Clonal Dominance in Follicular Lymphoma: A Predictor of Poor Prognosis and Histological Transformation

Introduction

Follicular lymphoma (FL) exhibits a highly variable course, sometimes resulting in histological transformation (HT) to higher‐grade lymphoma, complicating clinical management and the prediction of individual outcomes. Each patient's tumor expresses a unique cell surface immunoglobulin (Ig) that may recognize antigens and/or transduce signals.

Methods

We performed high-throughput RNA sequencing (5’ RACE) to simultaneously target the B- and T-cell Ig repertoires (BCR and TCR) in order to assess clonal dominance from bulk RNA samples of diagnostic FFPE biopsies. The data were correlated retrospectively with clinical data, gene expression profiling (RT-MLPseq, Lymphosign^® signature), and next-generation sequencing data (custom targeted panel) from two independent monocentric FL patient cohorts treated with standard first-line immunochemotherapy. All patients provided written consent, and diagnoses were confirmed by Lymphopath network pathologists. Our aims were to predict progression-free survival (PFS), progression within 24 months (POD24), overall survival (OS), and HT risk using the preeminence of dominant clonotypes and the Simpson diversity index to measure intraclonal heterogeneity. The Simpson index, ranging from 0 to 1, measures the diversity of CDR3 sequences, with a value approaching 1 indicating uniformity. Log-rank tests were used to assess OS, PFS, and HT via Kaplan–Meier estimates. Univariate Cox regressions were conducted on relevant variables.

Results

The training cohort included 89 FL patients (grade 1-2: 88.8%; grade 3A: 11.2%, diagnosis between 2006 and 2019) with median age 62 [35-83] years; female sex, 49.4%; ECOG 0-1, 96.6%; stage III-IV, 89.9%; high tumor burden (GELF), 86.5%; R-CHOP-like treatment, 87.6%; FLIPI score 3-5, 48.3%; BCL2 t(14;18), 82.4% (missing data: n=4); HT, n=10 (4.5%); POD24, n=18 (20.2 We correlated B-cell clonal dominance with survival data and identified a subgroup of 10 out of 89 patients (11.2%) where the dominant clones represented more than 76.2% of all CDR3 sequences for IgVH, with a Simpson index greater than 0.89. Both criteria must be met to define high clonal dominance (HCD), with thresholds determined by receiver operating characteristic (ROC) analysis for OS events. These HCD patients, compared to others, had shorter OS (HR=5.1 [1.87-14.17], Cox p=0.002), a higher HT risk (log-rank p=0.042, HR=3.7 [0.9546-14.32], Cox p=0.06) but no significant difference in PFS and POD24 (PFS: HR=1.6 [0.66-3.75], p=0.304; POD24: HR= 0.51 [0.07-3.8], p= 0.51). Causes of deaths (n=18) between HCD and non-HCD patients were lymphoma (50% vs 75%, p=0.34), toxicity (0% vs 8.33%, p=1) or secondary malignancy (50% vs 16.67%, p=0.27). HCD patients had similar clinical characteristics, CD19 and MS4A1 gene expression, and median allele frequency for all somatic variants (26% vs. 25.7%), suggesting similar tumoral infiltration between HCD and non-HCD cases. There was no significant difference in the number of acquired IgVH/IgVL CDR3 N-glycosylation (N-gly) sites between HCD and non-HCD patients (median = 1 site, p=0.23). TP53 variants were more common in the HCD group (n=3, 30%, vs. n=8, 10.13%). HCD patients had only IgM isotype, lacking class switch recombination (CSR), unlike non-HCD patients who had IgG, IgE, or IgA isotypes (45.6%, p=0.005). All patients except one (a non-HCD patient with a BCL2 t(14;18) but no N-gly site) had mutated IgVH (<98% identity to the germline sequence).

The validation cohort consisted of 45 patients (grade 1-2: 91.3%; grade 3A: 8.7%, diagnosis between 2006 and 2021) with median age 64 [41-79] years; female sex, 43.5%; ECOG 0-1, 93.5%; stage III-IV, 95.7%; high tumor burden, 100%; R-CHOP-like treatment, 100%; FLIPI score 3-5, 67.4%. Using the same cutoffs as the training cohort, we identified 3/45 patients (6.7%, IgM isotype: n=2, mutated IgVH: n=3) with worse prognosis (PFS: HR [95% CI] = 5.7 [1.19-26.86], p=0.029; OS: HR=17 [2.39-122.8], p=0.005) and higher HT risk (log-rank p<0.001, HR not estimable) associated with HCD status.

Conclusions

Using the 5’ RACE assay, we identified a homogeneous subpopulation of highly proliferative lymphoma cells in two FL patient sets, associated with higher HT risk and death. HCD status may define a new entity of more aggressive FL, characterized by lack of CSR and lower clonal diversification, enhancing our understanding of FL biology and improving patient management.