AUA 2018: Duration of PSA Surveillance after Radical Prostatectomy: A Risk Adapted Approach  

San Francisco, CA (UroToday.com) How long do we check PSA levels after a patient undergoes a radical prostatectomy? With patients living longer than ever after definitive local therapy for prostate cancer, this is an important clinical question. In fact, a literature search revealed little/no definitive timeline to when we may safely refrain from continuing to check PSA after radical prostatectomy.  In clinical practice, most urologists will typically check a PSA once a year, whereas some will decrease this to every two years after a period of undetectable PSA. 


At AUA 2018, Dr. Westerman and colleagues from the Mayo Clinic discussed their risk adapted approach to duration of PSA surveillance after radical prostatectomy. Certainly, the risk of a detectable PSA following radical prostatectomy varies based on clinical and pathologic features, while risk of non-cancer death evolves as a function of age and comorbidities. Current guidelines recommend surveillance based on stage alone with no clear guidance on when to stop. As such, the author’s objective was to develop a risk adapted recommendation regarding postoperative surveillance duration.  

The authors identified 7,250 men with localized prostate cancer who underwent radical prostatectomy between 2001 and 2014 using the Mayo Clinic institutional prostatectomy registry. Using accelerated failure time (AFT) models, the authors estimated risk of (i) PSA ≥ 0.4, (ii) PSA ≥1, and (iii) systemic progression/local recurrence/treatment over 10 years stratified by risk group (low, intermediate, and high). Men were excluded if they had a detectable PSA at their first recheck. Low risk was defined as pT2a, PSA <10, and pathologic Gleason score ≤6, while ≥pT3, PSA > 20, pathologic Gleason score 8-10, or positive surgical margins were defined as high risk. The remainder were categorized as intermediate risk. Risk of non-cancer death was estimated by age (<55, 55-59, 60-64, 65-69, ≥70) and comorbidity index (CCI: 0 versus ≥1). Cumulative hazard rates for each endpoint were plotted along with risk of non-cancer death to determine points of intersection. Surveillance duration for each patient was calculated based on the hazard ratio of non-cancer specific mortality as determined by age and CCI. 

Surveillance duration for all stage, age, and comorbidity risk groups was generated for each endpoint. Based on their analysis, Dr. Westerman suggested that generally all men with CCI 0 and intermediate or high-risk disease should be surveilled throughout a 10-year follow-up period, whereas only men with high-risk disease and CCI ≥1 should have continual surveillance. The authors provided two examples to further delineate their results: 

  • Example #1: 60-year-old male with CCI 0, low risk disease, and a surveillance goal of detecting a PSA ≥0.4 would be followed for 6.5 years. After this time, his risk of death from any cause exceeds the likelihood of a PSA recurrence.  
  • Example #2: Male aged ≤55 years, with a CCI ≥ 1 and intermediate risk disease has a higher recurrence risk than all-cause mortality risk during the 10-year period and should continue surveillance.  
With this study, the authors provide easy to follow models for estimating risk of prostate cancer recurrence and risk of death using risk-adapted recommendations for how long to follow PSA following radical prostatectomy. Certainly, these should be used in clinical practice. However, several limitations of study were discussed with the authors, including the lack of granularity for adaptive screening among those with CCI ≥1 (ie. should risk-adaptive screening be further defined for a patient with CCI 2 vs CCI 6?), and we do not know how long we should follow PSA for patients >10 years after their radical prostatectomy, as this model was limited to 10-years of follow-up.  

Presented By: Mary Westerman, Mayo Clinic, Rochester, MN 
Co-Authors: Matthew T. Gettman, R. Jeffrey Karnes, Igor Frank, Rochester, MN 

Written by:  Zachary Klaassen, MD, Urologic Oncology Fellow, University of Toronto, Princess Margaret Cancer Centre Twitter: @zklaassen_md at the 2018 AUA Annual Meeting - May 18 - 21, 2018 – San Francisco, CA USA