Toward Automatic Detection of Gastric Lesion for Upper Gastrointestinal Endoscopy with Neural Network

In this paper, we compared the performance of several neural networks in the classification of early gastric cancer (EGC) images and proposed a method of converting the output value of the network into a calorific value to locate the lesion. The algorithm was improved using transfer learning and fine-tuning principles. The test set accuracy rate reached 0.72, sensitivity reached 0.67, specificity reached 0.77, and precision rate reached 0.78. The experimental results show the potential to meet clinical demands for automatic detection of gastric lesion.

A Gastric Metastatic Lesion That Resembled Early-Stage Gastric Cancer on Endoscopy during Treatment for Recurrent Breast Cancer: A Case Report

Breast cancer metastasis to the gastrointestinal tract is relatively rare. Patients with such disease often develop gastrointestinal symptoms, but it is sometimes asymptomatic. Endoscopic findings of gastric metastasis from breast cancer markedly vary from benign to malignant, and even in suspected malignant cases, it is often difficult to differentiate between primary and metastatic disease. We experienced a case in which an endoscopic examination performed during the treatment for metastatic breast cancer resembled an early-stage gastric cancer. A 71-year-old woman underwent curative surgery for right breast cancer 16 years previously. She underwent endoscopic submucosal dissection for early-stage gastric cancer 5 years ago. Two years ago, she developed metastatic disease in the lungs and mediastinal lymph nodes, and endocrine therapy was administered. At the same time, a follow-up endoscopy revealed a new elevated lesion, suspected to be an early-stage gastric cancer. However, histological diagnosis of the biopsy was metastasis of breast cancer. One and a half years later, a follow-up endoscopy revealed a gastric lesion that had reduced in size. She is still alive, having received a variety of systemic treatments. Patients with metastatic breast cancer are experiencing prolonged survival. Thus, follow-up endoscopy should be considered after the diagnosis of gastrointestinal metastasis considering the risk of lethal conditions, such as gastrointestinal bleeding and perforation. Our case serves as a reminder to clinicians how difficult it is to determine whether a gastric lesion is primary or metastatic based on endoscopic findings and the importance of communication with endoscopists and pathologists.

Gastric Eosinophilic Granuloma Related to Anisakiasis Resected by Laparoscopic and Endoscopic Cooperative Surgery

Background: Anisakiasis-related gastric eosinophilic granuloma is rare. Case Report: Herein, we report a patient with anisakiasis-related gastric eosinophilic granuloma who was treated with laparoscopic and endoscopic cooperative surgery (LECS). A 59-year-old woman was presented to our hospital for further examination of a gastric lesion that was initially diagnosed by a local medical doctor. Esophagogastroduodenoscopy showed a submucosal tumor-like lesion in the lower body of the stomach. Endoscopic ultrasonography showed a heterogeneous hypoechoic submucosal mass lesion in the submucosal layer measuring 10 mm, without evidence of deep involvement. Under a clinical diagnosis of gastrointestinal stromal tumor, the patient underwent LECS. Gross appearance of the resected specimen revealed a 1.5×1.0 cm submucosal tumor-like lesion. Microscopic examination revealed necrosed insects consistent with the characteristics of gastric anisakiasis, around which prominent eosinophilic infiltration and granulomas were observed. This prompted a diagnosis of gastric eosinophilic granuloma related to anisakiasis. Conclusion: To the best of our knowledge, this is the second case of gastric eosinophilic granuloma related to anisakiasis resected by LECS in the English medical literature. LECS might be a useful procedure for minimally invasive therapeutic diagnosis.