Reclassification prediction of first-year protocol biopsy on active surveillance of prostate cancer by p2PSA-related parameters: from PRIAS-JAPAN

Abstract

Background There is no useful predictive marker for reclassification on active surveillance. Thus, we aimed to investigate thresholds of [−2] proPSA (p2PSA)-related parameters to predict reclassification of the first-year protocol biopsy (1-year PBx) and evaluate the influence of clinical decision-making. Methods This was an observational, prospective cohort study conducted at 19 Japanese institutes. The inclusion criteria included clinical stage T1c/T2, prostate-specific antigen (PSA) levels ≤10 ng/mL, PSA density <0.2 ng/ml/cc, one or two positive biopsy cores, and Gleason score (GS) ≤6 (GS ≦7 for patients aged ≥70 years) at diagnostic biopsy. All participants were required to receive a blood-sampling test on a protocol visit at inclusion and at the 1-year PBx. PSA and PSA isoforms (free PSA, p2PSA) were measured, and parameters (%free PSA, %p2PSA, phi) were calculated. Multivariable logistic regression models were used to predict the reclassification risk. To assess the predictive power and thresholds for reclassification, we plotted Receiver Operating Characteristic (ROC) curves. Decision curve analysis (DCA) was used to evaluate the variables that yielded a net clinical benefit. Results A total of 135 patients were included, and 36 patients were reclassified on the 1-year PBx. Multivariate analyses showed that %p2PSA and phi at inclusion and p2PSA, %p2PSA, and phi before the 1-year PBx were significant predictors of reclassification at the 1-year PBx. The ROC analysis showed an optimal cutoff point, sensitivity, and specificity of %p2PSA and phi before the 1-year PBx of 1.64, 86%, 49% and 35.92, 89%, 47%, respectively. The DCA showed that phi before the 1-year PBx had the highest net benefit. The study limitation was its single-arm observational design. Conclusions %p2PSA and phi before the 1-year PBx had a good prediction power. phi is the most useful indicator for clinical decision-making on active surveillance. Trial registration This study is registered atthe Japan Trial Register with ID UMIN000009876 (https://upload.umin.ac.jp/cgi-open-bin/ctr_e/ctr_view.cgi?recptno=R000011573).