Radical prostatectomy versus external-beam radiotherapy for localized prostate cancer: long-term effect on biochemical control-in search of the optimal treatment

Carmen González-San Segundo,Felipe Herranz-Amo,Ana Alvarez-González,Pedro Cuesta-Álvaro,Marina Gómez-Espi,Eva Paños-Fagundo,Juan A Santos-Miranda
DOI: https://doi.org/10.1245/s10434-011-1660-0
Abstract:Background: The optimal management of patients with clinically localized prostate carcinoma remains undefined due in part to the absence of well-designed, randomized trials. Methods: This retrospective study comprised 505 patients diagnosed with low- or intermediate- risk prostate cancer in 1998-2005 and treated at Hospital Gregorio Marañón (Spain) with radical prostatectomy (RP) or external-beam radiotherapy (EBRT). No adjuvant therapy was administered. Biochemical relapse was defined as a prostate-specific antigen (PSA) level ≥0.4 ng/ml for RP cases and nadir + 2 for EBRT cases. RP was performed in 271 patients (53.6%) and EBRT in 234 patients (46.4%). The median follow-up was 60 months. The analysis end point was to compare the biochemical recurrence-free survival (bRFS) between the two groups. Results: The 5-year bRFS rates for RP and EBRT were 79 ± 2% and 86 ± 2%, respectively (P = 0.48). Multivariate analysis indicated that initial PSA (P = 0.00), perineural invasion in the biopsy specimen (P = 0.00), Gleason score (P = 0.04), EBRT dose (P = 0.02), and positive margins (P = 0.00) were independent predictors of relapse. A decision tree model was constructed with these variables. In the EBRT cohort, a nadir PSA of <0.3 ng/ml was associated with the best 5-year bRFS (96.6 vs. 56.5% if nadir PSA > 1.3 ng/ml). Late biochemical failure (>5 years) was more frequent in the RT group and with low-dose EBRT (≤72 Gy). Conclusions: The biochemical failure rates were similar between PR and EBRT in low- and intermediate-risk subgroups. Outcome was determined by classic pre-treatment features, perineural invasion, low-dose EBRT (≤72 Gy), and nadir PSA value in the RT cohort.
What problem does this paper attempt to address?