Despite considerable progress made in improving therapeutic strategies, the overall survival for patients diagnosed with various cancer types remains low. Further, patients often cycle through multiple therapeutic options before finding an effective regimen for the specific malignancy being treated. A focus on building enhanced computational models, which prioritize therapeutic regimens based on a tumor's complete molecular profile, will improve the patient experience and augment initial outcomes. In this study, we present an integrative analysis of multiple omic datasets coupled with phenotypic and therapeutic response profiles of Cytarabine from a cohort of primary AML tumors, and Olaparib from a cohort of Patient-Derived Xenograft (PDX) models of ovarian cancer. These analyses, termed Pharmaco-Pheno-Multiomic (PPMO) Integration, established novel complex biomarker profiles that were used to accurately predict prospective therapeutic response profiles in cohorts of newly profiled AML and ovarian tumors. Results from the computational analyses also provide new insights into disease etiology and the mechanisms of therapeutic resistance. Collectively, this study provides proof-of-concept in the use of PPMO to establish highly accurate predictive models of therapeutic response, and the power of leveraging this method to unveil cancer disease mechanisms.
RMSE all modelsRMSE excluding -2313 12.79219 11.1153 e CTG-2313 lacks RNAseq AML Blast LSC CD123+