Abstract
Abstract Background: Currently, the usefulness of conventional markers for breast carcinoma is limited to assess individual outcome. A statistical model was developed to improve the prognostic accuracy using multiple conventional and emerging prognostic biomarkers. Methods: A total of 305 node-negative breast carcinomas who underwent surgery (+/− radiotherapy) but no adjuvant treatment was selected. Putative prognosis factors including age, tumor size, ER, PR, SBR grading, urokinase plasminogen activator (uPA), plasminogen activator inhibitor type 1 (PAI-1) and thymidine kinase (TK) were evaluated. The developed model was internally validated using Harrell's concordance index and calibrated. An external validation of the new model is warranted. Results: Age (p<0.001), PR (p= 0.02), and PAI-1(p=<0.001) were included in the Cox regression model predicting overall survival at 5-years. Internal validation revealed a concordance index of 0.711 to 0.694 before and after calibration. Conclusion: A nomogram can be used to predict probability survival curves for individual breast cancer patients and the effect of treatment options can be evaluated using these models. Citation Information: Cancer Res 2011;71(24 Suppl):Abstract nr P4-09-29.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have