A Study of AK117 in Combination With Azacitidine in Patients With Myelodysplastic Syndromes

This is a reasonably well-constructed signal-seeking study: randomized, double-blind, placebo-controlled, multicenter, and in newly diagnosed higher-risk MDS, which reduces bias and gives a fair test of incremental benefit over azacitidine. But the bar for a positive readout is still meaningful. The primary endpoint is complete remission rate, a hard binary efficacy measure, and the control arm already receives active standard therapy, so AK117 must add a clear remission benefit rather than merely preserve baseline activity. Phase 2 oncology/hematology studies without trial-field evidence of strong prior efficacy usually deserve sub-50% success odds. The endpoint window of up to about 2 years also adds timing and data-maturity risk. Most importantly, the study is still listed as recruiting even though the estimated primary completion date has passed by about three months, which raises concern about slower enrollment, delayed analysis, or potential underpowering. Strong design helps, but operational and efficacy uncertainty keep intrinsic YES odds below even.

AK117 (ligufalimab) showed promising Phase 1b data: 48.1% CR rate and 85.2% ORR in 27 evaluable HR-MDS patients with a manageable safety profile (29.1% anemia rate), meaningfully better than first-gen CD47 antibody magrolimab. However, the broader context is deeply sobering. Every randomized phase 2/3 trial combining agents with azacitidine in HR-MDS has failed—magrolimab (ENHANCE, futility + increased death), eprenetapopt, pevonedistat, sabatolimab, tamibarotene, and venetoclax all missed primary endpoints. Phase 1b CR rates consistently inflated relative to randomized controls. AK117's safety differentiation (no hemagglutination, no priming dose) is genuine, but magrolimab's failure was primarily efficacy-driven, not safety-driven. The trial enrolls only 90 patients across three arms (~30 each), limiting statistical power. The trial is still recruiting with completion estimated June 2026, meaning results are unlikely to be disclosed imminently. The base rate for positive randomized HR-MDS trials is effectively zero, warranting significant skepticism despite encouraging single-arm data.

This is a Phase 2 randomized, double-blind, placebo-controlled trial evaluating AK117 (a novel agent) combined with azacitidine in higher-risk MDS. The primary endpoint is complete remission rate per IWG 2023 criteria, which is clinically meaningful. Azacitidine is standard of care for HR-MDS, providing a solid backbone. However, Phase 2 trials carry inherent uncertainty, and adding novel agents to standard care has mixed historical success rates in this difficult-to-treat population. The randomized design strengthens reliability versus single-arm studies. Given the scientific rationale but Phase 2 risk profile, intrinsic probability of positive outcome is moderate, around 50%.

AK117 targets CD47, but this Phase 2 trial in higher-risk MDS faces significant hurdles. The primary endpoint is complete remission rate (CRR) per IWG 2023 criteria, a high bar. The control is azacitidine plus placebo; while azacitidine is standard of care, achieving a statistically significant CRR improvement is challenging. The sponsor, Akeso, has limited late-phase oncology track record. The trial is still recruiting, with the primary completion date estimated to have passed 91 days ago, indicating potential enrollment or operational delays. The endpoint assessment is up to 2 years, suggesting longer-term data is needed. Overall, the trial design is standard but the endpoint is stringent, and the lack of recent completion update adds uncertainty, tilting the intrinsic odds slightly below 50%.

Well-designed Phase 2: randomized, double-blind, placebo-controlled multicenter trial of AK117 (PD-L1/TGF-β bispecific) + azacitidine vs placebo + azacitidine in frontline higher-risk MDS. Primary endpoint CRR per IWG 2023 is standard, reliable, and clinically meaningful surrogate. Patient population newly diagnosed HR-MDS ideal for testing add-on to SOC azacitidine (historical CRR ~18%). Rationale strong: PD-L1 blockade enhances anti-tumor immunity; TGF-β trap counters fibrosis/suppression in MDS niche. No prior AK117 MDS data elevates uncertainty on magnitude of benefit. Operational risks notable: primary completion 91 days overdue despite recruiting status, signaling potential enrollment/execution delays. Sponsor Akeso (Chinese) has solid pipeline execution but disclosure may lag. Balanced, ~40% chance of statistically significant CRR improvement sufficient for positive readout.

AK117 is a CD47 monoclonal antibody being evaluated in HR-MDS combined with azacitidine. While the open-label Phase 1b data for AK117 showed a promising 48% Complete Remission Rate (CRR), this is nearly identical to the early Phase 1b results for Magrolimab (42% CRR), a first-in-class CD47 inhibitor. Magrolimab subsequently failed spectacularly in the Phase 3 ENHANCE trial, which was double-blind and placebo-controlled. In that pivotal trial, Magrolimab's CRR collapsed to 21.3%, effectively identical to the placebo arm's 23.6%, demonstrating that early efficacy signals were primarily driven by investigator bias in an open-label setting. Furthermore, multiple other CD47 programs (e.g., evorpacept, lemzoparlimab) have been discontinued in myeloid indications due to lack of efficacy. Because NCT06196203 is also a randomized, double-blind, placebo-controlled Phase 2 study, it will strip away the open-label assessment bias that artificially inflated AK117's Phase 1b results. The trial's relatively small sample size (~30 patients per arm) further reduces the statistical power to detect any marginal benefit. Given the consistent, class-wide failure of CD47 inhibitors to demonstrate true efficacy against placebo in HR-MDS, AK117's CRR will very likely regress to the azacitidine baseline of ~20-24%, resulting in a negative trial.