Endoscopic ultrasound (EUS) has emerged as a widely utilized tool in the diagnosis of digestive diseases. In recent years, the potential of artificial intelligence (AI) in healthcare has been gradually recognized, and its superiority in the field of EUS is becoming apparent. Machine learning (ML) and deep learning (DL) are the two main AI algorithms. This paper aims to outline the applications and prospects of artificial intelligence-assisted endoscopic ultrasound (EUS-AI) in digestive diseases over the past decade. The results demonstrated that EUS-AI has shown superiority or at least equivalence to traditional methods in the diagnosis, prognosis, and quality control of subepithelial lesions, early esophageal cancer, early gastric cancer, and pancreatic diseases including pancreatic cystic lesions, autoimmune pancreatitis, and pancreatic cancer. The implementation of EUS-AI has opened up new avenues for individualized precision medicine and has introduced novel diagnostic and treatment approaches for digestive diseases.

Histomorphologic types of gastric cancer (GC) have significant prognostic values that should be considered during treatment planning. Because the thorough quantitative review of a tissue slide is a laborious task for pathologists, deep learning (DL) can be a useful tool to support pathologic workflow. In the present study, a fully automated approach was applied to distinguish differentiated/undifferentiated and non-mucinous/mucinous tumor types in GC tissue whole-slide images from The Cancer Genome Atlas (TCGA) stomach adenocarcinoma dataset (TCGA-STAD). By classifying small patches of tissue images into differentiated/undifferentiated and non-mucinous/mucinous tumor tissues, the relative proportion of GC tissue subtypes can be easily quantified. Furthermore, the distribution of different tissue subtypes can be clearly visualized. The patch-level areas under the curves for the receiver operating characteristic curves for the differentiated/undifferentiated and non-mucinous/mucinous classifiers were 0.932 and 0.979, respectively. We also validated the classifiers on our own GC datasets and confirmed that the generalizability of the classifiers is excellent. The results indicate that the DL-based tissue classifier could be a useful tool for the quantitative analysis of cancer tissue slides. By combining DL-based classifiers for various molecular and morphologic variations in tissue slides, the heterogeneity of tumor tissues can be unveiled more efficiently.

The use of endoscopic submucosal dissection (ESD) for treating undifferentiated-type early gastric cancer is controversial. The objective of this study was to perform a meta-analysis to compare the long-term outcomes of ESD and surgery for undifferentiated-type early gastric cancer.

The PubMed, Cochrane Library, and EMBASE databases were used to search for relevant studies comparing ESD and surgery for undifferentiated-type early gastric cancer. The methodological quality of the included publications was evaluated using the Risk of Bias Assessment tool for Nonrandomized Studies. The rates of overall survival, recurrence, adverse event, and complete resection were determined. Odds ratios (ORs) and 95% confidence intervals (CIs) were also evaluated.

This meta-analysis enrolled five studies with 429 and 1,236 participants undergoing ESD and surgery, respectively. No significant difference was found in the overall survival rate between the ESD and surgery groups (OR, 2.29; 95% CI, 0.98-5.36; p=0.06). However, ESD was associated with a higher recurrence rate and a lower complete resection rate. The adverse event rate was similar between the two groups.

ESD with meticulous surveillance esophagogastroduodenoscopy may be as effective and safe as surgery in patients with undifferentiated-type early gastric cancer. Further large-scale, randomized, controlled studies from additional regions are required to confirm these findings.

Size, ulcer, differentiation, and location are known to be factors affecting the T stage accuracy of EUS in gastric cancer. However, whether an interaction exists among recognized variables is poorly understood. The aim of this study was to identify the combinatorial characteristics of group with high overestimation rate to determine which group should be considered carefully for EUS-based treatment plans.

We retrospectively analyzed early gastric cancer patients who underwent EUS from 2005 to 2016. The accuracy of EUS T stage and factors affecting over-/underestimation were examined by using decision tree analysis, the CHAID method.

The most significant factor affecting the accuracy of the EUS T stage was the size. The rate of overestimation was higher in lesions > 3 cm (37.2% vs. 28.8% vs. 17.1%, p < 0.001). In lesions > 3 cm, the rate of overestimation was higher in lesions with an ulcer (62.1% vs. 35.0%, p < 0.001). Moreover, for lesions ≤ 3 cm, the accuracy of the EUS T stage was more affected by differentiation and location. The rate of overestimation was higher in undifferentiated-type lesions ≤ 2 cm (24.5% vs. 13.9%, p < 0.001) and 2-3 cm (33.3% vs. 25.7%, p = 0.011). In the differentiated type, the location affected the accuracy of the EUS T stage.

In this hierarchical analysis, the rate of overestimation was higher in lesions > 3 cm with ulcer, lesions > 3 cm irrespective of ulcer, and undifferentiated-type lesions measuring 2-3 cm.

Endoscopic submucosal dissection (ESD) is an effective treatment for early gastric cancer (EGC) that has demonstrated a minimal risk of lymph node metastasis in retrospective studies. We sought to prospectively evaluate the short-term outcomes of ESD treatment in EGCs.

A prospective multicenter cohort study of neoplasms 3 cm or less in diameter at endoscopic size evaluation was performed in 12 Korean ESD study grouprelated university hospitals and the National Cancer Center. Resected specimens were evaluated by the central pathologic review board.

A patient cohort (n=712) with a total of 737 EGCs was analyzed. The margin-free en bloc resection rate was 97.3%, and curative resection of 640 lesions (86.8%) was achieved. Lower curative resection rates were associated with lesions 2 to 3 cm in size prior to ESD compared with lesions 2 cm or less in size (78.6% vs 88.1%, respectively, p=0.009). Significant factors associated with noncurative resection were moderately or poorly differentiated histological type, posterior wall tumor location, tumor size larger than 3 cm, ulceration, and submucosal invasion. Delayed bleeding occurred in 49 patients (6.9%), and 12 patients (1.7%) exhibited perforations.

ESD is an effective treatment with a high curative resection rate for EGCs that meets relatively conservative pre-ESD indications. Long-term survival outcomes should be evaluated in followup studies.

Early gastric cancer cases that are estimated to meet indications for treatment before endoscopic submucosal resection are often revealed to be out-of-indication after the treatment. We investigated the short-term treatment outcomes in patients with early gastric cancer according to the pretreatment clinical endoscopic submucosal resection indications.

We retrospectively reviewed the medical records of patients with early gastric cancer that met the pretreatment endoscopic submucosal resection indications, from 2004 to 2011. Curative resection rate and proportion of out-of-indication cases were compared according to the pre-endoscopic submucosal resection indications. Pre-endoscopic submucosal resection factors associated with out-of-indication in the final pathological examination were analyzed.

Of 756 cases, 660 had absolute and 96 had expanded pre-endoscopic submucosal resection indications. The curative resection rate was significantly lower in the patients with expanded indications (64.6%) than in those with absolute indications (81.7%; P<0.001). The cases with expanded indications (30.2%) were revealed to be out-of-indication more frequently than the cases with absolute indications (13.8%; P<0.001). Age of >65 years, tumor size of >2 cm, tumor location in the upper-third segment of the stomach, and undifferentiated histological type in pre-endoscopic submucosal resection evaluations were significant risk factors for out-of-indication after endoscopic submucosal resection.

Non-curative resection due to out-of-indication occurred in approximately one-third of the early gastric cancer cases that clinically met the expanded indications before endoscopic submucosal resection. The possibility of additional surgery should be emphasized for patients with early gastric cancers that clinically meet the expanded indications.

There are often discrepancies between the pretreatment evaluation of gastric neoplasms by endoscopy with biopsy and the final diagnosis of resected specimen in terms of pathology and depth of invasion. We evaluated the spectrum of discrepancies between pretreatment and posttreatment diagnosis which may deliver significant differences on clinical practice.

A total of 2041 patients with gastric dysplasia or cancer who underwent curative endoscopic resections or surgeries in 2012 were enrolled. Patients were classified into five different diagnostic groups: low-grade dysplasia (LGD), high-grade dysplasia (HGD), absolute indication early gastric cancer (AI-EGC), beyond absolute indication early gastric cancer (BAI-EGC), and advanced gastric cancer (AGC). The choice of initial treatment and final pathologic diagnosis was analyzed.

The study patients belonged to the following pretreatment diagnostic groups: LGDs in 162, HGDs in 164, AI-EGCs in 396, BAI-EGCs in 824, and AGCs in 495 cases. Posttreatment diagnostic groups were LGDs in 140, HGDs in 121, AI-EGCs in 322, BAI-EGCs in 947, AGCs in 505, and no residual tumor in 6 cases. In general, 6.9 % (141/2041) of cases were downgraded and 15.9 % (324/2041) were upgraded. Thirty-four percent of pretreatment HGDs (56/164) were changed to cancers after endoscopic resection. Thirty-three percent of pretreatment AI-EGCs (131/396) were regrouped as posttreatment BAI-EGCs. The additional surgery rate in each pretreatment group was 0.6 % in LGD, 4.3 % in HGD, 15.7 % in AI-EGC, 23.6 % in BAI-EGC among the patients with initial endoscopic resection (p < 0.01).

Twenty-three percent of gastric neoplasms changed in their final diagnostic group after endoscopic resection or surgery. This discrepancy should be considered when the initial treatment strategy is being selected.

While limited endoscopic submucosal dissection (ESD) is increasingly applied in the treatment of early gastric cancer, preoperative prediction of lymph node metastasis is very critical for determining treatment strategies preoperatively. Thus, the aim of this study was to accurately assess the prevalence and pattern of lymph node metastasis in early gastric cancer patients and to identify the best candidates for ESD.

From September 2008 to December 2013, a total of 539 patients with early gastric cancer were retrospectively analyzed in the present study. Of them, 503 patients underwent radical gastrectomy and 36 patients underwent ESD. The clinicopathological features were collected and correlations with lymph node metastasis were analyzed. The survival rates of patients were also analyzed.

Lymph node metastasis was observed in 80 of 503 patients (15.9 %). Among these, the rate for mucosal cancer was 8.3 %, and 20.1 % for submucosal cancer. By univariate analysis, risk factors for lymph node metastasis were growth pattern, tumor size, pathological type, depth of invasion, lymphatic-vascular invasion, and neural invasion. By multivariate analysis, risk factors for lymph node metastasis were tumor size, pathological type, depth of invasion, and lymphatic-vascular invasion. The incidence of lymph node metastasis was 0 % in the well-differentiated mucosal cancers, irrespective of tumor size. For the well-differentiated mucosal cancers, the overall survival rates were comparable between patients underwent gastrectomy with lymph node dissection and patients underwent ESD (100 vs 100 %).

The most important factors for predicting lymph node metastasis in early gastric cancer are tumor size, pathological type, depth of invasion, and lymphatic-vascular invasion. Well-differentiated mucosal gastric cancers could be candidates for ESD.

To evaluate the efficacy and safety of endoscopic submucosal dissection (ESD) for early gastric cancer (EGC) with undifferentiated-type histology.

A systematic literature review was conducted using the core databases. Complete resection, curative resection, en bloc resection, recurrence and adverse event rate were extracted and analyzed. A random effect model was applied. The methodological quality of the enrolled studies was assessed using the Newcastle-Ottawa Scale. Publication bias was evaluated using a funnel plot, the trim and fill method, Egger's test, and a rank correlation test.

Fourteen retrospective studies between 2009 and 2014 were identified (972 EGC lesions with undifferentiated-type histology). The total en bloc and complete resection rates were estimated as 92.1% (95%CI: 87.4%-95.2%) and 77.5% (95%CI: 69.3%-84%), respectively. The total curative resection rate was 61.4% (95%CI: 44.5%-75.9%). The overall recurrence rate was 7.6% (95%CI: 3.4%-16%). Limited to histologically diagnosed expanded-criteria lesions, the en bloc and complete resection rates were 91.2% and 85.6%, respectively. The curative resection rate was 79.8%.

In this analysis, ESD is a technically feasible treatment modality for EGC with undifferentiated-type histology. Long-term studies are needed to confirm these therapeutic outcomes.

A small differentiated gastric cancer with minute submucosal invasion after endoscopic resection is regarded as the curative resection criterion of an expanded indication. However, recent studies have shown a high incidence of lymph node metastasis in those meeting the expanded criteria. The aim of this study was to evaluate the validity of the 500 μm criterion and to determine an optimal cutoff value for the expanded indication.

We reviewed the clinicopathological data of 1,322 patients who underwent curative resection for submucosal gastric cancer between 2001 and 2013. The full thickness of the submucosa and the incidence of lymph node metastasis according to the depth of the submucosal invasion were evaluated. The sensitivity, specificity, and positive and negative predictive values were calculated with receiver operating characteristic (ROC) curve analysis.

The mean full thickness of the submucosa was 2,605 ± 1,760 μm, and the incidence of lymph node metastasis in 103 tumors meeting the expanded indication was 3.9 % (4/103). In the ROC analysis, the area under the curve was 0.664 (95 % confidence interval 0.538-0.791; p = 0.017), and the highest negative predictive value of 98 % was observed when the cutoff value was 300 μm.

This study suggests that a range reduction in the depth of submucosal invasion is required to obtain a high negative predictive value. Further large-scale studies are required to validate the optimal cutoff value proposed in this study.