Investigative clinical study on prostate cancer part VIII: prolactin hormone and the pituitary-testicular-prostate axis at the time of initial diagnosis and subsequent cluster selection of the patient population after radical prostatectomy.

AIM

To evaluate the prolactin hormone (PRL) physiopathology along the pituitary testicular prostate axis at the time of initial diagnosis of prostate cancer and the subsequent cluster selection of the patient population after radical prostatectomy in relation to clinical and pathological variables.

PATIENTS AND METHODS

Ninety-two operated prostate cancer patients were retrospectively reviewed. No patient had previously received hormonal treatment. The investigated variables included PRL, follicle stimulating hormone (FSH), luteinizing hormone (LH), total testosterone (TT), free testosterone (FT), total prostate specific antigen (PSA), percentage of positive cores at transrectal ultrasound scan biopsy (TRUSB) (P+), biopsy Gleason score (bGS), pathology Gleason score (pGS), estimated tumor volume in relation to percentage of prostate volume (V+), overall prostate weight (Wi) and age. Empirical PRL correlations and multiple linear predictions were investigated along the pituitary testis prostate axis in the different groups of the prostate cancer population and clustered according to pT (2a/b, 3a, 3b/4) status. The patient population was classified according to the log(10) PRL/V+ ratio and clustered as follows: group A (log(10) PRL/V+ ≤1.5), B (1.5< log(10)PRL/V+ ≤2.0) and C (log(10) PRL/V+ >2.0). Simple linear regression analysis of V+ predicting PRL was computed for assessing the clustered model and analysis of variance was performed for assessing significant differences between the groups.

RESULTS

PRL was independently predicted by FSH (p=0.01), LH (p=0.008) and P+ (p=0.06) in low-stage prostate cancer (pT2a/b). Interestingly, PRL was independently predicted by LH (p=0.03) and FSH, TT, FT, PSA, bGS, pGS, V+, Wi and age (all at p=0.01) in advanced stage-disease (pT3b/4). V+ was also significantly correlated (r=0.47) and predicted by P+ (p<0.0001) in the prostate cancer population. PRL was significantly correlated and predicted by V+ when the patient population was clustered according to the log(10)PRL/V+ ratio in group A (p=0.008), B (p<0.0001) and C (p<0.0001). Moreover, the three groups had significantly different mean values of PRL (p<0.0001), PSA (p=0.007), P+ (p=0.0001), V+ (p<0.0001), Wi (p=0.03), bGS (p=0.008), pGS (p=0.003); also, groups A, B and C had significant different pGS (p=0.03), pT (p=0.0008) and pR (p=0.01) frequency distributions.

CONCLUSION

At diagnosis, in an operated prostate cancer population, PRL was significantly correlated and independently predicted along the pituitary testis prostate axis in high-stage disease; V+ was also significantly correlated and predicted by P+. Because of the high correlation and prediction of PRL by both V+ and P+, the prostate cancer population at diagnosis was clustered according to the log(10)PRL/V+ ratio into groups A, B and C that, in theory, might be models with prognostic potential and clinical applications in the prostate cancer population. However, confirmatory studies are needed.