Comparison of endoscopic ultrasonography and conventional endoscopy for prediction of depth of tumor invasion in early gastric cancer

A novel artificial intelligence-based endoscopic ultrasonography diagnostic system for diagnosing the invasion depth of early gastric cancer

BACKGROUND

We developed an artificial intelligence (AI)-based endoscopic ultrasonography (EUS) system for diagnosing the invasion depth of early gastric cancer (EGC), and we evaluated the performance of this system.

METHODS

A total of 8280 EUS images from 559 EGC cases were collected from 11 institutions. Within this dataset, 3451 images (285 cases) from one institution were used as a development dataset. The AI model consisted of segmentation and classification steps, followed by the CycleGAN method to bridge differences in EUS images captured by different equipment. AI model performance was evaluated using an internal validation dataset collected from the same institution as the development dataset (1726 images, 135 cases). External validation was conducted using images collected from the other 10 institutions (3103 images, 139 cases).

RESULTS

The area under the curve (AUC) of the AI model in the internal validation dataset was 0.870 (95% CI: 0.796-0.944). Regarding diagnostic performance, the accuracy/sensitivity/specificity values of the AI model, experts (n = 6), and nonexperts (n = 8) were 82.2/63.4/90.4%, 81.9/66.3/88.7%, and 68.3/60.9/71.5%, respectively. The AUC of the AI model in the external validation dataset was 0.815 (95% CI: 0.743-0.886). The accuracy/sensitivity/specificity values of the AI model (74.1/73.1/75.0%) and the real-time diagnoses of experts (75.5/79.1/72.2%) in the external validation dataset were comparable.

CONCLUSIONS

Our AI model demonstrated a diagnostic performance equivalent to that of experts.

Value of quantitative microsurface structure analysis for evaluating the invasion depth of type 0-II early gastric cancer

Background and Aim

The microsurface structure reflects the degree of damage to the glands, which is related to the invasion depth of early gastric cancer. To evaluate the diagnostic value of quantitative microsurface structure analysis for estimating the invasion depth of early gastric cancer.

Methods

White-light imaging and narrow-band imaging (NBI) endoscopy were used to visualize the lesions of the included patients. The area ratio and depth-predicting score (DPS) of each patient were calculated; meanwhile, each lesion was examined by endoscopic ultrasonography (EUS).

Results

Ninety-three patients were included between 2016 and 2019. Microsurface structure is related to the histological differentiation and progression of early gastric cancer. The receiver operating characteristic curve showed that when an area ratio of 80.3% was used as a cut-off value for distinguishing mucosal (M) and submucosal (SM) type 0-II gastric cancers, the sensitivity, specificity, and accuracy were 82.9%, 80.2%, and 91.6%, respectively. The accuracies for distinguishing M/SM differentiated and undifferentiated early gastric cancers were 87.4% and 84.8%, respectively. The accuracy of EUS for distinguishing M/SM early gastric cancer was 74.9%. DPS can only distinguish M-SM1 (SM infiltration <500 μm)/SM (SM infiltration ≥500 μm) with an accuracy of 83.8%. The accuracy of using area ratio for distinguishing 0-II early gastric cancers was better than those of using DPS and EUS (P < 0.05).

Conclusion

Quantitative analysis of microsurface structure can be performed to assess M/SM type 0-II gastric cancer and is expected to be effective for judging the invasion depth of gastric cancer.

Ultrathin endoscope equipped with ultrasonic miniprobe for upper GI US in a porcine model

BACKGROUND AND AIMS

Ultrathin EGD (UT-EGD) is an ideal tool for unsedated upper GI examination and pediatric gastroenterology but is rarely competent for EUS miniprobe (EUS-MP). We developed a UT-EGD US method (UT-EUS) and verified its clinical application value through animal experiments.

METHODS

Five Bama miniature pigs were selected. Using an acoustic medium, we performed US on the duodenum, stomach, and esophagus, respectively, with conventional 20-MHz EUS miniprobe (EUS-MP-20), 20-MHz UT-EUS (UT-EUS-20), and 30-MHz UT-EUS (UT-EUS-30). The times to acquire 5 consecutive stable US images, number of identifiable wall layers, and quality and penetration depth of the images were recorded.

RESULTS

No significant differences were found in the time required to obtain images between EUS-MP-20 and UT-EUS-20 at each site (P > .05). UT-EUS-30 showed more wall levels than UT-EUS-20 (P < .05). No significant differences were noted between EUS-MP-20 and UT-EUS-20 in imaging quality and penetration depth (P > .05).

CONCLUSIONS

The UT-EUS is easy to use with a satisfactory image quality and has potential clinical application value.

Whether the use of intravenous contrast-enhanced ultrasound can improve the accuracy of T stage diagnosis of gastric cancer?

OBJECTIVE

To explore whether the use of intravenous contrast-enhanced ultrasound (CEUS) can improve the diagnostic accuracy of T stage of gastric cancer on the basis of oral contrast-enhanced ultrasound (OCEUS).

METHODS

Eighty-one patients with gastric cancer who underwent preoperative CEUS. Criteria for T staging was AJCC Stage 8th Edition. To compare the diagnostic accuracy, sensitivity and specificity of OCEUS and CEUS for T staging of gastric cancer using pathological results as the gold standard.

RESULTS

There were differences in the diagnostic accuracy of T2 (P = 0.048), T3(P = 0.002) and T4 (P = 0.000) between OCEUS and CEUS. And there was no difference in diagnostic accuracy for T1 (P = 0.118). CEUS significantly improved diagnostic sensitivity and specificity in T3 (42.86% to 61.90% for sensitivity, 86.96% to 93.75% for specificity), and T4 (64.71% to 76.47% for sensitivity, 78.33% to 95.74% for specificity).

CONCLUSION

Compared with OCEUS, CEUS did improve the diagnostic accuracy of T2, T3, and T4 stages of gastric cancer.

Application of artificial intelligence in endoscopic gastrointestinal tumors

With an increasing number of patients with gastrointestinal cancer, effective and accurate early diagnostic clinical tools are required provide better health care for patients with gastrointestinal cancer. Recent studies have shown that artificial intelligence (AI) plays an important role in the diagnosis and treatment of patients with gastrointestinal tumors, which not only improves the efficiency of early tumor screening, but also significantly improves the survival rate of patients after treatment. With the aid of efficient learning and judgment abilities of AI, endoscopists can improve the accuracy of diagnosis and treatment through endoscopy and avoid incorrect descriptions or judgments of gastrointestinal lesions. The present article provides an overview of the application status of various artificial intelligence in gastric and colorectal cancers in recent years, and the direction of future research and clinical practice is clarified from a clinical perspective to provide a comprehensive theoretical basis for AI as a promising diagnostic and therapeutic tool for gastrointestinal cancer.

Coexistence of early gastric cancer and benign submucosal lesions mimic invasive cancer: a retrospective multicenter experience

OBJECTIVE

To present a study to identify the characteristics of coexisting early gastric cancer (EGC) and benign submucosal lesions, with the aim of reducing the adverse consequences of overdiagnosis and overtreatment.

METHODS

In this retrospective study, we searched the endoscopic databases of three tertiary centers. We screened of patients suspected of early gastric cancer submucosal infiltration by conventional endoscopy and ultimately selected for endoscopic submucosal dissection treatment after endoscopic ultrasonography and magnifying endoscopy with narrow-band imaging examination. Patients with coexisting EGC and benign submucosal lesions in histological sections were included. Clinical data and endoscopic images were reviewed. To evaluate the precision of endoscopists' diagnoses for this type of lesion, eight endoscopists with different experiences were recruited to judge the infiltration depth of these lesions and analyze the accuracy rate.

RESULTS

We screened 520 patients and retrospectively identified 18 EGC patients with an invasive cancer-like morphology. The most common lesion site was the cardia (12/18, 66.67%). The coexisting submucosal lesions could be divided into solid (5/18, 27.78%) and cystic (13/18, 72.22%). The most common type of submucosal lesion was gastritis cystica profunda (12/18, 66.67%), whereas leiomyoma was the predominant submucosal solid lesion (3/18, 16.67%). Ten (55.56%) patients < underwent endoscopic ultrasonography; submucosal lesions were definitively diagnosed in 6 patients (60.00%). The accuracy of judgement of the infiltration depth was significantly lower in cases of coexistence of EGC with benign submucosal lesions (EGC-SML) than in EGC (38.50% versus 65.60%, P = 0.0167). The rate of over-diagnosis was significantly higher within the EGC-SML group compared to the EGC group (59.17% versus 10.83%, P < 0.0001).

CONCLUSIONS

We should be aware of the coexistence of EGC and benign submucosal lesions, the most common of which is early cardiac-differentiated cancer with gastritis cystica profunda.

Endoscopic submucosal dissection for early gastric cancer: It is time to consider the quality of its outcomes

Endoscopic resection, particularly endoscopic submucosal dissection (ESD), is widely used as a standard treatment modality for early gastric cancer (EGC) when the risk of lymph node metastasis is negligible. Compared with surgical gastrectomy, ESD is a minimally invasive procedure with additional advantages, such as preservation of the entire stomach and maintenance of the patient's quality of life. However, not all patients achieve curative resection after ESD of EGC. Several patients require surgical gastrectomy after ESD to achieve a curative treatment. Additional surgery after ESD, owing to non-curative resection, places considerable emotional and financial burdens on patients. Recently, as the number of endoscopists performing ESD has increased, the rate of non-curative resection after ESD has increased correspondingly. In order to decrease the non-curative resection rate, as well as determine the ideal rate of non-curative resection after ESD, it is time to consider quality indicators for the outcomes of ESD for EGC.

American Society for Gastrointestinal Endoscopy guideline on endoscopic submucosal dissection for the management of early esophageal and gastric cancers: methodology and review of evidence

This document from the American Society for Gastrointestinal Endoscopy (ASGE) provides a full description of the methodology used in the review of the evidence used to inform the final guidance outlined in the accompanying Summary and Recommendations document regarding the role of endoscopic submucosal dissection (ESD) in the management of early esophageal and gastric cancers. This guideline used the Grading of Recommendations, Assessment, Development and Evaluation framework and specifically addresses the role of ESD versus EMR and/or surgery, where applicable, for the management of early esophageal squamous cell carcinoma (ESCC), esophageal adenocarcinoma (EAC), and gastric adenocarcinoma (GAC) and their corresponding precursor lesions. For ESCC, the ASGE suggests ESD over EMR for patients with early-stage, well-differentiated, nonulcerated cancer >15 mm, whereas in patients with similar lesions ≤15 mm, the ASGE suggests either ESD or EMR. The ASGE suggests against surgery for such patients with ESCC, whenever possible. For EAC, the ASGE suggests ESD over EMR for patients with early-stage, well-differentiated, nonulcerated cancer >20 mm, whereas in patients with similar lesions measuring ≤20 mm, the ASGE suggests either ESD or EMR. For GAC, the ASGE suggests ESD over EMR for patients with early-stage, well or moderately differentiated, nonulcerated intestinal type cancer measuring 20 to 30 mm, whereas for patients with similar lesions <20 mm, the ASGE suggests either ESD or EMR. The ASGE suggests against surgery for patients with such lesions measuring ≤30 mm, whereas for lesions that are poorly differentiated, regardless of size, the ASGE suggests surgical evaluation over endosic approaches.

The Role of Artificial Intelligence in Gastric Cancer: Surgical and Therapeutic Perspectives: A Comprehensive Review

Stomach cancer has a high annual mortality rate worldwide necessitating early detection and accurate treatment. Even experienced specialists can make erroneous judgments based on several factors. Artificial intelligence (AI) technologies are being developed rapidly to assist in this field. Here, we aimed to determine how AI technology is used in gastric cancer diagnosis and analyze how it helps patients and surgeons. Early detection and correct treatment of early gastric cancer (EGC) can greatly increase survival rates. To determine this, it is important to accurately determine the diagnosis and depth of the lesion and the presence or absence of metastasis to the lymph nodes, and suggest an appropriate treatment method. The deep learning algorithm, which has learned gastric lesion endoscopyimages, morphological characteristics, and patient clinical information, detects gastric lesions with high accuracy, sensitivity, and specificity, and predicts morphological characteristics. Through this, AI assists the judgment of specialists to help select the correct treatment method among endoscopic procedures and radical resections and helps to predict the resection margins of lesions. Additionally, AI technology has increased the diagnostic rate of both relatively inexperienced and skilled endoscopic diagnosticians. However, there were limitations in the data used for learning, such as the amount of quantitatively insufficient data, retrospective study design, single-center design, and cases of non-various lesions. Nevertheless, this assisted endoscopic diagnosis technology that incorporates deep learning technology is sufficiently practical and future-oriented and can play an important role in suggesting accurate treatment plans to surgeons for resection of lesions in the treatment of EGC.

Artificial intelligence in theranostics of gastric cancer, a review

Gastric cancer (GC) is one of the commonest cancers with high morbidity and mortality in the world. How to realize precise diagnosis and therapy of GC owns great clinical requirement. In recent years, artificial intelligence (AI) has been actively explored to apply to early diagnosis and treatment and prognosis of gastric carcinoma. Herein, we review recent advance of AI in early screening, diagnosis, therapy and prognosis of stomach carcinoma. Especially AI combined with breath screening early GC system improved 97.4 % of early GC diagnosis ratio, AI model on stomach cancer diagnosis system of saliva biomarkers obtained an overall accuracy of 97.18 %, specificity of 97.44 %, and sensitivity of 96.88 %. We also discuss concept, issues, approaches and challenges of AI applied in stomach cancer. This review provides a comprehensive view and roadmap for readers working in this field, with the aim of pushing application of AI in theranostics of stomach cancer to increase the early discovery ratio and curative ratio of GC patients.

Recent advances in endoscopic management of gastric neoplasms

The development and clinical application of new diagnostic endoscopic technologies such as endoscopic ultrasonography with biopsy, magnification endoscopy, and narrow-band imaging, more recently supplemented by artificial intelligence, have enabled wider recognition and detection of various gastric neoplasms including early gastric cancer (EGC) and subepithelial tumors, such as gastrointestinal stromal tumors and neuroendocrine tumors. Over the last decade, the evolution of novel advanced therapeutic endoscopic techniques, such as endoscopic mucosal resection, endoscopic submucosal dissection, endoscopic full-thickness resection, and submucosal tunneling endoscopic resection, along with the advent of a broad array of endoscopic accessories, has provided a promising and yet less invasive strategy for treating gastric neoplasms with the advantage of a reduced need for gastric surgery. Thus, the management algorithms of various gastric tumors in a defined subset of the patient population at low risk of lymph node metastasis and amenable to endoscopic resection, may require revision considering upcoming data given the high success rate of en bloc resection by experienced endoscopists. Moreover, endoscopic surveillance protocols for precancerous gastric lesions will continue to be refined by systematic reviews and meta-analyses of further research. However, the lack of familiarity with subtle endoscopic changes associated with EGC, as well as longer procedural time, evolving resection techniques and tools, a steep learning curve of such high-risk procedures, and lack of coding are issues that do not appeal to many gastroenterologists in the field. This review summarizes recent advances in the endoscopic management of gastric neoplasms, with special emphasis on diagnostic and therapeutic methods and their future prospects.

Prospective Comparison of Oral Contrast-Enhanced Transabdominal Ultrasound Imaging With Contrast-Enhanced Computed Tomography in Pre-operative Tumor Staging of Gastric Cancer

The aim of this prospective study was to compare the diagnostic accuracy of oral contrast-enhanced transabdominal ultrasound imaging (OCTU) with that of contrast-enhanced computed tomography (CT) for the pre-operative tumor staging of gastric cancer, with post-operative pathology as the standard. We included 108 cases of gastric cancer with simultaneous OCTU and enhanced CT pre-operative tumor staging diagnoses. Results were compared with post-operative pathology based on the eighth edition of the American Joint Committee on Cancer tumor-node-metastasis staging guidelines for gastric cancer. The accuracy of each tumor stage was obtained by comparing OCTU and enhanced CT diagnoses with post-operative pathology. The McNemar test was used to compare the overall accuracy of the two methods. There was no statistical difference in accuracy between OCTU (72.2%) and enhanced CT (75.9%, p = 0.644) for overall pre-operative tumor staging diagnosis. For stages T1 to T4, the accuracy rates of OCTU were 84.2%, 81.8%, 69.4% and 65.5%, respectively, and those for enhanced CT were 52.6%, 72.7%, 87.8% and 72.4%, respectively. OCTU is comparable to enhanced CT in the preoperative overall T-stage diagnosis of gastric cancer.

The factors influencing the accuracy of pre-operative endoscopic ultrasonography assessment in endoscopic treatments for gastrointestinal tumors

BACKGROUND

This retrospective study aimed to evaluate the factors influencing the accuracy of Endoscopic Ultrasonography (EUS) as a preoperative assessment for gastrointestinal tumors.

METHODS

A total of 261 patients with 264 gastrointestinal tumors were enrolled in the study. The parameters of the gastrointestinal lesions examined under EUS and their pathology were recorded and analyzed.

RESULTS

The accuracy of EUS for detecting intramucosal lesions and subepithelial lesions (SELs) were 83.6% and 91.4%, respectively. One hundred and ninety-four (73.5%) lesions originated from the mucous layer, as determined by pre-operation EUS examinations. The accuracy of EUS in predicting the correct T stage for intramucosal lesions in the gastric region, esophagus, and colorectum was 77%, 71.8%, and 84.6%, respectively. According to the Paris endoscopic classification, the distribution of macroscopic patterns was different between the EUS-pathology conformity and nonconformity groups (p = 0.018). In the nonconformity group, 48.6% of erosive lesions were classified as 0-IIc, 0-IIa + IIc, 0-IIc + IIa or 0-III macroscopic patterns compared with 26% patients in the conformity group (p = 0.025). Univariate analyses demonstrated that ulcerative lesions (OR = 7.516, 95% Confidence Interval [CI] 2.574-21.952, p < 0.001), location at the cardia of the stomach (OR = 3.619, 95%CI 1.076-12.168, p = 0.038), malignant tumor (OR = 2.920, 95%CI 1.339-6.368, p = 0.007) were significantly associated with EUS inaccuracy. Multivariate logistic regression analyses showed that ulcer was an independent risk factor associated with EUS inaccuracy, with odds ratios of 5.094 (95% CI: 1.641-15.807, p = 0.005).

CONCLUSIONS

Our findings suggested that EUS is a reliable and easy-to-use diagnostic tool in decision-making regarding appropriate endoscopic treatment for gastrointestinal tumors. However, the diagnostic accuracy of EUS appeared questionable in the presence of ulceration.

Preoperative diagnosis of tumor depth in gastric cancer using transabdominal ultrasonography compared to using endoscopy and computed tomography

BACKGROUND

An accurate evaluation method for preoperative diagnosis has not yet been established in patients with gastric cancer (GC), though it is essential for optimal treatment. Current standard modalities are endoscopy and contrast computed tomography (CT). In this study, we investigated the efficacy and limitations of transabdominal ultrasonography (TUS) for the assessment of tumor invasion.

METHODS

We enrolled 178 consecutive patients with GC evaluated by TUS, endoscopy, and contrast CT before gastrectomy. For the TUS examination, patients ingested water to fill their stomachs. The clinical staging determined using these modalities was compared to the pathological staging.

RESULTS

The overall accuracy of clinical T staging using TUS was 47.8% (pT1a: 5.8% (2/35); pT1b: 58.8% (20/35); pT2: 69.6% (16/23); pT3: 66.7% (22/33); pT4a: 46% (23/50); pT4b: 100% (2/2)). Using endoscopy, contrast CT, and TUS, the overall accuracy was 60.7%. The accuracy of TUS was associated with the tumor region (U region: 50% (14/28); M: 31.8% (14/44); L: 53.7% (57/106); P = 0.048), but not with the cross-sectional parts (P = 0.49). Multivariate analysis identified inaccurate TUS as independently correlating with tumor region (M vs. U/L, odds ratio (OR) = 3.11, 95% confidence interval (CI) 1.41-6.87; P = 0.005) and pT (pT1 vs. pT2-4, OR = 3.00, 95%CI 1.31-6.87; P = 0.009).

CONCLUSIONS

The present study demonstrated the importance of TUS in evaluating GC. Thus, TUS may be useful for clinical T staging in certain circumstances, leading to treatment optimization.

Cooperation between artificial intelligence and endoscopists for diagnosing invasion depth of early gastric cancer

BACKGROUND AND STUDY AIMS

The diagnostic ability of endoscopists to determine invasion depth of early gastric cancer is not favorable. We designed an artificial intelligence (AI) classifier for differentiating intramucosal and submucosal gastric cancers and examined it to establish a diagnostic method based on cooperation between AI and endoscopists.

PATIENTS AND METHODS

We prepared 500 training images using cases of mainly depressed-type early gastric cancer from 250 intramucosal cancers and 250 submucosal cancers. We also prepared 200 test images each of 100 cancers from another institution. We designed an AI classifier to differentiate between intramucosal and submucosal cancers by deep learning. We examined the performance of the AI classifier and the majority vote of the endoscopists as high confidence and low confidence diagnostic probability, respectively, and cooperatively combined them to establish a diagnostic method providing high accuracy.

RESULTS

Internal evaluation of the training images showed that accuracy, sensitivity, specificity, and F1 measure by the AI classifier were 77%, 76%, 78%, and 0.768, and those of the majority vote of the endoscopists were 72.6%, 53.6%, 91.6%, and 0.662, respectively. A diagnostic method based on cooperation between AI and the endoscopists showed that the respective values were 78.0%, 76.0%, 80.0%, and 0.776 for the test images. The value of F1 measure was especially higher than those by AI or the endoscopists alone.

CONCLUSIONS

Cooperation between AI and endoscopists improved the diagnostic ability to determine invasion depth of early gastric cancer.

A Framework to Predict Gastric Cancer Based on Tongue Features and Deep Learning

Gastric cancer has become a global health issue, severely disrupting daily life. Early detection in gastric cancer patients and immediate treatment contribute significantly to the protection of human health. However, routine gastric cancer examinations carry the risk of complications and are time-consuming. We proposed a framework to predict gastric cancer non-invasively and conveniently. A total of 703 tongue images were acquired using a bespoke tongue image capture instrument, then a dataset containing subjects with and without gastric cancer was created. As the images acquired by this instrument contain non-tongue areas, the Deeplabv3+ network was applied for tongue segmentation to reduce the interference in feature extraction. Nine tongue features were extracted, relationships between tongue features and gastric cancer were explored by using statistical methods and deep learning, finally a prediction framework for gastric cancer was designed. The experimental results showed that the proposed framework had a strong detection ability, with an accuracy of 93.6%. The gastric cancer prediction framework created by combining statistical methods and deep learning proposes a scheme for exploring the relationships between gastric cancer and tongue features. This framework contributes to the effective early diagnosis of patients with gastric cancer.

Diagnostic value of endoscopic ultrasonography for the depth of gastric cancer suspected of submucosal invasion: a multicenter prospective study

BACKGROUND

Although the combination of conventional endoscopy (CE) and endoscopic ultrasonography (EUS) is useful for predicting the depth of early gastric cancer (EGC), the diagnostic value of EUS for submucosal (SM) invasive cancer has not been fully investigated.

METHODS

We conducted a multicenter prospective study from May 2017 to January 2021 to evaluate the validity of a diagnostic strategy combining CE and EUS and to clarify the additional value of EUS for EGC suspected of SM invasion. In each case, the diagnosis was first made using CE, followed by EUS, and finally confirmed using a combination algorithm.

RESULTS

A total of 180 patients with EGC were enrolled from 10 institutions, of which 175 were analyzed. The histopathological depths were M, SM1, SM2, and ≥ MP in 72, 16, 64, and 23 lesions, respectively. Treatment included 92 endoscopic submucosal dissection cases and 83 surgical cases. The overall diagnostic accuracy classified by M-SM1 or SM2-MP was 58.3% for CE, 75.7% for EUS, and 78.9% for the combination of CE and EUS; the latter two were significantly higher than that of CE alone (P < 0.001). The CE, EUS, and combination accuracy rates in 108 differentiated-type lesions were 51.9%, 77.4%, and 79.6%, respectively; the latter two were significantly higher than CE alone (P < 0.001). A significant additive effect of EUS was observed in CE-SM2 low-confidence lesions but not in CE-M-SM1 lesions or in CE-SM2 high-confidence lesions. Among the nine CE findings, irregular surface, submucosal tumor-like elevation, and non-extension signs were significant independent markers of pSM2-MP. Poorly delineated EUS lesions were misdiagnosed.

CONCLUSIONS

EUS provides additional value for differentiated-type and CE-SM2 low-confidence EGCs in diagnosing invasion depth.

CLINICAL REGISTRATION NUMBER

UMIN000025862.

Current Status of Artificial Intelligence-Based Computer-Assisted Diagnosis Systems for Gastric Cancer in Endoscopy

Artificial intelligence (AI) is gradually being utilized in various fields as its performance has been improving with the development of deep learning methods, availability of big data, and the progression of computer processing units. In the field of medicine, AI is mainly implemented in image recognition, such as in radiographic and pathologic diagnoses. In the realm of gastrointestinal endoscopy, although AI-based computer-assisted detection/diagnosis (CAD) systems have been applied in some areas, such as colorectal polyp detection and diagnosis, so far, their implementation in real-world clinical settings is limited. The accurate detection or diagnosis of gastric cancer (GC) is one of the challenges in which performance varies greatly depending on the endoscopist's skill. The diagnosis of early GC is especially challenging, partly because early GC mimics atrophic gastritis in the background mucosa. Therefore, several CAD systems for GC are being actively developed. The development of a CAD system for GC is considered challenging because it requires a large number of GC images. In particular, early stage GC images are rarely available, partly because it is difficult to diagnose gastric cancer during the early stages. Additionally, the training image data should be of a sufficiently high quality to conduct proper CAD training. Recently, several AI systems for GC that exhibit a robust performance, owing to being trained on a large number of high-quality images, have been reported. This review outlines the current status and prospects of AI use in esophagogastroduodenoscopy (EGDS), focusing on the diagnosis of GC.

Artificial intelligence in gastric cancer: applications and challenges

Gastric cancer (GC) is one of the most common malignant tumors with high mortality. Accurate diagnosis and treatment decisions for GC rely heavily on human experts' careful judgments on medical images. However, the improvement of the accuracy is hindered by imaging conditions, limited experience, objective criteria, and inter-observer discrepancies. Recently, the developments of machine learning, especially deep-learning algorithms, have been facilitating computers to extract more information from data automatically. Researchers are exploring the far-reaching applications of artificial intelligence (AI) in various clinical practices, including GC. Herein, we aim to provide a broad framework to summarize current research on AI in GC. In the screening of GC, AI can identify precancerous diseases and assist in early cancer detection with endoscopic examination and pathological confirmation. In the diagnosis of GC, AI can support tumor-node-metastasis (TNM) staging and subtype classification. For treatment decisions, AI can help with surgical margin determination and prognosis prediction. Meanwhile, current approaches are challenged by data scarcity and poor interpretability. To tackle these problems, more regulated data, unified processing procedures, and advanced algorithms are urgently needed to build more accurate and robust AI models for GC.

Clinical factors associated with noncurative endoscopic submucosal dissection for the expanded indication of intestinal-type early gastric cancer: Post hoc analysis of a multi-institutional, single-arm, confirmatory trial (JCOG0607)

OBJECTIVES

The multi-institutional, single-arm, confirmatory trial JCOG0607 showed excellent efficacy of endoscopic submucosal dissection (ESD) for the expanded indication of intramucosal intestinal-type early gastric cancer (EGC), which consists of two groups: lesions >2 cm if clinical finding of ulcer (cUL)-negative, or those ≤3 cm if cUL-positive because of the expected low risk of lymph node metastasis. However, the proportion of noncurative resections (NCR) requiring additional surgery was high (32.4%). This post hoc analysis aimed to explore the clinical factors associated with NCR.

METHODS

As the expanded indication includes two different groups, we explored the clinical factors associated with NCR separately in cUL-negative (>2 cm) and cUL-positive (≤3 cm) groups using the log-linear model.

RESULTS

Two hundred and sixty cUL-negative and 206 cUL-positive EGCs were analyzed. The proportions of NCR were 33.8% in the cUL-negative group and 29.6% in the cUL-positive group. A multivariable analysis demonstrated that moderately differentiated predominant histology diagnosed in pretreatment biopsy (risk ratio [RR] 1.93, 95% confidence interval [CI] 1.34-2.77, P < 0.001) and lesion in the upper stomach (RR 1.75, 95% CI 1.03-2.96, P = 0.038) in the cUL-negative EGCs, and tumor size >2 cm (RR 1.78, 95% CI 1.22-2.58, P = 0.003) and female sex (RR 1.62, 95% CI 1.07-2.44, P = 0.021) in the cUL-positive EGCs were independent factors associated with NCR.

CONCLUSIONS

Clinical risk factors associated with NCR were different between cUL-negative and cUL-positive EGCs. To avoid NCR, we need to take these factors into account when deciding expanded indications for ESD.

Machine Learning Improves the Prediction Rate of Non-Curative Resection of Endoscopic Submucosal Dissection in Patients with Early Gastric Cancer

Non-curative resection (NCR) of early gastric cancer (EGC) after endoscopic submucosal dissection (ESD) can increase the burden of additional treatment and medical expenses. We aimed to develop a machine-learning (ML)-based NCR prediction model for EGC prior to ESD. We obtained data from 4927 patients with EGC who underwent ESD between January 2006 and February 2020. Ten clinicopathological characteristics were selected using extreme gradient boosting (XGBoost) and were used to develop a ML-based model. Dataset was divided into the training and internal validation sets and verified using an external validation set. Sensitivity, specificity, and area under the receiver operating characteristic curve (AUROC) were evaluated. The performance of each model was compared by using the Delong test. A total of 1100 (22.1%) patients were identified as being treated non-curatively with ESD. Seven ML-based NCR prediction models were developed. The performance of NCR prediction was highest in the XGBoost model (AUROC, 0.851; 95% confidence interval, 0.837–0.864). When we compared the prediction performance by the Delong test, XGBoost (p = 0.02) and support vector machine (p = 0.02) models showed a significantly higher performance among the NCR prediction models. We developed an ML model capable of accurately predicting the NCR of EGC before ESD. This ML model can provide useful information for decision-making regarding the appropriate treatment of EGC before ESD.

A potential decision-making algorithm based on endoscopic ultrasound for staging early gastric cancer: a retrospective study

BACKGROUND

Clinical staging of gastric cancer (GC) before treatment is essential. Endoscopic ultrasound (EUS) is a recommended staging tool, but its efficacy remains controversial. Our previous prospective study evaluated the potential value of EUS for T staging and presented discrepancies. In this study, we aimed to evaluate the efficacy of EUS in T staging by comparing it with pathological staging. We analyze the factors that can potentially affect accuracy to identify suitable subgroups for EUS staging.

METHODS

Data from a total of 1763 consecutive patients with GC from January 2015 to December 2017 were analyzed. Results from EUS and pathological T staging were compared. The factors that might affect EUS's accuracy were analyzed.

RESULTS

The sensitivity, specificity, positive predictive value, and negative predictive value of EUS in patients with early GC were 62.08%, 96.13%, 90.94%, and 80.21%, respectively. The accuracy rates of uT1, uT2-uT4, and uT3-uT4 were 90.94%, 79.02%, and 78.39%, respectively. In multivariate analysis, underestimation was more likely to be observed in patients with tumors located in the middle or upper third of the stomach. Overestimation was more likely to be observed in patients with tumors located in the lower third or those without ulcer. Other factors affecting accuracy included ulcer, differentiation, larger size and undergoing surgery.

CONCLUSION

Our findings highlight the role of EUS in determining the T staging of GC. Overestimation and underestimation in T-staging were significantly associated with the tumor location in early GC, and a decision-making algorithm was proposed for clinical practice in early cancers based on these findings.

Comparison of Endoscopic Ultrasonography and Conventional Endoscopy for Prediction of Tumor Depth in Superficial Nonampullary Duodenal Epithelial Tumors

BACKGROUND

The utility of endoscopic ultrasonography (EUS) in predicting tumor depth among superficial nonampullary duodenal epithelial tumors (SNADETs) is unclear. The aim was to compare EUS with conventional endoscopy (CE) for the evaluation of tumor invasion of SNADETs.

METHODS

A retrospective analysis was performed on consecutive 174 lesions/169 patients with duodenal dysplasia or adenocarcinoma with invasion up to submucosa who underwent both CE and EUS before endoscopic (n = 133) or surgical (n = 41) treatment. Endoscopic staging by CE was performed based on the characteristic endoscopic criteria of submucosal invasion (irregular surface, submucosal tumor [SMT]-like marginal elevation, and fusion of converging folds). The diagnostic performance of each test was compared with the final histology.

RESULTS

The sensitivity and accuracy of estimating the depth were higher for CE compared to that of EUS (99.4% vs. 89.4%, p < 0.01 and 97.7% vs. 87.9%, p < 0.01, respectively). Univariate analysis of endoscopic factors revealed that tumor diameter, red color, SMT-like appearance, and hypoechogenicity were factors related to advanced histology. Multivariate analysis revealed that the presence of SMT-like appearance based on CE was an independent factor to predict submucosal invasion (p = 0.025). Gross morphology of the combined type was associated to incorrect diagnosis of EUS (p = 0.007). Among 3 cases in which EUS overestimated the tumor depth, carcinoma extension in submucosal Brunner's gland or nontumorous submucosal cystic dilation was observed.

CONCLUSION

EUS may not be necessary, and CE may be sufficient for determining the optimal therapeutic strategy for SNADETs.

Endoscopic Treatment of Superficial Gastric Cancer: Present Status and Future

Although the mortality rates of gastric cancer (GC) are gradually declining, gastric cancer is still the fourth leading cause of cancer-related death worldwide. This may be due to the high rate of patients who are diagnosed with GC at advanced stages. However, in countries such as Japan with endoscopic screening systems, more than half of GCs are discovered at an early stage, enabling endoscopic resection (ER). Especially after the introduction of endoscopic submucosal dissection (ESD) in Japan around 2000, a high en bloc resection rate allowing pathological assessment of margin and depth has become possible. While ER is a diagnostic method of treatment and may not always be curative, it is widely accepted as standard treatment because it is less invasive than surgery and can provide an accurate diagnosis for deciding whether additional surgery is necessary. The curability of ER is currently assessed by the completeness of primary tumor removal and the possibility of lymph node metastasis. This review introduces methods, indications, and curability criteria for ER of EGC. Despite recent advances, several problems remain unsolved. This review will also outline the latest evidence concerning future issues.

Long-term outcomes after endoscopic resection for late-elderly patients with early gastric cancer

BACKGROUND AND AIMS

With the population aging, the incidence of early gastric cancer (EGC) is increasing. We aimed to clarify the indications for endoscopic resection (ER) in late-elderly patients with EGC in terms of life expectancy.

METHODS

Patients aged ≥75 years who underwent ER for EGC at our institution from January 2007 to December 2012 were enrolled. Clinical data, including Eastern Cooperative Oncology Group performance status (ECOG-PS), Charlson comorbidity index, and Prognostic Nutritional Index (PNI), were collected at the time of ER. Overall survival (OS) was the main outcome measure.

RESULTS

Four hundred consecutive patients were enrolled. Mean patient age was 79.3 years (range, 75-93). The 5-year follow-up rate was 89.0% (median follow-up period, 5.6 years). Five-year OS was 80.8% (95% confidence interval [CI], 76.4-84.4), and 5-year net survival standardized for age, sex, and calendar year was 1.09 (95% CI, 1.03-1.15). With a multivariate analysis, ECOG-PS 2 to 4 (hazard ratio, 8.84; 95% CI, 3.07-25.4), PNI <49.1 (hazard ratio, 2.49; 95% CI, 1.53-4.06), and eCura C-2 (hazard ratio, 1.79; 95% CI, 1.11-2.88) were independent prognostic factors. When none of these factors was met, the 5-year OS rate was 90.4% (95% CI, 84.0-94.3).

CONCLUSIONS

ER for EGC in late-elderly patients may improve life expectancy. ER is recommended in patients with a good ECOG-PS and PNI and in whom ER is expected to be non-eCura C-2.

Application of convolutional neural networks for evaluating the depth of invasion of early gastric cancer based on endoscopic images

BACKGROUND AND AIM

Recently, artificial intelligence (AI) has been used in endoscopic examination and is expected to help in endoscopic diagnosis. We evaluated the feasibility of AI using convolutional neural network (CNN) systems for evaluating the depth of invasion of early gastric cancer (EGC), based on endoscopic images.

METHODS

This study used a deep CNN model, ResNet152. From patients who underwent treatment for EGC at our hospital between January 2012 and December 2016, we selected 100 consecutive patients with mucosal (M) cancers and 100 consecutive patients with cancers invading the submucosa (SM cancers). A total of 3508 non-magnifying endoscopic images of EGCs, including white-light imaging, linked color imaging, blue laser imaging-bright, and indigo-carmine dye contrast imaging, were included in this study. A total of 2288 images from 132 patients served as the development dataset, and 1220 images from 68 patients served as the testing dataset. Invasion depth was evaluated for each image and lesion. The majority vote was applied to lesion-based evaluation.

RESULTS

The sensitivity, specificity, and accuracy for diagnosing M cancer were 84.9% (95% confidence interval [CI] 82.3%-87.5%), 70.7% (95% CI 66.8%-74.6%), and 78.9% (95% CI 76.6%-81.2%), respectively, for image-based evaluation, and 85.3% (95% CI 73.4%-97.2%), 82.4% (95% CI 69.5%-95.2%), and 83.8% (95% CI 75.1%-92.6%), respectively, for lesion-based evaluation.

CONCLUSIONS

The application of AI using CNN to evaluate the depth of invasion of EGCs based on endoscopic images is feasible, and it is worth investing more effort to put this new technology into practical use.

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.

Deep learning system compared with expert endoscopists in predicting early gastric cancer and its invasion depth and differentiation status (with videos)

BACKGROUND AND AIMS

We aimed to develop and validate a deep learning-based system that covers various aspects of early gastric cancer (EGC) diagnosis, including detecting gastric neoplasm, identifying EGC, and predicting EGC invasion depth and differentiation status. Herein, we provide a state-of-the-art comparison of the system with endoscopists using real-time videos in a nationwide human-machine competition.

METHODS

This multicenter, prospective, real-time, competitive comparative, diagnostic study enrolled consecutive patients who received magnifying narrow-band imaging endoscopy at the Peking University Cancer Hospital from June 9, 2020 to November 17, 2020. The offline competition was conducted in Wuhan, China, and the endoscopists and the system simultaneously read patients' videos and made diagnoses. The primary outcomes were sensitivity in detecting neoplasms and diagnosing EGCs.

RESULTS

One hundred videos, including 37 EGCs and 63 noncancerous lesions, were enrolled; 46 endoscopists from 44 hospitals in 19 provinces in China participated in the competition. The sensitivity rates of the system for detecting neoplasms and diagnosing EGCs were 87.81% and 100%, respectively, significantly higher than those of endoscopists (83.51% [95% confidence interval [CI], 81.23-85.79] and 87.13% [95% CI, 83.75-90.51], respectively). Accuracy rates of the system for predicting EGC invasion depth and differentiation status were 78.57% and 71.43%, respectively, slightly higher than those of endoscopists (63.75% [95% CI, 61.12-66.39] and 64.41% [95% CI, 60.65-68.16], respectively).

CONCLUSIONS

The system outperformed endoscopists in identifying EGCs and was comparable with endoscopists in predicting EGC invasion depth and differentiation status in videos. This deep learning-based system could be a powerful tool to assist endoscopists in EGC diagnosis in clinical practice.

Role of endoscopic ultrasound in anticancer therapy: Current evidence and future perspectives

The digestive system is one of the most common sites of malignancies in humans. Since gastrointestinal tumors represent a massive global health burden both in terms of morbidity and health care expenditures, scientists continuously develop novel diagnostic and therapeutic methods to ameliorate the detrimental effects of this group of diseases. Apart from the well-established role of the endoscopic ultrasound (EUS) in the diagnostic course of gastrointestinal and hepatobiliary malignancies, we have recently become acquainted with a vast array of its therapeutic possibilities. A multitude of previously established, evidence-based methods that might now be guided by the EUS emerged: Radiofrequency ablation, brachytherapy, fine needle injection, celiac plexus neurolysis, and endoscopic submucosal dissection. In this review we endeavored to provide a comprehensive overview of the role of these methods in different malignancies of the digestive system, primarily in the treatment and symptom control in pancreatic cancer, and additionally in the management of hepatic, gastrointestinal tumors, and pancreatic cysts.

Diagnostic Efficacy and Decision-Making Role of Preoperative Endoscopic Ultrasonography in Early Gastric Cancer

Objectives: Endoscopic ultrasonography (EUS) is the most commonly used method for T staging of early gastric cancer (EGC). However, the studies pertaining to EUS for staging EGC reported widely varied sensitivities and specificities. This study aimed to estimate the overall diagnostic accuracy of EUS for staging the depth of EGCs and to explore the influential factors. Methods: We retrospectively reviewed data from 208 consecutive patients with EGC, and all patients underwent EUS for estimating tumor invasion depth, followed by either curative surgery or endoscopic submucosal dissection (ESD). The diagnostic accuracy of EUS was evaluated by comparing the final histologic results of resected specimens. The correlation between accuracy of EUS and characteristics of EGC lesion was analyzed. Results: A total of 211 EGC lesions in 208 patients were included. The overall diagnostic accuracy of EUS in assessing the tumor invasion depth of EGC was 55.9%. Multivariate analysis showed that submucosal invasion (OR 2.615; 95% CI 1.203-5.684, P = 0.015) was independently associated with misdiagnosis of the depth of EGC and 0-III type lesions (OR 31.495; 95% CI 2.083-476.256, P = 0.013) were an independent risk factor for over-diagnosis of invasion depth by EUS. However, EUS was only suitable for lesions within absolute indications for endoscopic resection. Conclusions: The overall accuracy of EUS in diagnosing invasion depth of EGC was relatively low. Thus, EUS is not necessary routinely for determining the therapeutic strategy for EGC.

Effect of administration of a proton pump inhibitor for ulcerative differentiated early gastric cancer prior to endoscopic submucosal dissection

OBJECTIVES

Predicting the depth of invasion for early gastric cancer (EGC) with ulcerative findings may prove difficult. The purpose of this study was to evaluate the extent of ulcer healing and clinical effectiveness of proton pump inhibitor (PPI) administration for ulcerative differentiated EGC prior to endoscopic submucosal dissection (ESD).

METHODS

A total of 214 patients with ulcerative differentiated EGC who were treated by ESD were retrospectively analyzed, among whom 129 had been administered a PPI prior to ESD and 85 patients had not. The degree of ulcer healing following administration of PPI and the clinical outcome were compared between the PPI and non-PPI groups.

RESULTS

The ulcer healing rate (80.7 vs. 30.6%, P < 0.001) and complete ulcer healing rate (45.0 vs. 16.5%, P < 0.001) differed significantly between the PPI and non-PPI groups, respectively. The ESD procedure time was 86.0 (73) vs. 100.0 (68.0) min (P = 0.044). However, no significant differences were observed in the en bloc resection rate, complete resection rate, or the incidence of adverse events. Based on the multivariate regression analysis, PPI administration [odds ratio (OR) = 4.909, 95% confidence interval (CI) 2.459-9.799, P < 0.001] and mucosal cancer (OR = 9.336, 95% CI 2.693-32.362, P < 0.001) were independent predictive factors for complete ulcer healing. Partial or complete ulcer healing following PPI administration had a better predictive ability than EUS for mucosal invasion (kappa value 0.553 vs. 0.101; Youden index 0.553 vs. 0.132).

CONCLUSIONS

Partial or complete ulcer healing following PPI administration predicts differentiated EGC that is confined to the mucosa.

Updates in artificial intelligence in gastroenterology endoscopy in 2020

PURPOSE OF REVIEW

Artificial intelligence is becoming rapidly integrated into modern technology including medicine. Artificial intelligence has a wide range of potential in gastroenterology, particularly with endoscopy, given the required analysis of large datasets of images. The aim of this review is to summarize the advances of artificial intelligence in gastroenterology (GI) endoscopy over the past year.

RECENT FINDINGS

Computer-aided detection (CADe) systems during real-time colonoscopy have resulted in increased adenoma detection rate with no significant increase in procedure times. Deep learning techniques have been utilized to accurately assess bowel preparation quality, which would impact surveillance colonoscopy recommendations. For the upper GI tract, CADe systems have been developed to aid in improving the diagnosis of Barrett's neoplasia during real-time endoscopy. Artificial intelligence-assisted real-time endoscopy has been shown to reduce blind spots during EGD.

SUMMARY

The application of artificial intelligence in gastroenterology endoscopy remains promising. Advances over the past year include improved detection of GI neoplasia during endoscopy and characterization of lesions. Further research including randomized, multicenter trials are needed to further evaluate the use of artificial intelligence for real-time endoscopy.

Endoscopic diagnosis and treatment of gastric dysplasia and early cancer: Current evidence and what the future may hold

Gastric cancer accounts for a significant proportion of worldwide cancer-related morbidity and mortality. The well documented precancerous cascade provides an opportunity for clinicians to detect and treat gastric cancers at an endoscopically curable stage. In high prevalence regions such as Japan and Korea, this has led to the implementation of population screening programs. However, guidelines remain ambiguous in lower prevalence regions. In recent years, there have been many advances in the endoscopic diagnosis and treatment of early gastric cancer and precancerous lesions. More advanced endoscopic imaging has led to improved detection and characterization of gastric lesions as well as superior accuracy for delineation of margins prior to resection. In addition, promising early data on artificial intelligence in gastroscopy suggests a future role for this technology in maximizing the yield of advanced endoscopic imaging. Data on endoscopic resection (ER) are particularly robust in Japan and Korea, with high rates of curative ER and markedly reduced procedural morbidity. However, there is a shortage of data in other regions to support the applicability of protocols from these high prevalence countries. Future advances in endoscopic therapeutics will likely lead to further expansion of the current indications for ER, as both technology and proceduralist expertise continue to grow.

Clinical TNM staging for esophageal, gastric, and colorectal cancers in the era of neoadjuvant therapy: A systematic review of the literature

AIM

Clinical staging is vital for selecting appropriate candidates and designing neoadjuvant treatment strategies for advanced tumors. The aim of this review was to evaluate diagnostic abilities of clinical TNM staging for gastrointestinal, gastrointestinal cancers.

METHODS

We conducted a systematic review of recent publications to evaluate the accuracy of diagnostic modalities on gastrointestinal cancers. A systematic literature search was performed in PubMed/MEDLINE using the keywords "TNM staging," "T4 staging," "distant metastases," "esophageal cancer," "gastric cancer," and "colorectal cancer," and the search terms used in Cochrane Reviews between January 2005 to July 2020. Articles focusing on preoperative diagnosis of: (a) depth of invasion; (b) lymph node metastases; and (c) distant metastases were selected.

RESULTS

After a full-text search, a final set of 55 studies (17 esophageal cancer studies, 26 gastric cancer studies, and 12 colorectal cancer studies) were used to evaluate the accuracy of clinical TNM staging. Positron emission tomography-computed tomography (PET-CT) and/or magnetic resonance imaging (MRI) were the best modalities to assess distant metastases. Fat and fiber mode of CT may be useful for T4 staging of esophageal cancer, CT was a partially reliable modality for lymph node staging in gastric cancer, and CT combined with MRI was the most reliable modality for liver metastases from colorectal cancer.

CONCLUSION

The most reliable diagnostic modality differed among gastrointestinal cancers depending on the type of cancer. Therefore, we propose diagnostic algorithms for clinical staging for each type of cancer.

Usefulness of artificial intelligence in gastric neoplasms

Recently, studies in many medical fields have reported that image analysis based on artificial intelligence (AI) can be used to analyze structures or features that are difficult to identify with human eyes. To diagnose early gastric cancer, related efforts such as narrow-band imaging technology are on-going. However, diagnosis is often difficult. Therefore, a diagnostic method based on AI for endoscopic imaging was developed and its effectiveness was confirmed in many studies. The gastric cancer diagnostic program based on AI showed relatively high diagnostic accuracy and could differentially diagnose non-neoplastic lesions including benign gastric ulcers and dysplasia. An AI system has also been developed that helps to predict the invasion depth of gastric cancer through endoscopic images and observe the stomach during endoscopy without blind spots. Therefore, if AI is used in the field of endoscopy, it is expected to aid in the diagnosis of gastric neoplasms and determine the application of endoscopic therapy by predicting the invasion depth.

[Clinical Practice Guideline for Endoscopic Resection of Early Gastrointestinal Cancer]

Although surgery was the standard treatment for early gastrointestinal cancers, endoscopic resection is now a standard treatment for early gastrointestinal cancers without regional lymph node metastasis. High-definition white light endoscopy, chromoendoscopy, and image-enhanced endoscopy such as narrow band imaging are performed to assess the edge and depth of early gastrointestinal cancers for delineation of resection boundaries and prediction of the possibility of lymph node metastasis before the decision of endoscopic resection. Endoscopic mucosal resection and/or endoscopic submucosal dissection can be performed to remove early gastrointestinal cancers completely by en bloc fashion. Histopathological evaluation should be carefully made to investigate the presence of risk factors for lymph node metastasis such as depth of cancer invasion and lymphovascular invasion. Additional treatment such as radical surgery with regional lymphadenectomy should be considered if the endoscopically resected specimen shows risk factors for lymph node metastasis. This is the first Korean clinical practice guideline for endoscopic resection of early gastrointestinal cancer. This guideline was developed by using mainly de novo methods and encompasses endoscopic management of superficial esophageal squamous cell carcinoma, early gastric cancer, and early colorectal cancer. This guideline will be revised as new data on early gastrointestinal cancer are collected.

Application of artificial intelligence-driven endoscopic screening and diagnosis of gastric cancer

The landscape of gastrointestinal endoscopy continues to evolve as new technologies and techniques become available. The advent of image-enhanced and magnifying endoscopies has highlighted the step toward perfecting endoscopic screening and diagnosis of gastric lesions. Simultaneously, with the development of convolutional neural network, artificial intelligence (AI) has made unprecedented breakthroughs in medical imaging, including the ongoing trials of computer-aided detection of colorectal polyps and gastrointestinal bleeding. In the past demi-decade, applications of AI systems in gastric cancer have also emerged. With AI’s efficient computational power and learning capacities, endoscopists can improve their diagnostic accuracies and avoid the missing or mischaracterization of gastric neoplastic changes. So far, several AI systems that incorporated both traditional and novel endoscopy technologies have been developed for various purposes, with most systems achieving an accuracy of more than 80%. However, their feasibility, effectiveness, and safety in clinical practice remain to be seen as there have been no clinical trials yet. Nonetheless, AI-assisted endoscopies shed light on more accurate and sensitive ways for early detection, treatment guidance and prognosis prediction of gastric lesions. This review summarizes the current status of various AI applications in gastric cancer and pinpoints directions for future research and clinical practice implementation from a clinical perspective.

Examination of Endoscopic Ultrasonographic Diagnosis for the Depth of Early Gastric Cancer

Background

Endoscopic ultrasonography (EUS) is one of the helpful tools to diagnose depth of early gastric cancer (EGC). In this study, we examined efficiencies of EUS for EGC such as overall accuracy, risk factors of over/under-staging, and accuracies of each invasive distance.

Methods

A total of 403 EGC lesions that could be investigated by EUS during pre-operation and histological diagnosis after endoscopic submucosal dissection (ESD) or surgery were enrolled in this study. For the 403 cases, we analyzed the accuracies of depth by conventional endoscopy (CE) and EUS retrospectively. We evaluated the clinical survey items of CE and EUS which will be described later to compare the differences between “accuracy group” and “over-staging group”, and between “accuracy group” and “under-staging group”, retrospectively. Additionally, 78 EGC lesions which were confined to the submucosa and for which it was possible to measure accurate invasive distance from the muscularis mucosae were examined for the relationship between preoperative diagnosis of depth by CE and EUS and invasive distance retrospectively.

Results

The overall accuracies of both CE and EUS in predicting EGC invasion depth were 87.3%. For CE staging, histological classification was the factor which influenced over-staging. Gastric regions and tumor area were the factors which influenced under-staging of CE. For EUS staging, tumor area was the factor which influenced over-staging, and gastric regions were the factors which influenced under-staging. Both CE and EUS were not sufficient for predicting the lesions confined to < 500 µm from the muscularis mucosae because the accuracies of both in predicting depth were less than 50%. However, EUS has a higher accuracy than CE for the lesions confined to 500 - 2,000 µm.

Conclusions

The overall accuracies of both CE and EUS in predicting EGC invasion depth were equal, but the contributing factors for over/under-staging were different. Both CE and EUS are not sufficient at present to predict the lesions confined to < 500 µm from the muscularis mucosae. However, the accuracy of EUS in predicting them may increase if high-performance EUS systems are developed in the future.

A Novel Model Based on Deep Convolutional Neural Network Improves Diagnostic Accuracy of Intramucosal Gastric Cancer (With Video)

Background and Aims

Prediction of intramucosal gastric cancer (GC) is a big challenge. It is not clear whether artificial intelligence could assist endoscopists in the diagnosis.

Methods

A deep convolutional neural networks (DCNN) model was developed via retrospectively collected 3407 endoscopic images from 666 gastric cancer patients from two Endoscopy Centers (training dataset). The DCNN model’s performance was tested with 228 images from 62 independent patients (testing dataset). The endoscopists evaluated the image and video testing dataset with or without the DCNN model’s assistance, respectively. Endoscopists’ diagnostic performance was compared with or without the DCNN model’s assistance and investigated the effects of assistance using correlations and linear regression analyses.

Results

The DCNN model discriminated intramucosal GC from advanced GC with an AUC of 0.942 (95% CI, 0.915–0.970), a sensitivity of 90.5% (95% CI, 84.1%–95.4%), and a specificity of 85.3% (95% CI, 77.1%–90.9%) in the testing dataset. The diagnostic performance of novice endoscopists was comparable to those of expert endoscopists with the DCNN model’s assistance (accuracy: 84.6% vs. 85.5%, sensitivity: 85.7% vs. 87.4%, specificity: 83.3% vs. 83.0%). The mean pairwise kappa value of endoscopists was increased significantly with the DCNN model’s assistance (0.430–0.629 vs. 0.660–0.861). The diagnostic duration reduced considerably with the assistance of the DCNN model from 4.35s to 3.01s. The correlation between the perseverance of effort and diagnostic accuracy of endoscopists was diminished using the DCNN model (r: 0.470 vs. 0.076).

Conclusions

An AI-assisted system was established and found useful for novice endoscopists to achieve comparable diagnostic performance with experts.

A review on recent advancements in diagnosis and classification of cancers using artificial intelligence

Artificial intelligence has illustrated drastic changes in radiology and medical imaging techniques which in turn led to tremendous changes in screening patterns. In particular, advancements in these techniques led to the development of computer aided detection (CAD) strategy. These approaches provided highly accurate diagnostic reports which served as a "second-opinion" to the radiologists. However, with significant advancements in artificial intelligence strategy, the diagnostic and classifying capabilities of CAD system are meeting the levels of radiologists and clinicians. Thus, it shifts the CAD system from second opinion approach to a high utility tool. This article reviews the strategies and algorithms developed using artificial intelligence for the foremost cancer diagnosis and classification which overcomes the challenges in the traditional method. In addition, the possible direction of AI in medical aspects is also discussed in this study.

Atypical Scar Patterns after Gastric Endoscopic Submucosal Dissection

Background/Aims: Endoscopic submucosal dissection (ESD) for gastric neoplasms is a widely performed procedure. Local recurrence is rare, but various post-ESD scars are encountered during follow-up endoscopy. Therefore, we investigated atypical scar patterns and evaluated the associated factors.Materials and Methods: Clinicopathologic and endoscopic reviews of gastric neoplasms treated with ESD from January 2009 to December 2015 were conducted. Atypical scar patterns were classified as irregular erythema, nodularity, or mucosal defect.Results: A total of 264 patients with 274 gastric neoplasms, including 201 adenomas and 73 early gastric cancers, were enrolled. The key endoscopic findings at the resection scar were defined on the basis of gross morphology as follows: irregular erythema, mucosal defect (erosion or ulcer), and nodularity. An irregular erythema scar pattern was associated with male sex, a nodularity scar pattern with smoking, and a mucosal defect scar pattern with infra-angle location (angle and antrum) and cancer. An irregular erythema with nodularity scar pattern was also associated with male sex. An irregular erythema with nodularity and mucosal defect scar pattern was associated with liver disease and chronic kidney disease.Conclusions: The atypical scar patterns after gastric ESD are associated with various clinicopathologic factors.

Clinical feasibility and oncologic safety of primary endoscopic submucosal dissection for clinical submucosal invasive early gastric cancer

PURPOSE

This study aimed to evaluate the clinical feasibility and oncologic safety of primary endoscopic submucosal dissection for cT1bN0M0 gastric cancer by identifying patients who had undergone curative resection after endoscopic submucosal dissection and patients who had undergone primary surgery but could be treated with endoscopic submucosal dissection.

METHODS

A single-center retrospective cohort study was conducted on patients with cT1bN0M0 gastric cancer (size: ≤ 30 mm) and differentiated histology from January 2007 to May 2017. Submucosal invasion was evaluated using conventional endoscopy and endoscopic ultrasonography. Patients were divided into the primary endoscopic submucosal dissection and primary surgery groups according to initial treatment.

RESULTS

Curative resection rate was 65.0% (91/140) in the primary endoscopic submucosal dissection group. Of patients in the primary surgery group, 49.1% (82/167) were considered eligible for endoscopic submucosal dissection. No differences in en bloc resection rate were observed between pT1a and pT1b gastric cancers after endoscopic submucosal dissection (97.8 vs. 97.9%). However, the negative horizontal margin rate was higher in pT1a gastric cancer than pT1b gastric cancer (98.9 vs. 91.3%). There was no difference in overall survival among the curative treatment groups (log rank P = 0.310).

CONCLUSION

In this study, 173 (56.4%) out of 307 patients with cT1bN0M0 gastric cancer could avoid unnecessary surgery if primary endoscopic submucosal dissection was used. Based on these findings, it is necessary to add an appropriate indication for endoscopic submucosal dissection in patients with cT1b gastric cancer. Further management could be determined based on pathologic findings after primary endoscopic submucosal dissection.

Endoscopic Submucosal Dissection versus Surgery for Undifferentiated-Type Early Gastric Cancer: A Systematic Review and Meta-Analysis

Background/Aims

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.

Methods

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.

Results

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.

Conclusions

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.

Quality in diagnostic upper gastrointestinal endoscopy for the detection and surveillance of gastric cancer precursor lesions: Position paper of AEG, SEED and SEAP

This position paper, sponsored by the Asociación Española de Gastroenterología [Spanish Association of Gastroenterology], the Sociedad Española de Endoscopia Digestiva [Spanish Gastrointestinal Endoscopy Society] and the Sociedad Española de Anatomía Patológica [Spanish Anatomical Pathology Society], aims to establish recommendations for performing an high quality upper gastrointestinal endoscopy for the screening of gastric cancer precursor lesions (GCPL) in low-incidence populations, such as the Spanish population. To establish the quality of the evidence and the levels of recommendation, we used the methodology based on the GRADE system (Grading of Recommendations Assessment, Development and Evaluation). We obtained a consensus among experts using a Delphi method. The document evaluates different measures to improve the quality of upper gastrointestinal endoscopy in this setting and makes recommendations on how to evaluate and treat the identified lesions. We recommend that upper gastrointestinal endoscopy for surveillance of GCPL should be performed by endoscopists with adequate training, administering oral premedication and use of sedation. To improve the identification of GCPL, we recommend the use of high definition endoscopes and conventional or digital chromoendoscopy and, for biopsies, NBI should be used to target the most suspicious areas of intestinal metaplasia. Regarding the evaluation of visible lesions, the risk of submucosal invasion should be evaluated with magnifying endoscopes and endoscopic ultrasound should be reserved for those with suspected deep invasion. In lesions amenable to endoscopic resection, submucosal endoscopic dissection is considered the technique of choice.

Mixed histology poses a greater risk for noncurative endoscopic resection in early gastric cancers regardless of the predominant histologic types

OBJECTIVES

Clinicopathologic characteristics and treatment outcomes of mixed-histological-type (MT) early gastric cancers (EGCs) treated with endoscopic submucosal dissection (ESD) have not been sufficiently elucidated. We aimed to clarify them in comparison with pure-histological-type EGCs.

METHODS

We used 3022 consecutive EGCs in 2281 patients treated with ESD from our prospectively maintained database. Cases were stratified into four groups according to the final diagnosis of the resected specimen are as follows: 2780 pure differentiated-type (DT), 127 DT-predominant MT (D-MT), 87 pure undifferentiated-type (UDT), and 28 UDT-predominant MT (U-MT). Clinicopathologic characteristics and treatment outcome were compared between pure DT and D-MT, and between pure UDT and U-MT separately. Risk factors for deep submucosal invasion, lymphovascular invasion, and a final diagnosis of MT were identified using multivariate analysis.

RESULTS

Both D-MT (41.7 vs. 92.0%; P < 0.0001) and U-MT (35.7 vs. 75.9%; P = 0.0002) showed a significantly lower curative resection rate than their pure histologic counterparts. Multivariate analysis revealed that MT was an independent risk factor for deep submucosal (OR 6.55; 95% CI, 4.18-10.14) and lymphovascular (OR 4.74; 95% CI, 2.72-8.29) invasion. Preoperative biopsy results that did not show well-differentiated tubular adenocarcinoma (OR 28.2; 95% CI, 18.9-42.9) were an independent risk factor for a final diagnosis of MT.

CONCLUSIONS

MT poses a greater risk for noncurative resection regardless of the predominant histologic types, reflecting more aggressive malignant potential. Although a biopsy examination rarely shows MT, clinicians should consider the possibility of MT when a biopsy examination does not show well-differentiated tubular adenocarcinoma.

Hierarchical Analysis of Factors Associated with T Staging of Gastric Cancer by Endoscopic Ultrasound

BACKGROUND

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.

METHODS

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.

RESULTS

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.

CONCLUSION

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.

Clinico-pathologic determinants of non-e-curative outcome following en-bloc endoscopic submucosal dissection in patients with early gastric neoplasia

Background

Endoscopic submucosal dissection (ESD) is gaining enormous popularity in the treatment of early gastric cancers (EGCs) in many institutions across the world. However, appropriate selection of candidates for endoscopic resection is crucial to sufficiently mitigate non-e-curative (NEC) resection. This study aims at identifying the various clinico-pathologic factors that independently predict the NEC outcome and depth of submucosal invasion following ESD procedure in patients with EGC.

Methods

Multiple logistic regression analysis was applied to investigate factors that independently predict both non-curability phenomenon and the level of submucosal invasion in patients with early gastric neoplasia. Statistical Packages for the Social Sciences version 23 was used for analysis.

Results

A total of 153 patients (162 EGC lesions) underwent en-bloc ESD after which the rate of complete resection and non-e-curative outcome were 95% and 22.2%, correspondingly. Multivariate analysis depicted that tumor location in the upper two third of stomach (odds ratio [OR], 5.46; 95% confidence interval [95% CI], 1.65–18.12; p = 0.006), tumor size > 2 cm (OR, 7.63; 95% CI, 2.29–25.42; p = 0.001), histologically undifferentiated tumor (OR, 15.54; 95% CI, 1.65–146.22; p = 0.001), and tumors with 0-IIa/0-IIc or their mixed variants with predominant 0-IIa/0-IIc (OR, 9.77; 95% CI, 1.23–77.65; p = 0.031) were all independent predictors of NEC resection for early gastric tumors. Additionally, location in the upper two third of the stomach (OR, 8.88; 95% CI, 2.90–27.17; p < 0.001), ulcerated lesions (OR, 3.70; 95% CI, 1.15–11.90; p = 0.028), lesions with > 2 cm (OR, 2.94; 95% CI, 1.08–8.02; p = 0.036) and those with poor differentiation (OR, 6.51; 95% CI, 2.23–18.98; p = 0.001) were found to have significant association with submucosal invasion.

Conclusions

Tumors located in the upper two third of the stomach having a larger size (> 2 cm), poor histo-differentiation and a gross type of 0-IIa/0-IIc or their mixed variants with predominant 0-IIa/0-IIc were significantly associated with a risk of NEC after ESD procedure. Thus, early gastric tumors displaying these features need to be handled carefully during endoscopic resection. Our findings may shed light on the pre-procedural detection of clinicopathologic factors that determine non-e-curability in patients with EGC.