我们评估了源自胰腺导管腺癌 (PDAC) 的双能计算机断层扫描 (DECT) 参数在区分高级别肿瘤和低级别肿瘤以及预测患者总生存期 (OS) 方面的价值。数据从 169 名连续患者中回顾性收集2017年1月至2023年3月期间经病理证实的PDAC，术前接受第三代双源DECT增强双时相扫描。排除既往接受过治疗、其他恶性肿瘤、小肿瘤或扫描质量较差的患者。两名放射科医生评估了 3 项临床特征和 7 项放射学特征，并测量了 16 项 DECT 衍生参数。应用单变量和多变量分析来选择独立的预测因子。开发了预测模型和相应的列线图，并评估了曲线下面积 (AUC)、校准和临床适用性。使用Kaplan-Meier生存和Cox回归分析评估因素与OS之间的相关性。169名患者被随机分为训练组（n = 118）和验证组（n = 51），其中43例（36.4%）经病理证实的高级别 PDAC 分别为 19 例（37.3%）和 19 例（37.3%）。血管侵犯、静脉期标准化碘浓度和静脉期有效原子序数是组织学分级的独立预测因子。构建列线图来预测 PDAC 中高级别肿瘤的风险，训练组和验证组的 AUC 分别为 0.887 和 0.844。列线图表现出良好的校准，并且比两个队列中的单个参数更有益。病理学和 nomoscore 预测的高级别 PDAC 与较差的 OS 相关（均 p< 0.05）。结合 DECT 参数和放射学特征的列线图可以预测 PDAC 患者术前的组织学分级和 OS。问题 术前确定PDAC 的组织学分级对于指导治疗至关重要，但目前的方法是侵入性的且有限。结果 基于 DECT 的列线图结合了血管侵犯、标准化碘浓度和有效原子序数，可以准确预测 PDAC 患者的组织学分级和 OS。临床相关性 基于 DECT 的列线图是预测 PDAC 组织学分级和 OS 的可靠、非侵入性工具。它提供了指导个性化治疗策略的重要信息，可能会改善患者管理和结果。© 2024。作者，获得欧洲放射学会的独家许可。

We evaluated the value of dual-energy computed tomography (DECT) parameters derived from pancreatic ductal adenocarcinoma (PDAC) to discriminate between high- and low-grade tumors and predict overall survival (OS) in patients.Data were retrospectively collected from 169 consecutive patients with pathologically confirmed PDAC who underwent third-generation dual-source DECT enhanced dual-phase scanning before surgery between January 2017 and March 2023. Patients with prior treatments, other malignancies, small tumors, or poor-quality scans were excluded. Two radiologists evaluated three clinical and seven radiological features and measured sixteen DECT-derived parameters. Univariate and multivariate analyses were applied to select independent predictors. A prediction model and a corresponding nomogram were developed, and the area under the curve (AUC), calibration, and clinical applicability were assessed. The correlations between factors and OS were evaluated using Kaplan-Meier survival and Cox regression analyses.One hundred sixty-nine patients were randomly divided into training (n = 118) and validation (n = 51) cohorts, among which 43 (36.4%) and 19 (37.3%) had high-grade PDAC confirmed by pathology, respectively. The vascular invasion, normalized iodine concentration in the venous phase, and effective atomic number in the venous phase were independent predictors for histological grading. A nomogram was constructed to predict the risk of high-grade tumors in PDAC, with AUCs of 0.887 and 0.844 in the training and validation cohorts, respectively. The nomogram exhibited good calibration and was more beneficial than a single parameter in both cohorts. Pathological- and nomoscore-predicted high-grade PDACs were associated with poor OS (all p < 0.05).The nomogram, which combines DECT parameters and radiological features, can predict the histological grade and OS in patients with PDAC before surgery.Question Preoperative determination of histological grade in PDAC is crucial for guiding treatment, yet current methods are invasive and limited. Findings A DECT-based nomogram combining vascular invasion, normalized iodine concentration, and effective atomic number accurately predicts histological grade and OS in PDAC patients. Clinical relevance The DECT-based nomogram is a reliable, non-invasive tool for predicting histological grade and OS in PDAC. It provides essential information to guide personalized treatment strategies, potentially improving patient management and outcomes.© 2024. The Author(s), under exclusive licence to European Society of Radiology.