目的 构建并验证胃癌患者术后生存概率的列线图预测模型。 方法 回顾性分析 2017 年 1 月至 2021 年 5 月辽宁省 肿瘤医院 370 例直接接受根治性手术治疗的胃癌患者,采用分层抽样的方法,将数据按照 2 ∶ 1 比例分为训练队列(n = 246)和 验证队列(n = 124),以患者基本信息、预后营养指数(PNI)、病理特征、炎症因子及肿瘤标志物为自变量,以总生存(OS)期为 因变量,通过单因素和多因素 Cox 比例风险回归模型分析,确定了胃癌患者预后的独立危险因素,并构建了列线图模型。 与传 统 TNM 分期对比,通过受试者操作特征(ROC)曲线并计算曲线下面积(AUC)、校准曲线评估模型准确性与区分度,并结合决 策曲线分析(DCA)和 Kaplan-Meier 生存曲线进一步验证列线图的临床效果。 结果 Cox 比例风险回归模型分析,确定了胃癌 患者术后的独立预后因素,包括 PNI、淋巴结转移、浸润深度和糖类抗原 125。 通过列线图直观展示了数据分析结果,训练队列 和验证队列中 AUC 分析、校准曲线及 DCA 均显示模型对胃癌患者预后的预测效能显著优于传统 TNM 分期。 结论 基于营养 指数构建的预测模型可精准预测胃癌患者的预后,为临床个性化决策提供依据。

Objective This study aims to construct and validate a nomogram for predicting postoperative survival probability in patients with gastric cancer. Method A retrospective analysis was conducted on 370 patients diagnosed with gastric cancer who underwent radical surgery at Liaoning Cancer Hospital between January 2017 and May 2021. Stratified sampling method was used to divide the data into a training cohort n = 246 and a validation cohort n = 124 at a 2 ∶ 1 ratio. Clinical data including baseline characteristics prognostic nutritional index PNI pathological features inflammatory factors and tumor markers were collected as candidate variables. Univariate and multivariate Cox proportional hazards regression analyses were performed to identify independent risk factors associated with postoperative survival dependent variables . Subsequently a nomogram prediction model was developed based on these identified factors. The performance of the nomogram—including its accuracy and discriminative ability—was evaluated using the area under the receiver operating characteristic curve AUC calibration curves and compared against that of the traditional TNM staging system. Additionally decision curve analysis DCA and Kaplan-Meier survival curves were employed to further assess its clinical utility. Result Cox regression analyses identified four independent prognostic factors PNI lymph node metastasis depth of invasion and CA125 levels. These factors were incorporated into the nomogram for visual representation of survival predictions. The nomogram demonstrated significantly superior prognostic performance compared to traditional TNM staging this was evidenced by higher AUC values improved calibration through calibration plots and greater net clinical benefit as determined by DCA. Conclusion The nomogram incorporating nutritional indicators—including PNI—can accurately predict postoperative survival in patients with gastric cancer thereby providing a reliable tool for personalized clinical decision-making.