Kim Cherry, Choi Hyuck Jae, Cho Kyoung-Sik
Department of Radiology and Research Institute of Radiology, University of Ulsan College of Medicine, Asan Medical Center, 388-1 Poong-nap-dong, Songpa-gu, Seoul, Republic of Korea.
Eur J Radiol. 2014 Jun;83(6):914-918. doi: 10.1016/j.ejrad.2014.02.025. Epub 2014 Mar 22.
Although renal sinus fat invasion has prognostic significance in patients with renal cell carcinomas (RCCs), there are no previous studies about the value of multidetector computed tomography (MDCT) about this issue in the current literature.
A total of 863 consecutive patients (renal sinus fat invasion in 110 patients (12.7%)) from single institutions with surgically-confirmed renal cell carcinoma who underwent MDCT between 2010 and 2012 were included in this study. The area under the curves (AUCs) of the receiver operating characteristic (ROC) analysis was used to compare diagnostic performance. Reference standard was pathologic examination. Weighted κ statistics were used to measure the level of interobserver agreement. Multivariate logistic regression model was used to find the predictors for renal sinus fat invasion. Image analysis was first performed with axial-only CT images. A second analysis was then performed with both axial and coronal CT images. A qualitative analysis was then conducted by two reviewers who reached consensus regarding tumor size, decreased perfusion, tumor margin, vessel displacement, and lymph node metastasis. The reference standard was pathologic evaluation.
The AUCs of the ROC analysis were 0.881 and 0.922 for axial-only images and 0.889 and 0.902 for combined images in both readers. The AUC of tumor size was 0.884, a similar value to that of the reviewers. In multivariate analysis, tumor size, a linear-nodular or nodular type of fat infiltration, and an irregular tumor margin were independent predicting factors for perinephric fat invasion.
MDCT shows relatively high diagnostic performance in detecting perinephric fat invasion of RCC but suffers from a relatively low PPV related to low prevalence of renal sinus fat invasion. Applying tumor size alone we could get similar diagnostic performance to those of radiologists. Tumor size, fat infiltration with a nodular appearance, and an irregular tumor margin were predictors for perinephric invasion.
尽管肾窦脂肪浸润对肾细胞癌（RCC）患者具有预后意义，但目前文献中尚无关于多排螺旋计算机断层扫描（MDCT）在该问题上价值的先前研究。
本研究纳入了2010年至2012年间在单一机构接受MDCT检查且手术确诊为肾细胞癌的863例连续患者（110例患者（12.7%）存在肾窦脂肪浸润）。采用受试者操作特征（ROC）分析的曲线下面积（AUC）来比较诊断性能。参考标准为病理检查。采用加权κ统计量来衡量观察者间的一致性水平。使用多变量逻辑回归模型来寻找肾窦脂肪浸润的预测因素。首先仅对轴位CT图像进行图像分析。然后对轴位和冠状位CT图像进行第二次分析。随后由两位阅片者进行定性分析，他们就肿瘤大小、灌注降低、肿瘤边缘、血管移位和淋巴结转移达成了共识。参考标准为病理评估。
两位阅片者中，仅轴位图像的ROC分析AUC分别为0.881和0.922，联合图像的AUC分别为0.889和0.902。肿瘤大小的AUC为0.884，与阅片者的结果相似。在多变量分析中，肿瘤大小、线性结节状或结节状脂肪浸润类型以及不规则肿瘤边缘是肾周脂肪浸润的独立预测因素。
MDCT在检测RCC的肾周脂肪浸润方面显示出相对较高的诊断性能，但由于肾窦脂肪浸润的患病率较低，其阳性预测值相对较低。仅应用肿瘤大小我们就能获得与放射科医生相似的诊断性能。肿瘤大小、结节状脂肪浸润和不规则肿瘤边缘是肾周浸润的预测因素。
