Identification of a genetic signature enriching for response to ibrutinib in relapsed/refractory follicular lymphoma in the DAWN phase 2 trial

Abstract

Abstract Background The single‐arm DAWN trial (NCT01779791) of ibrutinib monotherapy in patients with relapsed/refractory follicular lymphoma (FL) showed an overall response rate (ORR) of 20.9% and a median response duration of 19.4 months. This biomarker analysis of the DAWN dataset sought to determine genetic classifiers for prediction of response to ibrutinib treatment. Methods Whole exome sequencing was performed on baseline tumor samples. Potential germline variants were excluded; a custom set of 1216 cancer‐related genes was examined. Responder‐ versus nonresponder‐associated variants were identified using Fisher's exact test. Classifiers with increasing numbers of genes were created using a greedy algorithm that repeatedly selected genes, adding the most nonresponders to the existing “predicted nonresponders” set and were evaluated with 10‐fold cross‐validation. Results Exome data were generated from 88 patient samples and 13,554 somatic mutation variants were inferred. Response data were available for 83 patients (17 responders, 66 nonresponders). Each sample showed 100 to >500 mutated genes, with greater variance across nonresponders. The overall variant pattern was consistent with previous FL studies; 75 genes had mutations in >10% of patients, including genes previously reported as associated with FL. Univariate analysis yielded responder‐associated genes FANCA, HISTH1B, ANXA6, BTG1, and PARP10, highlighting the importance of functions outside of B‐cell receptor signaling, including epigenetic processes, DNA damage repair, cell cycle/proliferation, and cell motility/invasiveness. While nonresponder‐associated genes included well‐known TP53 and CARD11, genetic classifiers developed using nonresponder‐associated genes included ATP6AP1, EP400, ARID1A, SOCS1, and TBL1XR1, suggesting resistance to ibrutinib may be related to broad biological functions connected to epigenetic modification, telomere maintenance, and cancer‐associated signaling pathways (mTOR, JAK/STAT, NF‐κB). Conclusion The results from univariate and genetic classifier analyses provide insights into genes associated with response or resistance to ibrutinib in FL and identify a classifier developed using nonresponder‐associated genes, which warrants further investigation. Trial registration: NCT01779791.

Keywords

• biomarkers
• genetic variants
• lymphoma
• mutations