目的 分析胃肠道恶性肿瘤(GIC)患者术后腹腔感染(IAI)危险因素,构建列线图模型并进行验证。 方法 选取 2018 年 1 月至 2023 年 12 月江苏大学附属医院收治的 230 例 GIC 患者作为研究对象。 按照 7 ∶ 3 比例,将其随机分为建模组(161 例)和验证组(69 例)。 基于建模组数据,利用 Lasso-logistic 回归分析 GIC 术后 IAI 危险因素,构建列线图模型并进行验证。 结果 Lasso 回归筛选出 5 项非零系数指标,分别为糖尿病、美国麻醉医学协会(ASA)分级>Ⅱ级、联合脏器切除、手术时间、术 前控制营养状态(CONUT)评分。 多因素 Logistic 回归分析显示,ASA 分级>Ⅱ级、联合脏器切除、手术时间≥4. 5 h、术前 CONUT 评分≥5 分是 GIC 患者术后发生 IAI 的独立危险因素(均 P<0. 05)。 基于上述指标构建列线图模型,受试者操作特征 (ROC)曲线表明,该模型预测建模组、验证组患者发生 IAI 的曲线下面积为 0. 829(95%CI = 0. 759 ~ 0. 899)、0. 862(95%CI = 0. 725~ 0. 999)。 校准曲线表明,该模型在建模组、验证组中的预测概率与实际概率一致性较好。 决策曲线表明,该模型在建 模组、验证组中均具有较广的临床净收益。 结论 本研究开发的列线图能较好识别 GIC 术后 IAI 高风险患者。

Objective To analyze the risk factors for postoperative intra-abdominal infection IAI in gastrointestinal cancer GIC patients then construct and validate the nomogram model . Method A total of 230 GIC patients who admitted to our hospital from January 2018 to December 2023 were selected as the study group. According to a ratio of 7 to 3 patients were randomly divided into a modeling group of 161 cases and a validation group of 69 cases. Based on modeling group data Lasso-logistic regression was used to analyze the risk factors of postoperative IAI in GIC then construct and validate the nomogram model. Result Lasso regression screened five non - zero coefficient indicators which were diabetes ASA grade > grade Ⅱ combined organ resection operation time and preoperative CONUT score. Multivariate logistic regression analysis showed that ASA grade >Ⅱ combined organ resection surgery time≥4. 5 hours and preoperative CONUT score≥5 were independent risk factors for postoperative IAI in GIC patients all P<0. 05 . Based on the above indicators a nomogram model was constructed. The ROC curve showed that the AUC for predicting IAI in the modeling group and validation group patients were 0. 829 95% CI = 0. 759 - 0. 899 and 0. 862 95% CI = 0. 725 - 0. 999 . The calibration curve indicated that the model had good consistency between the predicted probability and the actual probability in the modeling group and validation group. The decision curve indicated that the model had a wide range of clinical net benefits in the modeling and validation groups. Conclusion The nomogram developed in this study can effectively identify high-risk postoperative IAI patients in GIC and can be used to guide clinical practice.