Maruyama Hiroki, Takahashi Kazuya, Kojima Kosuke, Nakajima Nao, Sato Hiroki, Mizuno Ken-Ichi, Sugitani Soichi, Terai Shuji
Division of Gastroenterology & Hepatology Graduate School of Medical and Dental Sciences, Niigata University Niigata Japan.
Division of Gastroenterology Koseiren Murakami General Hospital Niigata Japan.
JGH Open. 2025 Jun 29;9(7):e70203. doi: 10.1002/jgh3.70203. eCollection 2025 Jul.
Endoscopic submucosal dissection (ESD) is a minimally invasive treatment for early gastric cancer (GC); however, post-ESD bleeding remains a serious and unpredictable complication. This study aimed to develop machine-learning (ML) models to predict post-ESD bleeding and identify associated risk factors, ensuring accurate and interpretable predictions.
A retrospective, multicenter clinical database was constructed for patients who underwent ESD for early GC. An ML model was developed using patient characteristics and perioperative findings to predict bleeding within 28 days post-ESD. Its performance was compared with that of a logistic regression-based non-ML model. Feature importance analysis was performed to aid interpretation.
Among 1084 patients (median age: 75 years), post-ESD bleeding occurred in 63 (5.8%). The area under the curve of the ML model was better than that of the non-ML model (0.80 vs. 0.71, = 0.03). Furthermore, the ML model demonstrated a trend toward higher sensitivity compared with the non-ML model (0.74 vs. 0.58, = 0.58). When stratified by ML-estimated bleeding probability, observed bleeding rates were 2.3%, 8.8%, and 28.6% in the low- (< 50%), intermediate- (50%-80%), and high-probability (≥ 80%) groups, respectively. The odds of bleeding were significantly higher in the intermediate- (OR 4.03, = 0.03) and high-probability (OR 16.7, < 0.01) groups compared to the low-probability group. Anticoagulant use with atrial fibrillation emerged as a key predictor.
The ML model effectively rules out post-ESD bleeding and identifies clinically meaningful risk factors, supporting its use in personalized prophylactic strategies.
内镜黏膜下剥离术（ESD）是早期胃癌（GC）的一种微创治疗方法；然而，ESD术后出血仍然是一种严重且不可预测的并发症。本研究旨在开发机器学习（ML）模型来预测ESD术后出血并识别相关危险因素，以确保准确且可解释的预测。
为接受早期GC的ESD治疗的患者构建了一个回顾性多中心临床数据库。利用患者特征和围手术期结果开发了一个ML模型，以预测ESD术后28天内的出血情况。将其性能与基于逻辑回归的非ML模型进行比较。进行特征重要性分析以辅助解释。
在1084例患者（中位年龄：75岁）中，63例（5.8%）发生了ESD术后出血。ML模型的曲线下面积优于非ML模型（0.80对0.71，P = 0.03）。此外，与非ML模型相比，ML模型显示出更高敏感性的趋势（0.74对0.58，P = 0.58）。根据ML估计的出血概率分层时，低（<50%）、中（50%-80%）和高概率（≥80%）组的观察到的出血率分别为2.3%、8.8%和28.6%。与低概率组相比，中（OR 4.03，P = 0.03）和高概率（OR 16.7，P < 0.01）组的出血几率显著更高。房颤患者使用抗凝剂成为关键预测因素。
ML模型有效地排除了ESD术后出血，并识别出具有临床意义的危险因素，支持其在个性化预防策略中的应用。
