DOI: 10.4103/iju.iju_340_26 ISSN: 0970-1591

Biochemical recurrence, salvage treatment, and long-term oncological outcomes after radical prostatectomy for International Society of Urological Pathology grade group 2 prostate cancer

Lucas Diéguez-Álvarez, Sara Gil-Carballal, Ana Castelló-Porcar, José María Martínez-Jabaloyas

ABSTRACT

Introduction:

International Society of Urological Pathology (ISUP) grade group (GG) 2 is the most common grade at radical prostatectomy (RP), yet its long-term natural history is poorly defined. We evaluated the predictors of biochemical recurrence (BCR) and the relationship between BCR, salvage treatment, and oncological outcomes.

Methods:

We analyzed 224 consecutive patients with confirmed GG2 prostate cancer after RP between 1999 and 2015. BCR was defined as two consecutive prostate-specific antigen ≥0.2 ng/mL. Kaplan–Meier and competing-risk analyses were performed, and predictors of BCR were assessed using multivariable Cox regression analysis.

Results:

Median follow-up was 209 months. BCR occurred in 91 patients (40.6%) after a median of 47 months. Ten-year BCR-free survival was 62.0%, 72.5% with negative surgical margins, and 51.9% with positive margins ( P < 0.001). Positive surgical margin (PSM) was the only independent predictor of BCR (hazard ratio: 2.29, 95% confidence interval 1.40–3.75; P = 0.001). No patient received adjuvant radiotherapy. Among 91 patients with BCR, 35 (38.5%) received no salvage treatment; after 107 months’ median follow-up from recurrence, none developed metastasis or died from prostate cancer. Metastasis developed in 8 patients (3.6%) and prostate cancer-specific death occurred in 2 (0.9%). Twenty-year metastasis-free and cancer-specific survival were 95.4% and 97.9%, respectively.

Conclusions:

In our cohort, PSM was the main predictor of BCR, whereas metastatic progression and prostate cancer-specific death after BCR were uncommon. Favorable outcomes were observed among recurrent patients managed without salvage treatment, although nonrandom treatment allocation and the limited number of progression events preclude causal inference.

More from our Archive