Legend Biotech’s CAR T-cell Therapy Shows Early Promise in SCLC, Other Tumors

LB2102, an autologous chimeric antigen receptor (CAR) T-cell therapy targeting delta-like ligand 3 (DLL3), has shown promise in treating small cell lung cancer (SCLC) and some neuroendocrine tumors,¹ based on phase 1 findings presented at the 2026 American Society of Clinical Oncology Annual

Meeting. The data were presented by Zhonglin Hao, MD, PhD, coleader of the Thoracic Oncology Program, Markey Cancer Center, University of Kentucky HealthCare. The novel therapy, developed by Legend Biotech, has been exclusively licensed to Novartis.

LB2102 takes aim at DLL3, a protein highly expressed on the surface of SCLC and other neuroendocrine tumors and largely absent in healthy tissue. What distinguishes LB2102 from conventional CAR T-cell designs is a built-in “armor” component: a dominant-negative transforming growth factor β (TGF-β) receptor (dnTGFBR2) that blocks immunosuppressive TGF-β signaling within the tumor microenvironment, theoretically enhancing CAR T cell activation, proliferation, and antitumor function.²

The first-in-human dose-escalation study (NCT05680922) enrolled 20 heavily pretreated patients with relapsed or refractory SCLC (n=17) or large cell neuroendocrine carcinoma (LCNEC, n=3).^1,2 As Hao explained at the outset, the study population was notable for its clinical complexity. “Most had extensive-stage small cell lung cancer at diagnosis; 70% had a history of brain metastasis,” he said.

In addition, 95% of the patients required bridging therapy during the CAR T cell manufacturing period, which ran approximately 8 to 10 weeks from apheresis to infusion. The study group had received a median of 1 prior line of therapy, but some had as many as 7. All patients who had previously received tarlatamab (Imdelltra; Amgen), a DLL3-targetingbispecific T-cell engager, were excluded from the study.

Seven dose levels were planned, but dose level 6 was skipped based on safety data. Enrollment started at dose level 1 (0.3 × 10⁶ CAR+ T cells/kg) and continued through dose level 7 (16.0 × 10⁶ CAR+ T cells/kg). The therapy demonstrated a manageable safety profile, with no dose-limiting toxicities and no treatment-related deaths. Cytokine release syndrome (CRS) occurred in 30% of the patients, all grade 2 or lower, and immune effector cell–associated neurotoxicity syndrome (ICANS) was observed in 3 patients, with 2 patients experiencing grade 1 ICANS and 1 experiencing grade 3; all cases resolved.¹

Efficacy signals emerged starting at dose level 3 and at higher dose levels (≥2.0 × 10⁶ CAR+ T cells/kg). Among response-evaluable patients at those dose levels, the objective response rate was 28.6%, and the disease control rate was 78.6%, Hao said. Responses were durable; at a median follow-up of 12.5 months, median duration of response (DOR) was 6.5 months with 2 ongoing responses, and median duration of disease control was 6.1 months.¹

Pharmacokinetic analyses showed CAR T cell expansion in all patients at dose level 3 and above, with patients achieving partial responses showing higher peak CAR T cell exposure compared with those with stable or progressive disease. Hao said that a patient treated off-protocol at dose level 3 achieved a partial response lasting more than 12 months, with CT imaging showing approximately 80% tumor reduction. “The findings support the level 3 targeted CAR T cell [therapy] in patients with small cell lung cancer and large cell neuroendocrine carcinoma,” he said during the presentation.

In response to a request from The American Journal of Managed Care® (AJMC®), Legend Biotech CEO Ying Huang, PhD, answered questions about the results:

AJMC: What is the clinical rationale for targeting DLL3 with an autologous CAR T-cell therapy? Why does this approach make sense for a relapsed/refractory population in SCLC or LCNEC?

Huang: DLL3 is highly expressed on small cell lung cancer and large cell neuroendocrine carcinoma tumor cells, making it a rational antigen for targeted approaches. The rationale for an autologous CAR T approach in this population reflects both the biology of SCLC and the limitations of existing therapies. SCLC is characterized by an immunosuppressive tumor microenvironment driven in part by TGF-β signaling, which can impede T-cell activity in the tumor. LB2102 includes a dnTGFBR2 construct designed to reduce TGF-β signaling in the engineered T cells; this is intended to improve resistance to TGF-β–mediated immunosuppression, although clinical benefit remains to be confirmed.

In the relapsed/refractory setting specifically, the rationale is reinforced by the absence of meaningful treatment options. Patients with extensive-stage small cell lung cancer who progress after first-line platinum-based chemotherapy have few effective choices, and responses to subsequent therapies are typically short-lived.

The approval of a DLL3-directed bispecific T-cell engager in the second-line setting has validated the target clinically but does not address all patient populations, particularly those with LCNEC histology, those who are ineligible for or intolerant of bispecific therapy, or those whose disease progresses after DLL3-directed treatment. CAR T offers a mechanistically distinct form of DLL3-directed activity that may be relevant in these settings.

AJMC: The percentage of patients who experienced CRS was far less than what is typically seen in hematological cancers. How is this explained?

Huang: In this study, at a median follow-up of 12.5 months, CRS occurred in 30% (6/20) of patients, all grade 1 and 2; this lower incidence likely reflects differences in tumor anatomy and trafficking compared with hematologic malignancies.

In B-cell lymphomas and multiple myeloma, target antigens are expressed on cells that circulate freely or are readily accessible in bone marrow and lymph nodes. CAR T cells engage tumor cells rapidly and at high frequency after infusion, producing the cytokine release that characterizes CRS with approved hematologic CAR T products.

Solid tumors present a different kinetic environment entirely; CAR T cells must traffic through vascular endothelium, penetrate a physically and immunologically hostile stroma, and engage tumor cells across a substantially smaller surface area. The result is likely due to a more gradual, anatomically constrained pattern of T-cell activation and correspondingly lower systemic cytokine peaks.

AJMC: Will lymphodepletion approaches be similar to what we are familiar with in CAR T-cell therapy in hematological cancers? Why or why not?

Huang: The LB2102 phase 1 study used standard fludarabine and cyclophosphamide (Flu/Cy) lymphodepletion, which is consistent with the conditioning approach established across approved autologous CAR T programs in hematologic malignancies. Flu/Cy remains the most extensively characterized regimen for creating an optimal biological environment for CAR T expansion, and there is no current data-supported rationale to deviate from this backbone in early-phase solid tumor CAR T development.

We will be further exploring the tolerability profile in this population as we move forward. Most grade ≥3 adverse events in the LB2102 study were hematologic and contributed to lymphodepletion rather than to LB2102 itself, an important distinction when characterizing the therapy’s independent safety profile, and a consideration for conditioning regimen optimization as the program advances.

AJMC: The disease control rate of 70% median and duration of response of 6.5 months are on par with several later-stage therapies in SCLC (except for tarlatamab, which has a median DOR of 9 months). Where would a DLL3-targeting CAR T-cell therapy fit into treatment sequencing?

Huang: The phase 1 study was a dose-escalation study with a primary objective of characterizing the safety and tolerability of LB2102 and determining the recommended dose for expansion, not optimizing efficacy nor treatment sequencing, and the results should be interpreted in that context.

The data demonstrate that LB2102 produces meaningful antitumor activity across a broad patient population, including SCLC and LCNEC, and patients with multiple prior lines of therapy, where treatment options are limited.

Where the therapy ultimately fits in the treatment paradigm will be determined by ongoing and future studies, including the separate development program being conducted by Novartis with an optimized manufacturing platform and a more defined patient population.

AJMC: Is there potential to use DLL3 CAR T-cell therapy in SCLC in combination with other therapies, given reduced toxicity relative to other CAR T-cell therapies?

Huang: The manageable safety profile of LB2102 does make combination approaches more clinically feasible to explore than would be the case with therapies carrying heavier toxicity burdens. The rationale for any combination strategy should be driven by biology rather than tolerability alone, and no combination data exist at this stage of development.

Formal evaluation of combination approaches will require further studies. The scope and direction of that investigation will be determined as part of the broader development program.

References

1. Hao Z, Chiappori A, Creelan BC, et al. Preliminary results from an ongoing phase 1 study of LB2102, a dnTGFBR2-armored DLL3-targeted autologous CAR-T cell therapy, in patients with relapsed or refractory SCLC or LCNEC. J Clin Oncol. 2026;44(suppl 16):8012. doi:10.1200/JCO.2026.44.16_suppl.8012
2. Schoenfeld AJ, Chiappori A, Creelan B, et al. Phase 1 study of DLL3-targeted CAR-T cells armored with dnTGFBR2 in small-cell lung and large-cell neuroendocrine cancers. J Thorac Oncol. 2025;20(10 suppl 1):S99-S100. doi:10.1016/j.jtho.2025.09.182