Purpose:
PSA monitoring is the main tool for urologists and practitioners to detect prostate cancer relapse after radical prostatectomy, and is recommended in current guidelines for the management of corresponding patients. PSA doubling time (PSADT) is a useful concept to assist PSA results interpretation. However, several calculation options compete, and there is no definite consensus regarding how to best follow up PSA in this condition. The aim of this retrospective observational study was to describe prognostic markers (TNM grade, Gleason score, capsular penetration), PSA trajectories and cancer relapse after radical prostatectomy in a series of patients, and to assess critically the current PSA monitoring approaches, in particular PSADT calculation.
Methods:
Patients were selected from a PSA measurements database of a medical laboratory, collected over 10 years. Inclusion criteria were PSA levels monitored 3 times over 1 year or 4 times over 2 years in the aftermath of a radical prostatectomy for prostate cancer. The clinical data were collected from the medical records of practitioners (remission versus relapse status; dates of interventions; post-surgery pTNM stage, Gleason score and capsular penetration; subsequent biopsies; imaging; chemotherapy, radiotherapy or hormonal treatment). Relapse was defined by distant metastasis or biopsy or imaging showing recurrence, or by the initiation of secondary anticancer treatment. PSADT calculation was made using either the Log-slope method or the 2-points method, once including all available values of PSA, and once selecting only PSA values above 0.1 ng/mL. PSA trajectories were described by longitudinal non-linear mixed effect modelling and a variogram analysis, while their prognostic value was assessed by Kaplan-Meier survival analysis and Cox proportional hazard models.
Results:
PSA trajectories are highly variable and divergent between patients, even after taking into account their T and N grade, individual clinical markers which are strong predictors of relapse. At about 2 years distance, a change in PSA becomes indicative of clinically relevant trend in a fraction of patients (5%), which increases along the time. PSADT calculated with the Log-slope method predicts relapse better than with the 2 point method. Taking into account all available PSA values is also more efficient than restricting the calculation to those above 0.1 ng/mL. Already when estimated at 6 months or 1 year postoperatively, PSADT definitely improves the prediction of relapse in addition to grade T and Nodular Invasion..
Conclusion:
PSA trajectories are characterized by large variability, even accounting for known prognostic markers. Our results fully support the regular follow-up of PSA and the calculation of PSADT for prediction and detection of cancer relapse after radical prostatectomy.