Predictive model for capsule penetration in prostate cancer using biomarkers.

Introduction: Since the beginning, when proteomics has demonstrated it's ability to characterize a large number of proteins, their measurement and post-translational modifications, this has been applied in various fields of science. Objective: The objective of the present work is to demonstrate the prognostic value of biomarkers and to select those biomarkers with an impact on the evolution of patients after prostatectomy with radical intent Protocol design: A total number of 218 patients were considered, prospectively, from June 2011 - January 2014, this group fully respecting the inclusion and exclusion criteria. Data extracted from the selected patient records were organized first in an Excel date base. Through successive filtering and sorting were identified and completed the missing or incorrectly transcribed data. After ensuring data consistency, they were imported into SPSS Statistics 17.0.program Results: We can observe that for Telomerase and Hepsin confidence intervals for the odds ratio are between 0,490 and 1,964 between 0,570 and 3,549 respectively. Containing the value 1, the two parameters have no prognostic power. Interleukin 6 has the highest prognostic power, when exceeding the threshold value with an odds ratio of 3.269 , then with close values, the Vascular Endothelial Growth Factor with Odds Ratio of 2.24 and Transforming Growth Factor beta fraction 1 with Odds Ratio of 2.21. By removing the two prognostic powerless biomarkers the correct classification model of this capsular penetration is improved from 68.3% to 69.3%. Conclusions: We found a strong link between increasing the value of a single biomarker above the cuoff value and the presence of capsular penetration and a very strong link between increasing values above the cutoff of 3 parameters, the Odds Ratio being equal with 10.65 and an interval between 3.62 and 31.31. We consider that this paper has a consistent dose of originality, the scientific context being favorable to the development of predictive models and concept of "individual medicine". Acknowledgement: This paper is supported by the Sectorial Operational Programme Human Resources Development (SOP HRD) 2007-2013, financed from the European Social Fund and by the Romanian Government under the contract number POSDRU/107/1.5/S/82839. [ABSTRACT FROM AUTHOR]