A machine learning model for predicting the lymph node metastasis of early gastric cancer not meeting the endoscopic curability criteria.

BACKGROUND: We developed a machine learning (ML) model to predict the risk of lymph node metastasis (LNM) in patients with early gastric cancer (EGC) who did not meet the existing Japanese endoscopic curability criteria and compared its performance with that of the most common clinical risk scoring system, the eCura system. METHODS: We used data from 4,042 consecutive patients with EGC from 21 institutions who underwent endoscopic submucosal dissection (ESD) and/or surgery between 2010 and 2021. All resected EGCs were histologically confirmed not to satisfy the current Japanese endoscopic curability criteria. Of all patients, 3,506 constituted the training cohort to develop the neural network-based ML model, and 536 constituted the validation cohort. The performance of our ML model, as measured by the area under the receiver operating characteristic curve (AUC), was compared with that of the eCura system in the validation cohort. RESULTS: LNM rates were 14% (503/3,506) and 7% (39/536) in the training and validation cohorts, respectively. The ML model identified patients with LNM with an AUC of 0.83 (95% confidence interval, 0.76-0.89) in the validation cohort, while the eCura system identified patients with LNM with an AUC of 0.77 (95% confidence interval, 0.70-0.85) (P = 0.006, DeLong's test). CONCLUSIONS: Our ML model performed better than the eCura system for predicting LNM risk in patients with EGC who did not meet the existing Japanese endoscopic curability criteria. We developed a neural network-based machine learning model that predicts the risk of lymph node metastasis in patients with early gastric cancer who did not meet the endoscopic curability criteria.

Zinc transporter 1 expression in hepatocellular carcinoma correlates with prognosis: A single-center retrospective study.

OBJECTIVES: Zinc is crucial in the pathogenesis of hepatocellular carcinoma; however, no reports have examined its association with clinical parameters and zinc transporter 1 (ZNT1) expression intensity. This study aimed to assess the association between ZNT1 expression and prognosis in patients with hepatocellular carcinoma. METHODS: This retrospective study included 65 patients who underwent surgical hepatocellular carcinoma resection at a single center between January 2011 and June 2015. ZNT1 expression on hepatocellular carcinoma cells from specimens was assessed using immunohistochemistry, and the relationship between its intensity and various clinical indexes was examined with univariate and multivariable analyses and the Mann-Whitney U, Kruskal-Wallis, Bonferroni, and log-rank tests. RESULTS: ZNT1 expression on the hepatocellular carcinoma cell membrane was negative in 31 patients and positive in 34 patients, including nine patients showing strongly positive expression. Patients with and without ZNT1 expression had similar blood zinc concentrations, α-fetoprotein levels, protein induced by vitamin K absence-antagonist-II levels, gross classification, maximal tumor diameters, and background liver disease. The blood zinc concentrations were significantly lower in patients with strongly positive ZNT1 expression (57.0 ± 22.1 µg/dL) than in those with positive ZNT1 expression (71.1 ± 14.2 µg/dL; P = 0.015) or those with no ZNT1 expression (72.9 ± 14.1 µg/dL; P = 0.043). Overall survival was significantly shorter in ZNT1-expressing patients than in non-expressing patients (log-rank test, P = 0.024). Multivariable analysis using the Cox proportional hazards model identified maximal tumor diameter (hazard ratio, 1.018; 95% confidence interval, 1.002-1.034; P = 0.026) and ZNT1 expression status (hazard ratio, 2.082; 95% confidence interval, 1.196-3.621; P = 0.010) as prognostic contributing factors.

Utility of an artificial intelligence system for classification of esophageal lesions when simulating its clinical use.

Previous reports have shown favorable performance of artificial intelligence (AI) systems for diagnosing esophageal squamous cell carcinoma (ESCC) compared with endoscopists. However, these findings don't reflect performance in clinical situations, as endoscopists classify lesions based on both magnified and non-magnified videos, while AI systems often use only a few magnified narrow band imaging (NBI) still images. We evaluated the performance of the AI system in simulated clinical situations. We used 25,048 images from 1433 superficial ESCC and 4746 images from 410 noncancerous esophagi to construct our AI system. For the validation dataset, we took NBI videos of suspected superficial ESCCs. The AI system diagnosis used one magnified still image taken from each video, while 19 endoscopists used whole videos. We used 147 videos and still images including 83 superficial ESCC and 64 non-ESCC lesions. The accuracy, sensitivity and specificity for the classification of ESCC were, respectively, 80.9% [95% CI 73.6-87.0], 85.5% [76.1-92.3], and 75.0% [62.6-85.0] for the AI system and 69.2% [66.4-72.1], 67.5% [61.4-73.6], and 71.5% [61.9-81.0] for the endoscopists. The AI system correctly classified all ESCCs invading the muscularis mucosa or submucosa and 96.8% of lesions ≥ 20 mm, whereas even the experts diagnosed some of them as non-ESCCs. Our AI system showed higher accuracy for classifying ESCC and non-ESCC than endoscopists. It may provide valuable diagnostic support to endoscopists.

[A Case of Early Gastric Cancer in the Remnant Stomach after Successful Treatment with ESD].

We report a case of early gastric cancer in the remnant stomach after successful treatment with endoscopic submucosal dissection(ESD). A 64-year-old woman had undergone distal gastrectomy, D2 dissection, and Billroth Ⅰ reconstruction for advanced gastric cancer 11 years previously. During a routine upper gastrointestinal endoscopy, an elevated lesion was detected at the lesser curvature of the upper gastric body of the remnant stomach, and biopsy indicated a Group 4 tumor. Curative en bloc resection of the lesion was achieved via ESD, although there was severe fibrosis along the suture line. The pathological result was 0-I, pT1a, tub1, 3×3 mm, UL(－), ly(－), v(－), HM0(8 mm), VM0(800 µm), indicating curative resection. Surveillance of the upper gastrointestinal tract 5 years after gastric cancer surgery enabled the early detection of the gastric cancer and curative resection with ESD.

Real-time assessment of video images for esophageal squamous cell carcinoma invasion depth using artificial intelligence.

BACKGROUND: Although optimal treatment of superficial esophageal squamous cell carcinoma (SCC) requires accurate evaluation of cancer invasion depth, the current process is rather subjective and may vary by observer. We, therefore, aimed to develop an AI system to calculate cancer invasion depth. METHODS: We gathered and selected 23,977 images (6857 WLI and 17,120 NBI/BLI images) of pathologically proven superficial esophageal SCC from endoscopic videos and still images of superficial esophageal SCC taken in our facility, to use as a learning dataset. We annotated the images with information [such as magnified endoscopy (ME) or non-ME, pEP-LPM, pMM, pSM1, and pSM2-3 cancers] based on pathologic diagnosis of the resected specimens. We created a model using a convolutional neural network. Performance of the AI system was compared with that of invited experts who used the same validation video set, independent of the learning dataset. RESULTS: Accuracy, sensitivity, and specificity with non-magnified endoscopy (ME) were 87%, 50%, and 99% for the AI system and 85%, 45%, 97% for the experts. Accuracy, sensitivity, and specificity with ME were 89%, 71%, and 95% for the AI system and 84%, 42%, 97% for the experts. CONCLUSIONS: Most diagnostic parameters were higher when done by the AI system than by the experts. These results suggest that our AI system could potentially provide useful support during endoscopies.

[A Case of Superficial-Type Gastric Cancer with Metastatic Ovarian Cancer Diagnosed by Exploratory Laparotomy].

Here, we report a case of superficial-type gastric cancer with metastatic ovarian cancer(Krukenberg tumor)diagnosed by exploratory laparotomy. Chemotherapy was initiated at an early stage in this patient. A 43-year-old woman with superficialtype gastric cancer(0-Ⅱb plusⅡa), an ovarian tumor, and a solitary sclerotic bone lesion underwent exploratory laparotomy and bilateral salpingo-oophorectomy. Pathological findings showed that the resected ovarian tumor specimen contained the same type of signet ring cell carcinoma as the biopsy gastric cancer specimen; hence, the patient was diagnosed with superficial- type gastric cancer with metastatic ovarian cancer. She was treated with first-line chemotherapy(capecitabine plus oxaliplatin)15 days after exploratory laparotomy, followed by second-line chemotherapy(ramucirumab plus paclitaxel), thirdline chemotherapy(nivolumab), and fourth-line chemotherapy(irinotecan). Twenty-two months after the start of first-line chemotherapy, she finally died due to bone metastasis.

Endoscopic detection and differentiation of esophageal lesions using a deep neural network.

BACKGROUND AND AIMS: Diagnosing esophageal squamous cell carcinoma (SCC) depends on individual physician expertise and may be subject to interobserver variability. Therefore, we developed a computerized image-analysis system to detect and differentiate esophageal SCC. METHODS: A total of 9591 nonmagnified endoscopy (non-ME) and 7844 ME images of pathologically confirmed superficial esophageal SCCs and 1692 non-ME and 3435 ME images from noncancerous lesions or normal esophagus were used as training image data. Validation was performed using 255 non-ME white-light images, 268 non-ME narrow-band images/blue-laser images, and 204 ME narrow-band images/blue-laser images from 135 patients. The same validation test data were diagnosed by 15 board-certified specialists (experienced endoscopists). RESULTS: Regarding diagnosis by non-ME with narrow-band imaging/blue-laser imaging, the sensitivity, specificity, and accuracy were 100%, 63%, and 77%, respectively, for the artificial intelligence (AI) system and 92%, 69%, and 78%, respectively, for the experienced endoscopists. Regarding diagnosis by non-ME with white-light imaging, the sensitivity, specificity, and accuracy were 90%, 76%, and 81%, respectively, for the AI system and 87%, 67%, and 75%, respectively, for the experienced endoscopists. Regarding diagnosis by ME, the sensitivity, specificity, and accuracy were 98%, 56%, and 77%, respectively, for the AI system and 83%, 70%, and 76%, respectively, for the experienced endoscopists. There was no significant difference in the diagnostic performance between the AI system and the experienced endoscopists. CONCLUSIONS: Our AI system showed high sensitivity for detecting SCC by non-ME and high accuracy for differentiating SCC from noncancerous lesions by ME.

Endoscopic Balloon Dilation Followed By Intralesional Steroid Injection for Anastomotic Strictures After Esophagectomy: A Randomized Controlled Trial.

OBJECTIVE: Endoscopic balloon dilation (EBD) is a standard treatment for anastomotic strictures after esophagectomy, and requires multiple dilations. We conducted a randomized controlled trial to assess the efficacy of adding a steroid injection to EBD to reduce restricture. METHODS: Patients were randomized to receive EBD combined with either triamcinolone or placebo injection. The primary endpoint was the number of dilations required to resolve the stricture. The secondary endpoints were restricture-free survival and adverse events. Patients with a dysphagia symptom score of ≥2 after esophagectomy with an endoscopy-confirmed anastomotic stricture were included. A total of 50 mg of triamcinolone acetonide (50 mg/5 mL) or an identical volume of normal saline solution as a placebo was injected per site using a 25-gauge needle immediately after EBD. Both the patient and treating physician were blinded to the treatment given. RESULTS: During the 4-year study period, 65 patients were randomized to either the steroid group (n = 33) or placebo group (n = 32). The median number of EBDs required to resolve strictures was 2.0 (interquartile range, 1.0-2.5) in the steroid group and 4.0 (interquartile range, 2.0-6.8) in the placebo group (p < 0.001). After 6 months of follow-up, 39% of patients who had received steroid injections remained recurrence free compared with 16% of those who had received saline injections (p = 0.002). No adverse events occurred during follow-up. CONCLUSIONS: Steroid injection shows promising results for the prevention of stricture recurrence in patients who underwent EBD for anastomotic strictures.

Investigation of mucosal pattern of gastric antrum using magnifying narrow-band imaging in patients with chronic atrophic fundic gastritis.

BACKGROUND: Magnifying narrow-band imaging (M-NBI) can reportedly help predict the presence and distribution of atrophy and intestinal metaplasia in the gastric corpus. However, the micro-mucosal pattern of the antrum shown by M-NBI differs from that of the corpus. We studied the distribution and histology of the micro-mucosal pattern in the antrum based on magnifying endoscopy. METHODS: Endoscopic images of the greater curvature of the antrum were evaluated in 50 patients with chronic atrophic fundic gastritis (CAFG). The extent of CAFG was evaluated by autofluorescence imaging. The micro-mucosal pattern was evaluated by M-NBI and classified into groove and white villiform types. The localization of white villiform type mucosa was classified into three types in relation to the areae gastricae: null, central, and segmental types. Biopsies were taken from regions showing different micro-mucosal patterns. Associations among the extent of CAFG, micro-mucosal pattern, and histology were examined. RESULTS: As the extent of CAFG increased, the proportion of white villiform type mucosa increased, whereas that of groove type mucosa decreased (P=0.022). In patients with extensive CAFG, most of the areae gastricae was composed of the segmental or central type of white villiform type mucosa (P=0.044). The white villiform type mucosa had significantly higher grades of atrophy (P=0.002) and intestinal metaplasia (P<0.001) than did the groove type mucosa. CONCLUSION: White villiform type mucosa is indicative of atrophy and intestinal metaplasia in the gastric antrum. It extends to the whole or central part of the areae gastricae as CAFG becomes more extensive.

Incomplete resection rate of cold snare polypectomy: a prospective single-arm observational study.

Background and study aims Cold snare polypectomy (CSP) is considered to be safe for the removal of subcentimeter colorectal polyps. This study aimed to determine the rate of incomplete CSP resection for subcentimeter neoplastic polyps at our center. Patients and methods Patients with small or diminutive adenomas (diameter 1 - 9 mm) were recruited to undergo CSP until no polyp was visible. After CSP, a 1 - 3 mm margin around the resection site was removed using endoscopic mucosal resection. The polyps and resection site marginal specimens were microscopically evaluated. Incomplete resection was defined as the presence of neoplastic tissue in the marginal specimen. We also calculated the frequency at which the polyp lateral margins could be assessed for completeness of resection. Results A total of 307 subcentimeter neoplastic polyps were removed from 120 patients. The incomplete resection rate was 3.9 % (95 % confidence interval [CI] 1.7 % - 6.1 %); incomplete resection was not associated with polyp size, location, morphology, or operator experience. The polyp lateral margins could not be assessed adequately for 206 polyps (67.1 %). Interobserver agreement between incomplete resection and lateral polyp margins that were inadequate for assessment was poor (κ = 0.029, 95 %CI 0 - 0.04). Female sex was an independent risk factor for incomplete resection (odds ratio 4.41, 95 %CI 1.26 - 15.48; P  = 0.02). Conclusions At our center, CSP resection was associated with a moderate rate of incomplete resection, which was not associated with polyp characteristics. However, adequate evaluation of resection may not be routinely possible using the lateral margin from subcentimeter polyps that were removed using CSP.Trial registered at University Hospital Medical Information Network (UMIN 000010879).

Traction-assisted colonic endoscopic submucosal dissection using clip and line: a feasibility study.

BACKGROUND AND STUDY AIMS: Colonic endoscopic submucosal dissection (ESD) is a challenging procedure because it is often difficult to maintain good visualization of the submucosal layer. To facilitate colonic ESD, we designed a novel traction method, namely traction-assisted colonic ESD using clip and line (TAC), and investigated its feasibility. PATIENTS AND METHODS: We retrospectively analyzed 23 patients with large colonic superficial lesions who had undergone TAC. The main outcome was the procedural success rate of TAC, which we defined as successful, sustained application of clip and line to the lesion until the end of the procedure. RESULTS: The procedural success rate of TAC was 87 % (20/23). In all three unsuccessful cases, the lesions were in the proximal colon and the procedure times over 100 minutes. The overall mean procedure time was 61 min (95 % confidence interval, 18 - 172 min). We achieved en bloc resections of all lesions. There were no perforations or fatal adverse events. CONCLUSIONS: TAC is feasible and safe for colonic ESD and may improve the ease of performing this procedure.

Features of electrocoagulation syndrome after endoscopic submucosal dissection for colorectal neoplasm.

BACKGROUND AND AIMS: Endoscopic submucosal dissection (ESD) is a promising treatment for large gastrointestinal superficial neoplasms, although it is technically difficult, and perforation and delayed bleeding are well-known adverse events. However, there have been no large studies about electrocoagulation syndrome after colorectal ESD. The aim of this study was to evaluate the incidence and clinical significant risk factors of post-ESD coagulation syndrome (PECS). METHODS: This was a retrospective cohort study conducted in a referral cancer center. A total of 336 patients with colorectal neoplasms (143 adenomas or serrated lesions and 193 carcinomas) underwent ESD from January 2011 to June 2013. Incidence, outcome, and factors associated with occurrence of PECS were investigated. RESULTS: Occurred in 32 patients (9.5%). The median time until PECS was 15.5 h, and the median period of PECS was 32.5 h. Fever (≥37.6 °C) after ESD was found in 41% of the PECS group and 9% of the non-PECS group (P < 0.001). All PECS cases were managed conservatively. On multivariate analysis, female patients (odds ratio [OR] = 3.2, P = 0.002), lesion location at ascending colon and cecum (OR = 3.5, P = 0.001), and resected specimen ≥40 mm (OR = 2.1, P = 0.05) were independent risk factors for PECS. CONCLUSIONS: Occurred in 32 patients (9.5%) with colorectal ESD; however, all cases had a good outcome with conservative management. Female sex, tumor location at the ascending colon and cecum, and resected specimen ≥40 mm were independently significant risk factors for PECS.

Endoscopic submucosal dissection as minimally invasive treatment for superficial pharyngeal cancer: a phase II study (with video).

BACKGROUND AND AIMS: Although endoscopic submucosal dissection (ESD) has been applied for superficial pharyngeal cancer, no prospective trials have been reported. To investigate the efficacy and safety of ESD for superficial pharyngeal cancer, we conducted a prospective phase II trial. METHODS: Fifty-four patients with 73 lesions were enrolled from September 2010 to August 2014, and ESD was performed. The primary endpoint was the complete resection rate. Secondary endpoints were safety, recurrence-free survival, overall survival, and incidence of metachronous pharyngeal cancer. RESULTS: Fifty-four patients had stage 0-III cancer: stage 0, n = 22; stage I, n = 14; stage II, n = 17; and stage III, n = 1. The en bloc resection rate was 100%, and the complete resection rate was 79.5% (58/73 lesions; 95% confidence interval, 68%-88%). No serious adverse events related to ESD were encountered. Four patients required nasogastric intubation and feeding. No patients required percutaneous endoscopic gastrostomy and tracheotomy. Swallowing, speech, and airway functions were preserved in all patients. One of the 54 patients died of an unrelated illness. Median follow-up was 27 months (range 6-55 months). Local cervical lymph node metastasis was observed in 1 patient, and the case was salvaged successfully with lymph node dissection. The 3-year overall and recurrence-free survival rates were 97.7% and 98.1%, respectively. Cumulative development of multiple cancers of the pharynx at 3 years was 18.4%. CONCLUSIONS: ESD appears to be a safe and effective minimally invasive treatment in patients with superficial pharyngeal cancer. ( CLINICAL TRIAL REGISTRATION NUMBER: UMIN000003623.).

Long-term survival after endoscopic resection for early gastric cancer in the remnant stomach: comparison with radical surgery.

BACKGROUND: Endoscopic resection (ER) has recently become standard treatment, even for early gastric cancer (EGC) in the remnant stomach. We aimed to compare long-term survival after ER versus radical surgery for EGC in the remnant stomach. METHODS: We retrospectively compared overall and cause-specific survival of patients who had undergone ER or radical surgery for EGC in the remnant stomach from 1998 to 2012. RESULTS: During the study period, 32 patients with intramucosal (M), two with shallow submucosal (SM1) and eight with deep submucosal (SM2) cancers had undergone ER (ER group) whereas six with M and seven with SM2 cancers had undergone surgery (surgery group). All patients were followed up for a median of 60 months; during follow up, 15 patients died, including three in the ER group with SM2 cancer who died of gastric cancer. The overall 5-year survival rates of M-SM1 and SM2 cancer patients in the ER and surgery groups were 89%, 48%, 80%, and 67%, respectively (P=0.079). The disease-specific 5-year survival rates of M-SM1 and SM2 cancer patients in the ER and surgery groups were 100%, 48%, 100%, and 100%, respectively (P=0.000). Operation time and hospital stay were significantly shorter in the ER than the surgery group (P<0.001). Grade 2 perforation occurred in two patients in the ER group and Grade 3 anastomotic leakage in two patients in the surgery group. CONCLUSION: ER provides excellent outcomes, comparable with those of radical surgery, in patients with M-SM1 gastric cancer in the remnant stomach; however, patients with SM2 cancer require radical surgery.

Diagnostic features of sessile serrated adenoma/polyps on magnifying narrow band imaging: a prospective study of diagnostic accuracy.

BACKGROUND AND AIMS: The narrow band imaging classification system (NBI International Colorectal Endoscopic [NICE] classification) classifies colorectal polyps very accurately. However, sessile serrated adenoma/polyps (SSA/Ps) pathologically resembles hyperplastic polyp and has a possibility to be left in situ on NICE classification. The aim of this study was to establish and evaluate new simple diagnostic features for SSA/Ps using magnifying narrow band imaging (M-NBI). METHODS: We performed a single-arm observational study of diagnostic accuracy in two stages, as follows: seeking stage, development of simple diagnostic features for SSA/P and definition of diagnostic criteria based on retrospective assessments of M-NBI; and validation stage, prospective validation and evaluation of the new diagnostic criteria. RESULTS: In the seeking stage, we identified brownish, oval, expanded crypt openings and thick-branched vessels on the surfaces of SSA/Ps. We named these "expanded crypt openings" (ECOs) and "thick and branched vessels" (TBVs), respectively. In the validation stage, we enrolled 796 polyps in 261 patients, and classified 126 polyps as NICE type 1; all these lesions were endoscopically removed and assessed histopathologically. The sensitivity, specificity, and accuracy of ECOs for SSA/Ps were 84.3%, 81.1%, and 82.4%, whereas those of TBVs were 45.1%, 68.9%, and 59.2%, respectively. M-NBI provided a sensitivity of 98% and specificity of 59.5% for discrimination of SSA/Ps from other lesions classified as NICE type 1. CONCLUSIONS: Identification of ECOs, supplemented with TBVs, has high sensitivity for the diagnosis of SSA/P. These findings may facilitate the use of endoscopic optical diagnosis in clinical practice.

Endoscopic optical diagnosis provides high diagnostic accuracy of esophageal squamous cell carcinoma.

BACKGROUND: Recent technological advances have stimulated the development of endoscopic optical biopsy technologies. This study compared the accuracy of endoscopic diagnosis using magnifying narrow-band imaging (NBI) and histologic diagnosis of esophageal squamous lesions. METHODS: Patients at high risk for esophageal squamous cell carcinoma were examined with endoscopy and subsequent biopsy. The lesions diagnosed as cancer on NBI and the lesions diagnosed as cancer on biopsy were resected endoscopically or surgically. Histological diagnoses of resected specimens, the reference standards in this study, were made by a pathologist who was blind to both the endoscopic and biopsy diagnoses. The primary outcome was the accuracy of endoscopic and biopsy diagnosis. A noninferiority trial design with a noninferiority margin of -10% was chosen to investigate the accuracy of endoscopic diagnosis using magnifying NBI. RESULTS: Between November 2010 and October 2012, a total of 111 lesions in 85 patients were included in the analysis. The accuracy of endoscopic diagnosis and biopsy diagnosis for all lesions was 91.0% (101/111) and 85.6% (95/111), respectively. The difference in diagnostic accuracy was 5.4% (95% confidence interval: -2.9%-13.7%). The accuracy of endoscopic diagnosis and biopsy diagnosis of invasive cancers was 94.9% (74/78) and 84.6% (66/78), respectively. The difference was 10.3% (95% confidence interval: 1.6%-19.0%) for invasive cancers. The lower bound of the 95% confidence interval was above the prestated -10% in both cases. CONCLUSION: Noninferiority of endoscopic diagnosis by magnifying NBI to histologic diagnosis by biopsy was established in this study (p = 0.0001). TRIAL REGISTRATION: The study was registered on 9th November 2010 in the University Hospital Medical Network Clinical Trials Registry as number: UMIN000004529.