Liu Jinjin, Gu Qiuping, Li Zhonghua
Department of Oncology, Ganzhou People's Hospital, Ganzhou, Jiangxi, China.
Department of Gastroenterology, Ganzhou People's Hospital, Ganzhou, Jiangxi, China.
J Gastrointest Surg. 2025 Aug;29(8):102114. doi: 10.1016/j.gassur.2025.102114. Epub 2025 Jun 9.
To identify the risk factors for reflux esophagitis (RE) after proximal gastrectomy (PG) in patients with proximal gastric cancer and the construct of a nomogram prediction model.
This retrospective study included 421 patients with proximal gastric cancer who underwent PG at our hospital between January 2022 and April 2024. Patients were randomly grouped into a modeling group (295 cases) and a validation group (126 cases). The occurrence of postoperative RE was the primary outcome. Logistic regression was used to analyze potential risk factors. R software was used to construct the nomogram model. The receiver operating characteristic (ROC) curve was used to evaluate discrimination, the Hosmer-Lemeshow (H-L) test for calibration, and decision curve analysis (DCA) to evaluate its clinical utility.
Among the 421 patients, 88 cases (20.90%) developed RE, while among 295 patients in the modeling group, 62 cases (22.02%) developed RE. Multivariate logistic regression identified BMI, surgical time, intraoperative hemorrhage, length of esophagectomy, residual stomach volume, and stapler size were risk factors for postoperative RE (P<0.05). The AUC of the nomogram in the modeling group was 0.883, and in the validation group it was 0.851. H-L test showed good calibration in the modeling group (χ²=7.305, P=0.728) and the validation group (χ²=6.896, P=0.701). The DCA curve indicated that the nomogram offered clinical benefit when the threshold probability was between 0.08 and 0.97.
BMI, surgical time, intraoperative blood loss, length of esophagectomy, residual stomach volume, and stapler size are significant factors for postoperative RE. The nomogram model constructed based on these factors can effectively predict the risk of postoperative RE in patients with proximal gastric cancer, potentially aiding in personalized risk stratification and preventive strategies.
确定近端胃癌患者近端胃切除术后反流性食管炎（RE）的危险因素，并构建列线图预测模型。
本回顾性研究纳入了2022年1月至2024年4月在我院接受近端胃切除术的421例近端胃癌患者。患者被随机分为建模组（295例）和验证组（126例）。术后RE的发生为主要结局。采用逻辑回归分析潜在危险因素。使用R软件构建列线图模型。采用受试者工作特征（ROC）曲线评估辨别力，采用Hosmer-Lemeshow（H-L）检验进行校准，并采用决策曲线分析（DCA）评估其临床实用性。
421例患者中，88例（20.90%）发生RE，建模组295例患者中，62例（22.02%）发生RE。多因素逻辑回归分析确定体重指数、手术时间、术中出血、食管切除长度、残胃容积和吻合器型号是术后RE的危险因素（P<0.05）。建模组列线图的AUC为0.883，验证组为0.851。H-L检验显示建模组（χ²=7.305，P=0.728）和验证组（χ²=6.896，P=0.701）校准良好。DCA曲线表明，当阈值概率在0.08至0.97之间时，列线图具有临床益处。
体重指数、手术时间、术中失血量、食管切除长度、残胃容积和吻合器型号是术后RE的重要因素。基于这些因素构建的列线图模型可以有效预测近端胃癌患者术后RE的风险，可能有助于个性化风险分层和预防策略。
