. home.aspx

NEWS

home.aspx
   


GNS Healthcare presents verified use of AI to classify drivers of response to immune checkpoint inhibitor (ICI) therapy

July 22, 2020 / PR Newswire
SHARESHARESHARE

GNS Healthcare (GNS), a leading AI and simulation company, presents results that validate the use of AI to accurately classify tumors based on their immunogenicity and predict response to immune checkpoint inhibitor (ICI) therapy using real-world data. The study showcases the power of causal AI to capture biomarkers and mechanisms, in addition to PD(L)1 and tumor mutation burden (TMB), that are consistent with known immunology. These markers, including CXCL13 upregulation and STK11 mutation, are in line with the targets that are currently being explored for stratification of responders vs. non-responders to ICI therapy, cohort selection, enrichment of future immunoncology trials, or ICI efficacy improvement through combination therapy. The study applied AI to tumor data from The Cancer Genome Atlas (TCGA) to identify the drivers of immune response. The data from nearly 700 NSCLC and over 400 HNSCC patients were fed into REFS, GNS's causal AI and simulation platform, which reverse-engineered in silico patients which accurately classified tumors based on their response. Macrophage activation and polarization, which is driven in part by metabolic reprogramming, was identified as the primary driver of tumor immunogenicity which can allow for a more targeted approach to patient care and clinical trial design. "Over the past decade we have seen nearly a dozen immuno-oncology treatments approved but treatment protocols are still based only on a few biomarkers. The presentation of our work is not only a validation of how AI can extract critical insights from real-world data, but also a milestone in our mission to make precision oncology a reality," said Colin Hill, GNS Healthcare CEO and Co-Founder. These findings from these in silico patients can be used by biopharma companies to select optimal patient populations for clinical trials based on likelihood of response and discover novel biomarkers that make tumors more susceptible to immune therapy, irrespective of response to PD(L)1 therapy. The findings are also beginning to unlock the value of investments in real-world and clinical data to inform future trial design, enable discovery of novel drug targets, and better position drugs across global markets.