Evaluating false-positive detection in a computer-aided detection system for colonoscopy.

BACKGROUND AND AIM: Computer-aided detection (CADe) systems can efficiently detect polyps during colonoscopy. However, false-positive (FP) activation is a major limitation of CADe. We aimed to compare the rate and causes of FP using CADe before and after an update designed to reduce FP. METHODS: We analyzed CADe-assisted colonoscopy videos recorded between July 2022 and October 2022. The number and causes of FPs and excessive time spent by the endoscopist on FP (ET) were compared pre- and post-update using 1:1 propensity score matching. RESULTS: During the study period, 191 colonoscopy videos (94 and 97 in the pre- and post-update groups, respectively) were recorded. Propensity score matching resulted in 146 videos (73 in each group). The mean number of FPs and median ET per colonoscopy were significantly lower in the post-update group than those in the pre-update group (4.2 ± 3.7 vs 18.1 ± 11.1; P < 0.001 and 0 vs 16 s; P < 0.001, respectively). Mucosal tags, bubbles, and folds had the strongest association with decreased FP post-update (pre-update vs post-update: 4.3 ± 3.6 vs 0.4 ± 0.8, 0.32 ± 0.70 vs 0.04 ± 0.20, and 8.6 ± 6.7 vs 1.6 ± 1.7, respectively). There was no significant decrease in the true positive rate (post-update vs pre-update: 95.0% vs 99.2%; P = 0.09) or the adenoma detection rate (post-update vs pre-update: 52.1% vs 49.3%; P = 0.87). CONCLUSIONS: The updated CADe can reduce FP without impairing polyp detection. A reduction in FP may help relieve the burden on endoscopists.

Clinical outcomes and learning curve of Tip-in endoscopic mucosal resection for 15-25 mm colorectal neoplasms among non-experts.

BACKGROUND AND AIM: Tip-in endoscopic mucosal resection (EMR) has a high en bloc resection rate for large colorectal neoplasms. However, non-experts' performance in Tip-in EMR has not been investigated. We investigated whether Tip-in EMR can be achieved effectively and safely even by non-experts. METHODS: This retrospective study included consecutive patients who underwent Tip-in EMR for 15-25 mm colorectal nonpedunculated neoplasms at a Japanese tertiary cancer center between January 2014 and December 2020. Baseline characteristics, treatment outcomes, learning curve of non-experts, and risk factors of failing self-achieved en bloc resection were analyzed. RESULTS: A total of 597 lesions were analyzed (438 by experts and 159 by non-experts). The self-achieved en bloc resection (69.8% vs 88.6%, P < 0.001) and self-achieved R0 resection (58.3% vs 76.5%, P < 0.001) rates were significantly lower in non-experts with <10 cases of experience than in experts, but not in non-experts with >10 cases. Adverse event (P = 0.165) and local recurrence (P = 0.892) rates were not significantly different between experts and non-experts. Risk factors of failing self-achieved en bloc resection were non-polypoid morphology (OR 3.4, 95% CI 1.6-7.3, P = 0.001), lesions with an underlying semilunar fold (OR 3.6, 95% CI 1.6-7.3, P < 0.001), positive non-lifting sign (OR 3.1, 95% CI 1.2-8.0, P = 0.023), and non-experts with an experience of ≤10 cases (OR 3.6, 95% CI 2.1-6.3, P < 0.001). CONCLUSION: The clinical outcomes of Tip-in EMR for 15-25 mm lesions performed by non-experts were favorable.

A novel low-power pure-cut hot snare polypectomy for 10-14 mm colorectal adenomas: An ex vivo and a clinical prospective feasibility study (SHARP trial).

BACKGROUND AND AIM: Hot snare polypectomy using blend or coagulation current is widely used; however, it causes deeper tissue heat injury, leading to adverse events. We hypothesized that hot polypectomy using low-power pure cut current (PureCut, effect 1 10 W) could reduce deeper tissue heat injury. We conducted animal experiments to evaluate the deeper tissue heat injury and conducted a prospective clinical study to examine its cutting ability. METHODS: In a porcine rectum, hot polypectomy using Blend current (EndoCut, effect 3 40 W) and low-power pure cut current was performed. The deepest part of heat destruction and thickness of the non-burned submucosal layer were evaluated histologically. Based on the results, we performed low-power pure cut current hot polypectomy for 10-14 mm adenoma. The primary endpoint was complete resection defined as one-piece resection with negative for adenoma in quadrant biopsies from the defect margin. RESULTS: In experiments, all low-power pure-cut resections were limited within the submucosal layer whereas blend current resections coagulated the muscular layer in 13% (3/23). The remaining submucosal layer was thicker in low-power pure cut current than in blend current resections. In the clinical study, low-power pure-cut hot polypectomy removed all 100 enrolled polyps. For 98 pathologically neoplastic polyps, complete resection was achieved in 84 (85.7%, 95% confidence interval, 77-92%). The lower limit of the 95% confidence interval was not more than 15% below the pre-defined threshold of 86.6%. No severe adverse events occurred. CONCLUSIONS: A novel low-power pure-cut hot polypectomy may be feasible for adenoma measuring 10-14 mm. (UMIN000037678).

Investigation of vertical margin involvement in endoscopic resection for T1 colorectal cancer.

OBJECTIVES: The resection of vertical margin-negative submucosally invasive colorectal cancer (CRC) relies on the pathological risk assessment of lymph node metastasis. However, no large-scale study has clarified the endoscopic resection (ER) outcome for submucosally invasive CRC, focusing on the vertical margin status. This retrospective study aimed to examine vertical margin involvement in ER for submucosally invasive CRC and explore the treatment consequences associated with vertical margin status. METHODS: We analyzed 395 submucosally invasive CRC cases in 389 patients who underwent ER at our hospital between 2008 and 2020. The presence of residual tumors and simultaneous lymph node metastasis in patients who underwent additional surgery was assessed and compared between the vertical incomplete ER and the vertical margin-negative groups. RESULTS: Among the patients, 270 were men, with a median age of 69 years. The vertical incomplete ER rate was 21.5%, with positive vertical margins and unclear vertical margins identified in 12.2% and 9.3% of the cases, respectively. Among 154 patients who underwent additional surgery after ER, the vertical incomplete ER group had a significantly higher residual tumor rate than the vertical margin-negative group (P = 0.001). The vertical incomplete ER group had a significantly higher lymph node metastasis rate than the vertical margin-negative group (P = 0.029). CONCLUSION: This study clarified the substantial risk of vertical incomplete ER in submucosally invasive CRC and revealed the high risk of residual tumor and lymph node metastasis in vertical incomplete ER for submucosal CRC.

Ultrafast Ring Closure Reaction of Gaseous cis-Stilbene from S1(ππ*).

cis-Stilbene (cis-St) is a well-known benchmark system for cis-trans photoisomerization. cis-St also produces 4a,4b-dihydrophenanthrene (DHP) in solution with a quantum yield of less than 0.19. The ring closure reaction, however, has never been identified for gaseous cis-St, and a recent computational simulation predicted the quantum yield of DHP to be only 0.04. In the present study, we identified an ultrafast ring closure reaction of gaseous cis-St for the first time using extreme ultraviolet time-resolved photoelectron spectroscopy. Surface hopping trajectory calculations at the SA3-XMS-CASPT2(2,2) level of theory reproduce the features of the observed time-resolved photoelectron spectra and predict the cis-St:DHP:trans-St branching ratio to be 0.55:0.41:0.04, in contrast with previous estimates. The results indicate that photoexcited cis-St favors ring closure over cis-trans isomerization under the isolated condition.

Formation of Long-Lived Dark States during Electronic Relaxation of Pyrimidine Nucleobases Studied Using Extreme Ultraviolet Time-Resolved Photoelectron Spectroscopy.

Ultrafast electronic relaxation of nucleobases from 1ππ* states to the ground state (S0) is considered essential for the photostability of DNA. However, transient absorption spectroscopy (TAS) has indicated that some nucleobases in aqueous solutions create long-lived 1nπ*/3ππ* dark states from the 1ππ* states with a high quantum yield of 0.4-0.5. We investigated electronic relaxation in pyrimidine nucleobases in both aqueous solutions and the gas phase using extreme ultraviolet (EUV) time-resolved photoelectron spectroscopy. Femtosecond EUV probe pulses cause ionization from all electronic states involved in the relaxation process, providing a clear overview of the electronic dynamics. The 1nπ* quantum yields for aqueous cytidine and uracil (Ura) derivatives were found to be considerably lower (<0.07) than previous estimates reported by TAS. On the other hand, aqueous thymine (Thy) and thymidine exhibited a longer 1ππ* lifetime and a higher quantum yield (0.12-0.22) for the 1nπ* state. A similar trend was found for isolated Thy and Ura in the gas phase: the 1ππ* lifetimes are 39 and 17 fs and the quantum yield for 1nπ* are 1.0 and 0.45 for Thy and Ura, respectively. The result indicates that single methylation to the C5 position hinders the out-of-plane deformation that drives the system to the conical intersection region between 1ππ* and S0, providing a large impact on the photophysics/photochemistry of a pyrimidine nucleobase. The significant reduction of 1nπ* yield in aqueous solution is ascribed to the destabilization of the 1nπ* state induced by hydrogen bonding.

A Japanese case of ovarian mucinous adenocarcinoma with germline double variants of MSH2 and BRCA2.

Germline double heterozygosity (GDH) is rarely reported in cases of inherited cancer syndromes, and GDH of a mismatch repair gene and BRCA has never been reported in Japan. Nonetheless, the current report demonstrates a case of ovarian mucinous adenocarcinoma with initiated Lynch syndrome (LS)-related surveillance because of a known germline MSH2 variant. Six and a half years after oophorectomy, multiple tumors developed in the patient's lungs, bones, and lymph nodes, and histology results confirmed mucinous adenocarcinoma. Systemic chemotherapy including an anti-PD-L1 antibody was effective for >1 year, but brain metastases developed. Pathology of the brain tumors showed mucinous adenocarcinoma without expression of MSH2 and MSH6, while multi-gene panel testing demonstrated not only high microsatellite instability and a high tumor mutation burden, but also germline BRCA2 variants. Further, germline testing in relatives confirmed both variants were from the paternal line, from which many LS-related cancers develop, but not BRCA-related cancer.

Endoscopic causes and characteristics of missed gastric cancers after endoscopic submucosal dissection.

BACKGROUND AND AIMS: Because endoscopic submucosal dissection (ESD) for early gastric cancer (EGC) preserves the entire stomach, missed gastric cancers (MGCs) are often found in the remaining gastric mucosa. However, the endoscopic causes of MGCs remain unclear. Therefore, we aimed to elucidate the endoscopic causes and characteristics of MGCs after ESD. METHODS: From January 2009 to December 2018, all patients undergoing ESD for initially detected EGC were enrolled. According to a review of EGD images before ESD, we identified the endoscopic causes (perceptual, exposure, sampling errors, and inadequate preparation) and characteristics of MGC in each endoscopic cause. RESULTS: Of 2208 patients who underwent ESD for initial EGC, 82 patients (3.7%) had 100 MGCs. The breakdown of endoscopic causes of MGCs was as follows: 69 (69%) perceptual errors, 23 (23%) exposure errors, 7 (7%) sampling errors, and 1 (1%) inadequate preparation. Logistic regression analysis showed that the risk factors for perceptual error were male sex (odds ratio [OR], 2.45; 95% confidence interval [CI], 1.16-5.18), isochromatic coloration (OR, 3.17; 95% CI, 1.47-6.84), greater curvature (OR, 2.31; 95% CI, 1.121-4.40), and lesion size ≤12 mm (OR, 1.74; 95% CI, 1.07-2.84). The sites of exposure errors were around the incisura angularis (11 [48%]), posterior wall of the gastric body (6 [26%]), and antrum (5 [21%]). CONCLUSIONS: We identified MGCs in 4 categories and clarified their characteristics. Quality improvements in EGD observation, with attention to the risks of perceptual and site of exposure errors, can potentially prevent missing EGCs.

Clinical outcomes of endoscopic submucosal dissection for superficial circumferential esophageal squamous cell carcinoma.

BACKGROUND AND AIMS: Our aim was to elucidate the clinical outcomes of endoscopic submucosal dissection (ESD) for superficial circumferential esophageal squamous cell carcinoma (cESCC). METHODS: Consecutive patients who underwent ESD for cESCC between 2009 and 2020 were retrospectively reviewed. Short-term outcomes were en-bloc resection, R0 resection, procedure time, and adverse events, whereas long-term outcomes were overall survival (OS), disease-specific survival (DSS), cumulative recurrence rate (CRR), and clinical course. RESULTS: Fifty-two patients with 52 cESCCs (median tumor length, 5.0 cm; interquartile range [IQR], 4.0-6.3) were evaluated. The en-bloc resection and R0 resection rates were 100% (95% confidence interval [CI], 94.4-100) and 69.2% (95% CI, 54.9-81.3), respectively. The median procedure time was 112 minutes (IQR, 87-162). Intraoperative perforations and delayed bleeding occurred in 4 (7.7%) and 1 (1.9%) patients, respectively. Among the 42 patients who underwent ESD alone, 36 (85.7%) experienced esophageal strictures. Within a median follow-up of 49.1 months (IQR, 25.7-74.7), the 4-year OS, DSS, and CRR were 86.2% (95% CI, 71.6-93.6), 95.5% (95% CI, 83.1-98.9), and 11.5% (95% CI, 4.1-23.1), respectively. There was no significant difference in the OS between patients with low-risk cESCC (pT1a, negative lymphovascular invasion, and negative vertical margin) and high-risk lesions, regardless of undergoing additional treatment (P = .93). In 31 patients with low-risk cESCC who were treated with ESD alone, the 4-year OS, DSS, and CRR were 93.2%, 100%, and 0%, respectively. CONCLUSIONS: ESD is a highly curative treatment for cESCC with favorable long-term outcomes, especially in low-risk patients. Stricture-prevention techniques should be improved to optimize the benefits of ESD for cESCC.

Diagnostic performance of endoscopy for subsquamous extension of superficial adenocarcinoma of the esophagogastric junction.

OBJECTIVES: Extension of adenocarcinoma of the esophagogastric junction under the squamous epithelium may lead to errors when determining lateral margins. However, the characteristics of subsquamous extension are unclear. Herein, we evaluated the prevalence and characteristics of subsquamous extension of adenocarcinoma of the esophagogastric junction and the diagnostic performance of endoscopy for this condition. METHODS: Eighty-nine consecutive patients with superficial adenocarcinoma of the esophagogastric junction who underwent endoscopic or surgical resection at a tertiary cancer center between January 2010 and December 2017 were retrospectively evaluated. Endoscopic subsquamous extension was defined as a submucosal tumor-like elevation covered by squamous epithelium and/or a brownish area with abnormal microvessels on the squamous epithelium observed using narrow-band imaging. The diagnostic performance of endoscopy for subsquamous extension was evaluated using histological subsquamous extension as gold standard. RESULTS: Thirty-nine patients (44%) had histological subsquamous extension. Proton pump inhibitor use was significantly associated with histological subsquamous extension [odds ratio: 4.65; 95% confidence interval (CI): 1.77-12.2]. The sensitivity, specificity, positive predictive value, negative predictive value and accuracy of endoscopic subsquamous extension were 56% (95% CI: 40-72%), 96% (86-99%), 92% (73-99%), 74% (62-84%) and 79% (69-87%), respectively. The median length difference between histological and endoscopic subsquamous extension was 2 mm (range: -6 to 9 mm). CONCLUSIONS: The sensitivity of endoscopic diagnosis of subsquamous extension was unsatisfactory. The endoscopic length of subsquamous extension tended to be underestimated. An oral safety margin of one centimeter is reasonable during endoscopic resection of adenocarcinoma of the esophagogastric junction.IMPACT STATEMENT This study will contribute significantly to the literature because this is the first study to determine the difference between the lengths of subsquamous extension detected endoscopically and histologically. This study determines the prevalence of subsquamous extension and identifies characteristics associated with subsquamous extension. An understanding of the risk of subsquamous extension is important when choosing a treatment strategy and planning the resection margins in patients with adenocarcinoma of the esophagogastric junction. This study provides patients with subsquamous extension characteristics and suggests a method for accurately diagnosing this condition.

Complete endoscopic removal rate of detected colorectal polyps in a real world out-patient practical setting.

OBJECTIVES: Colonoscopy with adenomatous polypectomy reduces the incidence and mortality of colorectal cancer. We introduced a strategy of removing all neoplastic polyps in single-session out-patient colonoscopy using cold polypectomy. We aimed to investigate the achievement of single-session complete removal rate of detected colorectal polyps in clinical practice. MATERIALS AND METHODS: This retrospective study included colonoscopy-scheduled 40-79-year-old outpatients, with at least one colorectal neoplasm, between January 2015 and December 2016. Exclusion criteria were: colorectal neoplasms 21 mm or larger in size; pre-examination for colorectal surgery or endoscopic submucosal dissection; colonoscopy performed by health check program; ongoing antithrombotic treatment; inflammatory bowel disease; familial adenomatous polyposis. We defined 'clean colon' as the removal of all detected neoplastic polyps in a single-session colonoscopy. We evaluated clean colon rate, factors relating to clean colon failure and complications. RESULTS: We evaluated 2527 patients (mean age 68 years; 799 women) with 8203 colorectal polyps (7675 adenomas, 423 serrated lesions, 105 Tis and T1 cancers). In 1-4 mm polyps, cold snare polypectomy (CSP; 51.8%) and cold forceps polypectomy (CFP; 45.8%) were applied. Clean colon rates were 95.1% per patient and 97.1% per lesion. The significant factors denoting clean colon failure were inadequate bowel preparation, ≥5 lesions, and the most advanced estimated histology of adenocarcinoma, on multivariate analyses. Post-polypectomy bleeding requiring endoscopic hemostasis occurred in five patients (0.2%) who had undergone endoscopic mucosal resection (EMR) or hot snare polypectomy (HSP). Perforation occurred in one patient (0.04%) with EMR. CONCLUSIONS: The clean colon rates were satisfactory in single-session out-patient colonoscopy using cold polypectomy.

Local recurrence risk of esophageal squamous cell carcinoma due to intralesional damage during endoscopic submucosal dissection.

BACKGROUND AND AIM: It is unclear whether additional treatment should be considered given the recurrence risk after endoscopic submucosal dissection (ESD) for esophageal squamous cell carcinoma (ESCC) when the vertical margin is positive or unclear (VM1/VMX) due to intralesional damage. This study aimed to elucidate the local recurrence risk of ESCC caused by intralesional damage during ESD. METHODS: Among consecutive patients with pT1a ESCCs initially treated by ESD at our institution between January 2006 and December 2018, ESCCs diagnosed as VM1/VMX were retrospectively reviewed. Exclusion criteria were piecemeal resection and any additional treatment after ESD. Intralesional damage included the following three types: a macroscopic hole inside the lesion, an incision from the lateral margin of the specimen into the lesion, and crushing injury or burn effect into the deepest area of the lesion without an obvious hole. The local recurrence rate after ESD was primarily analyzed. RESULTS: Of 1174 pT1a ESCCs initially treated using ESD, 22 lesions were histopathologically diagnosed as VM1/VMX due to intralesional damage (1.9%; 95% confidence interval [CI], 1.2-2.8%). At a median follow-up period of 60.0 (interquartile range, 15.0-84.0) months, no local recurrence was observed (0.0%; 95% CI, 0.0-13.3%) among 21 lesions finally evaluated. CONCLUSIONS: The impact of intralesional damage during ESD for ESCC on local recurrence might be negligible. Follow-up without additional treatment may be acceptable even if intralesional damage occurs and results in VM1/VMX after ESD for pT1a ESCCs.

Optimization of endoscopic ultrasound-guided tissue sample acquisition for commercially available comprehensive genome profiling.

BACKGROUND AND AIM: Optimal tumor samples are crucial for successful analysis using commercially available comprehensive genomic profiling (CACGP). However, samples acquired by endoscopic ultrasound-guided tissue acquisition (EUS-TA) are occasionally insufficient, and no consensus on the optimal number of needle passes required for CACGP exists. This study aimed to explore the optimal number of needle passes required for EUS-TA to procure an ideal sample fulfilling the prerequisite criteria of CACGPs. METHODS: Patients who underwent EUS-TA for solid masses between November 2019 and July 2021 were retrospectively studied. The correlation between the acquisition rate of an ideal sample and the number of needle passes mounted on a microscope slide was evaluated. Additionally, the factors predicting a successful analysis were investigated in patients scheduled for CACGP using EUS-TA-obtained samples during the same period. RESULTS: EUS-TAs using 22- and 19-gauge (G) needles were performed in 336 and 57 patients, respectively. There was a positive correlation between the acquisition rate and the number of passes using a 22-G needle (38.9%, 45.0%, 83.7%, and 100% for 1, 2, 3, and 4 passes, respectively), while no correlation was found with a 19-G needle (84.2%, 83.3%, and 85.0% for 1, 2, and 3 passes, respectively). The analysis success rate in patients with scheduled CACGP was significantly higher with ideal samples than with suboptimal samples (94.1% vs 55.0%, P < 0.01). CONCLUSIONS: The optimal estimated number of needle passes was 4 and 1-2 for 22- and 19-G needles, respectively.

Distortion Correction of Low-Energy Photoelectron Spectra of Liquids Using Spectroscopic Data for Solvated Electrons.

Time-resolved photoelectron spectroscopy (TRPES) enables real-time observation of ultrafast electronic dynamics in solutions. When extreme ultraviolet (EUV) probe pulses are employed, they can ionize solutes from all electronic states involved in the dynamics. However, EUV pulses also produce a strong ionization signal from a solvent that is typically 6 orders of magnitude greater than the pump-probe photoelectron signal of solutes. Alternatively, UV probe pulses enable highly sensitive and selective observation of photoexcited solutes because typical solvents such as water are transparent to UV radiation. An obstacle in such UV-TRPES measurements is spectral distortion caused by electron scattering and a yet to be identified mechanism in liquids. We have previously proposed the spectral retrieval (SR) method as an a posteriori approach to removing the distortion and overcoming this difficulty in UV-TRPES; however, its accuracy has not yet been verified by comparison with EUV-TRPES results. In the present study, we perform EUV-TRPES for charge transfer reactions in water, methanol, and ethanol, and verify SR analysis of UV-TRPES. We also estimate a previously undetermined energy-dependent intensity factor and expand the basis sets for SR analysis. The refined SR method is employed for reanalyzing the UV-TRPES data for the formation and relaxation dynamics of solvated electrons in various systems. The electron binding energy distributions for solvated electrons in liquid water, methanol, and ethanol are confirmed to be Gaussian centered at 3.78, 3.39, and 3.25 eV, respectively, in agreement with Nishitani et al. [ Sci. Adv. 2019, 5(8), eaaw6896]. An effective energy gap between the conduction band and the vacuum level at the gas-liquid interface is estimated to be 0.2 eV for liquid water and 0.1 eV for methanol and ethanol.

Tip-in EMR as an alternative to endoscopic submucosal dissection for 20- to 30-mm nonpedunculated colorectal neoplasms.

BACKGROUND AND AIMS: Tip-in EMR, which includes anchoring the snare tip, has recently shown a favorable en-bloc and R0 resection rate for colorectal neoplasms. Thus, Tip-in EMR may be an alternative to endoscopic submucosal dissection (ESD). We aimed to compare clinical outcomes between Tip-in EMR and ESD for large colorectal neoplasms. METHODS: This retrospective study evaluated consecutive patients who underwent Tip-in EMR or ESD for 20- to 30-mm nonpedunculated colorectal neoplasms at a Japanese tertiary cancer center between January 2014 and December 2019. Baseline characteristics, treatment results, and long-term outcomes were analyzed using 1:1 propensity score matching. RESULTS: Seven hundred nine lesions were evaluated. The Tip-in EMR group included 1 lesion with a nonlifting sign but no lesions with fold convergence. After propensity score matching, each group included 140 lesions. The ESD group showed significantly higher en-bloc resection rates (99.3% vs 85.0%) and R0 resection rates (90.7% vs 62.9%). Procedure time was significantly shorter in the Tip-in EMR group (8 minutes vs 60 minutes). The Tip-in EMR and ESD groups did not differ significantly with respect to local recurrence rate (2.1% vs 0%). CONCLUSIONS: Tip-in EMR is comparable with ESD with respect to the local recurrence rate but has a shorter procedure time, despite the lower en-bloc and R0 resection rates for 20- to 30-mm nonpedunculated colorectal neoplasms without fold convergence or nonlifting sign. Thus, Tip-in EMR could be a feasible alternative to ESD in these lesions.

Predicting the depth of superficial adenocarcinoma of the esophagogastric junction.

BACKGROUND AND AIM: Preoperative determination of the invasion depth of superficial adenocarcinoma of the esophagogastric junction is important for appropriate endoscopic or surgical resection. There are no objective criteria regarding this; therefore, we investigated the factors associated with the invasion depth of superficial adenocarcinoma of the esophagogastric junction. METHODS: This retrospective study evaluated patients with superficial adenocarcinoma of the esophagogastric junction who had undergone endoscopic or surgical resection at a Japanese tertiary cancer center between April 2004 and December 2017. We analyzed endoscopic features of intramucosal to slight submucosal (M-SM1; < 500 µm) and deep submucosal (SM2; ≥ 500 µm) adenocarcinoma of the esophagogastric junction and extracted significant factors associated with and assessed the diagnostic performance of endoscopic features for SM2 lesion. RESULTS: A total of 106 cases were included in this study. Multivariate analysis indicated that depressed or protruded type (odds ratio [OR], 11.1), lesion size ≥ 15 mm (OR, 3.11), uneven surface (OR, 6.31), and subsquamous extension (OR, 5.41) were significantly associated with SM2 adenocarcinomas of the esophagogastric junction. When the macroscopic type was depressed or protruded, high sensitivity (97%) but fair specificity (46%) were observed for SM2 adenocarcinoma of the esophagogastric junction, whereas uneven surface and subsquamous extension showed high specificity (96% and 87%) but fair sensitivity (36% and 46%). CONCLUSIONS: Depressed or protruded type, lesion size ≥ 15 mm, uneven surface, and subsquamous extension were significantly associated with the invasion depth of superficial adenocarcinoma of the esophagogastric junction. These endoscopic features are useful in determining the treatment method preoperatively.

The safety and feasibility of endoscopic submucosal dissection using a flexible three-dimensional endoscope for early gastric cancer and superficial esophageal cancer: A prospective observational study.

BACKGROUND AND AIM: Endoscopic submucosal dissection (ESD) is performed as one of standard treatments for patients with early gastric cancer (EGC) and superficial esophageal squamous cancer (SESCC). A prototype of a flexible endoscope with a 3-D system has been recently developed. This study aimed to investigate the safety and feasibility of ESD using a 3-D flexible endoscope (3-D ESD) for EGC and SESCC. METHODS: This single-center, prospective, observational study enrolled patients who underwent planned 3-D ESD. The clinical outcomes, including the incidence of adverse events and treatment results, were analyzed. Visibility and manipulation during 3-D ESD were evaluated using a visual analog scale (VAS). We also evaluated the effect of the 3-D system on the endoscopist using VAS and the critical flicker fusion frequency (CFFF). RESULTS: We analyzed 47 EGC and 20 SESCC cases. There are no bleeding cases that required transfusion and perforation during 3-D ESD in both EGC and SESCC patients. However, the incidence of delayed bleeding and delayed perforation was 1.5% (one case) each. The mean VAS scores for recognizing the submucosal layer during the submucosal dissection, visual perception of blood vessel, and depth perception were 72.7 ± 22.2, 74.7 ± 21.8, and 78.2 ± 19.9, respectively. In contrast, the mean VAS score for manipulation was 25.4 ± 19.7. Among endoscopists, there was no significant difference in the VAS of eyestrain and headache before and after ESD, and there was no significant difference in the CFFF. CONCLUSION: The safety and feasibility of 3-D ESD for EGC and SESCC are acceptable in both patients and endoscopists.

Detecting colon polyps in endoscopic images using artificial intelligence constructed with automated collection of annotated images from an endoscopy reporting system.

BACKGROUND: Artificial intelligence (AI) has made considerable progress in image recognition, especially in the analysis of endoscopic images. The availability of large-scale annotated datasets has contributed to the recent progress in this field. Datasets of high-quality annotated endoscopic images are widely available, particularly in Japan. A system for collecting annotated data reported daily could aid in accumulating a significant number of high-quality annotated datasets. AIM: We assessed the validity of using daily annotated endoscopic images in a constructed reporting system for a prototype AI model for polyp detection. METHODS: We constructed an automated collection system for daily annotated datasets from an endoscopy reporting system. The key images were selected and annotated for each case only during daily practice, not to be performed retrospectively. We automatically extracted annotated endoscopic images of diminutive colon polyps that had been diagnosed (study period March-September 2018) using the keywords of diagnostic information, and additionally collect the normal colon images. The collected dataset was devised into training and validation to build and evaluate the AI system. The detection model was developed using a deep learning algorithm, RetinaNet. RESULTS: The automated system collected endoscopic images (47,391) from colonoscopies (745), and extracted key colon polyp images (1356) with localized annotations. The sensitivity, specificity, and accuracy of our AI model were 97.0%, 97.7%, and 97.3% (n = 300), respectively. CONCLUSION: The automated system enabled the development of a high-performance colon polyp detector using images in endoscopy reporting system without the efforts of retrospective annotation works.

Endoscopic submucosal dissection versus surgery in elderly patients with early gastric cancer of relative indication for endoscopic resection.

OBJECTIVES: Surgery is recommended for early gastric cancer (EGC) beyond the endoscopic resection (ER)-indication for the risk of lymph node metastasis; however, ER may be chosen as a "relative ER-indication" considering age and comorbidities. This study aimed to compare outcomes of endoscopic submucosal dissection (ESD) only and surgery (primary surgery and additional surgery after non-curative ESD) among elderly patients with relative ER-indication EGC and to further assess prognostic factors. METHODS: Outcomes of ESD and surgery (417 cases; 114 ESD, 303 surgery) in elderly patients (≥75 years) with relative ER-indication EGC were retrospectively analyzed. Prognostic factors were also examined. RESULTS: During the observation period (median; ESD, 34 months; surgery, 61 months), 29% of ESD and 35% of surgery patients died, including 4% and 5% from gastric cancer (GC), respectively. ESD showed lower overall survival (OS) than surgery (P = 0.027) but comparable disease-free survival (P = 0.916). OS-associated factors were age and prognostic nutritional index (PNI) in males (age ≥79, hazard ratio [HR] 2.21, P = 0.001; PNI <45, HR 2.06, P = 0.031) and age in females (age ≥82, HR 4.06, P = 0.004). Treatment was not a prognostic factor in either subgroup. Pathological category ≥pT1b2 (submucosal invasion ≥500 µm) and lymphovascular invasion (LVI) were significantly associated with GC death (mortality: ≥pT1b2, 7.7%, P = 0.002; LVI, 10.1%, P < 0.001). CONCLUSIONS: In elderly patients with relative ER-indication EGC, ESD may have comparable long-term efficacy to surgery, and treatment selection had a minor contribution to OS. For patients with poor preoperative prognostic factors, diagnostic ESD may be performed first, followed by additional surgery based on pathological results.

Comparison of five-phase computed tomography images of type 1 autoimmune pancreatitis and pancreatic cancer: Emphasis on cases with atypical images.

BACKGROUND/OBJECTIVES: International consensus diagnostic criteria (ICDC) include characteristic images of autoimmune pancreatitis (AIP); however, reports on atypical cases are increasing. The aims of this study were to compare CT findings between AIP and pancreatic cancer (PC), and to analyze type 1 AIPs showing atypical images. METHODS: Five-phase CT images were compared between 80 type 1-AIP lesions and 80 size- and location-matched PCs in the case-control study. Atypical AIPs were diagnosed based on the four ICDC items. RESULTS: ICDC items were recognized in most AIP lesions; pancreatic enlargement (87.7%), narrowing of the main pancreatic duct (98.8%), delayed enhancement (100%), and no marked upstream-duct dilation (97.5%). CT values of AIPs increased rapidly until the pancreatic phase and decreased afterward, while those of PCs gradually increased until the delayed phase (P < 0.0001). Atypical images were recognized in 14.8% of AIPs, commonly without pancreatic enlargement (18.5 mm) and sometimes mimicking intraductal neoplasms. The CT values and their ratios were different between atypical AIPs and size-matched PCs most significantly in the pancreatic phase, but similar in the delayed phase. CONCLUSIONS: Ordinary type 1 AIPs can be diagnosed with the ICDC, but atypical AIPs represented a small fraction. "Delayed enhancement" is characteristic to ordinary AIPs, however, "pancreatic-phase enhancement" is more diagnostic for atypical AIPs.