Portal de Periódicos da CAPES

Inclusion of corrected calcium and creatinine levels in a novel nomogram for prediction of pathologically localized prostate cancer.

2016; Lippincott Williams & Wilkins; Volume: 34; Issue: 2_suppl Linguagem: Inglês

10.1200/jco.2016.34.2_suppl.140

1527-7755

Okyaz Eminağa, Omran Al-Hamad, U. Engelmann, Sebastian Wille, Axel Heidenreich,

Cancer, Lipids, and Metabolism

140 Background: The role of routine parameters in the prediction of organ-confined Prostate Cancer (oPCa) remains unclear. Methods: We conducted a retrospective study of 246 patients with PCa diagnosed on biopsy, who underwent radical prostatectomy with extended lymph node dissection. None of these patients received a neoadjuvant therapy. Gleason score (xGLS), percent of positive cores involved by PCa (pCores), results of transrectal evaluation (TRUS, DRE), prostate volume (PVol), albumin-corrected Calcium (Ca), Hemoglobin (Hb), Creatinine (Crea), PSA, and age at RPE were considered. Postoperative variables including Gleason score (pGLS), TNM stages were evaluated. These patients were divided according to disease advancement into pathologically localized (pT2) and advanced PCa (pT3/4 or pN1). A L2 penalized regression model including all preoperative parameters was evaluated to design a nomogram for oPCa prediction. Furthermore, the internal validation of the nomogram was performed using bootstrapping and cross-validations. The precision and accuracy of the novel nomogram were evaluated by using AUC, F-Score, Brier-Score and the classification accuracy (CA). Results: The lowest probability for pathologically localized PCa was found in patients having the following features: PSA > 20 ng/ml, detection of PCa in all cores, high Ca [1.21 fold over upper normal limit (ULN)] and Crea levels (2.27 fold over ULN), xGls > 7 and normal PVol. A nomogram for oPCa was developed including the following parameter: Crea, Ca, PSA, xGls, PVol, TRUS, pCores. The sensitivity and the specificity of the nomogram for prediction of pathologically localized PCa were 92.2% and 35.9%, respectively. AUC of 0.811, CA of 80.1%, F-Score 0.879, a precision of 84.0%. and Brier-Score of 0.2599 were calculated. The predicted probability of the novel nomogram was associated with higher accuracy for prediction of oPCa compared to those of Partin table and Kattan nomogram. Conclusions: Serum creatinine and corrected calcium are indpendent predictors of oPCa. The novel nomogram can predict oPCa with high precision and accuracy. However, an external validation of the novel nomogram is needed.