目的·基于超声剪切波弹性成像（shear wave elastography，SWE）方法定量评估胰腺组织软硬度，改良现有术后临床胰瘘（clinically relevant post-operative pancreatic fistula，CR-POPF）的预测模型。方法·前瞻性纳入拟诊为胰腺肿瘤并计划于复旦大学附属中山医院接受胰腺外科手术的患者，收集患者相关临床信息。在手术前1周内使用声触诊组织弹性成像与量化技术（virtual touch tissue imaging and quantification technology，VTIQ）测量胰腺病灶及门静脉前方胰体部正常腺体实质弹性定量剪切波速度（shear wave velocity，SWV）值。术中外科医师通过对胰腺组织的直接触诊定性评估组织质地，并将胰腺触诊结果分为质地软或中等-硬。术后3周随访期内，根据2016年国际胰瘘研究组织的标准对CR-POPF进行诊断。通过单因素和多因素Logistic回归分析围手术期CR-POPF的危险因素，并建立预测模型。通过受试者工作特征曲线（receiver-operating characteristic curve，ROC曲线）与决策曲线分析（decision curve analysis，DCA）比较改良模型与现有临床预测模型的诊断效能及临床效益。结果·2021年9月到2022年3月，共有100名患者被纳入研究，包括接受胰十二指肠切除术者33例（33.0%）及胰体尾切除术者67例（67.0%）。通过术后3周的随访发现35例（35.0%）患者并发CR-POPF。通过Logistic回归分析发现，门静脉前方胰体部实质的低SWV值［lgOR=-2.934（95%CI -4.387~-1.479），P=0.000］，较窄的主胰管内径（≤3 mm）［lgOR=0.805（95%CI 0.274~1.335），P=0.003］为胰腺切除术后并发CR-POPF的独立危险因素。基于SWE定量参数改良的风险评估模型预测CR-POPF的ROC曲线下面积达0.842，敏感度、特异度、阳性预测值、阴性预测值及似然比分别为85.7%、64.6%、70.5%、81.8%及2.422。与现有临床预测模型［胰瘘风险评分（fistula risk score，FRS）和改良胰瘘风险评分（alternative fistula risk score，a-FRS）］相比，DCA结果显示基于SWE的改良模型具有更好的临床效益。结论·结合SWE定量参数与CR-POPF临床关键危险因素的预测模型具有较高的诊断效能和临床效益，且能在术前对CR-POPF风险进行无创、定量、客观的评估。

Objective ·To modify previous clinically relevant post-operative pancreatic fistula (CR-POPF) risk evaluation models with quantitative evaluation of pancreatic tissue stiffness by ultrasound shear wave elastography (SWE). Methods ·In this prospective study, the patients who were diagnosed as having pancreatic tumors and scheduled to undergo pancreatectomy at Zhongshan Hospital, Fudan University were initially enrolled, whose clinical information was collected. Virtual touch tissue imaging and quantification technology (VTIQ) assessment was applied to the patients within one week before the surgery to measure the shear wave velocity (SWV) of pancreatic lesions and the normal parenchyma of pancreatic body in the superficial layer of the portal vein. During the surgery, the surgeons qualitatively evaluated the stiffness of pancreases via direct palpation and divided them into soft pancreases and medium-hard pancreases. During the 3-week follow-up period after pancreatectomy, CR-POPF was diagnosed according to 2016 International Study Group of Pancreatic Fistula (ISGPF) standard. Peri-operative risk factors of CR-POPF were analyzed by univariate and multivariate Logistic regression to build the prediction model. Evaluation and comparison of diagnostic efficacy and clinical benefits among different models were then performed via receiver operating characteristic (ROC) curve and decision curve analysis (DCA). Results ·From September 2021 to March 2022, 100 patients were enrolled in this study, including 33 patients (33.0%) who received pancreaticoduodenectomy (PD) and 67 patients (67.0%) who received distal pancreatectomy. CR-POPF was diagnosed in 35 patients (35.0%) during the 3-week post-pancreatectomy follow-up. Multivariate Logistic regression analysis revealed that the SWV value of the body part of pancreatic parenchyma in the superficial layer of the portal vein [lgOR=-2.934 (95%CI -4.387?-1.479), P=0.000] and the presence of a non-dilated main pancreatic duct (≤3 mm) [lgOR=0.805 (95%CI 0.274?1.335), P=0.003] were independent risk factors that significantly correlated with the occurrence of CR-POPF after pancreatectomy. The modified model based on the SWE parameter achieved the area under the ROC curve of 0.842, with the sensitivity, the specificity, the positive predictive value, the negative predictive value and the likelihood ratio of 85.7%, 64.6%, 70.5%, 81.8% and 2.422 in predicting CR-POPF. DCA revealed a better clinical benefit of the modified model compared to the previous prediction models [fistula risk score (FRS) and alternative fistula risk score (a-FRS)]. Conclusion ·The modified model based on the SWE parameter and identified clinical risk factors can make non-invasive, quantitative and objective evaluation of CR-POPF risk before pancreatectomy, and provide sufficient diagnostic efficacy and clinical benefits.