Dual epigenetic and PI3K blockade produced responses in heavily pretreated DLBCL
Ifupinostat delivered measurable single-agent activity in relapsed or refractory diffuse large B-cell lymphoma after at least two prior systemic regimens, but the response came with frequent and often severe hematologic toxicity. In this multicenter phase 2 study across 28 hospitals in China, 128 patients entered the safety set and 115 entered the efficacy analysis set after central review. The primary endpoint was objective response rate assessed by blinded independent central review, and the treatment achieved an ORR of 36.5%, including a complete response rate of 11.3%.
The clinical signal matters because this was a population with few dependable options left. Median follow-up reached 19.58 months, yet the median progression-free survival was only 2.7 months and median overall survival 9.9 months, showing how quickly the disease can outrun treatment even when tumor shrinkage occurs. Responses also emerged fast, with a median time to response of 1.3 months. In aggressive lymphoma, that early movement on imaging usually reflects direct suppression of survival circuitry rather than slow immune-mediated effects.
Response depth was real, but durability stayed uneven
The disease control rate reached 60.0%, and the median duration of response was 6.2 months. Among patients who achieved complete or partial response, outcomes were far better than the overall cohort: median PFS extended to 8.1 months and median OS to 25.6 months. Those numbers make the central biological point of the study plain. Once ifupinostat induced a meaningful remission, the disease often stayed controlled long enough to matter clinically. The problem was not whether the drug could produce responses. The problem was how often and how long those responses could be sustained in a population already exposed to multiple prior lines of therapy.
Subgroup data suggested activity across several clinical strata. Among patients younger than 60 years, ORR was 31.8%; among those 60 years and older, 42.9%. Response rates were 31.0% in men and 42.1% in women. By International Prognostic Index category, ORR ranged from 25.0% to 43.6% across the reported groups. Patients with two prior regimens had a higher ORR than those with three or more prior regimens, 41.3% versus 30.8%, which fits the expectation that lymphoma burden and biologic resistance deepen as treatment lines accumulate.
Even patients who had already passed through more modern salvage platforms showed activity. ORR was 80.0% among those who progressed after autologous hematopoietic stem cell transplantation, 100% after bispecific antibodies, and 66.7% after CAR-T therapy, although the numbers in those subgroups were very small. Those isolated signals should be read cautiously, but they suggest the drug may retain activity after mechanisms of resistance to other classes have already emerged.
Why a dual-targeted agent may work where single-pathway drugs have struggled
The rationale for ifupinostat is embedded in the biology of aggressive B-cell lymphoma. Histone deacetylases help compress chromatin and silence tumor suppressor programs, while the PI3K/AKT/mTOR axis drives proliferation, metabolism, motility, and treatment resistance. Blocking one pathway alone often invites compensatory signaling through the other. A dual inhibitor avoids the logistical and pharmacokinetic problems of combining separate agents and may shut down survival signaling more completely than either class used in isolation.
That mechanism helps explain the fast response kinetics seen here. Ifupinostat combines HDAC inhibition with PI3Kα blockade in one molecule, and the observed early tumor shrinkage is consistent with rapid disruption of transcriptional support and intracellular survival signaling. The study also explored PIK3CA amplification in 83 patients. ORR was 33.3% in those with amplification and 41.1% in those without it, suggesting no obvious enrichment signal from that biomarker in this dataset.
Hematologic toxicity dominated the safety profile
Toxicity was substantial. Treatment-related adverse events occurred in 99.2% of patients, and grade 3 or higher treatment-related events occurred in 86.7%. The most common severe toxicities were thrombocytopenia in 75.8% of patients, neutropenia in 33.6%, and leukopenia in 22.7%. No treatment-related deaths occurred, which is reassuring, but the marrow suppression burden was unmistakable and clinically consequential.
That toxicity pattern fits the pharmacology of epigenetic and PI3K pathway inhibition. Both target axes influence proliferative tissues, and the bone marrow is especially vulnerable when cell-cycle regulation and survival signals are altered. In practice, the data point toward a drug that can be active in refractory lymphoma but may demand close blood count monitoring, dose modification, and careful supportive care. A therapy with response potential but pervasive thrombocytopenia is not a minor inconvenience for this patient group; it shapes whether treatment can be delivered at all.
Treatment discontinuation was common in the efficacy set, with progression the leading reason. Other discontinuations included adverse events, withdrawal of consent, investigator decision, and patient choice. Only 2.6% of patients in the efficacy set remained on treatment at data cutoff, a reminder that refractory DLBCL continues to outpace many salvage strategies even when an investigational agent is available.
COVID-era interruptions complicated the phase 2b experience
The study included phase IIa and phase IIb cohorts, and outcomes differed between them. ORR was 42.9% in phase IIa and 33.8% in phase IIb. Median DoR, PFS, and OS were also longer in phase IIa than in phase IIb. Investigators performed a sensitivity analysis after excluding 36 phase IIb patients whose treatment was delayed or interrupted during COVID-19 control periods, and response improved in that restricted group. Those interruptions introduce a real-world layer of noise into efficacy interpretation, especially in a disease where schedule intensity may influence disease control.
Even with those caveats, the central finding holds: ifupinostat produced objective responses in a population with relapsed or refractory DLBCL after multiple prior therapies, including some patients who had already progressed after transplant, CAR-T therapy, or bispecific antibodies. The challenge now is not proving that the agent has activity. The unresolved questions are how to refine patient selection, how to manage the hematologic toxicity without blunting efficacy, and whether combinations or sequencing strategies can convert these early remissions into longer disease control.
The trial registry number was ChiCTR1900028179.