From the Departments of Radiology (T.H., M.D.F.M., N.S., L.L., R.E.T.) and Anatomical Pathology (T.A.F.), University of Ottawa, 451 Smyth Rd, Ottawa, ON, Canada K1H 8M5; Departments of Medical Imaging (T.H., M.D.F.M., N.S., L.L., R.E.T.) and Anatomical Pathology (T.A.F.), The Ottawa Hospital, Ottawa, Ontario, Canada; and Clinical Epidemiology Program, Ottawa Hospital Research Institute, Ottawa, Ontario, Canada (M.D.F.M., N.S., R.E.T.).
Radiology. 2015 Sep;276(3):787-96. doi: 10.1148/radiol.2015142215. Epub 2015 Apr 23.
To determine the accuracy of texture analysis to differentiate fat-poor angiomyolipoma (fp-AML) from renal cell carcinoma (RCC) on unenhanced computed tomography (CT) images.
In this institutional review board-approved retrospective case-control study, patients with AML and RCC were identified from the pathology database: there were 16 patients with fp-AML (no visible fat at unenhanced CT) and 84 patients with RCC. Axial unenhanced CT images were contoured manually by two independent analysts. Texture analysis was performed for each lesion, and reproducibility was assessed. Texture features related to the gray-level histogram, gray-level co-occurrence, and run-length matrix statistics were evaluated. The most discriminative features were used to generate support vector machine (SVM) classifiers. Diagnostic accuracy of textural features was assessed and 10-fold cross validation was performed. Unenhanced CT images for each patient were independently reviewed by two blinded radiologists who subjectively graded lesion heterogeneity on a five-point scale. Differences in area under the receiver operating characteristic curve (AUC) between subjective heterogeneity ratings and textural features were evaluated by using the DeLong method.
There was lower lesion homogeneity and higher lesion entropy in RCCs (P ≤ .01). A model incorporating several texture features resulted in an AUC of 0.89 ± 0.04. The average SVM accuracy of textural features ranged from 83% to 91% (after 10-fold cross validation). An optimal subjective heterogeneity rating of 2 or higher was identified as a predictor of RCC for both readers, with no significant difference in AUC between readers (P = .06). Each of the three textural-based classifiers was more accurate than either radiologists' subjective heterogeneity ratings for the models incorporating a subset of the top three textural features (difference in AUC between textural features and subjective visual heterogeneity, 0.25; 95% confidence interval: 0.02, 0.47; P = .03).
CT texture analysis can be used to accurately differentiate fp-AML from RCC on unenhanced CT images.
确定纹理分析在鉴别乏脂性血管平滑肌脂肪瘤（fp-AML）与肾细胞癌（RCC）中的准确性，该方法基于平扫 CT 图像。
本研究为回顾性病例对照研究，经医院伦理委员会批准，在病理数据库中查找 AML 和 RCC 患者：共纳入 16 例 fp-AML 患者（平扫 CT 未见脂肪）和 84 例 RCC 患者。两名独立的分析员手动勾画肿瘤的轴位平扫 CT 图像。对每个病灶进行纹理分析，并评估其可重复性。评估与灰度直方图、灰度共生矩阵和游程长度矩阵统计相关的纹理特征。使用支持向量机（SVM）生成最具鉴别力的特征分类器。评估纹理特征的诊断准确性，并进行 10 折交叉验证。每位患者的平扫 CT 图像由两名盲法阅片的放射科医生独立评估，他们对病灶异质性进行五分制主观评分。采用 DeLong 方法评估受试者工作特征曲线（ROC）下面积（AUC）之间的差异。
RCC 的病灶均匀性更低，熵值更高（P ≤.01）。纳入多个纹理特征的模型 AUC 为 0.89 ± 0.04。纹理特征的 SVM 平均准确率在 83%至 91%（经 10 折交叉验证）之间。对于两位阅片者，当主观异质性评分为 2 或更高时，均提示为 RCC，且两者之间 AUC 无显著差异（P =.06）。对于纳入前三个最优纹理特征子集的模型，每个纹理分类器均比放射科医生的主观异质性评分更准确（纹理特征与主观视觉异质性之间 AUC 的差异，0.25；95%置信区间：0.02，0.47；P =.03）。
基于平扫 CT 图像的纹理分析可用于准确鉴别 fp-AML 与 RCC。
