Marya, Neil B.

Abstract Background and aims Endoscopic resections are associated with bleeding and perforation and may be managed with through-the-scope (TTS) clips, over-the-scope clips and endoscopic suturing. The aim of this preclinical study was to compare technical success of closure using a novel TTS tissue helix tack and suture device (X-Tack) to TTS clips in a porcine model. Materials and methods Four subjects underwent 40 mucosal resections, diameter range 25–50 mm, in the stomach (n = 24) and colon (n = 16). Closures were randomized to X-Tack (n = 24) or clip (n = 16). Animals underwent weekly endoscopic follow-up for 4 weeks. Results Technical closure with X-Tack was successful in 24 of 24 (100 %) cases and with clips in 13 of 16 cases (81.3 %) (P = 0.0001). One colonic perforation occurred and was successfully managed using X-Tack. The rate of healing was not statistically different between the groups, and all sites healed at 4 weeks including the perforation and were confirmed by histology. Conclusions Compared to TTS clip, X-Tack is superior for effecting large mucosal defect closure, including durable sealing of full-thickness perforation. There was no difference in rate of healing between devices.

Objective The diagnosis of autoimmune pancreatitis (AIP) is challenging. Sonographic and cross-sectional imaging findings of AIP closely mimic pancreatic ductal adenocarcinoma (PDAC) and techniques for tissue sampling of AIP are suboptimal. These limitations often result in delayed or failed diagnosis, which negatively impact patient management and outcomes. This study aimed to create an endoscopic ultrasound (EUS)-based convolutional neural network (CNN) model trained to differentiate AIP from PDAC, chronic pancreatitis (CP) and normal pancreas (NP), with sufficient performance to analyse EUS video in real time. Design A database of still image and video data obtained from EUS examinations of cases of AIP, PDAC, CP and NP was used to develop a CNN. Occlusion heatmap analysis was used to identify sonographic features the CNN valued when differentiating AIP from PDAC. Results From 583 patients (146 AIP, 292 PDAC, 72 CP and 73 NP), a total of 1 174 461 unique EUS images were extracted. For video data, the CNN processed 955 EUS frames per second and was: 99% sensitive, 98% specific for distinguishing AIP from NP; 94% sensitive, 71% specific for distinguishing AIP from CP; 90% sensitive, 93% specific for distinguishing AIP from PDAC; and 90% sensitive, 85% specific for distinguishing AIP from all studied conditions (ie, PDAC, CP and NP). Conclusion The developed EUS-CNN model accurately differentiated AIP from PDAC and benign pancreatic conditions, thereby offering the capability of earlier and more accurate diagnosis. Use of this model offers the potential for more timely and appropriate patient care and improved outcome.