Role of PSA levels and pathological stadiation before radiation therapy in predicting mp-MRI results in patients with prostate cancer recurrence after radical prostatectomy

Objective: To evaluate prostate-specific antigen (PSA) value in multiparametric magnetic resonance imagin (mp-MRI) results prediction, analyzing patients with high (Gleason Score ≥ 8, pT ≥ 3, pN1) and low grade (Gleason Score < 8, pT < 3, pN0) prostate cancer (PCa). Materials and methods: One hundred eighty-eight patients underwent 1.5-T mp-MRI after radical prostatectomy and before radiotherapy. They were divided into 2 groups: A and B, for patients with biochemical recurrence (BCR) and without BCR but with high local recurrence risk. Considering Gleason Score, pT and pN as independent grouping variables, ROC analyses of PSA levels at primary PCa diagnosis and PSA before radiotherapy were performed in order to identify the optimal cut-off to predict mp-MRI result. Results: Group A and B showed higher area under the curve for PSA before radiotherapy than PSA at PCa diagnosis, in low and high grade tumors. For low grade tumors the best area under the curve was 0.646 and 0.685 in group A and B; for high grade the best area under the curve was 0.705 and 1 in group A and B, respectively. For low grade tumors the best PSA cut-off was 0.565-0.58 ng/ml in group A (sensitivity, specificity: 70.5%, 66%), and 0.11-0.13 ng/ml in B (sensitivity, specificity: 62.5%, 84.6%). For high grade tumors, the best PSA cut-off obtained was 0.265-0.305 ng/ml in group A (sensitivity, specificity: 95%, 42.1%), and 0.13-0.15 ng/ml in B (sensitivity, specificity: 100%). Conclusion: Mp-MRI should be performed as added diagnostic tool always when a BCR is detected, especially in high grade PCa. In patients without BCR, mp-MRI results, although poorly related to pathological stadiation, still have a good diagnostic performance, mostly when PSA > 0.1-0.15 ng/ml.