6533b86ffe1ef96bd12ce86a

RESEARCH PRODUCT

An Algorithm Combining Patient Performance Status, Second Hit Analysis, PROVEAN and Dann Prediction Tools Could Foretell Sensitization to PARP Inhibitors in Digestive, Skin, Ovarian and Breast Cancers

subject

description

Simple Summary PARP inhibitors, a family of targeted cancer therapeutics, have been shown to be efficient in patients with some deficiencies in the homologous recombination machinery. However, a quick and reliable identification of patients who would benefit from such therapies remains a challenge. In particular, patients with tumors carrying variants of unknown significance (VUS) in homologous recombination genes do not currently benefit from PARP inhibitor treatments. In this study, we present an algorithm that may allow classification of these variants with regard to their impact on tumor responsiveness to PARP inhibitors. If validated on a larger patient sample, our algorithm would allow patients with tumors potentially responsive to PARP inhibitors to benefit from this therapy. Abstract PARP inhibitors yield interesting outcomes for patients with ovarian tumors harboring BRCA1 or BRCA2 mutation, but also with other tumors with homologous repair (HR) deficiency. About 40% of variants are variants of unknown significance (VUS), blocking the use of PARP inhibitors. In this study, we analyzed NGS data from 78 metastatic patients treated with PARP inhibitors. We tested NGS data and in silico predictions to classify VUS as potentially benign or deleterious. Among 41 patients treated with olaparib, three had tumors harboring benign and 26 pathogenic variants, while 12 had VUS. Progression-Free Survival (PFS) analysis showed that benign variants did not respond to olaparib whereas pathogenic variants were associated with a median PFS of 190 days. Surprisingly, median PFS of patients with VUS-carrying tumors suggested that some of them may be sensitive to PARP inhibitors. By testing different in silico predictions and variant allelic frequency, we obtained an algorithm predicting VUS sensitivity to PARP inhibitors for patients with a Performance Status below 3. Our work suggests that VUS in HR genes could be predicted as benign or deleterious, which may increase the number of patients eligible for PARP inhibitor treatment. Further studies in a larger sample are warranted to validate our prediction algorithm.