Diagnostic performance of deep-learning-based virtual chromoendoscopy in gastric neoplasms.

BACKGROUNDS: Cycle-consistent generative adversarial network (CycleGAN) is a deep neural network model that performs image-to-image translations. We generated virtual indigo carmine (IC) chromoendoscopy images of gastric neoplasms using CycleGAN and compared their diagnostic performance with that of white light endoscopy (WLE). METHODS: WLE and IC images of 176 patients with gastric neoplasms who underwent endoscopic resection were obtained. We used 1,633 images (911 WLE and 722 IC) of 146 cases in the training dataset to develop virtual IC images using CycleGAN. The remaining 30 WLE images were translated into 30 virtual IC images using the trained CycleGAN and used for validation. The lesion borders were evaluated by 118 endoscopists from 22 institutions using the 60 paired virtual IC and WLE images. The lesion area concordance rate and successful whole-lesion diagnosis were compared. RESULTS: The lesion area concordance rate based on the pathological diagnosis in virtual IC was lower than in WLE (44.1% vs. 48.5%, p < 0.01). The successful whole-lesion diagnosis was higher in the virtual IC than in WLE images; however, the difference was insignificant (28.2% vs. 26.4%, p = 0.11). Conversely, subgroup analyses revealed a significantly higher diagnosis in virtual IC than in WLE for depressed morphology (41.9% vs. 36.9%, p = 0.02), differentiated histology (27.6% vs. 24.8%, p = 0.02), smaller lesion size (42.3% vs. 38.3%, p = 0.01), and assessed by expert endoscopists (27.3% vs. 23.6%, p = 0.03). CONCLUSIONS: The diagnostic ability of virtual IC was higher for some lesions, but not completely superior to that of WLE. Adjustments are required to improve the imaging system's performance.

Randomized Controlled Trial of Remimazolam Compared with Placebo in Japanese Patients Undergoing Colonoscopy: A Phase III, Investigator-Initiated Trial.

INTRODUCTION: We conducted an investigator-initiated clinical trial in which remimazolam was used to achieve sedation in patients undergoing colonoscopies. METHODS: This multicenter, double-blind, placebo-controlled, phase III investigator-initiated trial included patients who underwent colonoscopy under sedation with remimazolam (initial dose: 3 mg; additional dose: 1 mg) or normal saline (placebo). The primary endpoint of the study was the successful sedation rate during colonoscopy, defined as achieving a Modified Observer's Assessment of Alertness/Sedation (MOAA/S) score of ≤4 before the procedure, maintaining this score throughout colonoscopy, and requiring no more than five additional drug doses per 15 min. RESULTS: The sedation success rate was 95.0% (38/40 patients) in the remimazolam group and 0.0% (0/11 patients) in the placebo group (p < 0.01). The time from the end of procedure to regaining consciousness was 0.0 (interquartile range: 0.0-0.0) min in both groups. The time from the end of the procedure to ambulation was 5.0 (interquartile range: 0.0-10.0) min in the remimazolam group and 0.0 (interquartile range: 0.0-0.0) min in the placebo group (p = 0.02). Serious adverse events were not observed. CONCLUSION: The use of remimazolam to achieve sedation in Japanese patients undergoing colonoscopy was more effective than placebo.

Development of a new endoscopy system to visualize bilirubin for the diagnosis of duodenogastroesophageal reflux.

OBJECTIVES: Reflux hypersensitivity (RH) is a form of refractory gastroesophageal reflux disease in which duodenogastroesophageal reflux (DGER) plays a role. This study aimed to determine the usefulness of an endoscopy system equipped with image-enhanced technology for evaluating DGER and RH. METHODS: The image enhancement mode for detecting bilirubin and calculated values were defined as the Bil mode and Bil value, respectively. First, the visibility of the Bil mode was validated for a bilirubin solution and bile concentrations ranging from 0.01% to 100% (0.002-20 mg/dL). Second, visibility scores of the Bil mode, when applied to the porcine esophagus sprayed with a bilirubin solution, were compared to those of the blue laser imaging (BLI) and white light imaging (WLI) modes. Third, a clinical study was conducted to determine the correlations between esophageal Bil values and the number of nonacid reflux events (NNRE) during multichannel intraluminal impedance-pH monitoring as well as the utility of esophageal Bil values for the differential diagnosis of RH. RESULTS: Bilirubin solution and bile concentrations higher than 1% were visualized in red using the Bil mode. The visibility score was significantly higher with the Bil mode than with the BLI and WLI modes for 1% to 6% bilirubin solutions (P < 0.05). The esophageal Bil value and NNRE were significantly positively correlated (P = 0.031). The area under the receiver operating characteristic curve for the differential diagnosis of RH was 0.817. CONCLUSION: The Bil mode can detect bilirubin with high accuracy and could be used to evaluate DGER in clinical practice.

Randomized controlled trial of remimazolam compared with placebo in Japanese patients undergoing upper gastrointestinal endoscopy: Phase III investigator-initiated clinical trial.

OBJECTIVES: To assess the effectiveness of remimazolam against normal saline (placebo) as a sedative agent for endoscopy in a multicenter, randomized, double-blind, investigator-initiated phase III controlled trial. METHODS: We included 48 Japanese patients undergoing upper gastrointestinal endoscopy. For the procedure, an initial remimazolam dose of 3 mg and additional doses of 1 mg were administered, as determined in the phase II clinical study. The primary study end-point was the successful sedation rate during gastrointestinal endoscopy, determined as a Modified Observer's Assessment of Alertness/Sedation score ≤4 before the start of endoscopy, the completion of gastrointestinal endoscopy, and two or fewer additional doses per 6 min. RESULTS: The successful endoscopy sedation rates were 91.9% and 9.1% in the remimazolam and placebo groups, respectively (P < 0.01). The time from the end of endoscopy to arousal was 0.0 (0.0-0.0) min for both groups. The number of additional doses required to achieve sedation was lower in the remimazolam group than that in the placebo group (P < 0.01). CONCLUSIONS: Remimazolam demonstrated a significantly higher sedation effect during upper gastrointestinal endoscopy in Japanese patients with safe and fast recovery compared with placebo.

Hybrid and Conventional Endoscopic Submucosal Dissection for Early Gastric Neoplasms: A Multi-Center Randomized Controlled Trial.

BACKGROUND & AIMS: Hybrid endoscopic submucosal dissection (H-ESD), which incorporates endoscopic submucosal dissection (ESD) with endoscopic mucosal resection, has been developed to make ESD technically easier. This study aimed to determine if H-ESD is superior to conventional ESD (C-ESD) for small early gastric neoplasms (EGNs). METHODS: We conducted a multi-center, prospective, open-label, randomized controlled trial to compare the treatment outcomes of H-ESD and C-ESD (Hybrid-G Trial). Patients with differentiated type intramucosal EGN ≤20 mm in diameter and without ulceration were randomly assigned (1:1) to groups that underwent H-ESD or C-ESD. A single multi-functional snare, SOUTEN (ST1850-20, Kaneka, Medix, Tokyo, Japan), was used for H-ESD. The primary outcome was procedure time. Secondary outcomes included mucosal incision time, time and speed of submucosal dissection, curability, and endoscopic procedural adverse events. RESULTS: A total of 39 and 40 patients underwent H-ESD and C-ESD, respectively. The procedure time of H-ESD was significantly shorter than that of C-ESD (33.16 min vs 62.46 min; H-ESD/C-ESD ratio: 0.53; 95% confidence interval, 0.41-0.69; P < .0001). There was no significant difference in mucosal incision time between the 2 groups; the time and speed of submucosal dissection of H-ESD were significantly shorter than those of C-ESD. No difference was observed between the 2 groups in other outcomes. CONCLUSIONS: H-ESD has significantly shorter procedure time than C-ESD, with high and comparable curability and safety for both H-ESD and C-ESD. H-ESD can be a good option for the endoscopic treatment of small EGNs. (UMIN Clinical Trials Registry, Numbers: UMIN000041244).

Auxiliary diagnosis of subepithelial lesions by impedance measurement during EUS-guided fine-needle biopsy.

BACKGROUND AND AIMS: EUS-guided FNA/biopsy (EUS-FNA/B) is the citerion standard for diagnosing subepithelial lesions (SELs); however, its diagnostic ability for SELs <20 mm is low. We developed a new diagnostic method to differentiate between GI stromal tumor (GIST) and non-GIST by measuring high-frequency impedance (H-impedance) using an EUS-FNB needle. METHODS: The H-impedance of gastric epithelial neoplasms from 16 cases were measured with a conventional impedance probe to confirm whether H-impedance is clinically useful for assessing cell density (study 1). The H-impedance values of exposed SELs from 25 cases with use of the conventional probe (study 2) and nonexposed SELs from 20 cases with use of the EUS-FNB needle probe (study 3) were measured to determine the diagnostic ability of H-impedance for differentiating GISTs from non-GISTs. RESULTS: H-impedance significantly positively correlated with cell density (P = .030) (study 1). The H-impedance of GIST (99.5) measured with a conventional probe was significantly higher than with those of the muscular layer (82.4) and leiomyoma (89.2) (P < .01) (study 2). The H-impedance of GIST measured with the EUS-FNB needle was also significantly higher than that of leiomyoma (GIST: 80.2 vs leiomyoma, 71.8; P = .015). The diagnostic yield of the impedance method for differentiating GISTs from non-GISTs had 94.4% accuracy, 88.9% sensitivity, 100% specificity, and 0.95 area under the curve. Diagnostic ability was not affected by lesion size (P = .86) (study 3). CONCLUSION: Auxiliary differential diagnosis between gastric GISTs and non-GISTs by the H-impedance measurement during EUS-FNB could be a good option, especially when the lesion is <20 mm.

Negligible procedure-related dissemination risk of mucosal incision-assisted biopsy for gastrointestinal stromal tumors versus endoscopic ultrasound-guided fine-needle aspiration/biopsy.

BACKGROUND: Mucosal incision-assisted biopsy (MIAB) is a valuable alternative to endoscopic ultrasound-guided fine-needle aspiration/biopsy (EUS-FNAB) for sampling gastric subepithelial lesions (SELs). This study aimed to evaluate the potential risk of dissemination and impact on postoperative prognosis associated with MIAB, which has not yet been investigated. METHODS: Study 1: A prospective observational study was conducted to examine the presence or absence and growth rate of tumor cells in gastric juice before and after the procedure in patients with SELs who underwent MIAB (n = 25) or EUS-FNAB (n = 22) between September 2018 and August 2021. Study 2: A retrospective study was conducted to examine the impact of MIAB on postoperative prognosis in 107 patients with gastrointestinal stromal tumors diagnosed using MIAB (n = 39) or EUS-FNAB (n = 68) who underwent surgery between January 2001 and July 2020. RESULTS: In study 1, although no tumor cells were observed in gastric juice in MIAB before the procedure, they were observed in 64% of patients after obtaining samples (P < 0.001). In contrast, no tumor cells were observed in the gastric juice in EUS-FNAB before and after the procedure. In study 2, there was no significant difference in 5-year disease-free survival between MIAB (100%) and EUS-FNAB (97.1%) (P = 0.27). CONCLUSION: MIAB is safe, with little impact on postoperative prognosis, although the procedure releases some tumor cells after damaging the SEL's pseudocapsule.

Comparison of treatment outcomes between endoscopic submucosal dissection with the needle-type knife and insulated-tip knife for superficial esophageal neoplasms.

Our study aimed to compare the treatment outcomes between endoscopic submucosal dissection (ESD) with an insulated-tip knife (ESD-IT) and a needle-type knife (ESD-N) for large superficial esophageal neoplasms, as no study of this kind has been previously reported. We used the dataset of a multicenter, randomized controlled trial that compared conventional ESD (C-ESD) and traction-assisted ESD (TA-ESD) for superficial esophageal neoplasms. We compared the procedural outcomes between ESD-IT and ESD-N in a post hoc analysis and conducted sub-analyses based on traction assistance and electrical knife type. We included 223 (EST-IT, n = 169; ESD-N, n = 54) patients with no significant differences in baseline characteristics. The operator handover rate due to ESD difficulties was significantly higher in ESD-N (ESD-IT = 0.6% vs. ESD-N = 13.0%, P = 0.001), while the injection volume was significantly higher in ESD-IT than in ESD-N (40.0 vs. 20.5 mL, P < 0.001). Other outcomes were comparable between both groups (procedural time: 51.0 vs. 49.5 minute, P = 0.89; complete resection: 90.5% vs. 90.7%, P > 0.99; and complication rate: 1.8% vs. 3.7%, P = 0.60 for ESD-IT and ESD-N, respectively). In the sub-analyses, the handover rate was significantly lower with TA-ESD than with C-ESD for ESD-N (3.2% vs. 26.1%, P = 0.034), and a significantly smaller injection volume was used in TA-ESD than in C-ESD for ESD-IT (31.5 vs. 47.0 mL, P < 0.01). ESD with either endoscopic device achieved favorable treatment outcomes with low complication rates. The handover rate in ESD-N and the injection volume in ESD-IT improved with the traction method.

Randomized controlled trial comparing conventional and traction endoscopic submucosal dissection for early colon tumor (CONNECT-C trial).

OBJECTIVES: Endoscopic submucosal dissection (ESD) is a widely used treatment for early gastrointestinal cancer. However, colon ESD remains challenging. Previous studies on colon ESD using the traction method used a small sample, single-center design, providing insufficient evidence of this procedure's efficacy. We thus aimed to investigate the efficacy and safety of the traction method in colon ESD in this multicenter randomized trial. METHODS: We conducted a prospective, multicenter, randomized, two-arm controlled trial at 10 facilities in Japan. A 1:1 allocation was conducted for the conventional ESD (C-ESD) and traction ESD (T-ESD) groups. The primary end-point was ESD procedure time. RESULTS: We included 128 C-ESD and 123 T-ESD cases from April 2020 to August 2021. The median procedure times for C-ESD and T-ESD were 61 (40-100) and 53 (40-76) min (P = 0.18), respectively, and no significant differences were observed between the groups. Subgroup analysis showed that the median procedure times for patients with a lesion diameter of ≥30 mm in the C-ESD and T-ESD groups were 89 (57-80) and 69 (50-104) min (P = 0.05), respectively, and for nonexpert operators were 81 (62-120) and 64 (52-109) min (P = 0.07), respectively. CONCLUSIONS: The traction method did not contribute to a significantly shortened ESD procedure time. However, this method may be useful when the tumor diameter is large or if the procedure is conducted by nonexpert endoscopists.

The treatment effects of acotiamide in esophagogastric outflow obstruction: a prospective longitudinal observational study.

OBJECTIVES: We have found that an altered lower esophageal sphincter (LES) accommodation response is an underlying cause of esophagogastric junction outflow obstruction (EGJOO). The objective of this study was to examine the treatment effect of acotiamide, a prokinetic agent which improves impaired gastric accommodation in functional dyspepsia, in patients with EGJOO. METHODS: A prospective observational longitudinal study was conducted between October 2014 and March 2020. Acotiamide (100 mg, 3 times a day) was administered to 25 patients with EGJOO for 4 weeks. High-resolution manometry (HRM) was performed just before and after 4 weeks of treatment. RESULTS: As the primary outcome, the extent of integrated relaxation pressure (IRP) after treatment (14.6, 12.1-22.0 mmHg) was significantly lower than that before treatment (19.4, 17.1-27.4 mmHg). The extent of LES accommodation index after treatment (32.7, 21.0-40.0 mmHg) was also significantly lower than that before treatment (39.3, 31.2-50.2 mmHg). Acotiamide normalized the IRP (< 15 mmHg) in 13 of 25 patients with EGJOO (52%), and the IRP was decreased in 20 of 25 patients with EGJOO (80%). As the secondary outcome, the total FSSG score in 25 patients with EGJOO before and after acotiamide treatment showed no significant difference. In a sub-analysis of 13 patients in whom EGJOO was normalized by acotiamide, however, dysphagia was reported to be significantly improved by acotiamide. CONCLUSIONS: Acotiamide has a treatment effect on patients with EGJOO via a reduction in the IRP level through the lowering of both the basal LES pressure and LES accommodation response. Dysphagia is a key symptom to be evaluated and treated in patients with EGJOO.

Efficacy of macrogol 4000 plus electrolytes in bowel preparation for colonoscopy in patients with chronic constipation.

BACKGROUND: Chronic constipation is a significant factor in poor bowel preparation for colonoscopy. Macrogol 4000 plus electrolytes (Movicol, EA Pharma, Tokyo, Japan), containing polyethylene glycol (PEG) and electrolytes, have been used recently to treat patients with constipation. However, prospective studies on the use of macrogol 4000 for bowel cleansing for colonoscopy are lacking. This study aimed to investigate the efficacy and safety of macrogol 4000 in addition to PEG administered in patients with chronic constipation. METHODS: This single-center, single-arm prospective study enrolled patients with chronic constipation who were scheduled to undergo colonoscopy. The primary endpoint was the proportion of good bowel preparation assessed using the Boston bowel preparation scale (BBPS) (6 or more points). The secondary endpoints were the time from when pPEG (MoviPrep, EA Pharma, Tokyo, Japan) was taken until colonoscopy could be started, amount of PEG taken, number of defecations, whether additional PEG doses were taken, and adverse events. Endoscopy-related endpoints included cecal intubation rate, insertion time, observation time, adenoma detection rate (ADR), and polyp detection rate (PDR). The tolerability of PEG and macrogol 4000 was assessed using a questionnaire. RESULTS: Forty patients were included in the analysis. The median BBPS was 7 (range 3-9) and ≥ 6 points in 37 cases (92.5%). The median time until colonoscopy can be started was 210 min (90-360 min), the median volume of PEG taken was 1500 mL (1000-2000 mL), and the median number of defecations was 7 (3-20). No adverse events were observed. Fourteen patients required an additional dose of PEG. Cecal intubation was achieved in all cases, the median insertion time was 6.0 min (range 2.3-22 min), and the median observation time was 8.8 min (range 4.0-16.0 min). The ADR and PDR were 60.0% and 75.0%, respectively. A proportion of patients rated the tolerability of macrogol 4000 and PEG as 95.0% and 50.0%, respectively. CONCLUSIONS: Intake of macrogol 4000 in addition to PEG is effective and safe for colonoscopy in patients with chronic constipation. Clinical trial registration statement This study was registered in the UMIN-CTR database (UMIN-ID000038315).

Self-Completion Method of Endoscopic Submucosal Dissection Using Endosaber without Any Other Device or Assistance: An ex vivo Porcine Model Study.

BACKGROUND: Endoscopic submucosal dissection (ESD) is a standard treatment for tumors of the gastrointestinal tract. We developed a self-completion method of ESD using Endosaber to eliminate the need for an additional device or human assistance during the procedure. The aim of this study was to evaluate the technical feasibility and outcomes of this method in an ex vivo porcine training model. METHODS: This was a pilot study, and the procedures were performed by 4 experts. Mock lesions measuring 15 mm in diameter were prepared at the posterior wall in the middle or lower esophagus obtained from domestic pigs. Each operator performed ESD on the mock lesions in 3 models. The primary outcome was ESD completion rate using the self-completion method. The secondary outcomes were procedure time, en bloc resection rate, perforation rate, and number of injections during the procedure. RESULTS: All 12 ESDs were completed using the self-completion method. The median procedure time (interquartile range) was 483.5 (399-619.3) s (median incision time: 240.4 [168.3-332.5] s; median dissection time: 222 [182.8-257] s). En bloc resection rate was 100%. No perforation was noted during any of the procedures. The median number of injections was 10.5 (9-14.3). The procedure time decreased significantly with increase in experience (p = 0.020). CONCLUSIONS: The self-completion ESD method using one Endosaber without any assistance achieved a 100% en bloc resection rate without any perforation. The need for an additional device or assistance was successfully eliminated. This method may prove to be a simple and cost-effective ESD procedure for lesions in humans.

Transrectal laparoscopy using flexible endoscopy with a submucosal tunneling method: Porcine survival model.

OBJECTIVES: Transrectal laparoscopy (TRLS) using a flexible endoscope was recently proposed for peritoneal observation. Although previous studies have reported its feasibility, follow-up durations were insufficient to ascertain technical safety. Moreover, knowledge about the technical feasibility of collecting peritoneal cytological lavage or ascites during TRLS is limited. Thus, this study aimed to confirm the safety and efficacy of TRLS in a porcine survival model. METHODS: After creating artificial ascites in 10 animals, TRLS was performed as follows: submucosal tunnel creation on the anterior wall of the rectum, intentional perforation at the distal end of the tunnel, endoscopic ascites collection and intraperitoneal observation, and clip closure at the mucosal incision site. The pigs were administered antibiotics orally for 7 days after TRLS and killed for histological evaluation and bacterial culture after 28 days of observation. RESULTS: The technical success rates of insertion into the abdominal cavity, ascites collection, and clip closure were 100%. All frequent anatomical sites for peritoneal dissemination including the stomach, subdiaphragmatic space, and pelvic space were fully observable without adverse events. The median procedure time was 36.3 min. Full 28-day survival was observed in all cases without any infection. The autopsies showed no infection, including abscess formation. Bacterial cultures of the peritoneal cavity were negative 28 days after TRLS in all cases. CONCLUSIONS: Transrectal laparoscopy enabled ascites collection and intraperitoneal observation without adverse events. All animals survived without peritonitis. Therefore, TRLS can be an option for intraperitoneal examination.

The effect of scissor-type versus non-scissor-type knives on the technical outcomes in endoscopic submucosal dissection for superficial esophageal cancer: a multi-center retrospective study.

The Clutch Cutter was invented as a scissor-type knife for endoscopic submucosal dissection (ESD) of gastrointestinal neoplasms. ESD with the scissor-type knife (ESD-S) may be considered a technically easier procedure than ESD with non-scissor-type knives (ESD-NS). Therefore, this study aimed to compare the technical outcomes of ESD-S with those of ESD-NS for superficial esophageal cancer. This was a multicenter retrospective study. Patients with superficial esophageal cancer treated with ESD between October 2015 and March 2018 at three hospitals were retrospectively reviewed. The ESD-S group had 48 patients and the ESD-NS group had 114 patients. A propensity score matching analysis was performed to compensate for the confounding bias between both groups. Multivariate analyses and propensity score matching were used to adjust for age, sex, the tumor size, tumor location, tumor depth, degree of tumor circumference, operator level, usage of the traction method, and the sedation method. The primary outcome was the procedure time of the ESD. Secondary outcomes were the rate of en-bloc/complete resection and the rate of complications including perforation, delayed bleeding, and stricture. Propensity score matching analysis provided 36 matched pairs. Median procedure time in the ESD-S group was significantly shorter than that in the ESD-NS group (44.0 min vs. 66.5 min, P = 0.020). In addition, the treatment outcomes were similar in both groups (en-bloc resection: 100% vs. 97.2%, P = 1; complete resection: 88.9% vs. 86.1%, P = 1; curative resection: 80.6% vs. 77.8%, P = 1; perforation: 0% vs. 5.6%, P = 0.49; delayed bleeding: 0% in both groups; stricture: 2.8% vs. 8.3%, P = 0.61). ESD-S was associated with a shorter procedure time than ESD-NS, without an increase in the incidence of complications. Therefore, the scissor-type knife should be considered as an endo-knife for ESD of superficial esophageal cancers.

Ex vivo porcine model study on the treatment outcomes of scissor-type knife versus needle-type knife in endoscopic submucosal dissection performed by trainees.

BACKGROUND: Endoscopic submucosal dissection (ESD) for gastrointestinal neoplasms can be technically difficult for trainee endoscopists. Presently, there is no consensus for trainees to select the endo-knife type in ESD. Therefore, we conducted a comparison study of treatment outcomes between scissors-type and needle-type knives in ESD performed by trainees in an ex vivo porcine model. METHODS: This study was conducted on trainee endoscopists who participated in ESD hands-on seminars held in August 2018 and September 2019. A total of 22 trainees from 13 institutions were divided into two groups according to their endoscopic experience. Under expert supervision, each trainee performed two ESDs in porcine models, namely, scissor-type knife (ESD-S) and needle-type knife (ESD-N). The efficacy and safety, including the procedure time and rates of self-completion, en bloc resection, and complications, were compared between ESD-S and ESD-N. In subgroup analysis, we also investigated the predictors associated with the difficulty of ESD for trainees using multivariate logistic regression analysis. RESULTS: Eight trainees had an experience of over 1000 endoscopies (senior trainee: S-Trainee), whereas the others had an experience of less than 1000 endoscopies (junior trainee: J-Trainee). Among the S-Trainees, no significant differences were observed in any treatment outcome between ESD-S and ESD-N. Among the J-Trainees, the total procedure and mucosal incision times were significantly shorter in ESD-S than in ESD-N [total procedure time: 16.5 min (range 10.0-31.0) vs. 22.3 min (range 10.0-38.0), P = 0.018; circumferential incision time: 10.0 min (range 6-16) vs. 17.0 min (range 5.0-31.5); P = 0.019]. Regarding complications, muscular injury occurred in two patients during ESD-N performed by J-Trainees; however, no muscular injury occurred during ESD-S. In subgroup analysis, ESD-N was an independent predictive factor of difficult ESD (odds ratio 5.28, 95% confidence interval 1.25-22.30; P = 0.024). CONCLUSIONS: This study revealed that trainees, particularly those who have experienced less than 1000 endoscopies, should opt for the scissor-type knife to perform ESD.

Is Additional Surgery Always Sufficient for Preventing Recurrence After Endoscopic Submucosal Dissection with Curability C-2 for Early Gastric Cancer?

BACKGROUND: When a lesion does not meet the curative criteria of endoscopic submucosal dissection (ESD) for early gastric cancer (EGC), referred to as non-curative resection or curability C-2 in the guidelines, an additional surgery is the standard therapy because of the risk of lymph node metastasis (LNM). OBJECTIVE: This study aimed to identify high-risk patients for recurrence after additional surgery for curability C-2 ESD of EGC. METHODS: This multicenter retrospective cohort study enrolled 1064 patients who underwent additional surgery after curability C-2 ESD for EGC. We evaluated the recurrence rate and the risk factors for recurrence after additional surgery in these patients. RESULTS: The 5-year recurrence rate after additional surgery was 1.3%. Multivariate Cox analysis revealed that the independent risk factors for recurrence after additional surgery were LNM (hazard ratio [HR] 32.47; p < 0.001) and vascular invasion (HR 4.75; p = 0.014). Moreover, patients with both LNM and vascular invasion had a high rate of recurrence after additional surgery (24.6% in 5 years), with a high HR (119.32) compared with those with neither LNM nor vascular invasion. Among patients with no vascular invasion, a high rate of recurrence was observed in those with N2/N3 disease according to the American Joint Committee on Cancer TNM staging system (27.3% in 5 years), in contrast with no recurrence in those with N1 disease. CONCLUSIONS: Patients with both LNM (N1-N3) and vascular invasion, as well as those with N2/N3 disease but no vascular invasion, would be candidates for adjuvant chemotherapy after additional surgery for curability C-2 ESD of EGC.

Different risk factors between early and late cancer recurrences in patients without additional surgery after noncurative endoscopic submucosal dissection for early gastric cancer.

BACKGROUND AND AIMS: Cancer recurrence is observed in some patients without additional radical surgery after endoscopic submucosal dissection (ESD) that does not fulfill the curability criteria for early gastric cancer (EGC), categorized as "noncurative resection" or "curability C-2" in the guidelines. However, time to cancer recurrence is different in such patients. Thus, we aimed to identify the risk factors of early and late cancer recurrences in these patients. METHODS: Between 2000 and 2011, this multicenter study analyzed 905 patients who were followed up without additional radical surgery after ESD for EGC categorized as curability C-2. We evaluated the risk factors for early and late cancer recurrences, separately, after ESD. The cut-off value was defined at 2 years. RESULTS: Time to cancer recurrence in the enrolled patients showed a bimodal pattern, and the 5-year cancer recurrence rate was 3.2%. Multivariate Cox analyses revealed that lymphatic invasion (hazard ratio [HR], 8.56; P = .003) was the sole independent risk factor for early cancer recurrence. Regarding late cancer recurrence, vascular invasion (HR, 4.50; P = .039) was an independent risk factor, and lymphatic invasion tended to be a risk factor (HR, 3.63; P = .069). CONCLUSIONS: This multicenter study with a large cohort demonstrated that lymphatic invasion is mainly associated with early cancer recurrence; however, vascular invasion was a risk factor only for late recurrence in patients without additional treatment after ESD for EGC categorized as curability C-2. This finding may contribute to decision making for treatment strategies after ESD, especially for patients with a relatively short life expectancy.