Introduction:TAVO412, a multi-specific antibody targeting epidermal growth factor receptor (EGFR), mesenchymal epithelial transition factor (c-Met), and vascular endothelial growth factor A (VEGF-A), is undergoing clinical development for the treatment of solid tumors. TAVO412 has multiple mechanisms of action for tumor growth inhibition that include shutting down the EGFR, c-Met, and VEGF signaling pathways, having enhanced Fc effector functions, addressing drug resistance that can be mediated by the crosstalk amongst these three targets, as well as inhibiting angiogenesis. TAVO412 demonstrated strong in vivo tumor growth inhibition in 23 cell-line derived xenograft (CDX) models representing diverse cancer types, as well as in 9 patient-derived xenograft (PDX) lung tumor models.
Methods:Using preclinical CDX data, we established transcriptomic biomarkers based on gene expression profiles that were correlated with anti-tumor response or distinguished between responders and non-responders. Together with specific driver mutation that associated with efficacy and the targets of TAVO412, a set of 21-gene biomarker was identified to predict the efficacy. A biomarker predictor was formulated based on the Linear Prediction Score (LPS) to estimate the probability of patients or tumor model response to TAVO412 treatment.
Results:This efficacy predictor for TAVO412 demonstrated 78% accuracy in the CDX training models. The biomarker model was further validated in the PDX data set and resulted in comparable accuracy.
Conclusions:In implementing precision medicine by leveraging preclinical model data, a predictive transcriptomic biomarker empowered by next-generation sequencing was identified that could optimize the selection of patients that may benefit most from TAVO412 treatment.