Patients with adverse pathological features (APF) at radical prostatectomy (RP) for prostate cancer (PC) are candidates for adjuvant treatment. Clinicians lack reliable markers to predict these APF preoperatively. Protein tyrosine phosphatase 1B (PTP-1B) is involved in migration and invasion of PC, and its expression could predict presence of APF. Our aim was to compare PTP-1B expression in patients with and without APF, and to explore PTP-1B expression as an independent prognostic factor.
Tissue microarrays (TMAs) were constructed using RP archival specimens for immunohistochemical staining of PTP-1B; expression was reported with a standardized score (0-9). We compared median PTP-1B score between cases with and without APF. We constructed two logistic regression models, one to identify the independence of PTP-1B score from biologically associated variables (metformin use and type 2 diabetes mellitus [T2DM]) and the second to seek independence of known risk factors (Gleason score and prostate specific antigen [PSA]).
A total of 73 specimens were suitable for TMA construction. Forty-four (60%) patients had APF. The median PTP-1B score was higher in those with APF: 8 (5-9) vs 5 (3-8) (p=0.026). In the logistic regression model including T2DM and metformin use, the PTP-1B score maintained statistical significance (OR 1.21, 95% CI 1.01-1.45, p=0.037). In the model including PSA and Gleason score; the PTP-1B score showed no independence (OR 1.68, 95% CI 0.97-1.41, p=0.11). The area under the curve to predict APF for the PTP-1B score was 0.65 (95% CI 0.52-0.78, p=0.03), for PSA+Gleason 0.71 (95% CI 0.59-0.82, p=0.03), and for PSA+Gleason+PTP-1B score 0.73 (95% CI 0.61-0.84, p=0.001).
Patients with APF after RP have a higher expression of PTP-1B than those without APF, even after adjusting for T2DM and metformin exposure. PTP-1B has a good accuracy for predicting APF but does not add to known prognostic factors.