Abstract Background CTC count before a new line of treatment and CTC count early changes under chemotherapy have been reported as an independent prognostic marker in metastatic breast cancer in a recent pooled analysis in 841 pts (Liu M. ASCO 2011). The aim of this study was to build a prognostic tool including CTC and other parameters to assesse its predictive value for progression-free survival (PFS) and overall survival (OS). Methods: Data from the IC 2006–04 study were used. This prospective multicentre study included 267 metastatic breast cancer patients treated by first line chemotherapy with or without targeted therapy, in whom appropriate pre-treatment prognostic variables (age, performance status, number of metastatic sites, disease-free interval, ER, PR and HER2 status, tumor grade, LDH, serum markers, CTC count by CellSearch technique before treatment and before cycle 2) were available for statistical analysis. We constructed a multivariate Cox regression model for PFS and OS prediction. A stepwise selection process was applied to achieve the most informative and parsimonious models. Performance was measured with the C-index statistic. Internal validation was performed using leave-two-out technique. Results: Four nomograms have been obtained, in two clinical settings: at inclusion (before the start of any treatment) taking into account the initial CTC count, and during treatment (before cycle 2) taking into account CTC changes under treatment. Their accuracy was good for PFS and OS prediction, with C-index ranging from 0.72 to 0.88. Internal validations allow considering a good accuracy of the models in an external population. Conclusion: These clinically relevant nomograms are a simple tool for a personalized prognostic assessment including CTC assessment. Validation on independent series of patients are ongoing. Citation Information: Cancer Res 2011;71(24 Suppl):Abstract nr P4-07-04.
Read full abstract