Investigative clinical study on prostate cancer part IX and X: estradiol and the pituitary-testicular-prostate axis at the time of initial diagnosis and subsequent cluster selection of the patient population after radical prostatectomy

Abstract AIM: To evaluate estradiol (E(2)) physiopathology along the pituitary-testicular-prostate axis at the time of initial diagnosis of prostate cancer (PC) and subsequent cluster selection of the patient population. PATIENTS AND METHODS: Records of the diagnosed (n=105) and operated (n=91) patients were retrospectively reviewed. Age, percentage of positive cores at-biopsy (P+), biopsy Gleason score (bGS), E(2), prolactin (PRL), luteinizing hormone (LH), follicle-stimulating hormone (FSH), total testosterone (TT), free-testosterone (FT), prostate-specific antigen (PSA), pathology Gleason score (pGS), estimated tumor volume in relation to percentage of prostate volume (V+), overall prostate weight (Wi), clinical stage (cT), biopsy Gleason pattern (bGP) and pathology stage (pT), were the investigated variables. None of the patients had previously undergone hormonal manipulations. E(2) correlation and prediction by multiple linear regression analysis (MLRA) was performed. At diagnosis, the log E(2)/log bGS ratio clustered the population into groups A (log E(2)/log bGS ≤ 2.25), B (2.25<log E(2)/log bGS ≤ 2.48) and C (2.48< log E(2)/log bGS ≤ 2.59). The operated population was clustered according to the log E(2)/log pGS ratio into groups A (log E(2)/log pGS ≤ 2.25), B (2.25< log E(2)/log pGS ≤ 2.48) and C (2.48< log E(2)/log pGS ≤ 2.59). Simple linear regression analysis of bGS and pGS predicting E(2) was computed; differences between the clusters were assessed by analysis of variance (ANOVA) and by contingency tables. RESULTS: At diagnosis, E(2) was correlated to TT (r=0.32, p=0.0006) and FT (r=0.25, p=0.0009); moreover, E(2) was independently-predicted by TT (p=0.009) and bGS (p=0.04) on MLRA. The bGS significantly predicted E(2) in all groups. Groups A, B and C differed in mean values for E(2) (p<0.0001), TT (p=0.005), FT (p=0.05), P+ (p=0.01) and bGS (p=0.003); moreover, the frequencies of the different bGPs were significantly different in the three groups (p=0.004). Interestingly, groups A, B, and C were associated with high-, intermediate- and low-bGS tumor grade, as well as with low-, intermediate- and high-serum levels of E(2), TT and FT, respectively. In the operated population, E(2) significantly correlated to FSH (r=-0.20, p=0.04), TT (r=0.34, p=0.0008), FT (r=0.29, p=0.003), bGS (r=0.22, p=0.03) and V+ (r=0.26, p=0.01); moreover, E(2) was independently-predicted by TT (p=0.05) and bGS (p=0.03) on MLRA. The pGS significantly predicted E(2) in all groups that differed for mean values of E(2) (p<0.0001), TT (p=0.004), FT (p=0.002) and pGS (p=0.007), as well as for pT (p<0.0001) and pGS (p=0.008) frequencies. Interestingly, clusters A, B, and C were associated with high-, intermediate- and low-pGS-pT frequencies as well as with low-, intermediate- and high-mean serum levels of E(2), TT and FT, respectively. CONCLUSION: In a diagnosed- and operated-PC population, E(2) serum levels were functionally related along the pituitary-testis-prostate cancer axis; also the log E(2)/log bGS and log E(2)/log pGS ratio, clustered the population in three groups where the risk of progression might be ranked as high (group A), intermediate (group B) and low (group C). However, further confirmatory studies are needed