Clinical value of the Histogram analysis of three-dimensional quantitative dynamic contrast enhanced MRI for diagnosing prostate cancer

Abstract

Objective To investigate the diagnostic efficiency of the quantitative dynamic contrast enhanced MRI (DCE-MRI) parameters using three dimensional (3D)-Histogram analysis for diagnosing prostate cancer. Methods A total of 50 patients (30 patients with prostate cancer and 20 patients with benign prostatic hyperplasia confirmed by biopsy results) undergoing routine prostate MRI and DCE-MRI scans were retrospectively analyzed. All patients were post-processed by DCE-MRI quantitative analysis, and calculated according to the 3D volume of interest (VOI) method and two dimensional (2D) ROI method to determine the Ktrans, Kep and Ve value for prostate cancer area, peripheral zone and the central gland area. Histogram analysis was used to analyze the 3D VOI result. Kruskal-Wallis H test was used to compare the differences between different parts of the prostate using the two methods. The diagnostic ability of the two methods for differentiating prostate cancer from noncancerous areas were determined by ROC. The correlations between the Histogram analysis of quantitative parameters and Gleason score were assessed with Spearman correlation. Results There were 33 prostate cancer areas derived from the 30 prostate cancer cases and got 20 peripheral zones and 20 central glands from the 20 benign prostate hyperplasia cases. The differences between the Histogram-based Ktrans (mean, 10th, 90th skewness and kurtosis), Kep (mean,10th, 90th) and Ve (skewness) of prostate cancer area, noncancerous area in the peripheral zone and central gland were statistically significant (P<0.05) . The differences between the ROI-based Ktrans and Kep of different prostate areas were statistically significant (P<0.05). The Histogram-based Ktrans (meas,90 th) had higher Az (0.92, 0.92) than that of the ROI-based Ktrans (0.85). There were 8 Gleason 6 areas, 19 Gleason 7 areas, 4 Gleason 8 areas and 2 Gleason 10 areas detected from the 33 prostate cancer areas, the mean Gleason value was (7.06±0.97). The Histogram-based Ktrans (10th, 90th,skewness) and Kep (mean,10th,90th) had correlation with Gleason score (r=0.53 to 0.68, P<0.05). Conclusions The 3D-Histogram analysis of quantitative DCE-MRI has a higher efficiency in diagnosing prostate cancer than 2D ROI-based approach, and it is feasible to stratify the pathological grade of prostate cancer by quantitative DCE-MRI with 3D-Histogram metrics. Key words: Prostate neoplasms; Magnetic resonance imaging; Histogram analysis