Radiogenomics of prostate cancer: Association between qunatitative multiparametric MRI features and PTEN.

Abstract

126 Background: Better methods are needed to assess prior to prostatectomy the risk of aggressive prostate cancer. Radiogenomics is a promising new paradigm that aims to gain molecular and genomic insights from clinical images. Loss of PTEN expression correlates with clinically aggressive disease and is associated with a 7-fold increase in the risk of prostate cancer death. Methods: From 38 patients who had undergone multi-parametric prostate MRI prior to prostatectomy, a pathologist and a radiologist simultaneously identified 45 peripheral-zone cancer regions of interest (ROIs). Histologic sections of the cancer foci underwent immunohistochemical analysis and were scored according to percent of tumor cells expressing PTEN as: negative (0-20%), mixed (20-80%), or positive (80-100%). From the MRI ROIs, the average and 10th percentile ADC values, T2-weighted signal-intensity histogram skewness, and quantitative perfusion parameters were calculated. Both dynamic perfusion two-compartment model and an empirical mathematical model (EMM) were used to fit the average contrast concentration curves within the ROIs as a function of time. Associations between the quantitative image features and PTEN expression were analyzed with Pearson's correlation coefficient (r). Results: The PTEN scores were: positive (n = 21, 47%), mixed (n = 12, 27%), and negative (n = 12, 27%). Two perfusion imaging contrast uptake parameters obtained from EMM correlated with PTEN scores (r = 0.25, p < 0.1 and r = 0.43, p < 0.01), and T2-weighted signal-intensity skewness also showed some correlation tendency (r = −0.25, p < 0.1). No correlation was seen between mean ADC and 10th percentile ADC values and PTEN score. Conclusions: This preliminary study of radiogenomics of prostate cancer suggests that fast contrast uptake of cancer on DCE-MR imaging and a T2-weighted imaging feature are potentially associated with prostate cancer PTEN expression, which warrants further studies and validation.