Vessel and tissue recognition during third-space endoscopy using a deep learning algorithm

In this study, we aimed to develop an artificial intelligence clinical decision support solution to mitigate operator-dependent limitations during complex endoscopic procedures such as endoscopic submucosal dissection and peroral endoscopic myotomy, for example, bleeding and perforation. A DeepLabv3-based model was trained to delineate vessels, tissue structures and instruments on endoscopic still images from such procedures. The mean cross-validated Intersection over Union and Dice Score were 63% and 76%, respectively. Applied to standardised video clips from third-space endoscopic procedures, the algorithm showed a mean vessel detection rate of 85% with a false-positive rate of 0.75/min. These performance statistics suggest a potential clinical benefit for procedure safety, time and also training.

Underwater vs Conventional Endoscopic Mucosal Resection of Large Sessile or Flat Colorectal Polyps: A Prospective Randomized Controlled Trial

BACKGROUND & AIMS

Conventional endoscopic mucosal resection (CEMR) with submucosal injection is the current standard for the resection of large, nonmalignant colorectal polyps. We investigated whether underwater endoscopic mucosal resection (UEMR) is superior to CEMR for large (20-40mm) sessile or flat colorectal polyps.

METHODS

In this prospective randomized controlled study, patients with sessile or flat colorectal polyps between 20 and 40 mm in size were randomly assigned to UEMR or CEMR. The primary outcome was the recurrence rate after 6 months. Secondary outcomes included en bloc and R0 resection rates, number of resected pieces, procedure time, and adverse events.

RESULTS

En bloc resection rates were 33.3% in the UEMR group and 18.4% in the CEMR group (P = .045); R0 resection rates were 32.1% and 15.8% for UEMR vs CEMR, respectively (P = .025). UEMR was performed with significantly fewer pieces compared to CEMR (2 pieces: 45.5% UEMR vs 17.7% CEMR; P = .001). The overall recurrence rate did not differ between both groups (P = .253); however, subgroup analysis showed a significant difference in favor of UEMR for lesions of >30 mm to ≤40 mm in size (P = .031). The resection time was significantly shorter in the UEMR group (8 vs 14 minutes; P < .001). Adverse events did not differ between both groups (P = .611).

CONCLUSIONS

UEMR is superior to CEMR regarding en bloc resection, R0 resection, and procedure time for large colorectal lesions and shows significantly lower recurrence rates for lesions >30 mm to ≤40 mm in size. UEMR should be considered for the endoscopic resection of large colorectal polyps.

Endoscopic treatment of Zenker's diverticulum with the stag beetle knife (sb knife) - feasibility and follow-up

BACKGROUND AND AIMS

Flexible endoscopic treatment of symptomatic Zenker's diverticulum (ZD) is an established treatment option. This study reports the first large cohort of ZD patients treated with the stag beetle knife (sb knife, a new scissor-like device) regarding feasibility, safety, and sustainability of mucomyotomy using this technique.

METHODS

From August 2013 to January 2016, n = 52 patients (pts) were treated at Klinikum Augsburg, a tertiary referral center, with the sb knife junior or standard. For stability and safety, the septum is fixed with a soft overtube before intervention. Symptoms were analyzed before and at 1 and 6 months past intervention using an extensive questionnaire of dysphagia, odynophagia, regurgitation, chronic cough, state of health, and complications.

RESULTS

The mean size of ZD was 3 cm (1-5 cm). Forty-seven out of 52 (90.4%) patients received one treatment session. The mean procedure time was 32 min (18-60 min). In 10 procedures (17%), a clip was placed at the bottom of the resection line. No major complications (e.g., perforation, mediastinitis) occurred. Five patients (9.6%) required a second treatment after a mean of 7 months (3-13) due to symptomatic recurrence. One patient was lost to further follow-up after one month with no or rare complaints. One patient had a third treatment (1.9%) without complications. During a mean follow-up of 16 months (2-31), the dysphagia score improved from 2 (1-4) prior of treatment to 1 (0-4), odynophagia, regurgitation, and chronic cough were no longer reported in the asymptomatic patients at all.

CONCLUSIONS

Flexible endoscopic treatment of ZD with the sb knife and overtube is effective, safe, and has lasting effects with a relatively low recurrence rate.

Endoscopic hemostasis state of the art - Nonvariceal bleeding

New endoscopic techniques for hemostasis in nonvariceal bleeding were introduced and known methods further improved. Hemospray and Endoclot are two new compounds for topical treatment of bleeding. Initial studies in this area have shown a good hemostatic effect, especially in active large scale oozing bleeding, e.g., tumor bleedings. For further evaluation larger prospective studies comparing the substanced with other methods of endoscopic hemostasis are needed. For localized active arterial bleeding primary injection therapy in the area of ​​bleeding as well as in the four adjacent quadrants offers a good method to reduce bleeding activity. The injection is technically easy to learn and practicable. After bleeding activity is reduced the bleeding source can be localized more clearly for clip application. Today many different through-the-scope (TTS) clips are available. The ability to close and reopen a clip can aid towards good positioning at the bleeding site. Even more important is the rotatability of a clip before application. Often multiple TTS clips are required for secure closure of a bleeding vessel. One model has the ability to use three clips in series without changing the applicator. Severe arterial bleeding from vessels larger than 2 mm is often unmanageable with these conventional methods. Here is the over-the-scope-clip system another newly available method. It is similar to the ligation of esophageal varices and involves aspiration of tissue into a transparent cap before closure of the clip. Thus a greater vascular occlusion pressure can be achieved and larger vessels can be treated endoscopically. Patients with severe arterial bleeding from the upper gastrointestinal tract have a very high rate of recurrence after initial endoscopic treatment. These patients should always be managed in an interdisciplinary team of interventional radiologist and surgeons.