Résumé
Introduction Approximately 20–50% of patients (pts) with diffuse large B-cell lymphoma (DLBCL) or follicular lymphoma (FL) experience R/R disease, which is associated with poor outcomes. BCL6 is a transcriptional repressor required for normal B-cell maturation. BCL6 is 1 of the most frequently genetically misregulated proteins in DLBCL and FL, leading to repressed apoptosis and increased proliferation. BMS-986458 is an oral, highly selective bifunctional cereblon-dependent ligand-directed degrader of BCL6. Here, we present updated clinical findings from the dose-escalation part of CA123-1000 (NCT06090539), a first-in-human, multicenter, open-label study of BMS-986458 in pts with R/R NHL.
Methods Pts with R/R DLBCL (de novo or transformed) or R/R FL (including FL 3b) with ≥2 prior regimens and met investigator-assessed treatment (tx) criteria were eligible. BMS-986458 was dosed at 40, 80, or 160mg twice daily (BID) or 80, 160, or 260mg once daily (QD). The efficacy-evaluable (EE) population included pts with ≥1 dose of BMS-986458, had undergone a baseline and post-baseline tumor assessment, had experienced clinical progression, or died. Primary objectives were to assess safety and tolerability and determine the recommended phase 2 dose of BMS-986458. Secondary objectives were to assess plasma pharmacokinetics and the preliminary efficacy (Lugano 2014 criteria) of BMS-986458. Flow cytometry was used for pharmacodynamic analyses to measure BCL6 expression in peripheral blood. Tumors from 22 pts (9 FL and 13 DLBCL) were tested by immunohistochemistry (IHC) for BCL6 expression; the BCL6-negative threshold was ≤30% positive cells. Changes in circulating tumor DNA (ctDNA) levels in plasma were analyzed using CAPP-seq technology (Roche Sequencing Solutions, Pleasanton, CA).
Results As of July 28, 2025, 52 pts were treated with BMS-986458: 23 with DLBCL (not otherwise specified), 8 with high-grade BCL (HGBCL) with MYC and BCL2 rearrangements, and 21 with FL. Cell of origin (local assessment) in the 31 pts with DLBCL/HGBCL included 7 ABC, 10 GCB, and 14 unknown. Median age was 63 y (range 26–85). Pts were heavily pre-treated; median prior lines of therapy was 4 (range 2–12). 32 pts (62%) had prior CAR T cell therapy, 30 (58%) had prior bi/trispecific monoclonal antibodies, and 31 (60%) had prior CELMoD™/IMiD® therapy. In all treated pts, the median (range) duration of therapy was 1.9 mo (0.02–14.4). Any-grade tx-emergent adverse events (TEAEs) and BMS-986458–related TEAEs (TRAEs) occurred in 44 (85%) and 29 (56%) pts, respectively. The most common TRAEs were musculoskeletal/connective tissue disorders (n=12, 23%), mainly arthralgia (n=6, 12%). 2 pts had a grade ≥3 hematologic TRAE (thrombocytopenia [160mg QD] and neutropenia [260mg QD]). 2 pts experienced dose limiting toxicities; 1 had grade 3 arthritis and bone pain (40mg BID), and 1 had grade 3 prolonged QTc (260mg QD). No tx discontinuation or death due to TRAEs occurred. 4 pts had a TRAE leading to dose reduction (myalgia [n=2], prolonged QTc [n=1], and arthritis and bone pain [n=1]). Maximum degradation of BCL6 in peripheral blood was observed at the first dose level and was rapid and profound. Levels of key pro-inflammatory cytokines, including IL-6, TNF-α, IL-8, and IL-10, remained stable throughout tx in most pts. In EE pts (n=33), objective response rate (ORR) was 73% (n=24); 21% (n=7) had a complete response (CR) and 52% (n=17) had a partial response (PR). In pts with DLBCL, ORR was 68% (13/19) with 11% CR (2/19, 1 with triple-hit HGBCL); and in pts with FL, ORR was 79% (11/14) with 36% (5/14) CR. By C1D15, there was significant decrease from baseline in ctDNA levels (P=0.001), reaching >65% in 14/20 pts, suggesting a rapid tumor cell–intrinsic activity. IHC analysis did not suggest an association between BCL6 expression and response; only 2/22 tumors analyzed were BCL6 negative and both were from responders. Median follow-up was 3.2 mo (range, 0.4–14.8). In pts with a PR or CR, median (range) duration of response (DOR) was 4.0 mo (0–11.7), median DOCR (range) was not reached (0–11.7), median (range) time to response was 1.8 mo (1.0–2.4). 27 pts (52%) remain on therapy, and 25 (48%) have discontinued (progressive disease, n=15; AE, n=7; advanced to transplant, n=2; withdrawal by pt, n=1).
Conclusions BMS-986458 was well tolerated and shows promising efficacy in pts with heavily pre-treated R/R DLBCL and FL, supporting its continued development for the tx of NHL.