Outcome prediction in estrogen-receptor positive, chemotherapy- and tamoxifen-treated patients with locally advanced breast cancer.

11104 Background: Recurrence Score (RS) assay (Oncotype DX) predicts recurrence and chemotherapy (CMF) benefit in node-negative, estrogen-receptor (ER) positive, tamoxifen-treated patients.,Fifty-seven ER positive, locally advanced breast cancer patients underwent therapy with a doxorubicin-paclitaxel regimen before surgery followed by adjuvant CMF and tamoxifen. RT-PCR assays were used by Genomic Health to quantify expression of 384 genes on paraffin-embedded core biopsies. Excluding RS genes, survival risk groups were constructed using the supervised principal component method implemented in BRB-ArrayTools. Three models were developed: one for expression data, one for RS, age and inflammatory status, and one for a combination of these covariates and expression data. For each model a Leave-One-Out-Cross-Validation was used to evaluate the predictive value. Statistical significance was determined by repeating the entire cross-validation process 1,000 random permutations of the survival data. For the combined covariate-gene model the p value addressed whether the expression data adds significantly to risk prediction compared to the covariates.,The median follow-up was 76 months (range 18-103). The identified models based on expression data only or combined with covariates shared 7 genes (BECN1, STS, IL10, ABCC4, ABCC1, DHPS, ERBB3) while the first model included also, IRS1, FOXA1, ERCC1, ZSCAN21, FUS, HPN, ECGF1. The table reported the predicted 5-years Distant-Event Free-Survival and Overall-Survival for the three models.,These findings suggested a prediction improvement by the combined model over RS and clinical covariates in these patients. An external validation is warranted. [Table: see text] No significant financial relationships to disclose.