Clonal architecture of relapsed or refractory follicular helper t‐cell lymphoma: an ancillary study of the oracle trial, a lysa study

Follicular helper T-cell lymphoma (TFHL) results from the oncogenic transformation of a TFH cell, driven by mutations in genes involved in epigenetic regulation (TET2, DNMT3A, IDH2) and T-cell signaling (RHOA). Demonstration of TET2 and DNMT3A mutations in B cells, myeloid cells, or hematopoietic progenitor cells have suggested that TFHL can emerge from clonal hematopoiesis (CH) in a multi-step process. However, the frequency, the extension, and the relevance of such CH in TFHL oncogenesis are unclear. To describe and characterize the clonal architecture of refractory/relapsed (R/R) TFHL, we collected bone marrow samples of patients included in the ORACLE trial (NCT03593018), a phase 3 trial evaluating the oral 5-azacitidine (CC-486) compared to single-agent chemotherapy in patients with R/R TFHL. Bone marrow (BM) cells from 32 patients at inclusion were cultured in methylcellulose supplemented with standard growth factors. BFU-E, CFU-GM, and CFU-GEMM were counted and individually harvested. A total of 517 colonies derived from the 32 patients were then individually genotyped by next-generation sequencing (NGS) using a capture-based panel (limit of detection: 1%) covering 43 genes recurrently mutated in CH and/or TFHL, including TET2, DNMT3A, IDH2, and RHOA. Meanwhile, bulk BM cells from 29 patients were sequenced with the same NGS panel and then compared to the mutations found in cfDNA (31 patients) and tumor biopsies (28 patients). Overall, the frequency of CH and the phylogenetic link between the myeloid progenitors and the TFHL was evaluable in 30 TFHL patients. We successfully genotyped 471/517 (91%) colonies. CH defined by at least one somatic mutation detected in at least one myeloid colony or total BM cells was found in 29/30 (97%) patients. Among the 26/29 patients with mutated myeloid colonies, the median percentage of mutated colonies was 21% (IQR 7%–53%) and 7 had subclonal architecture with a median of 3 different clones. The 2 most frequent mutations were DNMT3A and TET2, detected in 24/30 (80%) and 18/30 (60%) patients' BM, respectively. We confirmed the absence of RHOAG17V and IDH2R172 mutations in BM cells of TFHL patients. While 17/18 patients with a detectable TET2 mutation in CH had a TET2 mutated TFHL, only 13/23 patients with a DNMT3A mutated CH had a DNMT3A mutated TFHL. We thus identified 3 groups of patients according to the distribution of mutations found in the BM and the LN lymph nodes: 16/30 (53.3%) patients had a TFHL clonally derived from CH (CH-related TFHL), 13 (43.3%) had TFHL, and clonally distinct CH (CH-unrelated TFHL), and one patient (3.3%) had undetectable CH. No significant difference in patient outcome was observed regarding the presence of CH-related or unrelated TFHL, but our results are limited by the small number of patients and treatment heterogeneity. Encore Abstract—previously submitted to EHA 2023 Keywords: aggressive T-cell non-Hodgkin lymphoma, tumor biology and heterogeneity No conflicts of interests pertinent to the abstract.