Multiparametric Magnetic Resonance Imaging Predicts Postoperative Pathology but Misses Aggressive Prostate Cancers as Assessed by Cell Cycle Progression Score.

Abstract

PURPOSE: We identified prognostic biomarkers in prostate cancer by a radiogenomics strategy that integrates gene expression using the cell cycle progression score and medical images. MATERIALS AND METHODS: We obtained institutional review board approval and written informed consent from 106 men with prostate cancer, including 60% at low risk, who underwent multiparametric magnetic resonance imaging before radical prostatectomy was done and a cell cycle progression score was determined. The correlation between the results of multiparametric magnetic resonance imaging and Gleason grade or cell cycle progression score was assessed by logistic regression. RESULTS: Patients with primary Gleason grade greater than 3 had a longer median maximal tumor diameter (13 vs 10 mm) and a lower median apparent diffusion coefficient (0.745 vs 0.88×10(-3) mm2 per second, each p=0.0001) than those with primary Gleason grade 3 or less. Maximal diameter 10 mm or greater (OR 4.9, 95% CI 1.7 to 14.0, p=0.0012) and apparent diffusion coefficient 0.80×10(-3) mm2 per second or less (OR 7.5, 95% CI 3.0 to 18.7, p<0.0001) were significantly associated with primary Gleason grade greater than 3. The combined measure of maximal diameter less than 10 mm and apparent diffusion coefficient greater than 0.80×10(-3) mm2 per second identified only index lesions harboring primary Gleason grade 3. However, 7 of those lesions showed a molecular pattern of high risk lethal prostate cancer (cell cycle progression score greater than 0). CONCLUSIONS: Multiparametric magnetic resonance imaging is able to predict low and high risk Gleason scores in the tumor. However, the cell cycle progression score did not completely match the imaging result. These findings suggest that management of early stages prostate cancer could strongly benefit by performing magnetic resonance imaging targeted biopsy coupled with molecular analysis.