In prostate cancer (PCa) recurring after radical prostatectomy (RP), salvage radiotherapy (SRT) is recommended to be given at PSA <0.5 ng/ml. It has been speculated, that the advantage from early SRT is mainly caused by lead-time bias: Calculating from time of SRT, earlier treatment would per-se result in longer time to event/censoring compared with later treatment, but not extend the interval from RP to post-SRT failure.
In 603 consecutive PCa patients receiving SRT between 1997 and 2017, we compared outcomes, calculating from time of irradiation vs. time of surgery.
In multivariable analysis, tumor stage pT3-4, pathological Gleason score GS ≤6 vs. GS 7 vs. GS ≥8, post-RP PSA persistence (nadir ≥0.1 ng/ml), and the pre-SRT PSA (continuous or with cutoff 0.4 ng/ml) were significant risk-factors for biochemical progression (BCR) and progression-free survival (PFS) post-SRT and post-RP. A pre-SRT PSA <0.4 ng/ml was a significant discriminator for Kaplan-Meier rates of BCR and PFS. The Cox model for overall survival (OS) included age at RP (continuous), pT2 vs. pT3-4, and pre-SRT PSA (continuous) as significant predictors. However, no significant cutoff for the pre-SRT PSA could be identified to differentiate Kaplan-Meier estimates of OS, possibly because there were too few events, as 88% of the patients were still alive at last follow-up.
The pre-SRT PSA has a significant impact on BCR, PFS and potentially on OS, calculating either from RP or from SRT to event/censoring, respectively. This contradicts the hypothesis of lead-time bias falsifying the advantage from early SRT.
Radiotherapy and oncology : journal of the European Society for Therapeutic Radiology and Oncology. 2020 Sep 10 [Epub ahead of print]
Detlef Bartkowiak, Reinhard Thamm, Alessandra Siegmann, Dirk Böhmer, Volker Budachb, Thomas Wiegel
Dept. Radiation Oncology, University Hospital Ulm, Ulm, Germany., Dept. Radiation Oncology, Charité University Hospital, Berlin, Germany., Dept. Radiation Oncology, University Hospital Ulm, Ulm, Germany. Electronic address: .