The Effect of Histogram Analysis of DCE-MRI Parameters on Differentiating Renal Tumors.

Abstract

We aimed to assess the role of histogram analysis of DCE-MRI parameters for accurately distinguishing renal clear cell carcinoma from renal hamartoma with minimal fat. Patients with renal tumors were enrolled from January 2013 to December 2015, including renal clear cell carcinoma (n = 39) and renal hamartoma (n = 10). Preoperative DCE-MR Imaging was performed, and whole-tumor regions of interest were drawn to obtain the corresponding histogram parameters, including skewness, kurtosis, frequency size, energy, quartile, etc. Histogram parameters differences between renal clear cell car-cinoma and renal hamartoma with minimal fat were compared. The diagnostic value of each significant parameter in predicting malignant tumors was determined. Histogram parameters of the DCE map contributed to differentiating the benign from malignant renal tumor groups. Histogram analysis of DCE maps could effectively present the heterogeneity of renal tumors and aid in differentiating benign and malignant tumors. ROC analysis results indicated that when frequency size = 1,732 was set as the threshold value, favorable diagnostic performance in predicting malignant tumors was achieved (AUC - 0.964; sensitivity - 84.6%; specificity - 100%), followed by skewness, Energy, Entropy, Uniformity, quartile 5, quartile 50, and kurtosis. Histogram analysis of DCE-MRI shows promise for differentiating benign and malignant renal tumors. Frequency size was the most significant parameter for predicting renal clear cell carcinoma.