Presentation Authors: Vincent Gnanapragasam*, Tristan Barrett, Vineetha Thankappannair, David Thurtle, Cambridge, United Kingdom, Ola Bratt, Gothenburg, Sweden, Par Stattin, Uppsala, Sweden, Kenneth Muir, Artitaya Lophatananon, Manchester, United Kingdom
Introduction: Active surveillance (AS) is a well-established option for men diagnosed with early prostate cancer. There is however no consensus on the ideal inclusion and exclusion criteria, i.e. what defines disease that is unsafe to monitor, or how to optimally survey men. Here we tested whether using prognosis can guide a rational approach to AS.
Methods: We have previously developed and validated the 5 tiered-Cambridge Prognostics Groups (CPG) that predicts disease prognosis at diagnosis (Gnanapragasam et al 2018 doi: 10.1186/s12916-018-1019-5). The first 3 groups (CPG1-3) were of interest in this study defined as: _x000D_
CPG1: Grade Group (GG) 1 AND PSA < 10 AND Stages T1 to T2, CPG2: GG 2 OR PSA 10 to 20 AND Stages T1 to T2, CPG3: GG2 AND PSA 10 to 20 ng/ml AND Stages T1 to T2 OR GG3 AND Stages T1 to T2. _x000D_
These strata map closely to the new AUA low, favourable and unfavourable intermediate-risk groups respectively. The model was applied to UK and Swedish historical cohorts to test differences in 10-year prostate cancer mortality (PCM) in men managed conservatively or by upfront radical therapy. We then applied the model to a contemporary mpMRI biopsy diagnosed and monitored AS cohort to identify predictors of pathological progression.
Results: The UK cohort (n=3659) included 1299, 1413 and 947 men in CPG1, 2 and 3 respectively. The cumulative PCM in CPG1 was 2.3%. The corresponding rates for treated and untreated men was 1.5% & 3.5% in CPG2 and 1.9% & 8.6% in CPG3. Similar results were seen in a Swedish cohort (n=27,942) with 15,477, 8495 and 3970 men in CPG1, 2 and 3 respectively. 10-year PCM was 1.0%, 2.2% and 2.7% in CPG1, CPG2 untreated and CPG2 treated men respectively. In contrast, untreated men in CPG3 had a 10-year PCM of 12.5% versus 6.1% in treated men. Thus, men in CPG1&2 consistently had mortality rates < 5% regardless of intervention. We next tested using progression to CPG3 as a standardized endpoint for AS. In a well characterized mpMRI based AS cohort (n=133), only 8 men (6%) progressed to CPG3 over a median of 3.5 years. Key diagnostic predictors of progression were a PSA density (PSAd) â‰¥0.2 and/or a positive mpMRI scan at diagnosis (â‰¥Likert 3). No one with a PSAd < 0.2 and negative mpMRI progressed to CPG3. In contrast, the presence of both PSAdâ‰¥0.2 and positive mpMRI conferred a 27% progression risk. Men with only one of these had an intermediate risk of progression. This identities a 3-tiered surveillance strategy with escalating intensity of out-patient visits, surveillance mpMRI and biopsies informed by individual progression risk.
Conclusions: Our results suggest that AS can be stratified by prognosis and risk of progression. We propose a 3-tiered follow-up strategy, based on CPG, PSA density and mpMRI, that would reduce out-patient visits, surveillance scans and repeat biopsies. Prospective studies are needed to test whether such risk stratified AS can reduce over-monitoring without compromising safety.