Prediction of olaparib sensitivity for variants of unknown significance in homologous repair genes
| Autor: | Sandy Chevrier, Isabelle Desmoulins, Laure Favier, Romain Boidot, François Ghiringhelli, Laurent Arnould, Leila Bengrine |
|---|---|
| Rok vydání: | 2019 |
| Předmět: | |
| Zdroj: | Journal of Clinical Oncology. 37:3108-3108 |
| ISSN: | 1527-7755 0732-183X |
| DOI: | 10.1200/jco.2019.37.15_suppl.3108 |
| Popis: | 3108 Background: the recent use of PARP inhibitors in clinical practice gives very interesting outcome for ovary tumors with BRCA1 or BRCA2 mutation but also in other tumors with homologous repair deficiency. Nevertheless, no hotspot mutations are present, consequently, more than 85% of observed variants have unknown significance, blocking the use of PARP inhibitor. Methods: Exome analysis was performed on a cohort of 27 patients treated with olaparib. After bioinformatics analyses, variant interpretation was performed by interrogating different databases. For variants of unknown significance (VUS), PROVEAN software and allelic frequency normalized with tumor cellular content were used to classify VUS as potentially benign or potentially deleterious. Results: Among the 27 patients analyzed, 16 harbored already classified variants (3 benign and 13 pathogen variants) and 11 had VUS. The first Progression Free Survival (PFS) analysis showed that benign variants did not respond to olaparib with a median survival of 62 days, whereas pathogenic variants had a median of 109 days. Surprisingly, VUS had a median of 136 days, suggesting that some of them could be classified as potentially deleterious. On the subset of 11 patients with VUS, we applied PROVEAN prediction classifying 5 variants as benign and 6 variants as deleterious, with a median PFS of 54 days and 140 days ( p=0.3235), respectively. With the second prediction, based on variant allelic frequency, we obtained PFS of 73.5 months for benign variants and 210 days for deleterious ones ( p=0.29). By combining both predictions, we classified as benign, VUS predicted benign with both predictions, and as deleterious, VUS predicted as deleterious with at least one prediction. Consequently, we perfectly discriminated benign from deleterious variants with a median PFS of 36 days for predicted benign and 177 days for predicted deleterious ( p=0.0084). From all patients, PFS were significantly different ( p=0.0003) between benign (n=6, 56 days) and deleterious variants (n=21, 140 months). Conclusions: Our work tends to show that VUS of homologous repair genes could be predicted as benign or deleterious, and could increase the number of patients eligible for a treatment by PARP inhibitors. The number of patients needs to be increased in order to validate our prediction algorithm. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|