Oncobox, gene expression-based second opinion system for predicting response to treatment in advanced solid tumors.

Abstract

e13143 Background: Anticancer Targeted Drugs (ATDs) specifically target one or a few types of tumor-related molecules in a cell. More than two hundred of ATDs were approved worldwide. They have different mechanisms of action and are effective for different cohorts of patients. However, many individual cases remain poorly responsive and it is of great importance to identify predictive markers of ATD efficacy. Our aim was to develop a platform enabling smart selection of the most efficient ATD therapies. Methods: We generated a second-opinion platform for clinical oncologists termed Oncobox. It is based on the analysis of gene expression profile of a cancer sample in comparison with the corresponding normal tissue biosamples in order to personalize selection of targeted drugs for individual cancer patients. Based on RNA-seq gene expression data, pathway activation levels are calculated and along with the concentrations of molecular target genes products used as predictors of tumor response to ATDs. Results: The Oncobox method was tested first on the retrospective samples of advanced tumors: gastric, renal cell, ovarian and colorectal cancers, thus showing ROC AUC values > 0.7. In currently ongoing prospective trial for advanced solid tumor patients, the Oncobox tests were completed for 239 patients, primary feedback information received for 144 patients. 25 patients (17%) died before prescription of the therapy, 19% received palliative care treatment, 39% received Oncobox-recommended therapies and 25% received other therapies (February 2019). Tumor responses were estimated for 30 patients (breast, colorectal, ovarian and other cancers) receiving therapies, with disease control rate of 71% for Oncobox-recommended ATDs (Bevacizumab, Crizotinib, Trastuzumab and others) and 44% for other therapies. Our results also suggest that cancer metastases and primary tumors may have different gene expression, molecular pathway activation and drug scoring patterns, thus pointing to the importance of testing multiple tumor sites. Conclusions: Our study suggests that profound analysis of gene expression profiles of tumor tissues may be useful for ATDs treatment of advanced and metastatic cancers. Clinical trial information: NCT03724097.