基于能谱CT 120 kVp-like图像和碘图的细胞外体积分数在术前预测肾透明细胞癌分级的价值  

The Value of Extracellular Volume Fraction Based on 120 kVp-like Images and Iodine Maps in Preoperative Prediction of Renal Clear Cell Carcinoma Grading in Spectral CT

作　　者：何健 王诗耕 魏巍 胡梦婷 张竞颐 程启烨 董德硕 杜长谕 刘义军[1] HE Jian;WANG Shigeng;WEI Wei(Department of Radiology,The First Affiliated Hospital of Dalian Medical University,Dalian,Liaoning Province 116011,P.R.China)

机构地区：[1]大连医科大学附属第一医院放射科,116011

出　　处：《临床放射学杂志》2025年第4期684-688,共5页Journal of Clinical Radiology

摘　　要：目的在能谱CT三期增强扫描图像中比较基于120 kVp-like图像的CT值增值(ΔCT)和碘基物质图像的碘密度(ID)计算的细胞外体积分数(ECV)在术前预测肾透明细胞癌(ccRCC)病理分级的价值。方法回顾性分析2021年11月至2024年7月因肾脏肿瘤接受手术且经病理证实为ccRCC的患者100例,其中低级别组(WHO/ISUP 1、2级)64例,高级别组(WHO/ISUP 3、4级)36例。所有患者CT图像的原始数据完整,可重组120 kVp-like图像和碘基物质图。在120 kVp-like图像上测量平扫期、皮髓质期、实质期和排泄期肿瘤的CT值以获得ΔCT_(皮髓质期)、ΔCT_(实质期)、ΔCT_(排泄期),在碘基物质图上测量增强三期肿瘤的ID以获得ID_(皮髓质期)、ID_(实质期)和ID_(排泄期)。记录红细胞压积(HCT)指标,并根据公式分别得到基于ΔCT和ID的ECV。采用受试者工作特征曲线(ROC)对ECV的预测效能进行评估,并比较曲线下面积(AUC)的差异。结果由ΔCT和ID两种方法计算得到的ECV中,高级别组ccRCC均高于低级别组(P<0.05)。基于三期的ΔCT值和ID值计算得到的ECV在鉴别ccRCC病理分级中的AUC分别为0.780、0.901、0.697、0.790、0.923、0.743。在能谱CT 120 kVp-like图像中ΔCT_(实质期)与碘基物质图中ID_(实质期)计算的ECV预测效能较高,敏感度和特异度分别为86.11%、84.37%和83.33%、92.19%。ID_(实质期)计算的ECV的AUC要优于ΔCT_(实质期)(P<0.05)。结论基于实质期ID计算的ECV可以在术前准确地预测ccRCC病理分级。Objective To compare the value of extracellular volume fraction(ECV)calculated from CT enhancement values(ΔCT)based on 120 kVp-like images and iodine density(ID)from iodine maps in preoperatively predicting the pathological grading of renal clear cell carcinoma(ccRCC)in phase III enhanced spectral CT images.Methods A retrospective analysis was conducted on 100 patients who underwent surgery for renal tumors and were pathologically confirmed as ccRCC between November 2021 and July 2024.Among them,64 were in the low-grade group(WHO/ISUP grades 1 and 2)and 36 in the high-grade group(WHO/ISUP grades 3 and 4).The original CT image data of all patients were complete,allowing for the reconstruction of 120 kVp-like images and iodine maps.Tumor CT values during the non-contrast,corticomedullary phase,nephrographic phase,and excretory phase were measured on the 120 kVp-like images to obtainΔCT for the corticomedullary phase,nephrographic phase,and excretory phase.ID for the enhanced tumors during the three phases was measured on the iodine maps to obtain ID for the corticomedullary phase,nephrographic phase,and excretory phase.Hematocrit(HCT)values were recorded,and ECV based onΔCT and ID were calculated using respective formulas.The predictive performance of ECV was evaluated using the receiver operating characteristic curve(ROC),and differences in area under the curve(AUC)were compared.Results ECV calculated from bothΔCT and ID showed that the high-grade ccRCC group had higher values than the low-grade group(P<0.05).The AUC for ECV calculated fromΔCT and ID values in the three phases for distinguishing ccRCC pathological grades were 0.780,0.901,0.697,0.790,0.923,and 0.743,respectively.The ECV prediction performance calculated fromΔCT in the nephrographic phase and ID in the nephrographic phase was relatively high,with sensitivities and specificities of 86.11%,84.37%and 83.33%,92.19%,respectively.The AUC for ECV calculated from ID in the nephrographic phase was superior to that calculated fromΔCT in the nephrograph

关 键 词：肾透明细胞癌 细胞外体积分数 WHO/ISUP分级 能谱CT 

分 类 号：TN9[电子电信—信息与通信工程]