PS1138 TUMOR MUTATIONAL BURDEN ASSESSED BY TARGETED HIGH THROUGHPUT SEQUENCING PREDICTS TREATMENT FREE SURVIVAL IN CLL: RESULTS FROM A BICENTER STUDY

Background: Chronic lymphocytic leukemia (CLL) is an indolent lymphoid neoplasm with a very heterogeneous clinical evolution. Clinico‐biological classifications of Binet and Rai remain the reference for prognosis. At the molecular level, IGHV mutational status or some recurrent cytogenetic abnormalities (del(17p), del(11q), trisomy 12 or isolated del(13q)) have also a prognostic value. Recently, it was shown that multiple hit profile or mutational complexity may influence the CLL prognosis. Tumor mutational burden (TMB) is usually defined as the frequency of non‐synonymous mutations in the tumor genome. It was originally calculated from whole genome or exome sequencing data. Recently, various groups have shown that it can be calculated from targeted sequencing panels with a similar informativeness. Primarily used in solid cancers as a biomarker predicting response to immunomodulatory therapies, TMB has almost never been studied as a prognostic biomarker per se. Aims: In this work, we studied the prognostic impact of TMB in CLL. Methods: Two independent series from two different hospital university centers of 80 and 70 patients each were analyzed, the former being a training and the latter a validation series. TMB was determined from an Ampliseq panel covering 221.6 kb and targeting 70 genes mutated recurrently in CLL and B lymphomas for the training series. For the validation series, TMB was assessed from an Illumina panel covering 125 kb and targeting 66 genes recurrently altered in haematological neoplasms. Results: The median TMB of the training series was 2/221.6 kb. Among untreated patients, 20/46 had a TMB greater than or equal to 2 mutations (43%), compared to 26/34 (76%) among those who had previously received treatment (Chi2 test, p = 0.003). Also, 14/33 (42%) of patients with mutated IGHV gene had TMB greater than or equal to 2 versus 30/43 (70%) for patients with non‐mutated IGHV gene. Treatment‐free survival (TFS) was significantly decreased for patients with a TMB greater than or equal to 2 (median TFS = 1.1 year vs. 3.8 years, logrank test, p = 0.008). This was particularly true among Binet stage A patients (median TFS = 1.4 years vs. 4.3 years, log‐rank test, p = 0.007) and for patients with mutated IGHV gene (median TFS = 1.2 years vs. 8.2 years, log‐rank test, p = 0.019). Univariate Cox analysis showed that TMB was a significant prognostic marker (p = 0.011). After multivariate analysis, TMB and Binet stage were the only independent prognostic variables against IGHV status and cytogenetic abnormalities (isolated del(13q), trisomy 12, del(11q) or del(17p)). With the same threshold, analysis of the validation series confirmed the TFS prognosis value of TMB (logrank test: p = 0.003). Like for the training series, TMB was increased in patients with unmutated IGHV gene and was independent from cytogenetic abnormalities and IGHV status after multivariate analysis. In both training and validation series, TMB gradually increased with the Rossi's score in a similar manner. Summary/Conclusion: Overall, TMB was low in CLL, as expected from the literature. However, it was very heterogeneous between patients. TMB was mainly increased in previously treated patients and, consistently, was associated with decreased treatment‐free survival. TMB was also increased in patients with unmutated IGHV gene. Moreover, TMB was a significant prognostic variable independently of the IGHV mutation status and the ”classical“ cytogenetic data.