TP53, CD36 Mutations and CDKN2A Loss Predict Poor Outcome in Mantle Cell Lymphoma: Molecular Analysis of the FIL V-Rbac Phase 2 Trial

INTRODUCTION: The FIL V-RBAC trial (EudraCT: 2017-004628-31) is a phase 2 study which enrolled patients with previously untreated mantle cell lymphoma (MCL), stratifying them at initial diagnosis depending on risk factors. High-risk patients, defined as either Ki67≥30% and/or blastoid variant and/or TP53 mutations and/or 17p deletion, received an abbreviated course of R-BAC followed by consolidation and maintenance with venetoclax. Conversely, standard-risk patients with no risk factors received standard R-BAC. The aim of the project was to evaluate the prognostic role of gene mutations and copy-number analysis (CNV) in patients enrolled in the FIL V-RBAC trial.

METHODS: Patients enrolled in the trial with available gDNA extracted from lymph node biopsy were included in this mutational analysis. A CAPP-Seq assay including 146 genes relevant for MCL pathogenesis was used coupled with a robust and previously validated bioinformatic pipeline. Based on CAPP-Seq data, CNV analysis was performed using CNVkit (version 0.9.10) and GISTIC2.0 was used to identify statistically significant CNVs. The primary endpoint was progression-free survival (PFS) according to molecular lesions.

RESULTS: Among the 140 patients enrolled in the V-RBAC trial, 132 were analyzed (53 high-risk and 79 standard-risk). The median age of analyzed patients was 72 years and the median follow-up was 32.0 months. The CAPP-Seq analysis recapitulated the mutational landscape of MCL and identified ATM as the most frequently mutated gene (41.7%, N=55), followed by KMT2D (23.5%, N=31), TP53 (23.5%, N=31), WHSC1 (14.4%, N=19), and BIRC3 (9.8%, N=13). As expected, TP53 mutated patients displayed a significantly inferior PFS compared to TP53 unmutated patients (3-year PFS 30.8% versus 79.9%, p<0.001). Among other genes analyzed, only NOTCH1/2 mutations (10.2%, N=14) and CD36mutations (7.3%, N=7) were associated with significantly inferior PFS. In multivariate analysis adjusted for blastoid variant and Ki67>30%, only CD36 (HR 4.73, 95% CI 1.82-12.28, p=0.001) and TP53 (HR 4.75, 95% CI 2.35-9.60, p<0.001) mutations were independently associated with inferior PFS. Interestingly, CD36 mutations significantly stratified the outcome of standard risk patients (p<0.001), but not of high-risk patients (p=0.348), suggesting that CD36may capture additional high-risk patients that are not currently identified by standard prognostic markers. Regarding CNVs, the most frequently detected were amp3q (36.4%, N=51), del13q (35.0%, N=49), CDKN2A loss (30%, N=42) and del11q (27.9%, N=39). After Bonferroni correction, only CDKN2A loss retained prognostic value (HR 2.47, 95% CI 1.20-5.09, p=0.014), when adjusted for TP53 and CD36 mutations. Finally, non-negative matrix factorization based on gene mutations and CNVs identified 3 different clusters characterized by different molecular composition and with clinical relevance for both PFS (p<0.001) and OS (p=0.0037). Cluster 1 (N=41, 29.3%) was enriched in TP53 and CDKN2A loss and was associated with the worst prognosis (3-year PFS of 39.7% and OS of 50.0%); cluster 2 (N=50, 35.7%) was enriched in ATM mutations and associated with the best prognosis (3-year PFS of 87.5% and OS of 82.0%); cluster 3 (N=35, 25.0%) was enriched in KMT2D mutations and presented an intermediate prognosis (3-year PFS of 77.4% and OS of 76.1%).

CONCLUSIONS: Molecular analysis of the FIL V-RBAC trial indicates that, together with TP53 mutations, CD36mutations and CDKN2A loss are associated with adverse outcome in MCL patients treated with chemo-immunotherapy and the BCL2-inhibitor venetoclax.