Amgen's Imdelltra plus Imfinzi cuts death risk 58% in first-line ES-SCLC

What happened

On 24 October 2026, AstraZeneca and Amgen reported that adding tarlatamab (Imdelltra) to durvalumab (Imfinzi) as first-line maintenance therapy reduced the risk of death by 58% against Imfinzi alone in extensive-stage small cell lung cancer, from the Phase III DeLLphi-305 trial, at a median follow-up of roughly 16 months ^[1]. Nearly 80% of patients on the combination were alive at one year, against 53% on Imfinzi alone, and progression-free survival more than tripled ^[2]. The late-breaking results were presented at ESMO's Presidential Symposium (abstract LBA3) and published simultaneously in the New England Journal of Medicine ^[2]. Endpoints News covered the readout from the congress in Madrid ^[3]. The mature data follow the positive interim analysis AstraZeneca disclosed in an 8 September 2026 filing ^[4].

Why it matters

Tarlatamab is a DLL3-directed T-cell engager approved today only for patients whose ES-SCLC has progressed on or after platinum chemotherapy, the second-line-and-beyond setting ^[5]. DeLLphi-305 tests it one step earlier, as maintenance after first-line treatment, where the eligible population is far larger: AstraZeneca puts the 2026 global ES-SCLC treated population near 195,000, with median survival on current standard of care around one year ^[4]. A 58% reduction in the risk of death on the hardest endpoint is a large effect.

Durvalumab is already the backbone immunotherapy for ES-SCLC through the CASPIAN trial ^[4], so the combination extends a regimen AstraZeneca already sells rather than displacing it. For Amgen, the commercial read is direct: tarlatamab sales rose 115% year over year to $288 million in the second quarter of 2026 ^[6]. A move into first-line maintenance addresses a much larger patient pool, though the dossier contains no regulatory filing, acceptance or decision date for that expansion, so the timeline to a label change is not established here.

Who is exposed

Two challengers are developing in the same disease with different modalities. Zai Lab's DLL3 antibody-drug conjugate zocilurtatug pelitecan is in the registrational Phase 3 DLLEVATE study in second-line-plus SCLC, with enrollment expected to finish in the first half of 2027 and a possible US accelerated approval submission in 2027 ^[8], and the company flagged an ESMO 2026 data update ^[9]. No head-to-head data exist; the competitive read is Prognyx's inference from separate trials, not a direct comparison.

What to watch

• Amgen's own first-line programs beyond maintenance: a Phase 3 study of tarlatamab plus durvalumab plus chemotherapy against durvalumab plus chemotherapy (NCT07005128) and a Phase 1b first-line combination (NCT05361395), both active but no longer recruiting ^[10][11].
• A Phase 3 trial of subcutaneous versus intravenous tarlatamab in relapsed ES-SCLC (NCT07778316), recruiting toward 400 patients ^[12].
• Zai Lab's DLLEVATE enrollment completion in the first half of 2027 and any 2027 US filing ^[8].

The now-what

For a developer in ES-SCLC, a 58% reduction in the risk of death in first-line maintenance resets the comparator bar: single-agent or chemotherapy-only maintenance control arms now look weak, and trials reading out after this will be judged against a bispecific-plus-immunotherapy benchmark. Options on the table: design against the new combination rather than Imfinzi alone, prioritize mechanisms with activity after tarlatamab (as GSK is doing), or seek differentiation on route and tolerability where a subcutaneous option is in play.

What Prognyx could not verify

The NEJM article's direct link is not in the public record Prognyx retrieved; the exact DeLLphi-305 trial identifier is inferred (NCT06211036 fits the design but is not labelled as such in the sources, and its registry entry still shows an active, no-longer-recruiting status from a December 2025 snapshot) ^[13]; and no SEC filing dated on or after 24 October 2026 covering the full presentation was found.

Verdict: high relevance for ES-SCLC developers; act within the next two quarters to reset comparator assumptions before competing readouts land in 2027.