This article was originally published here
Radiol Oncol. 2021 Mar 25. doi: 10.2478/raon-2021-0017. Online ahead of print.
BACKGROUND: The aim of the study was to analyse the prognostic factors in postoperative prostate cancer irradiation and develop a nomogram for disease-free survival (DFS).
PATIENTS AND METHODS: This retrospective study included 236 consecutive prostate cancer patients who had radical prostatectomy followed by radiotherapy (RT) at a single tertiary institution between 2009 and 2014. The main outcome was DFS analysed through uni- and multivariable analysis, Kaplan-Meier curves, log-rank testing, recursive partitioning analysis, and nomogram development.
RESULTS: The median follow up was 62.3 (interquartile range [IQR] 38.1-79) months. The independent clinical factors associated with increased risk of recurrence or progression in the multivariate analysis (MVA) were prostate-specific antigen (PSA) level before RT, pT3 characteristic, and local failure as salvage indication. The value of PSA nadir had a significant impact on the risk of biochemical failure. Biochemical control and DFS were significantly different depending on treatment indication (p < 0.0001). The recursive partitioning analysis highlighted the importance of the PSA level before RT, Gleason Grade Group, PSA nadir, and local failure as a treatment indication. Finally, the nomogram for DFS was developed and is available online at https://apps.konsta.com.pl/app/prostate-salvage-dfs/.
CONCLUSIONS: The Pre-RT PSA level, pT3 characteristic and local failure as salvage indication are pivotal prognostic factors associated with increased risk of recurrence or progression. The Gleason grade group of 4-5 and PSA nadir value allow for further risk stratification. The treatment outcomes in postoperative prostate cancer irradiation are significantly different depending on treatment indication. An online nomogram comprising of both pre-treatment and current data was developed allowing for visualization of changes in prognosis depending on clinical data.