Oncology - Prostate
Introduction & Objective : Magnetic resonance imaging (MRI) is emerging as a highly-utilized tool for diagnosis and risk-stratification of prostate cancer. One area of interest has been in detection of extracapsular extension (ECE) of prostate tumors. The approach, however, has demonstrated limited accuracy with a 49% sensitivity and 74% specificity for ECE. The Decipher test is a genomic classifier that measures RNA expression of 22 different biomarkers using biopsy or prostatectomy specimens that is being used for prognostic analysis of prostate cancer. The aim of this study was to determine if the combined use of MRI and Decipher test would result in an improvement in risk stratification through an improved predictive value of ECE.
Methods : Between April 2015 and August 2017, 26 patients underwent robot-assisted radical prostatectomy following prostate MRI at the University of Maryland Medical Center. Specimens obtained were sent for analysis of Decipher scores, which reported 5 year metastasis risk of the cancer. MRI was evaluated for several criteria including imaging-based suggestion of ECE and Prostate Imaging Reporting and Data System (PI-RADS) score. Multivariate and receiver operating characteristic (ROC) curve analysis was used to assess the predictive values of Decipher scores and PIRADS scores for pathologic ECE.
Results : In total, 26 patients with complete data were analyzed of whom 14 had ECE (62%). 4 of these patients were PIRADS 3, 14 were PIRADS 4, and 8 were PIRADS 5. The 5 year metastasis risks obtained from Decipher tests ranged from 1% to 45.9%. Three ROC curves were created to compare the ability of PIRADS score to predict ECE, Decipher score to predict ECE, and Decipher and PIRADS scores combined to predict ECE. This demonstrated improved predictive value of ECE when Decipher and PIRADS scores are used together with the area under the curve (AUC) = 0.92 when compared to each score separately (Decipher score vs ECE AUC = 0.71; PIRADS score vs ECE AUC = 0.83, p < 0.05). Thus combining the genomic testing and MRI improved prediction of pathologic ECE in the prostatectomy sample.
Conclusions : Use of prostate MP-MRI PIRADS score and genomic testing scores may work synergistically to help predict pathologic ECE for prostate cancer at time of radical prostatectomy.