Tarlatamab Plus Durvalumab Maintenance Cuts the Risk of Death by 58% in Extensive-Stage Small Cell Lung Cancer: DeLLphi-305 at ESMO 2026
In this article
DeLLphi-305: Does adding tarlatamab to maintenance durvalumab prolong survival in extensive-stage small cell lung cancer?
Medical content reviewed by Prof. Dr. Mustafa Özdoğan · Last updated: October 2026
The DeLLphi-305 phase 3 trial, presented at ESMO 2026 Presidential Symposium I (LBA3) and published the same day in the New England Journal of Medicine, tested adding tarlatamab to maintenance durvalumab in extensive-stage small cell lung cancer (SCLC) that had not progressed after durvalumab, platinum and etoposide induction. Versus durvalumab alone, the risk of death fell by 58% (HR 0.42) and median overall survival rose from 12.7 to 21.1 months. This use is not yet approved.
- Primary endpoint (overall survival): HR 0.42 (95% CI 0.32–0.54; p<0.001); median 21.1 months (95% CI 20.8–not calculable) vs 12.7 months (11.2–14.8). At 12 months 79.4% vs 53.3% were alive: an absolute difference of 26 points.
- Progression-free survival and response (secondary): progression-free survival HR 0.34 (95% CI 0.28–0.42); median 7.4 vs 1.9 months; at 12 months 38.4% vs 9.5% were progression-free. Objective response rate 26.4% vs 4.1%.
- Safety: cytokine release syndrome (CRS) 49.6% (mostly grade 1–2; grade 3 in a single patient, 0.4%; no grade 4–5); ICANS 1.1%, none grade 3 or higher. Grade 3 or higher adverse events from any cause 46.8% vs 40.6%; discontinuation of tarlatamab for an adverse event 5.7%.
- Limits: open-label, planned interim analysis, median follow-up about 16 months (not the final median); overall survival was measured from maintenance randomization; the control arm was durvalumab alone, not lurbinectedin + atezolizumab. The combination is not yet approved.
Why does this matter?
Extensive-stage SCLC progresses quickly even after first-line platinum-etoposide plus a checkpoint inhibitor: in DeLLphi-305 the control arm, which received durvalumab alone, had a median progression-free survival of 1.9 months and a median overall survival of 12.7 months. According to the companies, median survival from the start of maintenance is about one year. Tarlatamab is a bispecific T-cell engager that binds DLL3, a protein found on roughly 85–96% of SCLC cells, and CD3 on T cells.
Tarlatamab is not a new drug: the FDA granted it accelerated approval in May 2024 (phase 2 DeLLphi-301) and full approval in November 2025 (phase 3 DeLLphi-304: median overall survival 13.6 vs 8.3 months; HR 0.60) for extensive-stage SCLC that progressed after platinum chemotherapy. In first-line maintenance, lurbinectedin + atezolizumab was approved on 2 October 2025 (IMforte: overall survival HR 0.73). DeLLphi-305 is therefore not the first trial to show an overall survival benefit in maintenance; Amgen's claim is that it is the first phase 3 trial including a BiTE to do so. The two approaches were not compared directly.
How was tarlatamab + durvalumab maintenance given?
In DeLLphi-305, patients with extensive-stage SCLC who had not progressed after durvalumab plus platinum-etoposide induction received maintenance durvalumab alone or with tarlatamab. Tarlatamab was step-dosed from 1 mg to 10 mg every 2 weeks and durvalumab every 4 weeks. The risk of death fell by 58%; the combination is not yet approved.
How was the DeLLphi-305 trial conducted?
DeLLphi-305 is a randomized, open-label, international phase 3 trial (NCT06211036). After induction therapy, patients were randomized 1:1 to one of two maintenance arms.
| Feature | Details |
|---|---|
| Design | Phase 3, randomized 1:1, open-label; 175 sites in 29 countries |
| Patients | 563 patients: 282 tarlatamab + durvalumab, 281 durvalumab alone |
| Population | Extensive-stage SCLC; no progression after 4 cycles of platinum-etoposide (at least 3 given concurrently with durvalumab); ECOG 0–1; treated or untreated asymptomatic brain metastases allowed, symptomatic brain metastases excluded |
| Tarlatamab + durvalumab | Tarlatamab 1 mg on cycle 1 day 1, 10 mg on days 8 and 15, then 10 mg every 2 weeks; durvalumab 1500 mg every 4 weeks; until progression or unacceptable toxicity |
| Control | Durvalumab alone (same schedule and duration) |
| Primary endpoint | Overall survival (from randomization at the start of maintenance) |
| Secondary endpoints | Progression-free survival (key secondary), then objective response rate (prespecified hierarchy) |
| Monitoring | After the first two tarlatamab infusions, 1–2 hours of observation in a health care setting (6–8 hours in some regions, including Europe) |
| Analysis | Prespecified interim analysis, data cutoff 7 July 2026; median follow-up 15.9 months (combination) and 16.1 months (control) |
Figure 1 · How were the two maintenance arms organized?
DeLLphi-305 trial flow (simplified)
What were the survival and response results?
The comparisons below are between arms of the same trial and therefore direct. Colored bars are tarlatamab + durvalumab and gray bars are durvalumab alone.
Survival and response rates
DeLLphi-305 · interim analysis, from maintenance randomization
Overall survival · 6 months
Overall survival · 12 months
Progression-free survival · 6 months
Progression-free survival · 12 months
Objective response rate
Objective response rate shows confirmed responses in patients with measurable disease at randomization. All values are between arms of the same trial. Median values are given in the table below.
| Endpoint | Tarlatamab + durvalumab | Durvalumab | Hazard ratio / difference |
|---|---|---|---|
| Overall survival, median (primary) | 21.1 months (95% CI 20.8–NC) | 12.7 months (11.2–14.8) | HR 0.42 (0.32–0.54); p<0.001 |
| Overall survival, 6 months | 91.5% | 74.6% | Difference 17 points |
| Overall survival, 12 months | 79.4% | 53.3% | Difference 26 points |
| Progression-free survival, median | 7.4 months (5.6–9.1) | 1.9 months (1.9–2.0) | HR 0.34 (0.28–0.42); p<0.001 |
| Progression-free survival, 6 months | 53.4% | 17.2% | Difference 36 points |
| Progression-free survival, 12 months | 38.4% | 9.5% | Difference 29 points |
| Objective response rate | 26.4% (20.9–32.5) | 4.1% (2.0–7.4) | p<0.001 |
Overall survival was measured from randomization at the start of maintenance therapy, not from diagnosis or the start of induction. This is a prespecified interim analysis; with a median follow-up of about 16 months, as the presenting investigator noted, this is not the final median survival. Median follow-up for progression-free survival was 13.9 months. CI = confidence interval; NC = not calculable.
How long did responses last?
The response analysis is based on patients with measurable disease at randomization (combination n=235, control n=244). Responses were more frequent and lasted longer with the combination, and many more patients were still on treatment at data cutoff.
| Measure | Tarlatamab + durvalumab | Durvalumab |
|---|---|---|
| Complete response | 0.9% (2 patients) | 0.4% (1 patient) |
| Partial response | 25.5% (60 patients) | 3.7% (9 patients) |
| Stable disease | 54.0% (127 patients) | 30.3% (74 patients) |
| Progressive disease | 18.3% (43 patients) | 61.5% (150 patients) |
| Duration of response, median | 12.7 months (7.4–NC) | 5.6 months |
| Responders still in response at 12 months | 52% | 0% |
| Still on treatment at data cutoff | 43.6% | 11.4% |
| Median duration of treatment | 46.9 weeks | 8.1 weeks |
What happened in patients with brain or liver metastases?
The overall survival and progression-free survival benefit was consistent across prespecified subgroups, including patients with brain or liver metastases and those with only stable disease after induction. Subgroup hazard ratios were not given in the sources, and subgroups are not powered to decide on their own. Reportedly, about 20% of patients had brain metastases (previous or current), about a third had liver metastases, and about 85% had a partial response to induction*; the results therefore reflect a group that left induction with a response.
What are the most common side effects during treatment?
The most prominent event in the combination arm was cytokine release syndrome (CRS), mostly grade 1–2. The companies described the safety profile as consistent with the known profiles of both drugs, with no new safety signals.
Adverse events
DeLLphi-305 · interim analysis
Cytokine release syndrome (any grade)
Grade 3 or higher adverse events (any cause)
Serious adverse events
Grade 3–4 treatment-related adverse events
Red bars show the tarlatamab + durvalumab arm and gray bars the durvalumab-alone arm. "Grade 3 or higher (any cause)" and "grade 3–4 treatment-related" are different definitions and should not be confused. Comparisons are within the same trial.
| Measure | Tarlatamab + durvalumab | Durvalumab |
|---|---|---|
| CRS, any grade | 49.6% | 1.4% |
| CRS grade 1 / grade 2 / grade 3 | 41.8% / 7.4% / 0.4% (1 patient); no grade 4–5 | Not reported |
| ICANS | 1.1% (single grade 2 event); none grade 3 or higher | Not reported |
| Broader neurologic events* | 9.6% | 6.8% |
| Grade 3 or higher adverse events (any cause) | 46.8% | 40.6% |
| Grade 3–4 treatment-related adverse events | 20.9% | 10.8% |
| Serious adverse events | 41.8% | 31.7% |
| Events with a fatal outcome* | 4.6% | 4.7% |
| Treatment-related deaths | Not reported (none) | Not reported |
| Tarlatamab discontinuation for an adverse event | 5.7% | Not applicable |
*Reported by a single secondary source. The company releases give grade 3–4 treatment-related events (20.9% and 10.8%); the 46.8% and 40.6% in secondary sources are grade 3 or higher events from any cause. The most common grade 3 or higher treatment-related events in the tarlatamab arm were decreased neutrophil count (3.2%), decreased appetite (2.1%), neutropenia (2.1%), asthenia (1.8%) and weight loss (1.8%). According to OncoDaily, about half of patients had taste disturbance, mostly grade 1*.
How is cytokine release syndrome managed in practice?
CRS began a median of about 22.5 hours after the most recent dose and occurred mainly after the first two doses. To manage it, 21.6% of patients received glucocorticoids, 3.9% tocilizumab, 1.4% low-flow oxygen and 0.4% vasopressors. Tarlatamab carries a boxed warning in the US label for CRS and neurologic toxicity including ICANS. According to Amgen, the current label calls for 6–8 hours of observation after the first two doses, a vital-signs check the next day, and staying within 1 hour of a health care setting with a caregiver for 48 hours (per OncLive, a shortening of the earlier 22–24 hour observation period). This treatment should be given at experienced centers able to recognize and manage CRS.
How do these results compare with other maintenance and later-line options?
The table puts side by side three trials that showed an overall survival benefit in extensive-stage SCLC, but they are not directly comparable.
| Trial | Setting and comparison | OS HR (95% CI) | Median OS | Regulatory status |
|---|---|---|---|---|
| DeLLphi-305 (n=563) | First-line maintenance; tarlatamab + durvalumab vs durvalumab (induction: durvalumab + platinum + etoposide) | 0.42 (0.32–0.54) | 21.1 / 12.7 months | Not approved |
| IMforte (n=483 randomized) | First-line maintenance; lurbinectedin + atezolizumab vs atezolizumab (induction: atezolizumab + carboplatin + etoposide) | 0.73 (0.57–0.95) | 13.2 / 10.6 months | FDA approved 2 Oct 2025 |
| DeLLphi-304 (n=509) | Second line; tarlatamab vs standard chemotherapy | 0.60 (0.47–0.77) | 13.6 / 8.3 months | FDA full approval Nov 2025 (accelerated approval May 2024) |
The comparison is indirect: induction regimens (durvalumab or atezolizumab), control arms, populations and follow-up differ; DeLLphi-304 covers second-line patients and is given only as context. No trial has compared tarlatamab + durvalumab with lurbinectedin + atezolizumab, so no conclusion about which is better can be drawn.
What is established and what is still uncertain?
Figure 2 · Evidence map
DeLLphi-305, ESMO 2026 LBA3
- Approval: Tarlatamab + durvalumab as first-line maintenance is not approved anywhere. In the US tarlatamab is approved only for extensive-stage SCLC that progressed on or after platinum-based chemotherapy. Amgen says it is moving to share the results with regulators worldwide; no submission status was reported.
- Approved alternatives: Lurbinectedin + atezolizumab has been FDA-approved for first-line maintenance since 2 October 2025 (after atezolizumab, carboplatin and etoposide induction). Durvalumab is approved with platinum-etoposide for first-line extensive-stage SCLC. Availability and reimbursement vary by country and must be checked locally.
- Access: The sources do not mention an access route for the combination outside clinical trials.
- What should patients do? Treatment of extensive-stage SCLC is planned with your oncologist according to response to induction, brain metastases and overall condition. Ask about the options currently available and suitable clinical trials.
Frequently asked questions
- Is tarlatamab + durvalumab maintenance available now?
- Not as an approved treatment. In the US tarlatamab is approved only for extensive-stage SCLC that progressed after platinum chemotherapy. Amgen said it will share the data with regulators worldwide; no submission status was reported.
- How much longer did patients live?
- In this interim analysis median overall survival was 21.1 vs 12.7 months, and 79.4% vs 53.3% were alive at 12 months. With a median follow-up of about 16 months, the final median is not yet known.
- What is the most important side effect?
- Cytokine release syndrome: 49.6%, mostly grade 1–2, with grade 3 in one patient. ICANS was 1.1%. Grade 3–4 treatment-related events were 20.9% vs 10.8%, and tarlatamab was discontinued for an adverse event in 5.7%.
- Is it better than lurbinectedin + atezolizumab?
- Unknown. No trial has compared the two regimens directly; induction regimen, control arm and follow-up differ, so the hazard ratios of 0.42 and 0.73 cannot be compared directly.
Glossary
- Maintenance therapy
- Treatment given after induction therapy, when the disease has not progressed, to keep it under control.
- Extensive-stage SCLC
- The stage of small cell lung cancer that has spread beyond one radiation field to distant organs.
- DLL3
- A surface protein found on most small cell lung cancer cells and at very low levels in healthy tissue.
- Bispecific T-cell engager (BiTE)
- An antibody that binds a protein on tumor cells and CD3 on T cells at the same time, directing the T cell to the tumor.
- Cytokine release syndrome (CRS)
- A reaction to immune cell activation that can cause fever, low blood pressure and low oxygen.
- ICANS
- Neurologic toxicity seen with T-cell engaging therapies that can affect consciousness and speech.
- Hazard ratio (HR)
- A value below 1 means lower risk of the event in the experimental arm; 0.42 equals roughly a 58% lower risk.
DROZDOGAN Academy Commentary
DeLLphi-305 shows that adding a bispecific T-cell engager to maintenance therapy for extensive-stage SCLC can markedly improve overall survival, but this is an interim analysis and independent confirmation from the full text is not yet available.
⚡ Why it matters
For years, maintenance therapy in extensive-stage SCLC was a checkpoint inhibitor alone and survival remained short. A rise in median overall survival from 12.7 to 21.1 months when tarlatamab is added is an effect of a size rarely seen in this disease, and it suggests that the DLL3-directed bispecific approach could move from the second line to the first line.
✅ Strengths
Randomized phase 3 design; 563 patients; overall survival as the primary endpoint; consistent improvement in progression-free survival and response rate; the same direction across subgroups including brain and liver metastases; CRS mostly low grade, ICANS rare, grade 3 or higher adverse event rates close to each other, and no treatment-related deaths reported.
⚠️ Limitations
Open-label design, interim analysis and roughly 16 months of follow-up; overall survival is not yet mature. The control arm was durvalumab alone; superiority over the approved lurbinectedin + atezolizumab was not tested. Eligibility criteria (ECOG 0–1, no progression after induction) limit the results to patients who responded well to induction. Quality of life, subsequent therapies and full adverse event details remain incomplete in the secondary sources.
🩺 In practice
The combination is not approved, so it is not an available standard today. Patients with extensive-stage SCLC who complete induction should discuss options and suitable clinical trials with their oncologist. Tarlatamab requires an observation infrastructure for CRS and neurologic toxicity.
❓ Open questions
What will the final median overall survival be? Does the benefit differ by DLL3 expression or response to induction? How does tarlatamab + durvalumab compare with lurbinectedin + atezolizumab? How practical and accessible will observation and CRS management be in real life? How is quality of life affected?
Bottom line: One of the strongest overall survival signals so far in extensive-stage SCLC maintenance; but an interim analysis, a single control arm and the lack of approval call for caution.
- This article was prepared in October 2026.
- Figures come from the Amgen (PR Newswire) and AstraZeneca (copy of the Business Wire release) press releases of 24 October 2026 and from CancerNetwork and OncoDaily reports. According to the companies, the results were published simultaneously in the New England Journal of Medicine with the ESMO 2026 LBA3 presentation; the full text was not accessible.
- Fatal-outcome event rates, broader neurologic events, taste disturbance and baseline patient characteristics (brain and liver metastases, response to induction) come from a single secondary source and are marked (*); they were not confirmed against the full publication.
- The sources label grade 3 or higher adverse events differently: the companies give grade 3–4 treatment-related events as 20.9% and 10.8%, while secondary sources give grade 3 or higher events from any cause as 46.8% and 40.6%; this article presents them as separate measures.
- The data come from a company-sponsored trial (sponsored by Amgen; AstraZeneca provided partial funding and durvalumab). The phrases "unprecedented", "new standard" and "first phase 3 including a BiTE" belong to the companies and investigator; lurbinectedin + atezolizumab previously showed an overall survival benefit in maintenance.
- Overall survival was measured from maintenance randomization; because this is an interim analysis, the median is not final. The sources call the analysis "interim" in some places and "primary" in others.
- The charts show only direct comparisons between arms of the same trial; the table with IMforte and DeLLphi-304 is indirect and given as context only.
- The figures are derived from verified statements in the text and contain no new data. The Academy commentary is our own.
References
- Amgen. IMDELLTRA in combination with IMFINZI reduced risk of death by an unprecedented 58% in first-line extensive-stage small cell lung cancer. Press release, 24 Oct 2026 (PR Newswire). https://www.prnewswire.com/news-releases/imdelltra-in-combination-with-imfinzi-reduced-risk-of-death-by-an-unprecedented-58-in-first-line-extensive-stage-small-cell-lung-cancer-302918652.html
- AstraZeneca. IMFINZI (durvalumab) plus tarlatamab-dlle reduced the risk of death by 58% in 1st-line extensive-stage small cell lung cancer in the DeLLphi-305 Phase III trial. Press release, 24 Oct 2026 (copy of the Business Wire release). https://www.financialcontent.com/article/bizwire-2026-10-24-imfinzi-durvalumab-plus-tarlatamab-dlle-reduced-the-risk-of-death-by-58-in-1st-line-extensive-stage-small-cell-lung-cancer-in-the-dellphi-305-phase-iii-trial
- Sands J, et al. Tarlatamab plus durvalumab vs durvalumab as first-line maintenance therapy in extensive-stage small cell lung cancer (DeLLphi-305). ESMO Congress 2026, Madrid, Presidential Symposium I, abstract LBA3, 24 Oct 2026; simultaneous publication in N Engl J Med 2026 (per Amgen and AstraZeneca). https://doi.org/10.1056/NEJMoa2609482
- CancerNetwork. Tarlatamab maintenance combination extends survival in extensive-stage small cell lung cancer. Oct 2026. https://www.cancernetwork.com/view/tarlatamab-maintenance-combo-extends-survival-in-extensive-stage-sclc
- OncoDaily. DeLLphi-305: Tarlatamab–Durvalumab Redefines First-Line Maintenance in Extensive-Stage Small Cell Lung Cancer. Oct 2026. https://oncodaily.com/oncolibrary/lung-oncology/dellphi-305-esmo26
- Amgen. IMDELLTRA in combination with IMFINZI demonstrated landmark improvement in overall survival in first-line extensive-stage small cell lung cancer. Press release, 8 Sep 2026. https://www.amgen.com/newsroom/press-releases/2026/09/imdelltra-in-combination-with-imfinzi-demonstrated-landmark-improvement-in-overall-survival-in-first-line-extensive-stage-small-cell-lung-cancer
- Amgen. FDA grants full approval to Amgen's IMDELLTRA in extensive-stage small cell lung cancer. Press release, 19 Nov 2025. https://www.amgen.com/newsroom/press-releases/2025/11/fda-grants-full-approval-to-amgens-imdelltra-in-extensive-stage-small-cell-lung-cancer
- OncLive. FDA approves lurbinectedin plus atezolizumab as first-line maintenance therapy for extensive-stage small cell lung cancer (IMforte). Oct 2025. https://www.onclive.com/view/fda-approves-lurbinectedin-plus-atezolizumab-as-first-line-maintenance-therapy-for-es-sclc
- OncLive. FDA reduces monitoring time for first two doses of tarlatamab in extensive-stage small cell lung cancer. Sep 2026. https://www.onclive.com/view/fda-reduces-monitoring-time-for-first-two-doses-of-tarlatamab-es-sclc
- Dana-Farber Cancer Institute. Dana-Farber presents phase 3 lung cancer data in ESMO Presidential Symposium. Oct 2026. https://www.dana-farber.org/newsroom/news-releases/2026/dana-farber-presents-phase-3-lung-cancer-data-in-esmo-presidential-symposium
- ClinicalTrials.gov. DeLLphi-305 (NCT06211036). https://clinicaltrials.gov/study/NCT06211036
Was this article helpful?
Your feedback helps us improve our content.
