本研究探讨了双能 CT (DECT) 延迟相细胞外体积 (ECV) 分数在预测接受术前免疫化疗的晚期胃癌 (FAGC) 患者肿瘤消退等级 (TRG) 方面的有效性。对2019年8月至2023年3月在我院接受术前免疫化疗的晚期胃腺癌患者进行了这项研究。根据TRG将患者分为病理完全缓解（pCR）组和非pCR组。 ECV 使用延迟相位碘图确定。此外，使用三相增强碘图对肿瘤碘密度和标准化碘比率进行了仔细分析。具有五重交叉验证和 Spearman 相关性的单变量分析确定了 DECT 参数和与 pCR 相关的临床指标。使用加权逻辑回归模型和五倍交叉验证来评估这些参数对 pCR 的预测准确性。在 88 名入组患者中（平均年龄 60.8±11.1 岁，63 名男性），21 名 (23.9%) 达到了 pCR。单变量分析表明 ECV 在区分 pCR 和非 pCR 组中具有显着作用（平均 p 值 = 0.021）。在逻辑回归模型中，ECV 独立预测 pCR，平均优势比为 0.911（95% 置信区间，0.798-0.994）。该模型结合了 ECV、肿瘤面积和 IDAV（碘密度从静脉期到动脉期的相对变化率），显示平均曲线下面积 (AUC) 为 0.780 (0.770-0.791) 和 0.766 (0.731-0.800)分别用于预测 pCR 的训练集和验证集。DECT 衍生的 ECV 分数是接受术前免疫化疗的 FAGC 患者中 TRG 的有价值的预测因子。这项研究表明，DECT 衍生的细胞外体积分数是病理完整的可靠预测因子接受术前免疫化疗的晚期胃癌患者的反应，为识别潜在治疗受益者提供了一种非侵入性工具。© 2024。作者，获得欧洲放射学会的独家许可。

This study examines the effectiveness of dual-energy CT (DECT) delayed-phase extracellular volume (ECV) fraction in predicting tumor regression grade (TRG) in far-advanced gastric cancer (FAGC) patients receiving preoperative immuno-chemotherapy.A retrospective analysis was performed on far-advanced gastric adenocarcinoma patients treated with preoperative immuno-chemotherapy at our institution from August 2019 to March 2023. Patients were categorized based on their TRG into pathological complete response (pCR) and non-pCR groups. ECV was determined using the delayed-phase iodine maps. In addition, tumor iodine densities and standardized iodine ratios were meticulously analyzed using the triple-phase enhanced iodine maps. Univariate analysis with five-fold cross-validation and Spearman correlation determined DECT parameters and clinical indicators association with pCR. The predictive accuracy of these parameters for pCR was evaluated using a weighted logistic regression model with five-fold cross-validation.Of the 88 patients enrolled (mean age 60.8 ± 11.1 years, 63 males), 21 (23.9%) achieved pCR. Univariate analysis indicated ECV's significant role in differentiating between pCR and non-pCR groups (average p value = 0.021). In the logistic regression model, ECV independently predicted pCR with an average odds ratio of 0.911 (95% confidence interval, 0.798-0.994). The model, incorporating ECV, tumor area, and IDAV (the relative change rate of iodine density from venous phase to arterial phase), showed an average area under curves (AUCs) of 0.780 (0.770-0.791) and 0.766 (0.731-0.800) for the training and validation sets, respectively, in predicting pCR.DECT-derived ECV fraction is a valuable predictor of TRG in FAGC patients undergoing preoperative immuno-chemotherapy.This study demonstrates that DECT-derived extracellular volume fraction is a reliable predictor for pathological complete response in far-advanced gastric cancer patients receiving preoperative immuno-chemotherapy, offering a noninvasive tool for identifying potential treatment beneficiaries.© 2024. The Author(s), under exclusive licence to European Society of Radiology.