为确定可预测胰腺导管腺癌（PDAC）隐匿转移（OM）的放射学特征和临床生物标志物。本回顾性研究纳入了2018年1月至2021年12月间接受手术探查、放射学诊断可切除（R）或边缘可切除（BR）的PDAC患者。根据手术探查期间是否发现远处转移，将患者分为OM组和非OM组。进行单变量和多变量logistic回归分析，确定隐匿转移的放射学和临床预测因素。模型的表现根据区分度和校准来确定。共纳入502名患者（中位年龄64岁，四分位距57-70岁，294名男性），其中68例（13.5%）患者发现远处转移，其中45例仅有肝转移、19例仅有腹膜转移，4例肝脏和腹膜都有转移。OM组中的边缘增强和胰周脂肪纹理比非OM组更常见。肿瘤大小（p = 0.028）、肿瘤可切除性（p = 0.031）、边缘增强（p < 0.001）、胰周脂肪纹理（p < 0.001）和CA125水平（p = 0.021）根据多变量分析是隐匿转移的独立预测因素，这些特征的曲线下面积（AUCs）分别为0.703、0.594、0.638、0.655、0.631。结合模型的AUC最高达到0.823。边缘增强、胰周脂肪纹理、肿瘤大小、肿瘤可切除性和CA125水平是PDAC隐匿转移的危险因素。放射学和临床特征的结合模型可能有助于预测PDAC的OM。Copyright © 2023 The Association of University Radiologists. Published by Elsevier Inc. All rights reserved.

To identify the radiological features and clinical biomarkers that could predict the occult metastasis (OM) of pancreatic ductal adenocarcinoma (PDAC).This retrospective study included PDAC patients who were radiologically diagnosed resectable (R) or borderline resectable (BR) and underwent surgical exploration from January 2018 to December 2021. Depending on whether distant metastases were found during the exploration, patients were divided into OM and non-OM groups. Univariate and multivariable logistic regression analyses were performed to determine the radiological and clinical predictive factors for occult metastasis. Model performance was determined by discrimination and calibration.A total of 502 patients (median age, 64 years; interquartile range, 57-70 years; 294 men) were enrolled, among which 68 (13.5%) patients were found with distant metastases, with 45 liver-only, 19 peritoneal-only, four patients had both liver and peritoneal metastases. Rim enhancement and peripancreatic fat stranding were more frequent in the OM group than in the non-OM group. Tumor size (p = 0.028), tumor resectability (p = 0.031), rim enhancement (p < 0.001), peripancreatic fat stranding (p < 0.001) and level of CA125 (p = 0.021) were independent predictors of occult metastasis according to the multivariable analyses, and the areas under the curve (AUCs) of these characteristics were 0.703, 0.594, 0.638, 0.655, 0.631, respectively. The combined model showed the highest AUC of 0.823.Rim enhancement, peripancreatic fat stranding, tumor size, tumor resectability and level of CA125 are risk factors for OM of PDAC. The combined model of radiological and clinical features may help the preoperative prediction of OM in PDAC.Copyright © 2023 The Association of University Radiologists. Published by Elsevier Inc. All rights reserved.