Detectable end of radiation prostate specific antigen assists in identifying men with unfavorable intermediate‐risk prostate cancer at high risk of distant recurrence and cancer‐specific mortality. Issue 8 (9th March 2018)

• Record Type: Journal Article
• Authors: Hayman, Jonathan; Phillips, Ryan; Chen, Di; Perin, Jamie; Narang, Amol K.; Trieu, Janson; Radwan, Noura; Greco, Stephen; Deville, Curtiland; McNutt, Todd; Song, Daniel Y.; DeWeese, Theodore L.; Tran, Phuoc T.
• Is Part Of: Prostate. Volume 78:Issue 8(2018)
• Page Start: 623
• Page End: 630
• Publication Date: 2018-03-09
• Subjects: cancer -- prostate -- risk stratification -- unfavorable intermediate risk; Prostate -- Diseases -- Periodicals; 616
• DOI: 10.1002/pros.23507
• Languages: English
• ISSNs: 0270-4137

Abstract

Abstract : Background: Undetectable End of Radiation PSA (EOR‐PSA) has been shown to predict improved survival in prostate cancer (PCa). While validating the unfavorable intermediate‐risk (UIR) and favorable intermediate‐risk (FIR) stratifications among Johns Hopkins PCa patients treated with radiotherapy, we examined whether EOR‐PSA could further risk stratify UIR men for survival. Methods: A total of 302 IR patients were identified in the Johns Hopkins PCa database (178 UIR, 124 FIR). Kaplan‐Meier curves and multivariable analysis was performed via Cox regression for biochemical recurrence free survival (bRFS), distant metastasis free survival (DMFS), and overall survival (OS), while a competing risks model was used for PCa specific survival (PCSS). Among the 235 patients with known EOR‐PSA values, we then stratified by EOR‐PSA and performed the aforementioned analysis. Results: The median follow‐up time was 11.5 years (138 months). UIR was predictive of worse DMFS and PCSS ( P = 0.008 and P = 0.023) on multivariable analysis (MVA). Increased radiation dose was significant for improved DMFS ( P = 0.016) on MVA. EOR‐PSA was excluded from the models because it did not trend towards significance as a continuous or binary variable due to interaction with UIR, and we were unable to converge a multivariable model with a variable to control for this interaction. However, when stratifying by detectable versus undetectable EOR‐PSA, UIR had worse DMFS and PCSS among detectable EOR‐PSA patients, but not undetectable patients. UIR was significant on MVA among detectable EOR‐PSA patients for DMFS ( P = 0.021) and PCSS ( P = 0.033), while RT dose also predicted PCSS ( P = 0.013). Conclusions: EOR‐PSA can assist in predicting DMFS and PCSS among UIR patients, suggesting a clinically meaningful time point for considering intensification of treatment in clinical trials of intermediate‐risk men.