Follicular Lymphoma Mutational Profile: Patients with mTOR Complex 1 (mTORC1) Mutations Present Differential Survival in First-Line R-CVP/R-CHOP Vs R-Bendamustine

Follicular lymphoma (FL) patients typically have an indolent behavior, but those patients who relapse or progress within 24 months of immunochemotherapy (POD24) have a poor survival. In order to improve assessment of individual risk at the time of diagnosis, molecular parameters are being incorporated in scores, such as M7-FLIPI. However, patients treated with rituximab-bendamustine (R-B) were not included in those series. The aim of this study was to analyze the mutational profile of patients with FL at the time of diagnosis and to investigate its correlation with outcome at first line therapy (exclusively R-CVP, R-CHOP and R-B).

This is a retrospective study from 3 centers that included patients with FL treated in first line with R-CVP, R-CHOP/CHOP-like or R-B. FFPE tissue from diagnostic lymph node biopsies were studied by Next Generation Sequencing (Illumina) using a QIAgen custom DNA panel covering 64 genes related to FL, involved in epigenetics and transcriptional regulation, BCR signaling, cell survival, immune response, migration and mTORC1 pathway. An accurate low-frequency variant caller for targeted sequencing data with unique molecular identifiers (UMIs) was used. All variants were manually curated and interpreted for pathogenicity using VariantStudio 3.0 and several population and mutation databases and in silico predictive tools. Further statistics and variant analysis were performed in R 3.6.2 version using maftools package and SPSS Statistics 25 (IBM).

A total of 191 patients with newly diagnosed FL have been included in this cohort analysis, with a final number of 109 cases with complete clinical data and successful characterization by NGS. Median age at diagnosis was 58 years (range 24-90) and 56% were males. Stage III-IV was seen in 93.6% and FLIPI score was: 0-1 in 16.5%, 2-3 in 68.8% and 4-5 in 11%. R-CVP was administered in 7.3%, R-CHOP in 65.1% and R-B in 27.5%. Overall response rate was 98.2% (CR 83.5%). With a median follow-up of 96 months, median PFS was 94 months and 5-y PFS was 54% in R-CHOP and 81% in R-B.

All patients presented mutations, with a median of 6 different genes mutated per patient (range 1-14). Very similar to previously described, top ten mutated genes were: CREBBP (75%), KMT2D (73%), BCL2 (46%), TNFRSF14 (43%), EZH2 (23%), STAT6 (21%), EP300 (19%), MEF2B (19%), IGLL5 (18%) and ARID1A (17%).

For all the cases, we detected at least one mutation in genes coding for epigenetic and transcriptional regulation, since it is a pivotal hallmark in FL. We did not find differences in survival, according to the mutational profile in this set of genes. However, when we compared the subset of “long survivors” (SLP> 110 moths -plateau reached-; n=14) against the rest, we observed that long survivors were enriched in MEF2B mutations (p=0.013) and presented less frequently with KMT2D mutations (p=0.022). Intriguingly, considering the entire cohort, the presence of MEF2B mutation was found to be strongly associated with bone marrow involvement (p<0.001).

Seven cases (6.4%) presented the hotspot gain of function mutation Y132D in the gene CTSS (cathepsin S), none of them had affected bone marrow at diagnosis (p=0.031). This mutation showed in our series a trend to better outcome, with 10-y PFS of 86% in mutated cases and 44% in those non-mutated (p=0.069). Recent investigations show that tumors with CTSS hyperactivation might benefit from a combination of immune checkpoint inhibitors and CTSS inhibition.

In our cohort, 32% of patients carried one or more mutations affecting the mTORC1 pathway (mTORC1^mut; genes analyzed: ATP6AP1, ATP6AP2, ATP6V1B2 and RRAGC). We found that mTORC1 mutation status at diagnosis showed differential prognostic impact depending on the treatment regimen, being R-CHOP/ mTORC1^mut cases those with the lower PFS and R-B/ mTORC1^mut those achieving best PFS (figure 1, p=0.0021). This is of interest, since mTORC1 pathway has been described to be involved quite selectively in the pathogenesis of FL, with promising therapeutic options.

We describe novel clinic-genetic correlations that might be useful for the management of FL and potentially lead to novel precision therapies. Of note, we found that there is one third of FL patients that present with mutations in genes affecting the mTORC1 pathway, showing poor outcomes when treated with R-CVP/R-CHOP but not if treated with R-B. Evaluation of mTORC1 targeted therapies might be of interest in this subset of FL.