Late biochemical recurrence after radical prostatectomy is associated with a slower rate of progression.

OBJECTIVE: To characterise the pattern of late biochemical recurrence (BCR) in the largest contemporary cohort of patients with localised prostate cancer treated with radical prostatectomy (RP) in the active surveillance era. PATIENTS AND METHODS: Consecutive patients who underwent RP for localised prostate cancer between 2003 and 2017 were identified from a prospectively recorded, dedicated prostate cancer database. Patients who received neoadjuvant androgen-deprivation therapy were excluded. These patients were categorised into the following groups: no BCR, BCR at <12 months (early), BCR at 12-60 months (intermediate), and BCR at >60 months (late), after RP. Clinicopathological characteristics were analysed using the Student's t-test, Mann-Whitney U-test, or chi-squared test where appropriate. Multivariable binomial logistic regression models were used to assess predictors of BCR at various time-points. RESULTS: In all, 2312 patients were included in the final analysis with up to 12 years of follow-up data. The average patient had clinically localised prostate cancer, an elevated PSA level, and International Society of Urological Pathology (ISUP) Grade Group 2 on biopsy. In all, 88.7% of patients had ISUP Grade Group ≥2 at RP. A subgroup of 446 patients had undetectable PSA levels at 5 years after RP; 11.7% of them progressed to experience BCR. In this subgroup, late recurrers had significantly higher-grade tumours on ISUP and Gleason sum (P <0.001 and P = 0.001, respectively), higher rates of extraprostatic extension (P = 0.022), and larger tumour volumes (P = 0.032). Logistic regression showed that RP ISUP Grade Group was a significant predictor of BCR (odds ratio 2.14, 95% confidence interval 1.43-3.20; P <0.001). CONCLUSION: This study characterises the pattern of late BCR in the largest contemporary active surveillance era cohort. We have identified that RP ISUP Grade Group is a strong predictive indicator for late BCR. We also propose that timing of BCR resides on a continuum of risk and that the potential concept of dormant micrometastatic involvement requires further research and evaluation.