Analytical validation of the PROphet computational model for clinical benefit prediction and decision-making tool in metastatic NSCLC

Abstract

The blood proteome, consisting of thousands of proteins engaged in various biological processes, acts as a valuable source of potential biomarkers for various medical applications. PROphet is a plasma proteomics-based test that serves as a decision-support tool for non-small cell lung cancer (NSCLC) patients. PROphet combines proteomic profiling using the SomaScan technology and subsequent computational algorithm. PROphet was implemented as a laboratory developed test (LDT). Under the Clinical Laboratory Improvement Amendments (CLIA) and Commission on Office Laboratory Accreditation (COLA) regulations, prior to releasing patient test results, a clinical laboratory located in the United States that employs an LDT must examine the performance characteristics concerning analytical validity. This study describes the experimental and computational analytical validity of the PROphet test, as required by CLIA/COLA. Experimental precision analysis displayed a median coefficient of variation (CV) of 3.9% and 4.7% for intra-plate and inter-plate examination, respectively, and the median accuracy rate between sites was 88%. Computational precision exhibited a high accuracy rate, with 93% of samples displaying complete concordance in results. A cross-platform comparison between SomaScan and other proteomics platforms yielded a median Spearman correlation coefficient of 0.51, affirming the consistency and reliability of the SomaScan platform. Our study presents a robust framework for evaluating the analytical validity of a platform that combines an experimental assay with subsequent computational algorithms. When applied to the PROphet test, strong analytical performance of the test was demonstrated.