Zhang Jia-Yi, Li Ding, Hu Guo-Jie
Institute of Integrated Medicine, Qingdao Medical College of Qingdao University, Qingdao University, Qingdao, Shandong, China.
Department of Traditional Chinese Medicine, The Affiliated Hospital of Qingdao University, Qingdao, Shandong, China.
Discov Oncol. 2025 May 8;16(1):688. doi: 10.1007/s12672-025-02453-y.
To construct a machine learning (ML) model to predict the progression of chronic atrophic gastritis (CAG) to gastric cancer (GC), given its precancerous significance.
Using medical records from the Affiliated Hospital of Qingdao University, common laboratory indicators were extracted. LASSO regression identified 10 core risk factors, which were further analyzed using binary logistic regression to develop a nomogram model in R. The model's performance was evaluated using receiver operating characteristic (ROC) curves, the concordance index (C-index), calibration curves, and decision curve analysis (DCA).
The model showed excellent performance, with a C-index of 0.887. The key factors included sex, coagulation, blood cell indexes, and blood lipid levels. The ROC areas were 0.892 (quantitative) and 0.853 (qualitative), confirming model reliability.
A new nomogram model for assessing GC risk in CAG patients was successfully developed. However, due to data collection and time limitations, future studies should expand the sample size, perfect the validation process, and optimize the model to achieve more accurate risk prediction.
鉴于慢性萎缩性胃炎（CAG）的癌前意义，构建一个机器学习（ML）模型来预测其向胃癌（GC）的进展。
利用青岛大学附属医院的病历，提取常见实验室指标。LASSO回归确定了10个核心危险因素，进一步使用二元逻辑回归在R中开发列线图模型。使用受试者工作特征（ROC）曲线、一致性指数（C指数）、校准曲线和决策曲线分析（DCA）评估模型性能。
该模型表现出色，C指数为0.887。关键因素包括性别、凝血、血细胞指标和血脂水平。ROC曲线下面积分别为0.892（定量）和0.853（定性），证实了模型的可靠性。
成功开发了一种用于评估CAG患者GC风险的新列线图模型。然而，由于数据收集和时间限制，未来研究应扩大样本量，完善验证过程，并优化模型以实现更准确的风险预测。
