Citations

Abstract: Background Artificial intelligence (AI)-assisted colonoscopy reduces costs in colorectal cancer screening, but its incremental value relative to non-AI resect-and-discard (RD) strategies for diminutive polyps (≤ 5 mm) remain unclear. We aimed to compare the clinical, economic, and environmental outcomes of AI-assisted and non-AI RD strategies with standard of care (SoC). Methods In our Markov model simulating 6 million individuals aged ≥45 years undergoing screening colonoscopy, 4 strategies were evaluated: SoC, non-AI RD, CADe-assisted RD (CADe-RD), and CADe+CADx-assisted RD (CADe+CADx-RD). The primary outcome was total quality-adjusted life years (QALYs), with secondary outcomes including cost and carbon dioxide (CO₂) emissions. Results Non-AI RD achieved the lowest total cost ($36.6B) and carbon emissions (0.90 million kg CO₂), while yielding the highest QALYs (51.37 million), although differences in QALYs across strategies were small (<0.01%). CADe-RD and CADe+CADx-RD remained cost-saving relative to SoC ($36.9B and $37.4B vs $37.8B, respectively) but were associated with higher costs than non-AI RD. CADe modestly increased adenoma detection without improving QALYs, while the addition of CADx increased costs and emissions and resulted in slightly lower QALYs. Sensitivity analysis showed that CADe+CADx-RD matched or exceeded CADe-RD QALYs only within a limited region of high diagnostic performance. At currently reported CADx performance levels, CADe+CADx-RD yielded lower QALYs and higher costs than CADe-RD. CADe+CADx-RD achieved non-inferiority in QALYs relative to CADe-RD only within a limited region of high diagnostic performance. Conclusions In our non-deterministic model, non-AI RD provided the optimal balance of clinical benefit, cost savings, and lowest environmental impact.

Artificial intelligence (AI) is transforming gastrointestinal (GI) endoscopy into a more standardized, data-driven, and workflow-integrated field. Advances in computer-assisted detection (CADe), diagnosis (CADx), quality assessment (CAQ), natural language processing (NLP), and multimodal deep learning have expanded AI applications across colonoscopy, upper endoscopy, endoscopic ultrasound (EUS), ERCP, cholangioscopy, and capsule endoscopy. These systems have demonstrated improvements in lesion detection, procedural quality assessment, workflow efficiency, and diagnostic support. However, current evidence remains largely focused on surrogate outcomes rather than patient-centered clinical benefits, while challenges related to generalizability, explainability, regulatory oversight, automation bias, and workflow integration continue to limit widespread adoption. Future progress will depend on prospective real-world validation, diverse datasets, explainable AI frameworks, and careful integration of human–AI interaction into clinical practice. Overall, AI is evolving from a supportive adjunct into an increasingly integrated component of gastrointestinal endoscopy with the potential to improve procedural quality, diagnostic consistency, and clinical efficiency.

Abstract Background: Early-onset pancreatic cancer (EOPC; age &lt;50 y at diagnosis) is rare (≈1.3/100,000), biologically distinct, and germline-enriched (BRCA1/2, PALB2, ATM, CDKN2A, Lynch). Despite NCCN and CAPS Consortium recognition, its pre-diagnostic electronic health record (EHR) trajectory is uncharacterized. We sought reproducible signatures to guide subtype-aware surveillance for this under-screened young-adult population, where conventional guidelines do not apply. Methods: We assembled an EOPC cohort from Mayo Clinic (3 sites, 2001–2023; n=169 EOPC, n=3,162 older-onset PDAC), identified by ICD-10 C25.0–C25.3, C25.7, C25.9 or ICD-9 157.x with age &lt;50 at first diagnosis. For each patient, 3-year pre-diagnostic trajectories of 10 routine laboratory analytes (bilirubin, CA 19-9, CEA, lipase, amylase, glucose, HbA1c, ALT, AST, alkaline phosphatase) were summarized over 30/90/180/365-day windows (mean, slope, delta; 1,706 features total; 131 retained at ≥10% non-zero). Features were winsorized (1–99 pct), standardized, and projected onto 15 PCA components capturing 81.7% variance. K-means clustering (k=2–6, min cluster fraction ≥10%) was assessed by silhouette and 100-resample bootstrap Adjusted Rand Index (ARI). Cluster signatures were compared against the older-onset PDAC cohort projected into the same component space. Results: Two reproducible pre-diagnostic clusters emerged (k=2; silhouette 0.469; bootstrap ARI 0.440). Cluster A: "occult prodrome" (87.0%; n=147; mean age 44.9 y) showed no clear signal across the 10 monitored analytes, with all analyte z-scores near zero through the 3-year window. Cluster B: "metabolic-biliary active" (13.0%; n=22; mean age 44.4 y) was characterized by elevated mean HbA1c (z=+1.08), rising bilirubin slope (z=+0.54), and rising ALT/alkaline phosphatase slopes (z=+0.51 each), consistent with combined diabetic-onset and biliary-obstructive prodrome. The two-cluster structure reproduced in older-onset Mayo PDAC at 87.3% / 12.7%, nearly identical to the EOPC distribution, indicating a fundamental pre-diagnostic dichotomy of pancreatic cancer rather than an age-specific phenomenon. The dominant occult phenotype across age strata indicates that lab-trajectory–anchored surveillance can only address the active minority and motivates complementary genetic-, demographic-, or imaging-anchored pathways for the occult majority. Conclusions: EOPC pre-diagnostic trajectories partition into a dominant occult prodrome (∼87%) without detectable signal in the 10 routine labs evaluated and a minority metabolic-biliary active prodrome (∼13%) with combined diabetic and obstructive features. The high occult prevalence motivates germline-guided active surveillance for genetically at-risk young adults rather than symptom-driven workup, anchoring next-generation surveillance trial designs in rare and hereditary pancreatic malignancies. External validation in independent institutional cohorts (Mayo Clinic Platform / MERCI, or comparable EHR networks) is the planned next step. Citation Format: Jianfu Li Li, Yingyun Yang, Wallace MIchael, Yan Bi. Pre-diagnostic clinical-trajectory signatures of early-onset pancreatic cancer [abstract]. In: Proceedings of the AACR Special Conference in Cancer Research: Breaking Barriers in the Fight against Rare Cancers; 2026 Jul 18-20; Philadelphia, PA. Philadelphia (PA): AACR; Cancer Res 2026;86(14_Suppl):Abstract nr A046.

Abstract Despite growing interest in AI for pancreatic cyst management, no prior study has systematically investigated how lesion characteristics influence model behavior or provided a comprehensive explainability analysis in IPMN malignancy risk stratification. We present the first multi-center study integrating explainable AI with uncertainty quantification to evaluate how cyst type, size, and location affect deep learning performance in IPMN malignancy prediction. Our retrospective study analyzed 170 IPMNs from seven centers using a radiomics-deep learning fusion model. Cases were stratified by dysplasia grade, IPMN type, size, and location; with model interpretability assessed using SHAP, LIME, Grad-CAM, and uncertainty quantification. Overall accuracy was 67.1%, with model uncertainty lower for correct versus incorrect predictions (0.72 vs 0.78, p < 0.001). On subgroup analysis: Low-Risk lesion accuracy exceeded High-Risk (81.3% vs. 42.9%, p < 0.001), BD accuracy was inversely correlated with cyst size while MD remained stable, and body and whole-pancreas cysts stratified more accurately than tail cysts (77.8%, 76.6% vs. 47.8%). AI models face significant, previously uncharacterized performance drops when evaluating complex high-risk IPMNs. Our study shows that integrating uncertainty quantification can successfully flag unreliable predictions, enabling a selective-prediction framework where clinicians can confidently rely on AI for low-risk, low-uncertainty lesions while deferring high-uncertainty cases to expert review.

INTRODUCTION: Colonoscopies reduce colorectal cancer incidence and mortality, but challenges remain in detecting all precancerous lesions. Operator fatigue, technique variability, and subtle lesions can lead to missed adenomas, highlighting the need for tools that standardize detection quality. Computer-aided detection (CADe) systems use artificial intelligence to enhance lesion detection and improve screening colonoscopy quality. This study evaluates the impact of CADe on key quality and efficiency metrics in a clinical setting. METHODS: A single-center retrospective study at an academic tertiary center analyzed 4,028 colonoscopies from October 2022 to December 2023. The CADe system was used in 4 of 8 endoscopy suites, creating a CADe and control group. Propensity matching accounted for age, sex, indication, and physician. Primary outcomes included adenoma detection rate (ADR) and polyp detection rate (PDR). Adenomas per colonoscopy (APC) and polyps per colonoscopy (PPC) were also measured. Per-polyp analyses were performed as a tertiary outcome. RESULTS: ADR was significantly higher in the CADe group (38.6%) vs control (34.2%) (rate ratio [RR] = 1.127, P < 0.05). PDR was significantly higher in the CADe group (67.2%) vs control (59.4%) (RR = 1.130, P < 0.05). APC was significantly higher in the CADe group (M = 0.406, SD 0.637) vs control (M = 0.359, SD 0.599) (RR = 1.131, P < 0.05). PPC was significantly higher in the CADe group (M = 1.297, SD 1.188) vs control (M = 1.046, SD 1.002) (RR = 1.240, P < 0.05). The increase in polyp resection with CADe was driven predominantly by ≤5 mm lesions, with comparatively smaller differences for larger polyps. DISCUSSION: Implementing CADe in clinical practice significantly improved ADR, PDR, APC, and PPC. However, the increase in ADR seems inflated by the detection of polyps ≤5 mm, whose clinical significance in reducing colorectal cancer remains uncertain.

Abstract Background Gastrointestinal (GI) cancer remains a major global health challenge. This study provides updated 2022 estimates of GI cancer incidence and mortality worldwide, along with projections of incident cases and deaths through 2050 to inform cancer control strategies. Methods In this population‐based study, data from the GLOBOCAN 2022 database were used to examine incidence and deaths associated with GI cancer. By using global demographic projections, this study estimated the projected number of incident cases and deaths until the year 2050. Results By 2050, the global burden of GI cancer is projected to reach 9.06 million incident cases (+85%) and 6.42 million deaths (+93%) from 2022. Increases in incidence are projected for both females (+87%) and males (+83%), as are increases in mortality for both females (+97%) and males (+91%). Whereas the Western Pacific region will have the highest absolute burden, with 3.88 million new cases and 2.79 million deaths, the greatest relative increase is expected in Africa, with incidence and deaths rising by 157% and 160%, respectively. As measured by the Human Development Index (HDI), the steepest increases are projected in low (+151% incidence; +152% deaths) and medium (+112% incidence; +114% deaths) HDI countries. Overall, all GI cancer types are projected to increase, led by pancreatic cancer in incidence (+95%) and colorectal cancer in deaths (+103%). Conclusions GI cancer incidence and deaths are expected to nearly double by 2050, with increases across sexes, cancer types, and regions, especially in lower HDI countries.

INTRODUCTION: A deeper understanding of the factors influencing survival in patients with pancreatic ductal adenocarcinoma (PDAC) is essential for optimizing treatment strategies. This study investigates the independent association of body composition parameters with overall survival in patients with PDAC. METHODS: This retrospective multisite cohort study included patients diagnosed with PDAC. Diagnostic computed tomography scans were retrieved, and body composition was evaluated using a validated deep learning-based segmentation algorithm that measured tissue volume and density in a 20 cm vertical abdominal section. RESULTS: A total of 1,666 patients with PDAC were included, 938 male (56.3%) and median age 69 years (interquartile range 61–76). In the subgroup of patients who underwent surgical resection (n = 509), myosteatosis (intramuscular infiltration of adipose tissue; hazard ratio [HR] 1.56, 95% confidence interval [CI] 1.16–2.11, P = 0.004), sarcopenic obesity (HR 1.75, 95% CI 1.06–2.91, P = 0.03), and less subcutaneous adipose tissue (HR 1.09, 95% CI 1.03–1.16, P = 0.002) were associated with higher mortality. In patients receiving palliative systemic therapy (n = 439), lower skeletal muscle density (HR 1.43, 95% CI 1.03–1.99, P = 0.03) was associated with higher mortality. In patients who did not undergo tumor-targeted treatment (n = 718), less visceral adipose tissue (HR 1.04, 95% CI 1.01–1.08, P = 0.02) was associated with higher mortality. DISCUSSION: Body composition parameters, derived from computed tomography scans at the time of PDAC diagnosis, particularly low skeletal muscle density, sarcopenic obesity, and low adipose tissue, are independently associated with overall survival in patients with PDAC. Evaluating body composition at diagnosis could enhance clinical decision-making and enable more personalized treatment strategies.

INTRODUCTION: Colorectal cancer remains a leading cause of cancer associated death in the United States and colonoscopy the primary screening strategy for prevention. Rates of adenomatous and serrated neoplasia detection are inversely associated with postcolonoscopy colorectal cancer. This crucial quality metric depends on accurate ascertainment of colorectal neoplasia findings from both endoscopy and histopathology records. We aimed to assess the feasibility of a random forest machine learning model to rapidly and accurately categorize colorectal neoplasia from electronic health record data. METHODS: A retrospective cohort study compared neoplasia detection rates among individuals undergoing colonoscopy at a large academic institution to develop a rule-based algorithm to categorize colorectal neoplasia from endoscopy reports and pathology systematized nomenclature of medicine – clinical terms. This cohort provided a large training set to develop a natural language processing system using a random forest approach to automatically classify unstructured pathology findings into adenoma, serrated, or advanced neoplasms. This system was manually validated through an independent holdout set. RESULTS: The training set comprised 35,953 unstructured pathology reports with matched systematized nomenclature of medicine – clinical terms from 95,188 unstructured colonoscopy reports. The final model was assessed on an independent holdout set of 337 manually annotated procedures obtaining an area under the receiver operating characteristic curve of 0.997 (confidence interval [CI] 0.994–1), 0.99 (CI 0.98–1), and 0.99 (CI 0.98–0.99) for prediction of adenoma, serrated, and advanced lesions, respectively. DISCUSSION: The random forest-based hybrid natural language processing system for classification of colonoscopy results was both accurate and explainable. NLP combined with effective machine learning algorithms can provide a scalable strategy for colonoscopy quality monitoring.

Abstract Pancreatic cancer, projected to become the second-deadliest malignancy in Western societies by 2030, requires urgent innovations in early detection and risk stratification. Intraductal papillary mucinous neoplasms (IPMNs) represent critical precursor lesions, but current clinical guidelines demonstrate suboptimal accuracy in malignancy prediction, leading to either unnecessary surgeries or missed opportunities for early intervention. In this paper, we present Cyst-X, an artificial intelligence (AI) framework that accurately predicts IPMN malignant transformation using multicenter magnetic resonance imaging (MRI) data. Unlike most previous approaches that rely on computed tomography (CT), our method capitalizes on the superior soft tissue contrast of MRI, allowing for more precise identification of subtle imaging biomarkers. We developed and validated deep learning models on 723 T1-weighted and 738 T2-weighted MRI scans from 764 patients among seven international institutions, demonstrating significantly superior performance (AUC=0.82) compared to current clinical (Kyoto) guidelines (AUC=0.75) and expert radiologists. AI-based imaging features correlate strongly with clinically recognized malignancy markers, providing potential biologically relevant insights. This approach holds promise to significantly refine clinical decision-making, reduce unnecessary surgeries, and better identify high-risk IPMN patients for timely intervention. Our approach integrates a novel pancreas segmentation network with robust classification models that identify subtle imaging biomarkers associated with malignancy risk. Importantly, we demonstrate that these models retain high performance in a privacy-preserving federated learning setting, where institutions collaboratively train AI models without exchanging patient data to address key regulatory and ethical barriers. We publicly release the Cyst-X dataset--the first large-scale, multi-center pancreatic cyst MRI collection--to accelerate research in this field. This study addresses a critical clinical need while establishing technical foundations for privacy-preserving AI in radiology that could transform pancreatic cancer management through earlier intervention and reduced unnecessary procedures. The dataset can be accessed at https://osf.io/74vfs/, and the source code for our deep learning segmentation and classification models is available at https://github.com/NUBagciLab/Cyst-X.

This review explores how multimodal foundation models (MFMs) are transforming biliary tract cancer (BTC) research. BTCs are aggressive malignancies with poor prognosis, presenting unique challenges due to difficult diagnostic methods, molecular complexity, and rarity. Importantly, intrahepatic cholangiocarcinoma (iCCA), perihilar cholangiocarcinoma (pCCA), and distal bile duct cholangiocarcinoma (dCCA) represent fundamentally distinct clinical entities, with iCCA presenting as mass-forming lesions amenable to biopsy and targeted therapies, while pCCA manifests as infiltrative bile duct lesions with challenging diagnosis and primarily palliative management approaches. MFMs offer potential to advance research by integrating radiological images, histopathology, multi-omics profiles, and clinical data into unified computational frameworks, with applications tailored to these distinct BTC subtypes. Key applications include enhanced biomarker discovery that identifies previously unrecognizable cross-modal patterns, potential for improving currently limited diagnostic accuracy—though validation in BTC-specific cohorts remains essential—accelerated drug repurposing, and advanced patient stratification for personalized treatment. Despite promising results, challenges such as data scarcity, high computational demands, and clinical workflow integration remain to be addressed. Future research should focus on standardized data protocols, architectural innovations, and prospective validation studies. The integration of artificial intelligence (AI)-based methodologies offers new solutions for these historically challenging malignancies. However, current evidence for BTC-specific applications remains largely theoretical, with most studies limited to proof-of-concept designs or related cancer types. Comprehensive clinical validation studies and prospective trials demonstrating patient benefit are essential prerequisites for clinical implementation. The timeline for evidence-based clinical adoption likely extends 7–10 years, contingent on successful completion of validation studies addressing current evidence gaps.