Prospective randomized controlled study of the role of PSA and PCA3 testing in a sequential manner in an opportunistic screening program for early diagnosis of prostate cancer

INTRODUCTION & OBJECTIVES: Prostate cancer (PCa) screening is under debate and clinical research is needed to optimize it. PCA3 at a cut-off-35 has not yet been fully validated in an opportunistic screening scenario for prostate cancer. Our objective is to evaluate the role of PCA3 as a second line biomarker in an opportunistic screening programme, with the aim of avoiding unnecessary biopsies (Bx). MATERIAL & METHODS: From September-2010 until December-2013, 3865 men, aged 40-75 years and with>10 years life expectancy, were initially screened with PSA and digital rectal examination (DRE). Men with normal DRE and PSA ≥ 3 ng/ml underwent PCA3 testing. Men with previous Bx or recent urinary tract infections were excluded. Written informed consent was obtained and the Local Ethics Committee approved the programme. All men with PCA3 ≥ 35 underwent an initial biopsy (IBx) -12cores-. Men with PCA3 < 35 were randomized 1:1 to either IBx or observation. Re-biopsy (16-18 cores) criteria were PSA increase > 0.5 ng/ml at 4-6months or PSAv > 0.75ng/ml/year. ERSPC risk calculator-3, which combines DRE, PSA and estimated prostate volume without using sonography, is taken as basal model to compare with our protocol. RESULTS: PCA3 was performed in 530 men (13.7%). In the PCA3+ arm (n=182, 34.3%), PCa was identified in 71 men at IBx (39.0%). In the randomized arm (n=348, 65.7%), 191 were observed and 157 underwent biopsy, finding 20 PCa (12.7%), (p<0.001). In PCA3+ and PCA3- arms, global PCa detection rates were 41.7% and 10.3% during a median of 19.3m of FU (p<0.001).The application of this dual protocol or ERSPC risk calculator-3 would have saved 65.7% and 31.9% IBx, respectively. The dual protocol would have missed 10.3% and 3.1% PCa and high grade PCa (HGPCa) during the mentioned FU among men with normal DRE and PSA ≥ 3ng/ml. To avoid the inherent verification bias, inverse probability weighting (IPW) method with the known verification probability in each arm is used. For the used cut-off of 35, PCA3 has a corrected sensitivity and specificity of 61.6% (CI95% 50.1-73.0) and 73.2% (CI95% 68.9-77.6) respectively. The corrected AUC for PCA3 was 0.740 (0.708-0.772). This is an ongoing prospective study limited by its medium follow-up period. Cost analysis is pending. CONCLUSIONS: We show real life results of using PCA3 in an opportunistic screening scenario as a single added variable. If PCA3 at a cut off 35 is used as a second line biomarker after PSA ≥ 3 ng/ml and a normal DRE, it can potentially save 33.8% more IBx than the ERSPC risk calculator-3, just missing 3.1% of HGPCa in the medium-short run.