P1245: parallel testing of liquid biopsy (ctdna) and tissue biopsy samples reveals a higher frequency of targetable ezh2 mutations in follicular lymphoma

Topic: 20. Lymphoma Biology & Translational Research Background: Recent genomic studies revealed EZH2 gain-of-function mutations, representing novel therapeutic targets in follicular lymphoma (FL) in around one quarter of patients. However, these analyses relied on single-site tissue biopsies and did not investigate the spatial heterogeneity and temporal dynamics of these alterations, potentially underestimating the prevalence of targetable EZH2 mutations during the disease course. Aims: We aimed to perform a systematic analysis of EZH2 mutations using paired tissue biopsies (TB), liquid biopsies (LB) and peripheral blood mononuclear cells (PBMNC) collected prior to treatment within the framework of a nationwide multicentric study. Methods: Pre-treatment LB, TB and PBMNC samples were collected from 123 patients treated in 10 Hungarian hematology centers. Among these, 114 had paired TB and LB, out of whom 78 patients were recruited at the time of diagnosis, 36 at the time of relapse and one patient had paired TB and LB samples at both timepoints. Additionally, in 39 patients paired diagnostic and relapse samples were characterized. The EZH2 mutation status and allele burden was assessed using an in-house-designed, highly sensitive multiplex droplet digital PCR assay (Bio-Rad Laboratories, CA, USA), which can detect the seven most prevalent EZH2 hotspot mutations in two reactions (p.Y646F/C/S, p.A682G and p.Y646N/H, p.A692V, respectively). Results:EZH2 mutation frequency was found to be 41.5% (51/123) in the entire cohort. In patients with paired TB and LB samples, 53 EZH2 mutations were identified in 37.8% (43/114) of the patients. Among these 51 mutations, 28 mutations were present in both compartments, with 9 and 14 variants that were only present in the TB or in the LB, respectively. Mutations were exclusively detected in TB samples in 6 patients (5.3%) and in LB samples in 9 patients (7.9%). Although there was no significant difference in the EZH2 mutation frequency between diagnosis and relapse (25.7% vs 41%, respectively, Chi-square test, p=0.15) we observed a switch in the EZH2 mutation status during the disease course in 14 patients out of the 39 (35.9%) with paired diagnostic and relapse samples available. Median variant allelic frequency (VAF) of EZH2 mutations in pretreatment LB samples was 2.3% (range 0.1 – 74%). There was no significant correlation between EZH2 VAF in LB samples and Follicular Lymphoma International Prognostic Index (FLIPI) score (p=0.051), LDH levels (p=0.16), bulky disease (p=0.18), presence of extranodal disease (p=0.08) or clinical stage (p=0.21, all Mann-Whitney U test). Fifty-three paired PBMNC and LB samples were available for analysis with only 17/53 (32%) PBMNC samples carrying the EZH2 mutation detected in the corresponding LB sample, with mean VAF being significantly higher in LB samples, than in PBMNC specimens with a detectable alteration (5.7% vs 0.04%, Wilcoxon test, p<0.0001). Summary/Conclusion: In this study, we performed an in-depth spatio-temporal analysis of EZH2 mutations in FL using tissue- and liquid biopsy samples and identified gain-of-function mutations in a considerably higher proportion of patients than previously reported (42% vs 27%) owing to the ability of this approach to resolve spatial heterogeneity. This expands the subset of FL patients who most likely would benefit from EZH2 inhibitor therapy and highlights the need for parallel EZH2 testing in LB and TB samples when considering EZH2 inhibitor therapy. Keywords: EZH2, ctDNA, Follicular lymphoma