Akeso's ligufalimab (AK117), a next-generation humanized IgG4 anti-CD47 monoclonal antibody, was tested in combination with azacitidine (a hypomethylating agent) and venetoclax (a BCL-2 inhibitor) — the current standard backbone for older, chemo-ineligible AML — in the randomized, double-blind, placebo-controlled Phase II AK117-206 trial. The combination delivered a 54% reduction in the relative risk of an EFS event (hazard ratio 0.46), pushed median EFS from 6.9 to 9.1 months, and lifted 9-month EFS from 14.1% to 53.2%. The composite complete response rate (CRc) was 56.7% versus 53.3%, and the share of patients who reached CRc with measurable residual disease (MRD) negativity — the deepest level of remission — was 46.7% versus 36.7%. The median CRc duration was 10.4 months in the ligufalimab arm, nearly double the 5.6 months seen in controls. Overall survival favored the ligufalimab arm as well, with 9-month OS of 78.7% versus 43.1%.
Acute myeloid leukemia (AML) is not a single disease. It is a category of aggressive bone-marrow cancers in which immature myeloid blasts crowd out normal blood production. Symptoms arrive quickly — fatigue, bleeding, infections, anemia — and the disease moves fast. The standard curative path for younger, fitter patients is intensive induction chemotherapy (the "7+3" cytarabine + anthracycline regimen), often followed by allogeneic stem-cell transplant.
The problem is age. The median age at AML diagnosis is about 68, and roughly 60% of newly diagnosed patients are too frail, have too many comorbidities, or have poor-risk cytogenetics to tolerate induction chemo. For them, the standard backbone since 2020 has been azacitidine (a hypomethylating agent) plus venetoclax — a regimen that produces complete remissions in roughly 50% of patients but leaves median overall survival in the 9-15 month range. Adding a third effective drug without adding intolerable toxicity has been the field's central pursuit.
CD47 — the "don't eat me" signal that cancer cells display to evade macrophages — has been one of the most pursued third-drug candidates. The original CD47 antibody, magrolimab from Gilead (acquired with the Forty Seven buyout in 2020), generated enormous early excitement but ran into two stubborn problems: red-blood-cell binding (causing anemia and the need for complex priming-dose regimens) and disappointing randomized data. The combination of magrolimab + AZA failed to beat AZA alone in higher-risk MDS in late 2024, and the broader program has been substantially de-prioritized.
CD47 sits on the surface of nearly every human cell. It binds SIRPα on macrophages and sends a "do not phagocytose" signal. Tumors overexpress CD47 to hide from innate immune surveillance. Blocking CD47 reactivates macrophage killing — but only if the antibody does not also bind the CD47 on healthy red blood cells and platelets, which triggers off-target clearance.
Magrolimab was an IgG4 antibody that bound CD47 on red cells and required a low-dose priming schedule to reticuloendothelial-adapt patients before therapeutic dosing. The drug worked in early-phase studies but the dose ramp complicated combinations and the binding profile made red-cell toxicity a recurring issue. Ligufalimab, by Akeso's design notes, is engineered with a different Fc-backbone mutation intended to minimize red-cell binding and hemolysis while preserving tumor-cell phagocytosis.
Akeso describes ligufalimab as a "next-generation" anti-CD47 antibody built specifically to address the hematologic toxicity that has dogged the class. The Phase II AK117-206 design — randomized, double-blind, placebo-controlled, against AZA + VEN alone — is the first rigorous Phase II in chemo-ineligible AML to test whether a CD47 antibody can deliver a clinically meaningful survival benefit on top of the current standard. The data presented at EHA 2026 suggests the answer is yes.
The trial enrolled treatment-naïve AML patients who were deemed ineligible for intensive chemotherapy. Patients were randomized 1:1 to receive AZA + VEN plus either ligufalimab or placebo. The primary endpoint was event-free survival; secondary endpoints included overall survival, composite complete response rate, MRD negativity, and duration of CRc.
| Endpoint | AZA + VEN + Ligufalimab | AZA + VEN + Placebo | Effect |
|---|---|---|---|
| Median event-free survival (months) | 9.1 | 6.9 | Hazard ratio 0.46 |
| 6-month EFS rate | 67.8% | 55.5% | +12.3 percentage points |
| 9-month EFS rate | 53.2% | 14.1% | +39.1 percentage points |
| 6-month OS rate | 83.3% | 73.2% | +10.1 percentage points |
| 9-month OS rate | 78.7% | 43.1% | +35.6 percentage points |
| Composite CR rate (CRc) | 56.7% | 53.3% | +3.4 percentage points |
| CRc with MRD negativity | 46.7% | 36.7% | +10.0 percentage points |
| Median duration of CRc (months) | 10.4 | 5.6 | +4.8 months |
The hazard ratio of 0.46 for EFS is the headline — it means that at any given time, a patient on the ligufalimab combination was roughly half as likely to have had a relapse, treatment failure, or death as a patient on AZA + VEN alone. The 9-month EFS gap (53.2% vs 14.1%) is striking because it points to durable separation rather than a short-term response blip. The MRD-negativity rate (46.7% vs 36.7%) suggests the depth of remission is greater with the triplet, which historically correlates with longer relapse-free intervals.
Safety, per Akeso's press release, was manageable. The frequency of treatment-emergent adverse events (TEAEs) and serious adverse events was comparable between the two arms. The most common TEAEs were consistent with what is expected in the AZA + VEN setting in older AML patients — cytopenias, nausea, fatigue, infections. Notably, there was no signal of the kind of hemolysis or red-cell agglutination that complicated magrolimab development, which lines up with Akeso's design claim of Fc-engineering to spare erythrocyte CD47.
The CRc improvement looks modest on paper — 56.7% versus 53.3% — but the response-rate endpoint has largely maxed out in AML. The remaining battleground is durability, and that is exactly where ligufalimab separated itself. The median duration of CRc nearly doubled (10.4 vs 5.6 months), the EFS hazard ratio cut by more than half, and the 9-month OS gap of 35.6 percentage points is the kind of survival signal that prompts regulatory submissions.
EHA (the European Hematology Association) is the world's largest hematology meeting. Oral presentations are reserved for abstracts the program committee judges to be the most practice-changing of the congress — typically late-stage randomized data with a clear implication for clinical care. The AK117-206 readout, with its hazard ratio of 0.46, the 9-month EFS gap, and the MRD-negative CRc benefit, fits the bill.
The data lands in the same session as updated readouts on menin inhibitors (revumenib, ziftomenib) and on next-generation FLT3 inhibitors, which means the international AML audience that convened in Milan was comparing all of these third-drug candidates head-to-head. Ligufalimab's CD47 mechanism is mechanistically distinct from menin inhibition and from FLT3 kinase blockade, which makes it complementary rather than competitive — a factor that should help Akeso position the drug as a partner to whatever targeted agent a given patient's tumor genetics call for.
Ligufalimab is not the only CD47 antibody in development, but it is the only one in late-phase AML randomized trials right now. The CD47 class has been bruised by magrolimab's MDS failure, by the discontinuation of several earlier programs, and by persistent concerns about red-blood-cell binding.
Akeso's commercial position is reinforced by the fact that the molecule is wholly owned — no global out-license to a Western partner, unlike the Antengene-UCB and InnoCare deals that have been common in the Chinese bispecific space. If the AK117-206 data holds in Phase III and Akeso files for NMPA approval, the company retains worldwide economics.
For an older AML patient outside China who has run through — or is ineligible for — intensive chemotherapy and AZA + VEN, the options have historically been thin. Some Western centers offer clinical-trial slots for menin inhibitors in KMT2A-rearranged or NPM1-mutant AML, or for FLT3 inhibitors in FLT3-mutated disease. For the 50% or more of patients who do not have those targetable mutations, CD47 plus the standard backbone is now the most promising active development path.
Three access routes exist for international patients interested in ligufalimab today:
NMPA breakthrough-therapy designation for ligufalimab in AML is plausible in 2026, with potential conditional approval by 2027 if the Phase III confirms the AK117-206 signal. A parallel FDA filing would likely follow the Phase III readout, with potential US approval in 2028 or later. International patients who enter the program now through clinical-trial or Lecheng pathways are likely to see drug supply before it is commercially available in their home countries.
The natural next step is a randomized Phase III trial of AZA + VEN with or without ligufalimab in chemo-ineligible frontline AML. Akeso has not yet publicly disclosed the design, but a 1:1 randomized, double-blind, placebo-controlled registration study in roughly 400-500 patients would be the standard regulatory template. Endpoint selection matters: if the FDA accepts EFS as a surrogate endpoint, the trial could read out in 2-3 years; if it insists on overall survival, expect 4-5 years.
Two combination strategies are also worth watching:
Akeso's broader pipeline — ivonescimab (PD-1/VEGF bispecific), cadonilimab (PD-1/CTLA-4 bispecific), the HER3 ADC AK138D1 — is large enough that combination programs with ligufalimab are plausible. Hematologic and solid-tumor pipelines at the same company are unusual but not unprecedented; the molecular logic (CD47 is expressed on many solid tumors as well) supports cross-tumor development.
The AML field has had a remarkable 18 months. FDA approvals of revumenib and ziftomenib in KMT2A-rearranged AML, expanded FLT3 inhibitor indications, and the gradual replacement of intensive chemo with targeted triplet backbones have all moved the needle. Ligufalimab's contribution would be unique: a CD47 antibody working in the largest AML subgroup (chemo-ineligible older patients) where there is currently no targeted therapy.
For inbound patients, the practical question is not whether ligufalimab is the best drug in its class — the data are too early to say — but whether it is available through a pathway they can access. The answer for many older AML patients outside China is: yes, but you have to do the legwork. The Hainan Lecheng zone, the cross-border clinical-trial mechanism, and the major academic hematology centers in Beijing, Shanghai, and Tianjin all have English-language intake coordinators who can assess eligibility.
Ligufalimab's AK117-206 Phase II readout at EHA 2026 is the strongest CD47 antibody data in chemo-ineligible AML we have seen to date. The hazard ratio of 0.46, the 9-month EFS gap of nearly 40 percentage points, and the doubling of median CRc duration are not the kind of improvements that happen by chance in a randomized, double-blind, placebo-controlled trial. They are the kind of improvements that change practice.
Whether the data hold in Phase III is the next 18-36 months of work. For international patients with older, chemo-ineligible AML, the immediate question is access — and the pathways exist, through Chinese clinical-trial enrollment, the Hainan Lecheng zone, and major academic hematology centers that participate in Akeso's program.
We will follow up with a Phase III design piece once Akeso publicly announces the registration trial, and with a separate article on Hainan Lecheng access pathways for international AML patients once ligufalimab's status there is confirmed.
Considering AML treatment in China?
China Hospitals Guide coordinates inbound AML care at Peking University People's Hospital, Ruijin Hospital Shanghai, the Institute of Hematology at the Chinese Academy of Medical Sciences, and Hainan Lecheng pilots.
Request a Patient Intake