CMStatistics 2018: Start Registration
View Submission - CMStatistics
Title: Patient-driven tumor xenograft based gene expression model to predict anti-cancer drug response in cancer patient Authors:  Youngchul Kim - H. Lee Moffitt Cancer Center and Research Institute (United States) [presenting]
Abstract: Responses of cancer patients to anticancer drugs vary because of the substantial heterogeneity in molecular characteristics of their tumors. A successful personalized anticancer therapy will then greatly depend on cancer biomarkers that can accurately select patients who will benefit from the drugs. Patient-derived tumor xenograft (PDX) has been widely recognized to inform therapeutic development strategies. We developed a pipeline, PDXGEM, to construct a gene expression model (GEM) for predicting response to anti-cancer drugs in cancer patients on the basis of data on gene expression and dose-response curve from a pan-cancer PDX cohort. As a proof-of-cancer study, we applied the PDXGEM to build GEMs of paclitaxel and cetuximab for breast cancer and colorectal cancer, respectively. For paclitaxel, 66-genes based GEM was built by training the data of 13 breast cancer PDXs and it had a consistently significant prediction performance in multiple breast cancer cohorts. PDXGEM resulted in 882-genes based GEM for cetuximab and its retrospective validation on 70 colorectal cancer patients also yielded a significant prediction (AUC=0.769, p=0.031). PDXGEM can be used to discovery predictive cancer biomarkers and improve therapeutic response rates and therapeutic quality in cancer patients by enriching highly responsive patients.