双能CT细胞外容积分数与肌少症联合模型用于胃切除术后吻合口漏术前风险分层
该回顾性研究纳入439例胃癌根治术患者,利用术前临床、双能CT及身体成分数据构建吻合口漏风险分层模型。最终模型包含贲门肿瘤位置、较高ECV和肌少症,表观AUC为0.853,Bootstrap校正AUC为0.843。模型将患者分为低、中、高风险组,高风险组占18.0%但涵盖66.7%的吻合口漏事件。
要点
- 贲门肿瘤、双能CT ECV升高和肌少症是吻合口漏的独立预测因素。
- 模型高风险组(≥10%)吻合口漏发生率22.78%,需术前重点评估。
- 模型仅需术前临床和双能CT数据,便于术前风险分层。
英文原文(摘要)· 点击展开
PURPOSE: To develop a purely preoperative risk stratification model for anastomotic leakage (AL) after radical gastrectomy using clinical, dual-energy CT (DECT), and body composition variables available before surgery.
METHODS: This retrospective study included 439 patients with gastric cancer who underwent radical gastrectomy. Preoperative clinical, DECT, and body composition data were collected. A domain-based predictor selection strategy was used, followed by Firth penalized logistic regression to account for the low event rate. Internal validation was performed using 1000 bootstrap resamples, with an additional full-pipeline bootstrap sensitivity analysis. Model performance was assessed by discrimination, calibration, decision curve analysis (DCA), and risk stratification.
RESULTS: AL occurred in 27 patients (6.15%). The final model included cardia tumor location (OR = 4.44), higher ECV (OR = 1.09 per 1%), and sarcopenia (OR = 3.66). The model showed good discrimination, with an apparent AUC of 0.853 and a bootstrap-corrected AUC of 0.843. Full-pipeline bootstrap sensitivity analysis yielded an optimism-corrected AUC of 0.833. Calibration was acceptable, with a Brier score of 0.050, and DCA suggested potential net benefit. Observed AL rates increased stepwise across low-risk (< 3%), intermediate-risk (3%-10%), and high-risk (≥ 10%) groups: 1.23%, 5.13%, and 22.78%, respectively. The high-risk group comprised 18.0% of the cohort but accounted for 66.7% of all AL events.
CONCLUSION: A concise model integrating tumor location, DECT-derived ECV, and sarcopenia showed favorable performance for preoperative AL risk stratification after gastrectomy. This internally validated model may help preoperative AL risk stratification, but multicenter external validation and impact analysis are required before implementation.
引用
- Ren T, You H, Liu X, et al. Preoperative risk stratification for anastomotic leakage after gastrectomy: a model based on dual‑energy CT extracellular volume fraction and sarcopenia. Abdom Radiol (NY). 2026. PMID: 42944910. doi:10.1007/s00261-026-05839-w
供执业医师工作参考,不替代临床判断;数值与分级以指南原文为准。