Introduction

Colorectal anastomotic dehiscence (CAD) is one of the most severe postoperative complications in colorectal surgery, significantly impacting patient morbidity and mortality. Current risk assessment methods rely on clinical and intraoperative factors, but no real-time predictive tool exists. This study aimed to develop an artificial intelligence model based on intraoperative recording of the anastomosis for CAD prediction.

Aims & Methods

A convolutional neural network (CNN) was trained using a dataset of annotated frames from colorectal surgery videos across three international high-volume centers: IPO Lisboa, Hospital das Clínicas de Ribeirão Preto, São Paulo, Brazil and Royal Liverpool University Hospital. The dataset included a total of 5,356 annotated frames from 25 patients, 2,007 frames classified as CAD-positive and 3,349 as CAD-negative. Four CNN architectures (EfficientNetB0, EfficientNetB7, ResNet50, and MobileNetV2) were tested. Performance was evaluated using sensitivity, specificity, accuracy, and area under the receiver operating characteristic (AUROC) curve. Heatmaps were generated to identify key regions influencing predictions.

Results

The best-performing model achieved an accuracy of 99.55%, AUROC of 99.60%, sensitivity of 99.21%, specificity of 100%, PPV of 100%, and NPV of 98.94%. The model reliably identified CAD-positive frames and provided visual explanations through heatmaps.

Conclusion

To our knowledge, this is the first AI model developed to predict CAD using intraoperative video analysis. Its accuracy suggests the potential to redefine surgical decision-making by providing real-time risk assessment. Further refinement with a larger dataset and diverse surgical techniques could enable intraoperative interventions to prevent CAD before it occurs, marking a paradigm shift in colorectal surgery.

Introduction

Endoscopic retrograde cholangiopancreatography (ERCP) with stent placement is the current standard for drainage of distal malignant biliary obstruction (MBO), but carries a significant risk of post-procedural pancreatitis.[1] EUS-guided choledochoduodenostomy (EUS-CDS) is increasingly used as a promising alternative, but comparative studies are still scarce.[2,3] While previous studies have mainly focused on stent patency, the current study aims to evaluate whether EUS-CDS is superior to ERCP in achieving a textbook outcome for the patients, indicating an optimal post-procedural course.

Aims & Methods

This single-center study included patients with distal MBO who underwent technically successful (1) EUS-CDS with a lumen apposing metal stent (LAMS) plus additional coaxial stent or (2) ERCP with fully covered self-expandable metal stent (FCSEMS) as part of the prospective SCORPION studies and SPHINX trial, respectively.[1,4] The primary endpoint was textbook outcome, defined as the absence of procedure- or stent-related adverse events (AEs), reinterventions, or hospital readmissions within 30 days of the procedure. Secondary endpoints included individual components of the primary outcome, mortality, time to initiation of treatment (e.g., systemic therapy or surgery), and textbook outcome at 90 days.

Results

A total of 144 patients were enrolled for EUS-CDS (n=49) or ERCP (n=95). Primary etiologies included pancreatic cancer (n=117, 81%) and distal cholangiocarcinoma (n=14, 10%). Coaxial stents used in the EUS-CDS group were FCSEMS (n=20) or double pigtail plastic stents (n=29). Textbook outcome was achieved in 44 patients (90%) in the EUS-CDS group and in 67 patients (71%) in the ERCP group at 30 days (RR 1.27, 95% CI 1.08-1.50). EUS-CDS led to fewer procedure- or stent-related AEs (RR 0.35, 95% CI 0.14-0.84) compared with ERCP, particularly post-procedure pancreatitis (0% vs. 23%). Other AEs, including perforation (2% vs. 0%), bleeding (0% vs. 0%), stent dysfunction with or without cholangitis (4% vs. 6%), and cholecystitis (4% vs. 3%) were similar between groups. Hospital readmission rate was lower in the EUS-CDS group (6% vs. 30%, RR 0.21, 95% CI 0.07-0.65), as was the length of hospital stay (mean 0.3 days [SD 1.3] vs. 1.8 days [SD 4.0], P=0.01). In-hospital mortality (8% vs. 5%), need for reinterventions (8% vs. 6%), and time to treatment initiation, either surgery or chemotherapy, did not differ (median 33 days [IQR 23-42] vs. 30 days [IQR 18-41]). At 90 days, textbook outcome was more often present in the EUS-CDS group when compared to ERCP however the difference was no longer significant (78% vs. 69%, RR 1.22, 95% CI 1.02-1.44).

Conclusion

EUS-CDS with LAMS and additional coaxial stent was superior to ERCP with FCSEMS in achieving textbook outcome at 30 days post-procedure. EUS-CDS resulted in less procedure- or stent-related AEs, particularly post-procedure pancreatitis, and consequently in fewer hospital readmissions and shorter hospital stay. Therefore, when expertise is available, EUS-CDS with coaxial stent placement could be considered as primary biliary drainage technique in patients with distal MBO.

References

1. Onnekink, A. M., Gorris, M., Bekkali, N. L., et al (2024). Endoscopic sphincterotomy to prevent post-ERCP pancreatitis after self-expandable metal stent placement for distal malignant biliary obstruction (SPHINX): a multicentre, randomised controlled trial. Gut, gutjnl-332695. https://doi.org/10.1136/gutjnl-2024-332695
2. Teoh, A. Y. B., Napoleon, B., Kunda, R et al (2023). EUS-Guided choledocho-duodenostomy using lumen apposing stent versus ERCP with covered metallic stents in patients with unresectable malignant distal biliary obstruction: a multicenter randomized controlled trial (DRA-MBO trial). Gastroenterology, 165(2), 473-482.e2. https://doi.org/10.1053/j.gastro.2023.04.016
3. Chen, Y., Sahai, A., Donatelli, G. et al (2023). Endoscopic Ultrasound-Guided Biliary Drainage of First Intent With a Lumen-Apposing Metal Stent vs Endoscopic Retrograde Cholangiopancreatography in Malignant Distal Biliary Obstruction: A Multicenter Randomized Controlled Study (ELEMENT Trial). Gastroenterology, 165(5), 1249-1261.e5. https://doi.org/10.1053/j.gastro.2023.07.024
4. Fritzsche, J. A., Fockens, P., Besselink, M. G. et al (2024). Optimizing EUS-guided choledochoduodenostomy with lumen-apposing metal stents for primary drainage of malignant distal biliary obstruction (SCORPION-IIp): a prospective pilot study. Gastrointestinal Endoscopy. https://doi.org/10.1016/j.gie.2024.10.012

Disclosure

Paul Fockens acted as a consultant for Cook Endoscopy and Olympus, outside the submitted work.
Roy L.J. van Wanrooij acted as a consultant for Boston Scientific. Jeanin E. van Hooft has received research support from Cook Medical and acted as lecturer for Cook Medical, Boston Scientific and Falk, and as consultant for Olympus. Rogier P. Voermans received a research grant and acted as consultant for Boston Scientific. All outside the submitted work. All other authors declare that they have no competing financial interests or personal relationships that could have appeared to influence the work reported in this abstract.