Does PSA at diagnosis correlate with Decipher Score?

Abstract

e17139 Background: The Decipher genomic classifier is a validated molecular biomarker that examines the activity of 22 genes and provides independent prognostic information for men with prostate cancer, supplementing standard clinicopathologic risk factors such as Gleason score, stage, and serum prostate-specific antigen (PSA). Multiple retrospective and prospective studies have demonstrated that Decipher improves risk stratification for adverse pathology, biochemical recurrence, metastasis, and cancer-specific survival, particularly in intermediate-risk and postprostatectomy settings. Methods: Analysis was conducted on data obtained from 422 males over the age of 18, in a community hospital in the last 5 years. We analyzed the PSA level closest to when the Decipher score was obtained to determine if there is any correlation between PSA level and Decipher score. Exclusion criteria: patients that had previous diagnosis of prostate cancer, Decipher or PSA result dates not available; a total of 72 patients were excluded. Pearson and Spearman rank correlation between PSA and Decipher was conducted on 350 patients. Results: After exclusion criteria, a total of 350 patients were included in the analysis. The Pearson correlation between PSA and Decipher was r = 0.13 with a p = 0.013, suggesting a statistically significant but weak linear association. In fact, after removing values above the 99th percentile, Pearson drops to r = 0.052, p = 0.33 illustrating that the signal is driven largely by outliers. Once ranks are considered, the spearman rank correlation was p≈0.07 with p≈0.17, indicating no significant monotonic relationship. Conclusions: In this community-based cohort of patients with prostate cancer, PSA level and Decipher score, had no monotonic relationship. In addition, despite reaching a statistically significant Pearson correlation, it remains very weak and largely influenced by extreme PSA outliers making it unlikely to be clinically significant. These findings support the established role of Decipher as an independent prognostic tool.