Is quantitative DCE-MRI useful in differentiation of indolent and significant prostate cancers?

Abstract

Aim: Direct visual assessment is recommended in prostate magnetic resonance imaging (MRI) for dynamic contrast enhancement (DCE), however, being a qualitative approach, it may cause inter-reader variability. The purpose of this study was to compare quantitative DCE parameters in the differentiation of clinically significant prostate cancer from indolent cancer using whole-mount histopathology. Methods: Seventy-six patients who underwent multiparametric MRI with suspicion of prostate cancer and subsequent radical prostatectomy were included. Index tumor location was determined with pathology reports. MRI findings of this location were evaluated by a different radiologist using prostate imaging-reporting and data system version 2.1 (PI-RADSv2.1) guideline. Gleason 3+3 tumors were considered indolent, and Gleason ≥ 3+4 tumors were considered significant cancers. Region-of-interests (ROI) were placed in the lesion and the normal peripheral zone. Lesion values and lesion/normal ratios of Ktrans, Kep, Ve, area under curve (iAUC) were calculated. T test was used in statistical analysis. Results: The numbers of cases with PI-RADSv2.1 scores of 2, 3, 4 and 5 were 5, 4, 24, and 43, respectively. There were 13 indolent cases and 63 patients with significant prostate cancer. Lesion/normal ratios of Ktrans, Kep, Ve, iAUC were 1.6, 1.59, 12, 2.1, respectively, in indolent cancers, and 3.1, 4.04, 1.39, 2.8, respectively, in significant cancers. Lesion/normal ratio of Ktrans was higher in significant cancers while lesion/normal ratio of Ve was higher in indolent cancers. Kep and iAUC were similar (P>0.05 for each). Conclusion: Quantitative DCE assessment may demonstrate more reproducible results. Lesion/normal tissue ratios of Ktrans and Ve were helpful in differentiation between indolent and significant prostate cancers.