Citations

Abstract Inflammatory bowel disease (IBD) is a chronic immune-mediated disorder of the gastrointestinal tract whose genetic basis is only partly resolved because most risk variants identified by genome-wide association studies (GWAS) lie in non-coding regions, limiting direct gene assignment and biological interpretation 1,2 . Here we analysed whole-exome and whole-genome sequencing data from 86,213 IBD cases and 478,363 controls of European ancestry. We identified 68 IBD genes directly implicated by conditionally independent protein-coding associations across the allele frequency spectrum. Many newly implicated IBD genes are supported by orthogonal genomic or pleiotropic evidence, pointing to disease-related pathways and nominating targets with therapeutic relevance. We further identified allelic series and non-additive effects at key loci such as NOD2 and TYK2. These results show that large-scale sequencing can resolve disease genes and pathways that remain ambiguous from non-coding association alone, providing a more direct route from human genetics to biological insight and therapeutic hypotheses.

Background: Pancreatitis significantly alters the microbial composition of the oral and intestinal compartments, causing dysbiosis that may contribute to disease mechanisms and potentially serve as a basis for diagnosis or treatment. Objective: To determine whether the oral or gut microbial signature can classify chronic pancreatitis (CP). Methods: Stool samples (n=707) were collected from participants in the Prospective Evaluation of Chronic Pancreatitis for Epidemiologic and Translational Studies (PROCEED). Samples were distributed among 200 healthy (HC), 310 CP, 49 acute pancreatitis (AP), and 148 recurrent acute pancreatitis (RAP). In addition, saliva samples were collected for a subset of participants (n=156). Whole genome sequencing was performed to assess microbiome composition. Machine learning algorithms were utilized to identify a signature with microbial features predictive of CP. Results: Gut alpha diversity was significantly decreased in AP, RAP, and CP compared with HC, with CP exhibiting the lowest diversity. In contrast, oral microbial diversity showed no significant variation across groups. Beta diversity analysis revealed distinct gut microbiome compositions between HC and pancreatitis subtypes, with CP showing the most pronounced differences. Random forest models using gut microbial species demonstrated robust predictive performance for CP using a minimum of 10 species (Area under the curve—AUC: 0.834; accuracy: 0.774). Despite similarities in gut microbiome composition across pancreatitis subtypes, a unique gut microbial signature for CP was identified highlighting the microbiome’s potential in CP diagnosis. Conclusion: Our study reveals a gut microbial signature predictive of CP using machine learning models in a large US multi-institutional cohort.

Pancreatic ductal adenocarcinoma (PDAC) is an aggressive malignancy with a 5-year survival rate of 13.3%. First-line treatment relies on two chemotherapy regimens, FOLFIRINOX (FOLFNX) or gemcitabine plus nab-paclitaxel (GEMPAC). However, direct clinical comparisons between these regimens have yielded inconsistent results across survival and toxicity endpoints, and the molecular basis of heterogeneous treatment responses remains poorly defined. To investigate regimen-specific tumor-cell-intrinsic mechanisms, we performed quantitative proteomic profiling of a primary PDAC-derived MIA PaCa-2 cell line following treatment with FOLFNX or GEMPAC. Differentially expressed proteins were analyzed using Gene Ontology, KEGG, and Ingenuity Pathway Analysis to define pathway-level alterations, and findings were contextualized using TCGA transcriptomic data. Proteomic analyses revealed that FOLFNX and GEMPAC engage in distinct cytotoxic programs. FOLFNX predominantly suppressed ribosome biogenesis and mitochondrial translation, consistent with sustained metabolic and biosynthetic stress, whereas GEMPAC preferentially disrupted mitotic cytokinesis and phosphatidylinositol phosphate biosynthesis, consistent with mitotic failure. Integration with TCGA data showed that FOLFNX-altered proteins aligned with favorable prognostic expression signatures, whereas GEMPAC-associated proteins were enriched among adverse profiles, reflecting engagement of distinct tumor-intrinsic programs. Together, these findings provide mechanistic insight into differential chemotherapy responses and establish a foundation for proteomics-based biomarkers to guide personalized chemotherapy selection in PDAC.

Abstract Objectives Traditional laboratory regulatory frameworks provide robust guidance for conventional clinical testing but lack requirements for data documentation for artificial intelligence and machine learning (AI/ML) solutions. This gap creates significant risks since AI solutions directly inherit patterns, biases, and limitations from their training data. Here, we describe the development and implementation of a comprehensive data documentation checklist with accompanying data lineage for use within an AI lifecycle framework for clinical laboratories. Methods Building on the previously established Clinical AI Readiness Evaluator (CARE) framework, we developed a comprehensive data checklist and data lineage methodology through a multiphase process, including (1) comprehensive review of existing AI/ML data documentation frameworks, (2) focused meetings with 3 institutional AI operations teams, and (3) 3 rounds of iterative refinement by our multidisciplinary team. The checklist’s effectiveness was then assessed using 3 diverse AI/ML projects moving through the CARE framework. Results The CARE Data Checklist and Data Lineage provide a structured approach to documenting critical aspects of datasets used in AI/ML projects, including data composition, demographics, collection methods, labeling processes, usage constraints, maintenance requirements, and a data readiness assessment. The checklist addresses unique data-centric challenges of AI/ML applications, facilitating transparency, reproducibility, and regulatory compliance. Conclusions The CARE Data Checklist and Data Lineage serve as both a technical guide and a communication tool bridging gaps between technical and clinical stakeholders. By working on these documents early in the AI lifecycle, laboratories can anticipate and address data-related challenges, ultimately saving time, optimizing resources, and improving the reliability of AI-augmented laboratory solutions.

Abstract Recent work on double descent has challenged classical bias–variance tradeoffs, showing that test error can decrease, increase sharply near the interpolation threshold, and then decrease again as model capacity grows. This phenomenon has been documented in regression and classification, but its relevance to survival analysis remains unclear. Survival data are subject to censoring, which obscures true event times, and widely used models such as the Cox proportional hazards model are optimized via partial likelihoods that emphasize ranking rather than calibrated risk estimation. It is therefore unknown whether double descent occurs in this setting, how censoring influences its manifestation, or how it interacts with standard performance metrics.We investigate these questions using synthetic survival data generated from Weibull hazards with controlled censoring, allowing systematic variation of model capacity from under to over parameterized regimes. While we verified double descent occurs in survival models, calibration plateaus and decouples from discrimination, even under strong $\ell_2$ regularization. This decoupling arises because the Cox partial likelihood optimizes rankings rather than magnitudes, producing extreme risk scores that break the Breslow estimator used to estimate survival probabilities. These results highlight limitations of discrimination-based model selection in survival analysis and underscore the need for calibration-aware evaluation in high-capacity prognostic models.

Abstract Hypertensive disorders of pregnancy (HDP) are a leading cause of maternal and perinatal morbidity, yet modifiable environmental risk factors remain poorly characterized. Prior studies typically have only examined a limited number of exposures and have rarely distinguished HDP subtypes (ie, gestational hypertension, preeclampsia, eclampsia, and chronic hypertension with or without superimposed preeclampsia) or accounted for residential mobility during pregnancy. To address these gaps, we conducted a spatial and contextual exposome study of HDP using linked electronic health records (EHR) and vital statistics data in Florida. We analyzed 686 412 singleton pregnancies conceived between 2013 and 2018, using computable phenotyping to distinguish HDP subtypes. A total of 245 spatial and contextual exposome measures spanning natural, built, and social environments were linked to residential histories from conception through gestational week 19. Using a two-phase double machine learning (DML) framework with exposure-specific, directed acyclic graph-guided confounder adjustment, we conducted discovery and replication analyses, followed by multi-treatment DML to estimate effect sizes. In Phase 1, 26 exposome measures replicated for gestational hypertension and 34 for overall HDP. In Phase 2, 12 measures remained associated with gestational hypertension and 11 with overall HDP, including air toxicants, meteorological factors, ultraviolet radiation, neighborhood crime indicators, environmental noise, and proximity to coastline. No exposures passed multiple-comparison thresholds for preeclampsia or eclampsia. These findings demonstrate that the spatial and contextual exposome contributes to HDP in a subtype-specific manner. Integrating EHR-linked phenotyping, residential mobility, and causal machine-learning methods offers a scalable framework for identifying environmental factors relevant to HDP prevention.

Abstract Background: While synoptic reports have standardized breast cancer data, a wealth of critical information remains locked in unstructured free-text, including nuanced biomarker details and procedural notes. This data is not easily computable, creating a significant barrier to understanding comprehensive population-level statistics and subtle clinical trends. Traditional data extraction methods are often insufficient to capture this complexity. We developed and validated a novel, flexible AI-powered system using an agent-based architecture to extract, normalize, and discretize this valuable unstructured data, enabling previously infeasible population-scale analytics. Methods: Our system employs a modular, agent-based framework where each "agent" is a specialized large language model (LLM) tasked with a specific function in a sequential workflow. The process begins with an initiator agent that parses the synoptic report. Subsequent agents then perform targeted tasks, such as identifying and normalizing tumor characteristics, procedural details, and biomarker information from free-text fields. A final agent aggregates this information into a structured, computable format. To rigorously test our approach, we first developed a synthetic data generator to create realistic, anonymized report snippets with known ground-truth values. We then evaluated the performance of several leading LLMs of various sizes (e.g., gemini, llama3.3, phi4, deepseek-r1, gemma3) on this synthetic dataset. Results: Our early analysis on simulated synoptic reports demonstrated high accuracy in data extraction and normalization. The gemini-1.5-pro-002 model achieved an overall accuracy of 96.9%, with a sensitivity of 99.3%, specificity of 94.3%, and an F1-score of 97.0%. The llama3.3:70b model also showed strong performance with a 98.7% accuracy, (F1: 98.8%, sensitivity: 99.8%, specifically: 97.6%). Gemma3:27b was the least performant model tested (F1: 66.2%), substantially worse than much smaller models (deepseek-r1:1.5b and phi4:14b). A detailed error analysis of false positives and negatives has allowed us to iteratively refine the agent logic. By the time of the symposium, we will have analyzed a large cohort of real, de-identified synoptic reports and will present comprehensive performance metrics and initial findings on population-level trends. Conclusion: This AI-powered, agent-based system offers a powerful and scalable solution to overcome the limitations of structured data in synoptic reports. By accurately extracting and normalizing previously inaccessible free-text information, our approach has the potential to provide clinicians and researchers with near real-time insights into population-level statistics and trends in breast cancer. This will facilitate a more profound understanding of disease patterns and ultimately support evidence-based advancements in patient care and clinical research. Citation Format: S. N. Hart. A Novel Agent-Based AI System to Unlock Unstructured Data in Synoptic Reports for Advanced Population Health Analysis in Breast Cancer [abstract]. In: Proceedings of the San Antonio Breast Cancer Symposium 2025; 2025 Dec 9-12; San Antonio, TX. Philadelphia (PA): AACR; Clin Cancer Res 2026;32(4 Suppl):Abstract nr PS3-06-11.

Abstract Objectives To develop and validate an agent-based Large Language Model (LLM) system for extracting structured data from breast cancer synoptic pathology reports and assess the performance gap between synthetic and real-world validation. Materials and Methods We developed a modular artificial intelligence (AI) agent-based framework employing sequential specialized LLMs for parsing pathology reports and extracting structured data. We normalized College of American Pathologists (CAP) cancer protocols into 8 sections, 86 subsections, and 229 discrete fields. Seven leading LLMs (gemini-2.5-pro, llama3.3-70b, phi4-14b, deepseek-r1 14B/70B, gemma3-27b, gemini-2.0-flash-lite) were validated using dual evaluation: synthetic validation (864 controlled test cases) and real-world ground truth (6651 annotated fields from 90 pathology reports). Results Synthetic validation demonstrated strong performance (accuracy: 93.8%-99.0%). Real-world evaluation revealed field extraction recall ranging from 61.8% to 87.7%, demonstrating a substantial “reality gap” with performance drops of 11-32 percentage points. The gemini-2.5-pro model achieved the highest real-world recall (87.7%). Model size did not predict performance: the 14B-parameter deepseek-r1 (77.6%) outperformed its 70B-parameter counterpart (70.4%). Discussion The substantial performance degradation from synthetic to real-world data underscores the complexity of authentic clinical documentation. Smaller models can achieve competitive or superior recall, reducing computational costs. With even the best models missing 12%-38% of annotated fields, mandatory human verification is essential for clinical deployment. Conclusion While LLM-based extraction systems show promise for pathology data extraction, synthetic validation alone provides false confidence. Rigorous real-world ground truth evaluation with expert annotation is essential before clinical deployment. These systems are best positioned as screening tools with mandatory human oversight rather than autonomous decision-making systems.

Abstract Background Icotrokinra is the first targeted oral peptide that selectively blocks the IL-23 receptor (IL-23R) and precisely inhibits the IL-23 pathway. In ANTHEM-UC—a Phase 2b, randomised, double-blind, placebo (PBO)-controlled, dose-ranging study in adults with moderately to severely active ulcerative colitis (UC)—all doses of icotrokinra achieved the primary endpoint of clinical response and delivered clinically meaningful improvements across key secondary endpoints at Week (W)12,1 which were maintained or improved at W28.2 Here we report the impact of icotrokinra on systemic and tissue biomarkers of inflammatory burden in UC. Methods Participants were randomised (1:1:1:1) to receive once-daily icotrokinra 100 mg, 200 mg, 400 mg, or PBO. Clinical inflammatory biomarkers, C-reactive protein (CRP) and faecal calprotectin (FCP), along with serum IL-23 pathway biomarkers (IL-22, IL-17A, and IL-17F) were used to determine the impact of icotrokinra on systemic inflammation. Gene expression changes in colonic biopsies were used to confirm tissue IL-23 pathway engagement and dampening of tissue inflammation with icotrokinra. Results At the W12 primary timepoint, all doses of icotrokinra attenuated levels of serum CRP and FCP (Figure 1). All doses of icotrokinra, but not PBO, markedly reduced serum levels of IL-22, IL-17A, and IL-17F at W12 relative to baseline. Greater reductions in these serum biomarkers of IL-23 pathway activation were observed with the 200 mg and 400 mg doses relative to the 100 mg dose (Figure 1, nominal p < 0.05). All doses of icotrokinra, but not PBO, markedly reduced expression of gene sets associated with the IL-23–driven inflammation in tissue at W12 relative to baseline (Figure 2, nominal p < 0.0001). With icotrokinra treatment, the tissue transcriptome more closely resembled that of healthy individuals, with decreased expression of gene sets associated with UC-relevant inflammatory cell types and increased expression of genes characteristic of healthy epithelial cells. Importantly, dampening of systemic and tissue inflammation with icotrokinra was sustained at W28 (Figures 1 & 2). Conclusion We show for the first time that blocking the IL-23R can inhibit IL-23-mediated inflammatory signaling in UC patients. Icotrokinra treatment induced a sustained (W28) reduction in systemic and tissue biomarkers of inflammatory burden in UC. Furthermore, icotrokinra induced transcriptional changes that are consistent with dampening of IL-23–driven inflammation and normalization of the tissue transcriptome. While all once-daily doses of icotrokinra attenuated systemic and tissue inflammation, greater reduction of systemic inflammatory burden was observed with the 200 mg and 400 mg doses of icotrokinra. References: 1. Abreu MT, et al. United European Gastroenterol J. 2025 10;13(suppl 8):172-3. 2. Jairath V, et al. Am J Gastroenterol. 2025 120;10(suppl 2):S312. Conflict of interest: Louis, Edouard: Education and Research Grants for my department: Abbvie, Takeda, Janssen, Pfizer, Fresenius-Kabi, Celltrion, EG pharma, Sandoz, Falk Personal Fees for conferences, advisory boards and consultancy: Abbvie, Takeda, Ferring, Pfizer, Janssen, Lilly, Galapagos, Celltrion, Arena, BMS, Falk, Biokuris, Fresenius-Kabi, Thabor Jairath, Vipul: Consulting fees from AbbVie, Alimentiv, Amgen, Anaptys Bio, Asahi Kasei, Asieris, Astra Zeneca, Attovia, Blackbird Labs, Bristol Myers Squibb, Boehringer Ingleheim, Biomebank, Caldera, Calluna, Catalytic Health, Celltrion, Ensho, Enthera, Exeliome Biosciences, Ferring, Fresenius Kabi, Gilead, Granite Bio, GSK, Johnson & Johnson, Lilly, Merck, Mountainfield, MRM Health, Nxera, Organon, OSE Immunotherapeutics, Pendopharm, Pioneering Medicine, Pfizer, Prometheus, Roche/Genentech, Sanofi, SCOPE, Shattuck Labs, Sorriso, Spyre, Synedgen, Takeda, Teva, Tillotts, Union Therapeutics, Ventus, Ventyx, Vividion, Xencor, Zealand Pharma. Siegmund, Britta: Consultant for AbbVie, Abivax, Boehringer Ingelheim, Bristol Myers Squibb, Dr. Falk Pharma, Eli Lilly, Endpoint Health, Galapagos, Johnson & Johnson, Materia Prima, MSD, Pfizer, Takeda, Wedbush Securities speaker for AbbVie, AlfaSigma, Bristol Myers Squibb, CED Service GmbH, Dr. Falk Pharma, Eli Lilly, Ferring, Galapagos, Johnson & Johnson, MD Education, MSD, Pfizer, Tr1xBio grant from Pfizer. Kannan, Arun K: Employee and shareholder of Johnson & Johnson Zeeman, Martha: Employee and shareholder of Johnson & Johnson Hart, Amy: Employee and shareholder of Johnson & Johnson Strawn, David: Employee and shareholder of Johnson & Johnson Venkat, Swati: Employee and shareholder of Johnson & Johnson Tomsho, Lynn: Employee and shareholder of Johnson & Johnson Mcrae, Bradford: Employee and shareholder of Johnson & Johnson Ruane, Darren: Employee and shareholder of Johnson & Johnson Surace, Lindsey: Employee and shareholder of Johnson & Johnson Erondu, Ngozi: Employee and shareholder of Johnson & Johnson Zhan, Joyce: Employee and shareholder of Johnson & Johnson Chen, Minhu: I have received speaker honoraria from Takeda China, Xian Janssen, and AbbVie China, as well as research funding from Takeda (China). Chachu, Karen: Advisory board for Johnson & Johnson consultant for OptumRx Matsuoka, Katsuyoshi: Grant: Abbvie Inc., Mochida Pharmaceutical Co., Ltd., ZERIA Pharmaceutical Co., Ltd., Nippon Kayaku Co., Ltd., JIMRO Co., Ltd. Personal Fees: Mitsubishi Tanabe Pharma, Janssen Pharmaceutical K.K., Abbvie Inc., EA Pharma Co., Ltd., Takeda Pharmaceutical Co., Ltd., Mochida Pharmaceutical Co., Ltd., ZERIA Pharmaceutical Co., Ltd., Nippon Kayaku Co., Ltd., Pfizer Inc., Kyorin Pharmaceutical Co., Ltd., Eli Lilly Japan K.K., Kissei Pharmaceutical Co., Ltd., Gilead Sciences, Celltrion Healthcare, Bristol-Myers Squibb Limdi, Jimmy: Speaker/consultancy fees from AbbVie, Abivax, Arena, BioHit, Bristol Myers Squibb, Celltrion, Eli Lilly, Ferring, Galapagos, Johnson & Johnson, MSD, Pfizer, Takeda research support from Galapagos, Takeda Rydzewska, Grazyna: Speaker and consultant fees from AbbVie, AlfaSigma, AstraZeneca, Bayer, Biocodex, Bristol Myers Squibb, Ferring, Johnson & Johnson, Lilly, PRO.MED.Pl, Recordati, Sanprobi, SOBI, Takeda travel/accommodation meeting expenses from AbbVie, Ferring, Takeda Abreu, Maria Teresa: Grant/Research Support: Janssen, Prometheus Bioscience, Takeda Consultant: Janssen, Prometheus Bioscience, Takeda Gilead, Eli Lilly, AbbVie Inc., Bristol Myers, Imedex, Loftus Jr, Edward V: Consultant/advisory board member for AbbVie, Amgen, Bristol Myers Squibb, Eli Lilly and Company, Genentech, Gilead Sciences, Johnson & Johnson, Pfizer, Takeda, UCB.