches to assess the impact of CR on first-line ICI in GI malignancies. Methods: This retrospective study included patients from three Mayo Clinic sites (N=455). Both previously reported IT cutoffs and a new cutoff derived from ROC analysis were evaluated. Patients were categorized by the median of all ITs, the median of the first four ITs, and whether ≥20% of ITs occurred before the identified cutoff. Univariate and multivariate Cox regression (adjusting for age, MSI status, type of cancer, metastasis status, number of cycles, and receipt of chemotherapy with ICI) assessed associations with OS. KM analyses were performed to illustrate survival differences. The primary endpoint was OS. Results: When patients were categorized based on whether ≥20% of infusions occurred before 14:00 (ROC-derived cutoff), a significant association with OS was observed. On univariate analysis, early infusion timing was associated with improved survival (HR 2.30, 95% CI 1.58–3.32, p <0.001). This remained significant after multivariable adjustment (HR 2.11, 95% CI 1.41–3.16, p < 0.001). Median OS was 14.2 months in the early group vs 3.8 months in the late group (log-rank p < 0.001). Using previously defined cutoffs, categorization by the median IT of all infusions or the median of the first four infusions did not demonstrate a significant association with OS in either univariate or multivariate Cox regression analyses. Conclusions: These findings highlight the potential role of infusion timing as a chronotherapy factor in GI cancers. Given the retrospective design and heterogeneous population, prospective studies are warranted to validate this signal and clarify its clinical relevance. Baseline demographics of patients by infusion timing (early vs late). Variables Early N=412 (%) LateN=43(%) Total N=455(%) p-value Age <70 249(60.4) 21(48.8) 270 (59.3) 0.1 >70 163(39.6) 22(51.2) 185 (40.7) Gender F 118(28.6) 16(37.2) 134 (29.5) 0.3 M 294(71.4) 27(62.8) 321 (70.5) Cancer Type Gastroesophageal 109 (26.5) 8 (18.6) 117 (25.7) 0.1 Hepatocellular 159 (38.6) 24 (55.8) 183 (40.2) Biliary 86 (20.9) 4 (9.3) 90 (19.8) MSI-H 58 (14.1) 7 (16.3) 65 (14.3) Metastasis No 150 (36.4) 22 (51.2) 172 (37.8) 0.1 Yes 262 (63.6) 21 (48.8) 283 (62.2) Immunotherapy Cycle Number <6 189 (45.9) 33 (76.7) 222 (48.8) <0.001 >6 223 (54.1) 10 (23.3) 233 (51.2)

830 Background: Immune checkpoint inhibitors (ICIs) are standard therapy for MSI-H GI cancers. Lynch syndrome, caused by germline mutations in mismatch repair (MMR) genes, leads to MSI-H tumors and robust ICI responses, but its effect on irAEs remains poorly defined. Clarifying this risk may inform treatment selection and toxicity monitoring. Methods: We retrospectively reviewed MSI-H or MMR-deficient (dMMR) GI cancer patients (pts) treated with ICIs across Mayo Clinic (AZ, MN, FL) from January 2017 to June 2025. Eligible pts were identified using Mayo Data Explorer. Demographics, treatment, and irAEs (graded per CTCAE v5.0) were collected. Lynch status was determined by germline testing when available. Logistic regression estimated odds ratios (ORs) for any grade and grade ≥3 irAEs, adjusting for age, cancer type, prior chemotherapy, and treatment regimen. Results: A total of 209 pts with dMMR/MSI-H GI cancer were included: 61 with Lynch syndrome and 148 without. Lynch pts were younger (median 58 vs 76 years, p < 0.001) and more likely to have received prior chemotherapy (62% vs 43%, p = 0.009). Colorectal cancer was the predominant cancer type (74%), but Lynch pts had proportionally more non-colorectal GI cancers (15% vs 2.7%, p = 0.007). Overall, 29% of pts experienced an irAE, with rates significantly higher in Lynch vs non-Lynch pts (43% vs 24%, p = 0.005). The most frequent irAEs were gastrointestinal, endocrine, and dermatologic, followed by hepatic and pulmonary events. Grade ≥3 irAEs occurred in 16% overall, with no significant difference between groups (p = 0.124). In adjusted models, Lynch syndrome was independently associated with a higher risk of any grade irAE (OR 2.14, 95% CI 1.09 - 4.19, p = 0.026), while risk of severe irAEs was not increased (OR 1.55, 95% CI 0.64 - 3.59, p = 0.316). Lynch syndrome pts were more likely to have combination therapy (17% vs 6.8%, p = 0.029). However, there were no differences observed in any grade irAE between monotherapy and combination immunotherapy (OR 1.36, 95% CI 0.49 - 3.54, p = 0.535). Conclusions: Lynch syndrome is associated with a higher prevalence of any-grade irAEs among dMMR/MSI-H GI cancer pts treated with ICIs, though severe events were not increased. Incorporating Lynch status into immunotherapy planning may improve risk stratification and guide regimen selection.

103 Background: Clinicians spend significant time extracting and synthesizing information to risk-stratify and make adjuvant chemotherapy (AC) recommendations for CRC. By automating parts of this process, LLMs could help reduce documentation burden and variability in decision-making. Methods: We selected 144 stage III, pMMR CRC patients treated at Mayo Clinic (2017–2021). HIPAA-compliant Gemini-2.0-Flash-001 and GPT-4o extracted clinical variables (age, neuropathy grade, pathological stage) from EHRs and generated AC recommendations. Two approaches were compared: (1) a rule-based algorithm with nested if/else statements translating variables into NCCN-based recommendations, and (2) a dynamic LLM agent generating recommendations using a structured prompt informed by NCCN guidelines. Concordance with treating physicians was assessed, with disagreements reviewed by a blinded GI oncologist. Results: The dataset included 144 patients with median age 67 (IQR 66–68), 48.6% female. The LLM pipeline extracted neuropathy grade with 100% accuracy and pathological T/N stages with 99.3% accuracy. Concordance with physician recommendations was 93.1% for the rule-based algorithm and 62.5% for the LLM agent. In discordant cases, blinded adjudication by a GI oncologist favored the algorithm and LLM agent over physicians in most cases (Table 1). Conclusions: This exploratory work shows LLM agents can reliably extract variables and generate evidence-based treatment recommendations for stage III CRC. The high rate of expert preference for LLM recommendations in discordant cases is intriguing but limited by small sample size and possible adjudicator bias. Discordant physician recommendations reflect the complexity of decision-making, including patient preferences, performance status, comorbidities, and subtle findings not fully captured in EHRs. Rather than replacing judgment, we envision LLMs as decision support tools—analogous to an advanced trainee reviewing charts and presenting evidence-based recommendations to the attending physician with ultimate authority. Larger multi-institutional studies with multiple adjudicators are needed to validate these findings and clarify AI’s role in oncology. Patient characteristics, LLM extraction accuracy, and concordance with physician recommendations. Metric Result Total patients 144 Median age (IQR) 67 (66-68) Female (%) 48.6% LLM extraction accuracy – Neuropathy grade 100% LLM extraction accuracy – Pathological T and N stages 99.3% Primary Concordance – Rule-based algorithm vs. physician 93.1% Adjudicator preference of rule-based algorithm over treating physician for discordant cases 100% Concordance – LLM agent vs. physician 62.5% Adjudicator preference of LLM agent over treating physician for discordant cases 82.0%

Abstract Background The incidence of type 2 diabetes mellitus among young adults has been increasing over the past few decades. There are limited data on contemporary national cardiovascular mortality rates in young adults with diabetes mellitus (DM). Methods We queried the Centers for Disease Control and Prevention's Wide‐Ranging Online Data for Epidemiologic Research database for patients aged 15 to 44 years from 1999 to 2019 in the primary analysis. A secondary analysis included data from 2020 to 2022 to examine the effect of the COVID‐19 pandemic. We analysed patients in whom cardiovascular diseases (CVD) were listed as the main cause of death and DM as a contributing cause of death. In calculating age‐adjusted mortality rates (AAMR) per 100 000 individuals, we used CVD mortality in young adults with comorbid diabetes as the numerator and the overall population as the denominator. We determined temporal trends by estimating the average annual percent change (AAPC) using the Joinpoint regression program. Results Among 3 309 079 individuals aged 15–44 years who died, 30 978 deaths were due to CVD with comorbid DM listed as a contributing cause. The overall AAMR increased from 1.08 per 100 000 individuals in 1999 to 1.23 per 100 000 individuals in 2019, with an average APC of 0.75 (95% CI, 0.31–1.20). Compared with 1999, the percentage increase in CVD deaths in 2019 among young adults with comorbid DM was +6.2%, whereas a decrease of 19.9% was observed among those without comorbid DM. Males had a higher AAMR than females (1.60 vs. 0.97). Non‐Hispanic Black individuals and non‐Hispanic American Indian/Alaska Native individuals had the highest AAMRs of 2.99 and 2.89, respectively. The AAMR was higher in rural regions compared to urban areas (1.63 vs. 1.23). The overall AAMR increased markedly from 1.23 in 2019 to 1.60 in 2022, with a similar pattern observed across demographic subgroups. Conclusion Our study reveals a rising trend in CVD mortality in young adults with DM as a contributing cause. Males, non‐Hispanic Black individuals, and individuals from rural regions had higher AAMR than their counterparts. This warrants the development of specific healthcare policies aimed at these at‐risk populations. Lay summary Our study shows that cardiovascular disease (CVD) deaths are increasing in young adults with diabetes mellitus listed as a contributing cause.