Background and Aims: Recent guidelines propose higher adenoma detection rate (ADR) benchmarks for colonoscopies performed for positive results for fecal immunochemical tests (FIT), but this is based on low-quality evidence. We aimed to compare ADR, advanced ADR (AADR), and number of adenomas per colonoscopy (APC) in direct screening colonoscopy (DSC) versus FIT-positive colonoscopy (FITC) in a multicenter Asia-Pacific cohort to justify differential targets. Methods: Asymptomatic average-risk patients >= 50 years of age who underwent screening colonoscopy directly or as follow-up for positive OC-Sensor FIT results were identified from 8 sites across the Asia-Pacific region. Overall, sex-specific ADR, overall AADR, and overall APC were compared between the 2 screening methods. Multivariable logistic regression was performed to adjust for confounding by differences in patient characteristics. Linear regression was used to correlate ADR with APC and to propose APC benchmarks. Results: A total of 2901 (mean age, 60.1 years; 57% men) individuals had DSC, and 2485 (mean age, 62.8 years; 57% men) underwent FITC. Overall ADR (53.6% vs 37.5%; odds ratio [OR], 1.93; P < .001), male-specific ADR (61.6% vs 44.6%; OR, 2; P < .001), female-specific ADR (43.2% vs 28.2%; OR, 1.94; P < .001) and overall AADR (29.9% vs 4.9%; OR, 8.2; P < .001) in FITC were significantly higher than the corresponding values for DSC. Differences remained significant after adjustment for patient characteristics. ADR was strongly and positively correlated to APC, with an ADR of 45% and 35% correlating to an APC of similar to 1 and similar to 0.65. Conclusions: Results from this international multicenter cohort study provide early evidence that newly proposed higher ADR targets are justified as quality indicators for FITC.
Background and Aims: Natural orifice transluminal endoscopic surgery (NOTES) has been established in animal models and human studies, but few clinical studies have investigated transvaginal NOTES in the diagnosis of unexplained refractory ascites. We aimed to assess the feasibility, efficacy, and safety of transvaginal NOTES for the diagnosis of unexplained ascites in female patients. Methods: A prospective study was done involving 3 female patients with unexplained ascites. After general anesthesia and disinfection, a 1.0-cm incision was made in the posterior fornix of the vagina. A gastroscope was inserted into the abdominal cavity through the transvaginal incision and an artificial pneumoperitoneum was created; NOTES peritoneoscopy was performed to scrutinize the pathologic changes. Endoscopic biopsy specimens were obtained for pathologic examination. The transvaginal incision was closed by direct suturing. Results: Transvaginal NOTES for diagnostic peritoneoscopy was successfully performed in 3 patients. The mean operative time was 61 minutes. The estimated blood loss was 5 to 10 mL. The pathologic diagnoses were tuberculosis for all patients, and the symptoms and ascites disappeared after antituberculosis therapy. During the 4-year follow-up, no clinically significant adverse events occurred in any patient after NOTES. No patient experienced an annex inflammation, vaginitis, dyspareunia, or sexual dysfunction. All patients were comfortable and satisfied with the nonscarring surgical procedure. Conclusions: Transvaginal NOTES for the diagnosis of unexplained ascites is feasible, effective, and safe. This method had no long-term effect on female sexual function and is particularly suitable for women who have special aesthetic requirements. (Clinical trial registration number: ChiCTR-TRC-10001053.)
Background and Aims: According to guidelines, endoscopic resection should only be performed for patients whose early gastric cancer invasion depth is within the mucosa or submucosa of the stomach regardless of lymph node involvement. The accurate prediction of invasion depth based on endoscopic images is crucial for screening patients for endoscopic resection. We constructed a convolutional neural network computer-aided detection (CNN CAD) system based on endoscopic images to determine invasion depth and screen patients for endoscopic resection. Methods: Endoscopic images of gastric cancer tumors were obtained from the Endoscopy Center of Zhongshan Hospital. An artificial intelligence based CNN-CAD system was developed through transfer learning leveraging a state-of-the-art pretrained CNN architecture, ResNet50. A total of 790 images served as a development dataset and another 203 images as a test dataset. We used the CNN-CAD system to determine the invasion depth of gastric cancer and evaluated the system's classification accuracy by calculating its sensitivity, specificity, and area under the receiver operating characteristic curve. Results: The area under the receiver operating characteristic curve for the CNN-CAD system was.94 (95% confidence interval [CI],.90-.97). At a threshold value of.5, sensitivity was 76.47%, and specificity 95.56%. Overall accuracy was 89.16%. Positive and negative predictive values were 89.66% and 88.97%, respectively. The CNN CAD system achieved significantly higher accuracy (by 17.25%; 95% CI, 11.63-22.59) and specificity (by 32.21%; 95% CI, 26.78-37.44) than human endoscopists. Conclusions: We constructed a CNN-CAD system to determine the invasion depth of gastric cancer with high accuracy and specificity. This system distinguished early gastric cancer from deeper submucosal invasion and minimized overestimation of invasion depth, which could reduce unnecessary gastrectomy.
Background and Aims: Mucosal injury (MI) is one of the most common perioperative adverse events of per-oral endoscopic myotomy (POEM). Severe undertreated MI may lead to contamination of the tunnel and even mediastinitis. This study explored the characteristics, predictors, and management approaches of intraoperative MI. Methods: A retrospective review of the prospectively collected database at a large tertiary referral endoscopy unit was conducted for all patients undergoing POEM between August 2010 and March 2016. MI was graded according to the difficulty of repair (I, easy to repair; II, difficult to repair). The primary outcomes were the incidence and predictors of intraoperative MI. Secondary outcomes were MI details and the corresponding treatment. Results: POEM was successfully performed in 1912 patients. A total of 338 patients experienced 387 MIs, for an overall frequency of 17.7% (338/1912). Type II MI was rare, with a frequency of 1.7% (39/1912). Major adverse events were more common in patients with MI than in those without MI (6.2% vs 2.5%, P<.001). On multivariable analysis, MI was independently associated with previous Heller myotomy (odds ratio [OR], 2.094; P=.026), previous POEM (OR, 2.441; P=.033), submucosal fibrosis (OR, 4.530; P<.001), mucosal edema (OR, 1.834; P=.001), and tunnel length >13 cm (OR, 2.699; P<.001). Previous POEM (OR, 5.005; P=.030) and submucosal fibrosis (OR, 12.074; P<.001) were significant predictors of type II MI. POEM experience >1 year was a protective factor for MI (OR,.614; P=.042) and type II MI (OR,.297; P=.042). Conclusions: MI during POEM is common, but type II injury is rare. Previous POEM and submucosal fibrosis were significant predictors of type II mucosal injury. POEM experience after the learning curve reduces the risk of MI.
Background and Aims: Radiofrequency ablation (RFA) is an accepted treatment for flat Barrett's neoplasia. Less is known about RFA for esophageal squamous cell neoplasia (ESCN). Our group has reported several prospective studies of RFA for ESCN in China with promising results through 12 months of follow-up. In this cohort study we aimed to evaluate longer term outcomes after RFA for ESCN. Methods: Patients with flat unstained lesions (USLs) on Lugol's endoscopy containing moderate-/high-grade intraepithelial neoplasia (MGIN/HGIN) or mucosal cancer were treated with RFA every 3 months until complete remission (CR; no MGIN or a worse histologic grade). Patients with CR at 12 months (CR12) were included for follow-up and underwent annual Lugol's endoscopy with biopsy sampling and re-RFA for flat USLs. The clinical course of patients with persistent ESCN at 12 months (treatment failures) is also reported. Results: Among the 78 patients in CR12, 67 (86%) had sustained CR during a median of 48 months (interquartile range, 48-48) of follow-up and 5 endoscopies (interquartile range, 4-6). Recurrence occurred in 7 of 78 patients (9%; MGIN, n = 6; HGIN, n = 1); all lesions were managed with RFA. Four other patients (5%) had progression (to HGIN, n = 1; submucosal esophageal squamous cell carcinoma, n = 3). During follow-up protocol violations occurred in 46 of 78 patients (59%). Of the 12 treatment failures, progression occurred in 6. Overall, 2 patients developed subepithelial disease that was not visible after Lugol's endoscopy. Based on post-hoc analysis, the pink color sign at baseline (a pink color change after Lugol's endoscopy) significantly predicted failure after RFA. Conclusions: RFA is relatively easy to apply and can efficiently treat large areas with ESCN. Despite protocol violations that may have interfered with the efficacy of RFA in 59% of patients, most patients with CR12 had sustained CR during follow-up. However, some patients progressed to advanced disease and 2 developed subepithelial disease, not visible after Lugol's endoscopy. Based on currently available data, we advise the restriction of the use of RFA for flat MGIN and HGIN without the pink-color sign on Lugol's chromoendoscopy. (Clinical trial registration number: NCT02047305.)
Background and Aims: Colonoscopy insertion is painful for some patients and is one of the main barriers to screening colonoscopy. Few studies have assessed the impact of the supine position (SP) on colonoscopy insertion, especially for unsedated patients. The aim of this randomized controlled trial was to clarify this issue. Methods: Unsedated patients were randomized to SP or left horizontal position (LHP) as the starting position of colonoscopy insertion. The primary outcome measure was cecal intubation time (CIT), and the secondary outcome measures were descending colon intubation time (DIT), pain score of patients, difficulty score according to the endoscopist, and patients' acceptance of unsedated colonoscopy. Results: A total of 347 patients were randomized to the SP group (175) or the LHP group (172). The CIT in the SP group was significantly shorter than that in the LHP group (275.0 seconds [interquartile range (IQR), 234.0-328.5 seconds] versus 316.0 seconds [IQR, 261.0-370.0 seconds], P < .001). The DIT was also shorter in the SP group (64.5 seconds [IQR, 52.0-86.3 seconds] versus 74.0 seconds [IQR, 62.0-92.0 seconds], P = .001). Compared with the LHP, the SP had a lower pain score (3.3 versus 3.9, P = .002), a lower difficulty score (3.1 versus 3.7, P < .001), a lower frequency of position change (7.1% versus 38.0%, P < .001), and less need for abdominal compression (39.1% versus 45.5%, P = .02). SP was the only modifiable and independent factor identified to reduce CIT and pain score and improve patients' acceptance of unsedated colonoscopy. Conclusions: As an economical and convenient method, SP can reduce CIT, ease pain, and improve patients' acceptance of unsedated colonoscopy.