Ki-67, topoisomerase IIα and miR-221 have a limited prostate cancer risk stratification ability on a medium-term follow-up: results of a high-risk radical prostatectomy cohort.

Background: Currently, no biomarkers are able to differentiate lethal from relatively indolent prostate cancer (PCa) within high-risk diseases. Nonetheless, several molecules are under investigation. Amongst them, topoisomerase-II-alpha (), and showed promising results. Our aim was to investigate their prognostic role in the context of biochemical recurrence (BCR), clinical recurrence (CR) and PCa-related death (PcD).

Methods: We included 64 consecutive cM0 high-risk PCa [prostate specific antigen (PSA) >20 ng/mL or Gleason Score (GS) >7 or cT >2] undergoing radical prostatectomy (RP). Changes in miR-221 expression and alternative splicing were determined using microarrays. Immunohistochemical determination of Ki67 and TOPIIa were performed using monoclonal antibody MIB-1 and 3F6 respectively. Cox proportional-hazards regression models were used to predict BCR and CR as multivariate analysis. BCR and CR were defined as three consecutive rises in PSA and PSA >0.2 ng/mL and histologically-proven local recurrence or imaging positive for distant metastasis respectively.

Results: We included 64 men. Mean pre-operative PSA was 26.53 (range, 1.3-135); all GSs were ≥7 and pT was ≥ T3 in 78.13%. Positive margins and lymph-nodes were present in 42.19% and 32.81% respectively. At a mean follow-up of 5.7 years (range, 1.8-12.5), 42.18% experienced BCR (n=27), 29.68% CR (n=19) and 7.81% PcD (n=5). On univariate analysis positive nodes (<0.01), seminal vesicle invasion (0.02) and downregulation (P=0.03), but not and (both P>0.5) were associated with BCR whereas only PSA (P<0.01), seminal vesicle invasion (P<0.01) and positive nodes (both P<0.01) were linked to CR. No parameters predicted PcD (all P>0.05) or BCR and CR on multivariate analysis (all P>0.05 - miR-221 HR 0.776; 95% CI: 0.503-1.196 for BCR and HR 0.673; 95% CI: 0.412-1.099 for CR). Limitation of the study include its small sample size and limited follow-up.

Conclusions: , and may not predict BCR, CR or PcD in high-risk PCa patients who underwent RP at a medium-term follow-up. Longer follow-up and larger cohorts are needed to confirm our findings.