New algorithm APhiGT for prostate cancer staging
The proportion of errors in preoperative prostate cancer (PC) staging in modern onco-urological practice is still high due to inaccuracies in tumor size (T) and Gleason score assessment on biopsy specimens. The aim of this study was to develop a new mathematical algorithm for PC staging at the start of the treatment.
The algorithm APhiGT was developed by using logistic regression method. APhiGT is calculated by the formula: APhiGT = 0.01*phi + 0.025*Age + Gleason + 0.6*T, where Age - number of full years; PHI - prostate health index; Gleason - rank (based on biopsy results), has value "1" for Gleason 5-6, "2" - for 7 (3+4), "3" - for 7 (4+3), "4" - for 8-10; T - rank (based on the TNM classification), has value "1" for T2N0 and "2" - for T3N0. The study included 337 primary patients with tPSA < 30.0 ng/ml (Hybritech calibration), for whom the serum levels of fPSA and [-2]proPSA (Beckman Coulter Access 2) were assessed. All patients were characterized by TNM and pTNM, and by Gleason score before and after prostatectomy.
Rate of errors in preoperative Gleason score assessment (<7 vs ≥7 divide) included 32% cases of hypodiagnosis and 8% of hyperdiagnosis. Estimation of the tumor size (T) parameter resulted in hypodiagnosis (T2-pT3) in 37% of cases, and hyperdiagnosis (T3-pT2) - in 44% of cases. "Grey zone" for APhiGT was 3.6 - 4.7. When APhiGT score was <3.6, pT2 status was confirmed in 95.3% of cases. Wherein, at PHI index using (with cut-off <25) pT2 status was confirmed in 88.9% of cases. At the values of APhiGT score >4.7, aggressive PC was diagnosed in 96.9% of cases. Based on ROC-analysis, for patients with PSA total levels 2.5-10.0 ng/ml, advantage of APhiGT over PHI was demonstrated. For discrimination of pGl <7 vs ≥7 AUC the following indicators were used: APhiGT - 0.85; PHI - 0.66; tPSA - 0.57. Ranks pT2 vs pT3 AUC were discriminated using APhiGT - 0.75; PHI - 0.65; and tPSA - 0.61. APhiGT was significantly more accurate than PHI and tPSA including for the range of tPSA <30.0 ng/ml. The calculator program for practical use of APhiGT model was developed.</p>
Our results indicate the prospects of preoperative PC staging improvement using new algorithm APhiGT.