BACKGROUND Magnetic resonance cholangiopancreatography (MRCP) depicts the biliary tree three-dimensionally but is used mainly for qualitative diagnosis rather than quantitative procedural planning. Endoscopic drainage of biliary strictures, particularly hilar and intrahepatic lesions, depends on an accurate understanding of stricture location, length, ductal curvature, and branch involvement, yet MRCP information is seldom converted into reproducible measurements. Artificial intelligence (AI) may bridge this gap by transforming diagnostic imaging into quantitative, procedure-oriented geometry. We hypothesized that MRCP-derived biliary anatomy could be converted into an interpretable three-dimensional (3D) model supporting biliary stricture assessment and candidate stent-planning geometry. AIM To develop and evaluate a proof-of-concept AI workflow that converts MRCP-derived biliary anatomy into a quantitative 3D stricture model. METHODS This single-center study analyzed 384 MRCP examinations (242 controls and 142 with suspected biliary stricture). Expert-corrected masks and 3D annotations were used to train or evaluate modules for region-of-interest localization, bile duct segmentation, projection-based multi-view classification, synthetic morphology restoration, and centerline-guided tubular mesh generation. Binary stricture detection and three-class severity classification were assessed with patient-level cross-validation after excluding unavailable or task-incompatible labels. Segmentation and restoration were evaluated using Dice scores. RESULTS In 380 evaluable cases, projection-based binary classification achieved an accuracy of 0.9770, balanced accuracy of 0.9721, macro F1 score of 0.9749, and receiver operating characteristic area under the curve of 0.9953. Three-class severity classification was less accurate (accuracy 0.8711), indicating that severity stratification is more difficult than binary detection. The best Dice score was 0.8113 in the segmentation comparison and 0.8944 in the extended normal-group experiment. In synthetic restoration, DiceFocalLoss produced the highest Dice score (0.7531), and lesion weighting improved reconstruction of interrupted duct segments. In anatomically simple cases, restored geometry was converted into a centerline-guided tubular mesh, whereas complex hilar or intrahepatic branching remained vulnerable to projection overlap. CONCLUSION MRCP-based AI modeling of biliary strictures is feasible for quantitative 3D assessment and exploratory stent-planning geometry. External validation and prospective clinical evaluation are required before procedural implementation.
BACKGROUND Systemic inflammation and nutritional status are closely related to tumor progression in colorectal cancer (CRC). Although systemic inflammation- and nutrition-related indicators (SINRIs) are readily available blood-based biomarkers, their associations with post-operative complications and overall survival (OS), as well as their incremental value for individualized OS prediction, remain insufficiently clarified. AIM To evaluate the associations of SINRIs with post-operative complications and OS in CRC and develop a machine learning (ML) prediction model. METHODS This multicenter cohort study included 10205 patients who underwent radical surgery for CRC at four tertiary hospitals in Southwest China between 2011 and 2024. Associations of 14 SINRIs with post-operative complications and OS were assessed using multivariable logistic and Cox regression models, while metabolic dysfunction indicators were evaluated in a sub-cohort analysis. Restricted cubic spline analyses were performed to assess dose-response relationships. Six ML models were compared for predictive performance using the Harrell’s concordance index (C-index) and interpreted using SHapley Additive exPlanations. RESULTS Most SINRIs were significantly associated with post-operative complications and OS. Higher levels of inflammatory indicators were associated with increased risks of complications and mortality, whereas protective markers such as lymphocyte-to-monocyte ratio, prognostic nutritional index, and hemoglobin, albumin, lymphocyte, and platelet score were associated with better outcomes. Exploratory sub-cohort analysis suggested that triglyceride-glucose (TyG), TyG-body mass index, and metabolic score for insulin resistance were associated with worse OS. Among six candidate models, gradient boosting survival analysis (GBSA) showed the most stable predictive performance, with C-index values of 0.808 in the training cohort and 0.739 and 0.723 in the two external validation cohorts. CONCLUSION SINRIs are closely associated with post-operative complications and OS in CRC, and the GBSA model demonstrated robust multicenter performance and may facilitate individualized survival prediction in CRC.
BACKGROUND Currently, early assessment of the biliary drainage response after therapeutic biliary endoscopic retrograde cholangiopancreatography (ERCP) commonly relies on bilirubin-based markers, but these may not fully reflect the earliest postoperative biochemical improvement. We hypothesized that the total bile acid (TBA) level may provide an earlier and more sensitive signal of biochemical recovery after biliary drainage. AIM To compare the early postoperative trajectories of TBA, bilirubin, and cholestatic enzymes after therapeutic biliary ERCP. METHODS This single-center retrospective cohort study included adults undergoing therapeutic biliary ERCP. Laboratory values were assessed at baseline within 24 hours before ERCP, at the first postoperative test within 24 hours, and at the second postoperative test within 5 days. Baseline comparisons, logistic regression, generalized estimating equations, paired within-patient comparisons, and unified longitudinal models were used to evaluate whether TBA served as an earlier and more sensitive biochemical indicator of drainage response than bilirubin-based markers after therapeutic ERCP. RESULTS The main analytic cohort comprised 896 patients, of whom 278 achieved an early biochemical drainage response while 618 did not. TBA showed the most pronounced postoperative decline, while the decreases in direct bilirubin and total bilirubin occurred more slowly. Also, in the 836 patients with complete three-time-point measurements, the relative decline in TBA was significantly greater than that of direct bilirubin and total bilirubin at both 24 hours and within 5 days postoperatively. Exploratory analyses further showed that although the levels of gamma-glutamyl transferase and alkaline phosphatase also declined significantly after drainage, their early relative declines remained significantly smaller than that of TBA. CONCLUSION TBA declines earlier and more markedly than bilirubin-based and conventional cholestatic markers after biliary drainage, supporting its use as an adjunctive early biochemical indicator after therapeutic biliary ERCP.
BACKGROUND Huge hepatocellular carcinoma (HCC) is associated with a high postoperative recurrence rate and poor long-term survival. The role of neoadjuvant therapy in patients with huge HCC eligible for curative resection remains uncertain. AIM To compare long-term outcomes after neoadjuvant therapy vs upfront surgery in huge HCC. METHODS This single-center retrospective surgical cohort study included patients with huge HCC who underwent curative liver resection from January 2019 to November 2024. Huge HCC was defined as a maximum tumor diameter ≥ 10 cm on pre-treatment enhanced computed tomography or magnetic resonance imaging. Patients were divided into neoadjuvant therapy and upfront surgery groups. The primary endpoint was recurrence-free survival (RFS), with overall survival (OS) and surgery-related outcomes as secondary endpoints. RESULTS A total of 103 patients were included, with 56 in the neoadjuvant therapy group and 47 in the upfront surgery group. After a median follow-up of 30 months, the RFS and OS rates in the neoadjuvant therapy group were superior to those in the upfront surgery group. The 1-year and 3-year RFS rates were 76.8% vs 55.3% and 59.2% vs 41.5%, respectively. The corresponding 1-year and 3-year OS rates were 98.2% vs 87.2% and 88.9% vs 71.7%, respectively. Multivariable analysis identified upfront surgery and lack of postoperative adjuvant therapy as independent risk factors for poorer RFS and OS. Surgery-related outcomes were comparable between groups. CONCLUSION In this retrospective surgical cohort study, neoadjuvant therapy was associated with improved survival without significantly increasing perioperative burden in patients with huge HCC.
BACKGROUND Magnetic resonance imaging (MRI) is contraindicated in patients with an ingested video capsule endoscopy (VCE) due to the theoretical risk of mechanical and thermal harm. However, isolated cases of inadvertent scanning have not shown clear harm. Manufacturers advise against performing VCE in patients with an MRI scheduled within the next 14 days, yet asymptomatic retention for up to 12 years has been reported. With the increasing use of VCEs, the likelihood of incidental MRI exposure is rising. To date, no systematic data have been published on the forces exerted on VCEs in an MRI environment. AIM To quantify magnetically induced forces on VCEs in a 3-Tesla (3-T) scanner to assess clinical safety risks. METHODS Displacement forces on 11 commercially available VCE models were measured in a 3-T MRI scanner according to the American Society of Testing Materials (ASTM) F2052-21. Following ASTM F2213-17, torque was measured by the pulley method for eight VCEs and calculated for three magnet-containing gastric capsules with high displacement forces. The lengths of internal metallic components of the VCEs, as indicators of potential thermal effects, were determined using X-ray images. RESULTS VCEs containing magnets showed significantly higher median displacement forces than those without magnets (2366 mN vs 293.92 mN, P = 0.012). Magnetically induced displacement forces were 50.7-fold higher than the individual gravitational force for magnet-containing devices (2366 mN vs 46.66 mN), compared to a 9.7-fold excess for capsules without a magnet (293.92 mN vs 30.20 mN). Median torque was 19.2 mNm for capsules without integrated magnets vs 289.02 mNm with magnets (P = 0.017). For all but one VCE, the forces exceeded the safety threshold defined by the capsule’s gravitational force. The mean length of internal metallic components was 21.3 mm (range: 18.2-25.4 mm); only two capsules slightly exceeded the 24-mm threshold for possible thermal effects. CONCLUSION In a 3-T MRI, all but one VCE, particularly magnet-containing capsules, exceeded accepted safety thresholds for displacement force and torque but remained below perforation-relevant levels.
