Tarlatamab extends small cell lung cancer survival by 8 months

People with extensive-stage small cell lung cancer who added tarlatamab to their maintenance immunotherapy lived a median of 21.1 months from the start of maintenance, against 12.7 months for those who stayed on durvalumab alone. That gap of roughly eight months comes from the phase 3 DeLLphi-305 trial, presented by Jacob Sands of Dana-Farber Cancer Institute in the Presidential Symposium at the ESMO Congress 2026 in Madrid and published the same day, 24 October 2026, in the New England Journal of Medicine.

Small cell lung cancer has been one of the hardest cancers in which to make progress. The NEJM paper opens by noting that median survival for extensive-stage disease sits at about 12 to 13 months with current treatment. A hazard ratio of 0.42 for death, in a randomised trial of 563 people, is a large effect for this disease. It's also an interim result from an open-label study, which shapes how far it can be pushed, and we cover that further down.

What was tested

Most people diagnosed with extensive-stage small cell lung cancer (where the cancer has spread beyond one side of the chest) start with four cycles of platinum chemotherapy and etoposide, given alongside an immunotherapy drug such as durvalumab (brand name Imfinzi). If the cancer hasn't grown by the end of those cycles, the chemotherapy stops and durvalumab carries on alone as "maintenance". The cancer usually comes back within months.

Tarlatamab (brand name Imdelltra, made by Amgen) works differently from conventional drugs. It's a bispecific T-cell engager: one end of the molecule grabs a protein called DLL3, which sits on the surface of most small cell lung cancer cells, and the other end grabs CD3 on the patient's own T cells. Pulling the two together prompts the T cell to attack the cancer cell. The NEJM authors point out that tarlatamab had already been shown to improve survival as a second-line treatment, after the cancer had progressed. DeLLphi-305 asked whether giving it earlier, while the disease was still under control, would do better.

In the trial, people in the experimental group received tarlatamab at 10 mg every two weeks, after a short step-up phase of smaller doses, plus durvalumab at 1,500 mg every four weeks. The control group had durvalumab alone on the same schedule. Treatment continued until the cancer progressed or side effects became too much. The trial was funded by Amgen and AstraZeneca, which makes durvalumab.

Who took part

The trial randomised 563 patients at 175 sites in 29 countries, according to the trial report summarised by CancerNetwork. Of these, 282 were assigned to tarlatamab plus durvalumab and 281 to durvalumab alone.

To get in, patients had to have completed four cycles of platinum and etoposide, with at least three of those cycles given alongside durvalumab, and their cancer had to be stable or shrinking at that point. People with brain metastases could join if they had no symptoms from them, whether or not the brain disease had been treated. According to Targeted Oncology's recap of the session, about 20% of participants had brain metastases and roughly a third had liver metastases, two groups that usually do badly.

That entry rule matters for anyone reading the numbers. Everybody in this trial had already got through first-line chemoimmunotherapy without their cancer growing. People whose cancer progressed during chemotherapy, or who were too unwell to finish it, weren't eligible, and their outlook is generally worse.

The registry entry is NCT06211036 on ClinicalTrials.gov. It lists a start date of June 2024, with final completion expected in 2028.

Results in plain numbers

The main measure was overall survival, counted from the day of randomisation (the start of maintenance). The key secondary measure was progression-free survival, the time until the cancer grew or the patient died.

This was a planned interim analysis with a data cut-off of 7 July 2026. By then 243 deaths had occurred, about 61% of the number the final analysis is designed around. Median follow-up was 15.9 months in the tarlatamab group and 16.1 months in the durvalumab group. Amgen had announced in September 2026 that the trial had met its overall survival goal at this point, as OncLive reported.

Sources: NEJM abstract (Pérol et al., 24 October 2026) for medians and hazard ratios; 6 and 12 month rates as reported by CancerNetwork and OncLive from the ESMO presentation.

A hazard ratio of 0.42 means that at any given point during follow-up, people on the combination were dying at about 42% of the rate of those on durvalumab alone. Put another way, the risk of death was 58% lower. In absolute terms the clearest figure is the one-year survival rate: about 79 in every 100 people on the combination were alive a year after starting maintenance, against about 53 in every 100 on durvalumab alone. That's roughly 26 extra people alive per 100 treated at that point.

The progression-free survival result is even more lopsided. On durvalumab alone, half the patients had seen their cancer grow within 1.9 months. With tarlatamab added, the median was 7.4 months, and at a year 38% were still free of progression against under 10%.

One caution on the 21.1 month figure. The lower end of its confidence interval is 20.8 months and the upper end couldn't be estimated, which tells you the median is sitting close to the edge of the available follow-up. It could move once more patients have been followed for longer.

Did it work for the harder-to-treat groups?

The benefit held across the subgroups that were presented. Among people with brain metastases the hazard ratio for death was 0.49 (95% CI 0.26 to 0.93), and among those with liver metastases it was 0.41 (95% CI 0.28 to 0.61), per Targeted Oncology. The brain metastasis group was small, which explains the wide interval, but both estimates point the same way as the main result.

Side effects

Adding a second active drug brought more side effects, though the gap was smaller than many clinicians might have expected.

Sources: NEJM abstract and Targeted Oncology's report of the ESMO presentation.

Cytokine release syndrome is the side effect people starting tarlatamab are warned about. When T cells are activated they release signalling proteins that can cause fever, low blood pressure and low oxygen levels. In DeLLphi-305 it affected about half of patients on the combination, nearly all at grade 1 or 2. There was a single grade 3 case (0.4%) and no deaths from it. Immune effector cell-associated neurotoxicity syndrome (ICANS), which can cause confusion or difficulty speaking, affected 1.1%, with none severe.

Cytokine release syndrome tends to happen around the first doses. That's why tarlatamab is started with step-up dosing and observation. OncLive reported on 14 September 2026 that the US Food and Drug Administration had approved a shorter monitoring period after the first two doses, which would ease the practical burden somewhat, though patients still need to be near a centre able to manage the reaction.

The near-identical discontinuation rates (5.7% against 5.8%) are notable. Despite the extra drug, people on the combination weren't stopping treatment because of side effects any more often than those on durvalumab alone.

Limitations

It's an interim analysis. The result crossed the pre-set statistical boundary at about 61% of the planned deaths. Trials stopped or reported early on a strong effect sometimes see that effect shrink with longer follow-up. The hazard ratio of 0.42 is very unlikely to vanish, but its exact size may change.

The trial was open-label. Patients and doctors knew who was getting tarlatamab. That has little effect on overall survival, which is hard to bias, but it can influence progression-free survival assessments and reports of side effects.

Survival is counted from maintenance, not diagnosis. As OncoDaily's write-up points out, the 21.1 month median starts at randomisation, after people had already been through about three months of chemoimmunotherapy. It shouldn't be lined up against first-line trials that count survival from the start of chemotherapy. A patient newly diagnosed can't read 21.1 months as their expected survival from diagnosis.

The comparison was with durvalumab alone. That's the right comparator for this setting, but it leaves open whether giving tarlatamab now beats keeping it in reserve for when the cancer returns. The trial reports don't give details in the sources we've seen about how many people in the control group went on to receive tarlatamab later, which bears directly on that question.

It only covers people who responded to first-line treatment. Roughly speaking, the trial selected people with somewhat better-behaved disease. Results for people who progressed during chemotherapy, or who couldn't complete it, are unknown.

Quality of life data aren't out yet. Patient-reported outcomes were a secondary measure, and we haven't seen them in the published abstract or press coverage.

What it means for patients

For someone with extensive-stage small cell lung cancer whose disease is stable after first-line chemoimmunotherapy, this is the first large randomised trial to show that adding a drug at the maintenance stage substantially extends life. The study authors concluded that the results support tarlatamab plus durvalumab as a new standard of maintenance treatment, and that conclusion is likely to feed into treatment guidelines and regulatory filings.

Whether it's available depends on where you live. Tarlatamab is already approved in several countries for small cell lung cancer that has come back after chemotherapy. The maintenance use tested in DeLLphi-305 would need regulators to extend the licence. The Limbic reported that tarlatamab has been recommended for public reimbursement in Australia for extensive-stage disease, though the exact scope of that listing isn't clear from the report. In the UK, any routine NHS use for maintenance would need a MHRA licence change and a NICE appraisal, and neither had been announced at the time of writing.

Practical points worth raising with an oncologist:

• Whether the cancer is stable after first-line treatment, since that was the gateway into this trial.
• How cytokine release syndrome would be monitored during the first doses, and how far you'd need to travel for that.
• Whether brain metastases, if present, are causing symptoms, as only people without symptoms from them took part.
• Whether a trial or early access programme is open locally. The ClinicalTrials.gov record shows DeLLphi-305 is no longer recruiting, but other tarlatamab studies are.

Final survival figures are due once the planned number of deaths is reached. The registry lists primary completion in July 2027.

This article explains published research and isn't medical advice. Talk to your own specialist team about treatment decisions.