DeLLphi-305 provides a convincing answer to the efficacy question surrounding earlier DLL3-directed therapy. It also creates a practical one: how quickly can oncology practices deliver tarlatamab safely outside specialized centers?

The phase 3 trial, published in The New England Journal of Medicine and presented at the ESMO Congress 2026, randomized 563 patients with extensive-stage small cell lung cancer without progression after four cycles of platinum-etoposide plus durvalumab to tarlatamab plus durvalumab or durvalumab alone. Patients had ECOG performance status 0 or 1. Asymptomatic brain metastases, treated or untreated, were allowed.

At a median follow-up of approximately 16 months, median overall survival from maintenance randomization was 21.1 months with the combination and 12.7 months with durvalumab alone. The hazard ratio for death was 0.42 (95% CI, 0.32–0.54; P<0.001).

Twelve-month survival was 79% versus 53%.

Median progression-free survival was 7.4 versus 1.9 months, with an HR for progression or death of 0.34 (95% CI, 0.28–0.42; P<0.001). Twelve-month PFS was 38% versus 10%.

Among patients with measurable disease at randomization, the objective response rate was 26.4% with tarlatamab-durvalumab and 4.1% with durvalumab. Median response duration was 12.7 and 5.6 months, respectively.

“What this shows is that by giving tarlatamab with durvalumab in the maintenance setting, patients have longer-term disease control and they live longer,” said lead investigator Dr. Jacob Sands, associate chief of the Lowe Center for Thoracic Oncology at Dana-Farber Cancer Institute.

The Clinical Meaning Is the Timing

Tarlatamab already improves survival when given after platinum-based therapy fails. DeLLphi-305 shows that waiting for relapse is probably no longer the best strategy for an eligible patient who has completed durvalumab-based induction without progression.

Dr. Sands described the result as establishing a new standard of care. The magnitude of the OS effect, its consistency across prespecified subgroups, and the separation of both OS and PFS curves support that conclusion.

Benefit was maintained in patients with baseline brain metastases, with an OS HR of 0.49, although this subgroup analysis should not be treated as independently powered. Patients with liver metastases also benefited.

For current practice, the most likely candidate resembles the trial population: ECOG 0/1, no progression after platinum-etoposide-durvalumab and no symptomatic CNS disease. Evidence remains limited for patients with poorer performance status or unresolved toxicity after induction.

The combination is not yet FDA-approved for first-line maintenance. Until the indication is expanded, clinicians must distinguish a practice-changing dataset from an approved treatment pathway.

Safety Considerations

Grade 3 or higher treatment-related adverse events occurred in 20.9% with tarlatamab-durvalumab and 10.8% with durvalumab. Grade 3 or higher treatment-emergent events from any cause occurred in 46.8% and 40.6%, respectively.

Treatment-related adverse events led to tarlatamab discontinuation in 3.2%. No fatal treatment-related events occurred in the combination arm.

CRS occurred in 49.6% of patients receiving tarlatamab. Most events were grade 1 (41.8%) or grade 2 (7.4%). One patient, or 0.4%, experienced grade 3 CRS. No grade 4 or fatal events occurred.

ICANS occurred in 1.1%, with no grade 3 or higher cases.

“Tarlatamab did not add much more from a side effect standpoint compared to durvalumab alone,” Dr. Sands said. He identified altered taste as the main longer-term concern for many patients.

The study’s abbreviated monitoring is clinically relevant. Patients were monitored on site for one to two hours after the first two tarlatamab doses, with six to eight hours required in certain regions. In September 2026, the FDA shortened labeled monitoring for the first two doses in relapsed disease to six to eight hours.

The eventual maintenance label will matter. If regulators accept a similar one- to two-hour observation schedule, delivery in community clinics becomes much more realistic. If longer monitoring requirements remain, geography, caregiver availability, and institutional CRS preparedness could limit access.

How Does It Compare With Lurbinectedin Maintenance?

Lurbinectedin plus atezolizumab is FDA-approved as maintenance after atezolizumab, carboplatin, and etoposide based on IMforte. DeLLphi-305 used a different induction and maintenance backbone.

There is no direct comparison between the regimens. Cross-trial efficacy comparisons would be unreliable. Selection may depend on prior checkpoint inhibitor, marrow reserve, CNS disease, expected toxicity, treatment frequency, and the clinic’s ability to monitor early tarlatamab doses.

Still, the DeLLphi-305 OS HR of 0.42 sets a high bar. For patients completing durvalumab-based induction, tarlatamab-durvalumab is likely to become the preferred maintenance strategy once regulatory and operational questions are resolved.

What Is Left at Progression?

Moving tarlatamab forward removes a proven second-line option. Subsequent treatment may include lurbinectedin, platinum rechallenge in selected patients, topotecan, or a clinical trial.

Post-progression treatment data will therefore be important. Clinicians need to know whether early tarlatamab changes sensitivity to later chemotherapy and whether another DLL3-directed approach retains activity.

“There are more promising drugs right now in clinical trials than all the research that has happened up to this point,” Dr. Sands said, emphasizing the importance of trial enrollment.

DeLLphi-305 is not simply another positive maintenance study; it moves an effective second-line treatment into the period when disease is still controlled. The next step is making sure the monitoring requirements and treatment sequence are workable for patients outside major academic centers.