Validation of a microRNA panel to pre-symptomatically predict the risk of radiation-induced fibrosis

Abstract ChromoLogic has identified a small panel of four circulating microRNA (miRNA) sequences that can be used to predict the onset of radiation-induced lung injury from plasma samples taken approximately 15–30 days post- exposure. These miRNAs are linked to genes associated with profibrotic signaling pathways (e.g., TGF-β and BMP signaling) that are differentially regulated in fibrotic lung tissue following radiation exposure. A Bayesian predictive model (the miDOS assay) was developed using this panel through the analysis of mouse (N = 128), non-human primate (N = 67), and human (N = 303) miRNA responses to irradiation and longitudinal assessment of pneumonitis and fibrosis. An initial assessment of the miDOS assay in human patients (N = 21) was performed at the Greater Los Angeles Veteran's Administration Medical Center (VA) in lung and breast cancer patients that received radiotherapy (RT) suggested the miDOS assay had good performance in this patient population (area under the ROC curve = 0.85). Based on feedback from the FDA following a pre- submission meeting (April 2022 Q230171) discussing the miDOS assay, ChromoLogic and Dr. Diana Gage, M.D. (VA/UCLA) propose to advance the development of this miRNA-based predictive assay for radiation- induced lung injury by (i) validating the performance of the assay in a larger patient population (N = 100); (ii) assessing the impact of cancer on baseline miRNA levels; (iii) assessing the longitudinal dynamics of miRNA abundance following radiation treatment; (iv) evaluating the impact of potential confounders on panel miRNA; and (v) evaluating the impact of an expanded miRNA panel on the specificity and sensitivity of the assay.