Patients with EGFR-mutated non-small-cell lung cancer who progress on a third-generation tyrosine kinase inhibitor have platinum doublet chemotherapy and not much else. HARMONi tested adding ivonescimab, a bispecific antibody, to that chemotherapy. It was double-blind and placebo-controlled across 114 centres in Asia, Europe and North America, randomising 438 patients with non-squamous stage IIIB to IV disease and ECOG performance status 0 or 1 to ivonescimab 20 mg/kg or placebo with pemetrexed and carboplatin every three weeks. Seventy per cent of participants were Asian.
Progression-free survival, assessed by blinded independent radiology review, was 6.8 months (95% CI 5.7 to 7.1) against 4.4 months (4.1 to 5.5), hazard ratio 0.52 (95% CI 0.41 to 0.66, p<0.0001). That is a substantial and well-measured effect. Overall survival, the co-primary endpoint, was 16.8 months (14.3 to 19.0) against 14.0 months (12.8 to 15.7), hazard ratio 0.79 with a confidence interval of 0.62 to 1.01 — pointing the right way and not yet significant, with 262 deaths among 438 patients at a median 29.7 months.
The safety cost is not trivial. Serious treatment-related adverse events occurred in 28% on ivonescimab against 15% on placebo — nearly double. Grade 3 to 4 thrombocytopenia was 12% against 6%. Treatment-related deaths were four against five, so mortality attributable to treatment did not rise, but the serious-event burden did.
Where this leaves prescribing: a real option in a setting with few, offered with the endpoint stated accurately. Just over two extra months before progression, a survival difference that has not been established, and roughly one additional serious adverse event for every eight patients treated. For a patient prioritising time without symptoms that may be worth taking; for one prioritising time out of hospital it may not. In Indian practice the additional consideration is cost, and a drug whose demonstrated benefit is currently radiological rather than survival-based faces a harder reimbursement argument. Seventy per cent Asian enrolment does at least mean the efficacy data apply to the population here.
- Consider ivonescimab with pemetrexed and carboplatin after third-generation EGFR-TKI failure, stating the endpoint accurately.
- The benefit shown is 2.4 months of progression-free survival; overall survival has not yet separated.
- Counsel on a near-doubling of serious treatment-related adverse events, 28% against 15%.
- Monitor platelets; grade 3 to 4 thrombocytopenia was twice as common with ivonescimab.
- Seventy per cent of participants were Asian, so the efficacy data apply well to Indian patients.
The statistics, in plain English
The two co-primary endpoints tell different stories and both are honest. A progression-free survival hazard ratio of 0.52 with an interval of 0.41 to 0.66 is precise and unambiguous — that effect is real. The overall survival hazard ratio of 0.79 with an interval of 0.62 to 1.01 just fails to exclude no difference; with 262 deaths already recorded this is not a trivially immature analysis, but it is also not a negative result. Watch for the updated survival readout, since a hazard ratio this close to significance can move either way as events accrue. The progression-free survival assessment used blinded independent radiology review, which removes the main bias that inflates this endpoint.
Read the rest in the app
You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.
Scan to keep reading on your phone. No account needed to start.
Tomorrow morning, before your first patient
One edition a day for oncology, written by the desk, every claim tied to its paper. Six minutes.
Get the app — free