Selection and management of men for active surveillance in low risk prostate cancer

Abstract

Aims:

To investigate: 1. Selection of men for active surveillance of prostate cancer a. Validation of risk calculators b. Suitability for inclusion of Gleason 3+4 disease. 2. Performance of prostate biopsy during AS a. Differences in quality of diagnostic biopsy between academic and referral centres. b. Optimization of biopsy templates c. Examination of prognostic indicators for disease progression

Methods:

Data were obtained from several difference sources: • Men suitable for AS on prostate biopsy but undergoing upfront radical prostatectomy were pooled from 3 international academic institutions in Cambridge (UK), Toronto (Canada) and Melbourne (Australia). • Prospectively maintained AS prostate cancer database at Princess Margaret Cancer Centre (PMCC) (1997-2012). Analyses performed: • Four risk calculators were assessed for their ability to predict different definitions of insignificant prostate cancer by area under the curve (AUC) of receiver operating characteristic curves and Brier scores for discrimination, calibration curves and decision curve analysis. • Men with biopsy Gleason 3+4 disease, suitable according to modified Royal Marsden, Sunnybrook Toronto and PRIAS selection criteria, were assessed for presence of adverse pathology at upfront radical prostatectomy. • Patients on AS at a tertiary referral centre (PMCC) were dichotomized depending on where their diagnostic biopsy was performed (interval versus external). Multivariate logistic regression was performed to examine for predictors of re-classification at the second, or confirmatory, biopsy. • Mapping of all patients with pathological progression at PMCC for location of disease progression enabled comparison of hypothetical biopsy templates (sextant and standard extended) to the institutional template used. • Men on AS at PMCC were evaluated for presence of disease progression at serial biopsy in the prostate transition zone (TZ). Multivariate Cox proportional hazards regression evaluated predictors of TZ progression. • At PMCC, men were dichotomized based on presence of cancer at their confirmatory biopsy. Pathological progression was investigated using a Cox proportional hazards regression model.

Results:

• All 4 models predicting presence of insignificant prostate cancer had weak discrimination at best (AUC 0.618-0.664). • Presence of Gleason 3+4 at biopsy, compared to 3+3 disease, increases risk of adverse pathology at radical prostatectomy if modified Sunnybrook Toronto criteria are used (19% versus 33%, p≤0.001). Using a stricter protocol such as PRIAS, there was no statistical difference between the groups. • External biopsy predicted both grade related re-classification (OR 4.14, C.I. 2.01-8.54, p<0.001) and volume related re-classification (OR 3.43, C.I. 1.87-6.25, p<0.001). • Sextant and standard extended biopsy templates were inferior to the institutional biopsy template in detecting presence of cancer (84% and 99% versus 100%), and pathological progression (47.9% and 81.9% versus 100%). • At each subsequent biopsy during AS, 2.7-6.7% of men had disease progression only in the TZ which would not have been detected if TZ biopsy was not performed. Predictors of TZ progression were maximum % single core (HR 1.99, C.I. 1.30-3.04, p=0.002), and MRI reporting cancer (HR 3.19, C.I. 1.23-8.27, p=0.02). • Men with no cancer at confirmatory biopsy were less likely to have pathological progression (HR 0.47, CI 0.29-0.77, p=0.003). Sub-analysis showed this was predictive of volume-related progression (HR=0.36, CI 0.20-0.62, p=0.0006) and not grade-related progression.

Conclusions:

• Utilization of models predicting suitability for AS should be used with caution as external validation in our cohort was weak. • If considering biopsy Gleason 3+4 disease for AS, a stricter protocol such as PRIAS must be utilized. • At PMCC, patients who had their initial diagnostic prostate biopsy for AS done externally, were more likely to have worse pathological features and re-classify on the second biopsy. • For men on AS, sextant and standard extended biopsy are less likely to detect prostate cancer or disease progression than the template used at PMCC. • TZ biopsy should be considered for all men having serial biopsy on AS, in particular those with high % core involvement or positive MRI findings. • Absence of cancer on B2 is associated with a significantly decreased risk of volume-related but not grade-related progression.