Paul Govern
A subset of cancer patients will respond well to immune checkpoint inhibitor (ICI) treatment, others not at all. As predictive biomarkers for ICI response are proposed, often they fail to replicate in later studies or generalize to different cancer types.
Seeking to address this impasse with greater statistical power, biostatisticians at Vanderbilt University Medical Center in a December 2021 paper introduced Cancer-Immu, an analysis portal for exploring associations across public data sets that link ICI response and clinical phenotypes to genetic and transcriptomics profiles.
Reported last month in Cancer Research, a meta-analysis enabled by Cancer-Immu uses data from 27 studies involving 3,037 ICI patients with 14 types of solid tumor. The analysis from VUMC researchers Jing Yang, PhD, Qi Liu, PhD, and Yu Shyr, PhD, replicates known biomarkers and identifies novel ones.
Combining eight patient features, ranging from tumor mutational burden to white blood cell interaction scores, the team finds that their ICI response model significantly outperforms existing predictive models.
The report delves into mechanisms behind a novel finding that figures significantly in the new model: Positive feedback loops between macrophages and T cells enhance activation and recruitment of these immune cells, thereby improving the overall effectiveness of the ICI response.
The study was supported by the National Institutes of Health (P50CA236733, P50CA098131, U2CCA233291, P01CA229123, U54CA274367, P30CA068485, P01AI139449).