Nikita Sushentsev, Anne Y. Warren, Richard Colling, Clare Verrill, Ekaterina Pazukhina, Oleg Blyuss, Alexey Zaikin, Tyler M. Seibert, Tristan Barrett, Jon Oxley, Ian G. Mills, Richard J. Bryant, J. Athene Lane, J Donovan, David E. Neal, Freddie C. Hamdy
OBJECTIVES: To retrospectively analyse the results of the Prostate Testing for Cancer and Treatment (ProtecT; ClinicalTrials.gov identifier: NCT02044172) trial to establish the association between cribriform-positive and -negative prostate cancer (PCa) and the 15-year risk of metastasis or death from PCa in patients who underwent radical prostatectomy (RP). PATIENTS AND METHODS: Between 1999 and 2009, the ProtecT phase 3 clinical trial enrolled 1643 men with clinically localised PCa who were randomised to receive active monitoring, RP, or radiotherapy. In this secondary analysis of the trial, a centralised histopathological review was conducted on available RP pathology slides to classify patients as cribriform-positive if they had invasive cribriform carcinoma and/or intraductal carcinoma. The primary outcome was a composite of progression to metastatic disease or death from PCa. Exposures included age, prostate-specific antigen density, RP Grade Group (GG), pathological T stage (pT), and cribriform status. Multivariable Cox proportional hazards regression models assessed 15-year risk. Cumulative incidence curves were compared using the Gray test. RESULTS: Of 480 men with RP specimens reviewed, 143 (30%) had cribriform-positive disease and 337 (70%) had cribriform-negative disease. All 21 metastatic or lethal events occurred exclusively in the cribriform-positive group (15-year cumulative incidence 14%). Within the cribriform-positive cohort, risk was concentrated in patients with pT3b stage and/or GG ≥3 (15-year cumulative incidence 27%). In multivariable analysis of cribriform-positive patients, pT3b stage (hazard ratio [HR] 8.19, 95% confidence interval [CI] 2.39-28.10; P < 0.001) and GG 3 disease (HR 5.12, 95% CI 1.59-16.40; P = 0.006) were independent predictors of adverse outcomes. Conversely, cribriform-positive patients with GG 2 and ≤pT3a had a 15-year event rate of only 3%. CONCLUSION: In the ProtecT trial, the 15-year risk of metastasis or death after RP was a binary outcome defined by cribriform status. The concentration of risk in men with cribriform-positive, high-grade and/or pT3b tumours identifies a target population for adjuvant therapy trials, while supporting management de-escalation for most RP patients.