Department of Radiology, Cancer Hospital of the University of Chinese Academy of Sciences (Zhejiang Cancer Hospital), 1 Banshan East Road, Hangzhou, 310022, Zhejiang Province, China.
Department of Radiology, Institute of Cancer and Basic Medicine, Chinese Academy of Sciences, 1 Banshan East Road, Hangzhou, 310022, Zhejiang Province, China.
Abdom Radiol (NY). 2020 Nov;45(11):3849-3859. doi: 10.1007/s00261-020-02542-2.
The purpose of this study was to investigate the difference between imaging features of ultra-small renal cell carcinoma (usRCC) and angiomyolipoma with minimal fat (mfAML) whose enhancement were both hypervascular by using multidetector computed tomography (MDCT).
Confirmed by pathology, 40 cases of hypervascular usRCC and 21 cases of hypervascular mfAML both with diameter of 2 cm or less were compared and analyzed retrospectively, including traditional imaging features and thin-section computed tomography (CT) dynamic enhanced parameters. Meanwhile, receiver operating characteristic (ROC) curves were constructed to evaluate the diagnostic efficacy of each significant parameter and the information with diagnostic value was selected to construct the prediction model.
Comparison of traditional imaging features: the features, included age, shape, location, central location of tumor, wedge sign, renal cortex lift sign, black star sign, enhanced homogeneity in cortical phase (CP) and enhancement pattern had no significant difference between usRCC and mfAML (P > 0.05); sex, cystic degeneration or necrosis, pseudocapsule sign, and enhanced homogeneity in nephrographic phase (NP) had significant differences between usRCC and mfAML (P < 0.05). Comparison of CT dynamic enhanced parameters: the CT value, NEV and REV of usRCC were all higher than mfAML in both CP and NP (P < 0.01). Respectively, the area under the ROC curve (AUC) were 0.74, 0.75, 0.78, 0.83, 0.81 and 0.78. The sensitivity and specificity for differentiating ucRCC from mfAML were 85.0% and 76.2% respectively when NEV_NP was 73.6 HU as the critical value. Multivariate analysis showed that male, cystic degeneration or necrosis, and NEV_NP higher than 73.6 HU as an independent risk factor for usRCC (P < 0.01). The AUC value of the prediction model constructed by the combination was 0.94, the accuracy was 86.89%, the sensitivity was 82.50%, and the specificity was 95.24%.
Morphological characteristics in traditional diagnosis of small renal carcinoma (diameter of 4 cm or less) have certain significance in differentiating hypervascular usRCC and mfAML in early stage, but the diagnostic efficacy was limited. Sex, cystic degeneration or necrosis, and quantitative parameters measured after enhancement play an important role in differential diagnosis of hypervascular usRCC and mfAML, and the prediction model constructed by the combination has a good diagnostic performance.
本研究旨在通过多排螺旋 CT(MDCT)探讨超微小肾细胞癌(usRCC)与富含脂肪血管平滑肌脂肪瘤(mfAML)增强后均为富血管型的影像学特征差异。
经病理证实的直径均为 2cm 或以下的 40 例富血供 usRCC 和 21 例富血供 mfAML 病例进行回顾性对比分析,包括传统影像学特征和薄层 CT(CT)动态增强参数。同时,构建受试者工作特征(ROC)曲线,评估各显著参数的诊断效能,并选择具有诊断价值的信息构建预测模型。
传统影像学特征比较:年龄、形状、位置、肿瘤中央位置、楔形征、肾皮质抬高征、黑星征、皮质期(CP)增强均匀性和增强模式等特征在 usRCC 和 mfAML 之间无显著差异(P>0.05);性别、囊性变性或坏死、假包膜征和肾实质期(NP)增强均匀性在 usRCC 和 mfAML 之间有显著差异(P<0.05)。CT 动态增强参数比较:CP 和 NP 时,usRCC 的 CT 值、NEV 和 REV 均高于 mfAML(P<0.01)。ROC 曲线下面积(AUC)分别为 0.74、0.75、0.78、0.83、0.81 和 0.78。当 NEV_NP 为 73.6HU 时,区分 ucRCC 与 mfAML 的敏感度和特异度分别为 85.0%和 76.2%。多因素分析显示,男性、囊性变性或坏死和 NEV_NP 高于 73.6HU 是 usRCC 的独立危险因素(P<0.01)。联合构建的预测模型的 AUC 值为 0.94,准确率为 86.89%,敏感度为 82.50%,特异度为 95.24%。
直径 4cm 或以下的小肾癌(直径 4cm 或以下)的传统诊断中的形态学特征在早期区分富血供 usRCC 和 mfAML 有一定意义,但诊断效能有限。性别、囊性变性或坏死和增强后测量的定量参数在鉴别富血供 usRCC 和 mfAML 中起重要作用,联合构建的预测模型具有良好的诊断性能。