Objectives: Our aim is to design a predictive model of seminal vesicle involvement. using clinical data.
Methods: We studied 1128 patients with clinically localized adenocarcinoma treated by radical prostatectomy (127 were pT3b). We identified (logistic regression) clinical variables related with pT3b. With the multivariate study influential variables a seminal vesicle involvement risk model is designed.
Results: Seminal vesicle involvement related factors: In univariate study: the influential variables are: Gleason 7 (OR:2);Gleason 8-10 (OR:4.5) T2 (OR:2.6); bilateral involvement in biopsy (OR:3.1); PSA 10-20 ng/ml ( OR:3.3); PSA >20 ng/ ml (OR:9.5). In the multivariate study are influential: Gleason 7 (OR:1.56) Gleason 8-10 ( OR: 3.4); T2 (OR:1.9); PSA 10-20 ng/ml (OR:3.1) and PSA >20 ng/,ml (OR:8.8). Predictive model: using multivariate logistic regression the weight of each variable is valued and a value between 1 and 4 is given. Gleason 2-6, T1; PSA<10 ng/ml value 1; Gleason 7; T2 y PSA 10-20 ng/ml value 2; Gleason 8-10 and PSA >20 ng/ml value 4. Each patient has a marker that fluctuates between 3 and 10. 5 Groups are designed with significantly different risks (p<0.05 in all cases ): Group 1 (3 points) (OR:1) (risk: 2.4% 95%IC 0.7%-4.3%) Group 2 (4 points) (OR:2.7) (risk: 6.5% 95%IC 5%-7.9%); Group 3(5-6 points) (OR:7.1)( risk:15% 95%IC 11%-19%) Group 4 ( 7--8 points) (OR:33.4) (risk: 45.5%; 95%IC 30%-59%) Group 5 (9-10 points) (OR:57.3) (risk: 58.8% 95%IC 35%- 82%).
Conclusion: The clinical model allows an accurate approximation to the seminal vesicles involvement risk.