One Missed, One Published the Arm That Failed, One Disclosed a Verb
- EyePoint reported LUGANO, the first of two pivotal Phase 3 trials of DURAVYU in wet AMD. The primary endpoint, change in best corrected visual acuity against on-label aflibercept, was not achieved in the full dataset. Non-inferiority appears only in an ad hoc analysis excluding 9 of 211 patients who lost 15 or more letters for reasons unrelated to wet AMD.
- argenx and Zai Lab reported ALKIVIA, in autoimmune myositis. Efgartigimod met the primary endpoint with a 15.4 point greater improvement in Total Improvement Score at Week 52 (47.95 against 32.56, p=0.0011). The Phase 3 portion enrolled 175 patients with a mandated corticosteroid taper.
- Daiichi Sankyo and AstraZeneca reported DESTINY-Lung04, first-line HER2-mutant non-squamous NSCLC. ENHERTU improved progression-free survival against platinum-pemetrexed plus pembrolizumab. HER2 mutations occur in roughly 2% to 4% of NSCLC, and ENHERTU has been approved in later lines since 2022.
- EyePoint’s second pivotal trial, LUCIA, reads out in Q4 2026, with an NDA possible in 1H 2027. Detailed ALKIVIA and DESTINY-Lung04 results go to future medical meetings.
Three toplines, 17 August 2026 · what each company released
| Trial | Claim | Figures disclosed | Not disclosed |
|---|---|---|---|
| LUGANO EyePoint | Primary not achieved | Ad hoc non-inferiority p=0.0096 excluding 9 of 211; 42% treatment burden reduction; 54% supplement-free to Week 56 | The full-dataset primary result itself |
| ALKIVIA argenx, Zai Lab | Primary met | TIS 47.95 vs 32.56, p=0.0011; IMNM 45.05 vs 30.24, p=0.0048; DM 51.51 vs 36.96, p=0.1093 | Full dataset, at a medical meeting |
| DESTINY-Lung04 Daiichi Sankyo, AstraZeneca | Primary met | None | Hazard ratio, median PFS, every number |
All three companies described their results as topline. EyePoint labels its p-values nominal throughout, meaning they are not adjusted for multiplicity and are not confirmatory. ALKIVIA’s dermatomyositis subtype was prespecified and did not reach significance.
CI read
Three Phase 3 toplines in one day, and the volume of detail released runs inversely to the strength of the result. EyePoint missed its primary and published pages of explanation. Daiichi Sankyo hit and published no number at all.
- EyePoint’s non-inferiority exists only after removing patients. Excluding 9 of 211 is an ad hoc decision made after seeing the data, and every p-value is labelled nominal. LUCIA in Q4, not this release, decides whether an NDA is filed.
- argenx published the arm that failed. Dermatomyositis improved 14.5 points at p=0.1093, disclosed beside IMNM’s 14.8 at p=0.0048. Volunteering your own non-significant subtype is a credibility purchase, and it is rare.
- Daiichi Sankyo disclosed a verb. Statistically significant PFS, no hazard ratio, no median, in a mutation present in 2% to 4% of NSCLC. Confidence shows as brevity, and there is nothing here to model until the meeting.
LUGANO topline (EyePoint) →
ALKIVIA topline (argenx and Zai Lab) →
DESTINY-Lung04 topline (Daiichi Sankyo and AstraZeneca) →