Data from Development of Candidate Genomic Markers to Select Breast Cancer Patients for Dasatinib Therapy

<div>Abstract<p>Patient selection is important for targeted therapies, yet phase I/II trials are often underpowered for developing predictors of drug response. The goal of this research was to define genomic predictors for dasatinib that could be prospectively tested in early-phase clinical trials.</p><p>Gene expression profiles of dasatinib-sensitive and dasatinib-resistant cell lines (<i>n</i> = 23) were compared to develop a dasatinib-sensitivity index (modified DS index). A Src pathway activity index (revised Src index) was defined using genes induced by the Src transfection of mammary epithelial cells and was optimized to be reproducible across cell lines and human specimens. A dasatinib target index was devised using the weighted sum of 19 kinases that bind to dasatinib with variable affinity. The performance of these prediction models was assessed in independent cell lines with known dasatinib sensitivity. The feasibility of applying these genomic tests to human samples was evaluated on 133 biopsies of primary breast cancers.</p><p>The modified DS index showed 90% accuracy in independent breast cancer cell lines (<i>n</i> = 12) and the target index, but not the revised Src index signature, also distinguished dasatinib-sensitive and dasatinib-resistant cells (<i>P</i> = 0.0024). The genomic predictors showed acceptable reproducibility in replicate cell line and human gene expression data. When all three predictors were applied to the same 133 patient samples, the predictors identified different patient subsets as potentially sensitive.</p><p>We defined three conceptually different potential predictors of dasatinib response that were reproducible across cell lines and human data. These candidate markers are being tested in a clinical trial to determine their utility. Mol Cancer Ther; 9(5); 1120–7. ©2010 AACR.</p></div>