ESMO Congress 2026

Tarlatamab is a DLL3×CD3 bispecific T-cell engager whose position in previously treated small-cell lung cancer (SCLC) was established by the phase 3 DeLLphi-304 trial: this multinational, open-label study enrolled 509 patients whose SCLC had progressed during or after platinum-based chemotherapy and randomized them to tarlatamab or chemotherapy (topotecan, lurbinectedin, or amrubicin), with overall survival (OS) as the primary endpoint; the prespecified interim analysis at the January 29, 2025 data cutoff showed significantly longer OS with tarlatamab and was published in the New England Journal of Medicine ^[1]. On these data, tarlatamab received traditional FDA approval in November 2025 for second-line extensive-stage SCLC, based on interim figures of median OS 13.6 vs 8.3 months and HR 0.60 ^[5][6]; a network meta-analysis also ranked it first among systemic therapies for platinum refractory/resistant SCLC (SUCRA 0.96) ^[3].

At ESMO26, Charles Rudin presented the extended follow-up of overall survival from DeLLphi-304: n=509, tarlatamab versus chemotherapy (topotecan, lurbinectedin, or amrubicin), primary endpoint OS, median follow-up of 23 months. Median OS was 13.6 vs 8.4 months, HR 0.62 (0.51–0.76); 18-month OS was 42% vs 24%; in the baseline brain metastases subgroup, OS HR was 0.52 (0.38–0.70); grade ≥3 TRAEs were 28% vs 63%, with no new safety signals. The investigators consider these results to reinforce tarlatamab as standard of care after first-line platinum-based chemotherapy.

Relative to existing evidence, the update advances the picture in two ways. First, with follow-up extended to 23 months the survival benefit remains robust — the HR of 0.62 is broadly consistent with the previously reported 0.60 — and the 18-month OS of 42% vs 24% adds a long-term characterization beyond median OS (13.6 vs 8.3 months) ^[5][6]. Second, the baseline brain metastases subgroup HR of 0.52 (0.38–0.70) provides subgroup-level evidence for DLL3-targeted therapy in patients with brain metastases, while the 28% vs 63% gap in grade ≥3 TRAEs quantifies the safety advantage in terms of high-grade toxicity, echoing the dual efficacy–safety superiority conclusion of the network meta-analysis ^[3].

Limitations: DeLLphi-304 is an open-label design ^[1]; the conference posts only updated OS, the brain metastases subgroup HR, and grade ≥3 TRAEs, while PFS, ORR, patient-reported outcomes, and the sample size and median OS of the brain metastases subgroup are not disclosed or remain to be confirmed; efficacy beyond the 23-month median follow-up also awaits further observation.