Recurrent Gene Mutations and Clinical Outcome in the Context of Lenalidomide Treatment of Relapsed/Refractory CLL — Results from Multicenter Phase II Trial CLL-009

INTRODUCTION: Mutations in genes such as TP53, SF3B1 and NOTCH1 are frequent in CLL and have been associated with outcome in previous studies. Whereas NOTCH1^mut appeared to be a negative predictive factor in the context of chemotherapy combined with Rituximab, other mutations, e.g. SF3B1^mut, TP53^mut and BIRC3^mut, were adverse prognostic factors in patient cohorts with different therapies. However, the impact on outcome among patients treated with non-chemotherapeutic agents is less clear.

METHODS: We assessed the incidence and impact of gene mutations in the CLL-009 trial (relapsed/refractory CLL, single-agent lenalidomide, Wendtner et al., Leuk Lymphoma 2016). Samples from study entry were available from 96 of 103 patients (93%). Analysis was performed via liquid chromatography (dHPLC) and sanger sequencing for TP53 and a customized targeted Next Generation Sequencing (tNGS) panel for ATM, BIRC3, EGR2, FBXW7, MYD88, TP53 (all exons), NOTCH1 (exon 34), SF3B1 (exon 14-16, 18) and XPO1 (exon 14, 15). All mutations with a variant allelic fraction > 5% were considered.

RESULTS: In total we identified 162 mutations across 11 genes in 79 of 96 patients with incidences of TP53^mut 37.5%, SF3B1^mut 33.3%, NOTCH1^mut 21.9%, ATM^mut 15.6%, XPO1^mut 9.4%, FBXW7^mut 7.3%, NFKBIE^mut 6.3%, POT1^mut 5.2%, BIRC3^mut 3.1%, EGR2^mut 2.0% and MYD88^mut 2.0%. Twelve of 36 (33.3%) TP53^mut patients had additional mutations within the same gene in contrast to only 2 of 32 (6.3%) SF3B1^mut patients and 1 of 20 (5.0) NOTCH1^mut patients. tNGS as compared to dHPLC and Sanger revealed 2 additional mutations in TP53 resulting in identification of 2 more TP53^mut patients.

We identified a variety of associations of subgroups defined by mutations with clinical and laboratory parameters, such as TP53^mut with del17p (p<0.01), FBXW7^mut with +12q (p<0.01) and SF3B1^mut with mutated IGHV (p=0.02). Patients <65 years more frequently had mutations in NOTCH1 (p<0.01), while >65 years more commonly had mutations in POT1 (p=0.02) and TP53 (p=0.02). Patients with NOTCH1^mut (p<0.01) and TP53^mut (p=0.03) had fewer prior lines of therapy (2.1 vs. 3.3 lines, p<0.01 for NOTCH1 and 2.8.vs. 3.3 lines, p<0.05 for TP53). Occurrence of tumor flare reaction was independent of all genetic subgroups.

Regarding response, OS and PFS, only the presence of NOTCH1^mut showed a trend for reduced OS (HR=1.66, p=0.08). All other mutational subgroups, in particular TP53^mut and SF3B1^mut, had no significant association with PFS and OS. Three or more prior lines of therapy had the strongest impact on OS (HR 1.4, p<0.01) as published for the whole clinical trial.

To identify factors of independent prognostic impact, we performed multivariable Cox regressions for PFS and OS including number of prior lines of therapy, del17p, del11q, IGHV status and gene mutations. Only del17p (HR=2.81, p=0.01) was associated with decreased PFS, while for OS unmutated IGHV (HR 2.36 p=0.04), NOTCH1^mut (HR 2.74 p<0.01) and >=3 therapy lines (HR 3.3, p<0.01) were identified as independent prognostic factors.

SUMMARY: In this cohort, only del17p was associated with decreased PFS, while for OS unmutated IGHV, NOTCH1^mut and >=3 therapy lines were identified as independent prognostic factors. The specific characteristics of the population under study and the treatment modality could explain not only the high incidence of recurrent mutations, but also uncommon associations of mutations to other characteristics or outcomes. Therefore, the associations of genetic prognostic factors may depend on the context of the particular trial population and the treatments administered, making a reassessment of prognostic markers and models mandatory in the era of novel treatments.

Figure

View largeDownload slide

Figure

View largeDownload slide

Close modal