We evaluate the Genomic Prostate Score in a prospective clinical setting and determine the cutoff point for likelihood of favorable pathology, below which definitive treatment should be advised.Methods:
Pathological data were recorded for men who had the Genomic Prostate Score performed and who ultimately underwent radical prostatectomy. Inclusion criteria were newly diagnosed prostate cancer, and NCCN classification as very low, low and low volume intermediate risk. Adverse pathology was defined as any pT3 stage and primary Gleason grade of 4 or any pattern 5. ROC analysis was used to determine the optimal cutoff point of likelihood of favorable pathology for each NCCN risk group.Results:
A total of 95 patients were enrolled and 50 patients (53%) underwent radical prostatectomy. Adverse pathology was found in 21 patients (42%), grouped as very low risk 0%, low risk 32.4% and low volume intermediate risk 71.4%. Among those with low risk disease, ROC analysis determined that a likelihood of favorable pathology cutoff of 76% or greater performed the best, correctly classifying 91.2% of patients with a sensitivity of 95.7%, specificity of 81.8% and AUC 0.95. For intermediate risk patients the optimal likelihood of favorable pathology cutoff was 68% or greater, with 92.3% correct, sensitivity 75%, specificity 100% and AUC 0.95.Conclusions:
NCCN low risk patients had the most meaningful information provided by the Genomic Prostate Score. Men with low risk disease with a likelihood of favorable pathology threshold greater than 75% are at very low risk for adverse pathology, whereas those with a likelihood of favorable pathology of 75% or less are at high risk. This likelihood of favorable pathology threshold is greater than 69% for men with low volume intermediate risk disease. These results should help clinicians use Genomic Prostate Score information when making decisions regarding active surveillance or intervention for prostate cancer.