Paper Details

ABSTRACT

Background

Relapsed/refractory T-cell lymphoma (R/R TCL), an aggressive form of non-Hodgkin lymphoma that is often resistant to chemotherapy and targeted therapies, remains an area of substantial unmet medical need, with limited effective treatment options after multiple prior therapies. APG-5918 is a novel, orally bioavailable, highly selective EED inhibitor that disrupts PRC2 function and suppresses histone H3 lysine 27 (H3K27) methylation, resulting in antitumor activity.

Methods

This is a multicenter, open-label, first-in-human study using APG-5918 for the treatment of patients with advanced malignancies in US and China (NCT05415098). APG-5918 was administered orally at doses ranging from 50 to 800 mg once daily (QD). Dose expansion cohorts in TCL were evaluated at 400, 600, and 800 mg QD. Responses were assessed according to International Society for Cutaneous Lymphomas (ISCL) and European Organisation for Research and Treatment of Cancer (EORTC) criteria for cutaneous TCL (CTCL) and Lugano criteria for other TCL subtypes. Safety was evaluated in all treated patients and efficacy assessed in response-evaluable patients.

Results

APG-5918 demonstrated superior single-agent antiproliferative activity in 4 TCL cell lines (IC50 range, 0.10-7.20 μM) compared with other PRC2-targeting agents in preclinical studies. As of July 30, 2026, a total of 95 patients with advanced malignancies (28 solid tumors and 67 R/R TCLs) had been enrolled. Patients with R/R TCLs, included angioimmunoblastic T-cell lymphoma (AITL; n=23), peripheral T-cell lymphoma not otherwise specified (PTCL-NOS; n=23), cutaneous T-cell lymphoma (CTCL; n=10), ALK-negative anaplastic large-cell lymphoma (ALCL; n=5), and other TCL subtypes (n=6).The median age was 55 years (range, 18-75), and patients had received a median of 2 prior systemic therapies (range, 1-9). APG-5918 exhibited a manageable safety profile. No dose-limiting toxicities were observed during dose escalation. The most common grade ≥3 treatment-related adverse events were hematologic toxicities, including neutrophil count decreased (22.2%), white blood cell count decreased (22.2%), platelet count decreased (21.0%), and anemia (16.0%). Treatment discontinuation because of adverse events was infrequent (9.9%). Among 47 efficacy-evaluable patients with R/R TCL treated at various doses of APG-5918, the objective response rate (ORR) was 44.7%, including 2 complete responses (CR; 4.2%) and 19 partial responses (PR; 40.4%). The disease control rate (DCR) was 74.5%, and only 2 patients (4.2%) experienced progressive disease as the best overall response. Among 20 efficacy-evaluable patients treated with APG-5918 600 mg QD, the ORR was 65.0%, including CR in 10.0% and PR in 55.0%, with a DCR of 95.0%. Responses occurred rapidly, with a median time to response of 1.7 months. In the cutaneous TCL subgroup (R/R CTCL, n=10), encouraging clinical activity was observed, with 40.0% ORR (95% CI, 9.9%–81.6%), including 1 CR (10%) and 3 PR (30%), and a DCR of 70.0% (95% CI, 29.0%–96.3%). These findings suggest promising activity of APG-5918 in the treatment of patients with R/R CTCL. Median duration of response (DOR) and progression-free survival (PFS) had not been reached; the lower limit of the 95% confidence interval for PFS was 6.9 months, suggesting durable clinical benefit. Pharmacodynamic analyses demonstrated sustained, marked suppression of H3K27 trimethylation to <10% of baseline, supporting the target engagement and proposed mechanism of action. Baseline mutations in T follicular helper cell–related genes, including TET2 (39%), RHOA (24%), and DNMT3A (20%), were associated with higher response rates.

Conclusions

APG-5918 demonstrated encouraging clinical activity and favorable tolerability in patients with heavily pretreated R/R TCLs. Encouraging activity was also observed in R/R CTCL, warranting further evaluation in a larger patient population. APG-5918 600 mg QD showed the highest and most consistent antitumor activity across TCL subtypes while maintaining an acceptable safety profile, supporting 600 mg QD as the recommended phase 2 dose (RP2D) for further clinical evaluation.