Early trajectory of inflammatory cytokines following tarlatamab administration in three advanced SCLC patients.

Introduction
Advanced small-cell lung cancer (SCLC) has one of the poorest prognoses, posing a major challenge to clinicians [1]. Recently, the bispecific T-cell engager (BiTE) tarlatamab, which redirects cytotoxic T-cells to delta-like ligand 3 expressed on SCLC cells, showed clinical benefit in previously treated SCLC patients in the DeLLphi-300 and DeLLphi-301 trials [2, 3]. Based on these findings, tarlatamab was approved for advanced SCLC after at least two prior treatment regimens.
Although BiTE therapy induces robust immune activation, it frequently causes serious adverse events such as cytokine release syndrome (CRS) and immune effector cell-associated neurotoxicity syndrome [2, 3]. CRS is particularly concerning, as it can be life-threatening in severe cases. It is thought to result from excessive T-cell responses and cytokine storms, similar to mechanisms observed in CAR-T therapy [4, 5]. However, the early in vivo cytokine dynamics following BiTE therapy remain unclear.
Here, we report early inflammatory cytokine trends in three advanced SCLC patients treated with tarlatamab.

Patients and methods
Three advanced SCLC patients who had completed second- or third-line therapy were included. Tarlatamab was approved in Japan in December 2024, and became available at our hospital in May 2025. The drug was administered between May and June 2025. Serum samples were collected at baseline, day 2, day 9, and additional time points during routine clinical testing. Samples were stored at − 80 °C. Cytokines were quantified using the Bio-Plex Pro Human Cytokine Screening Panel, 48-Plex (Bio-Rad, 12,007,283). Samples were diluted 1:4 in PBS and analyzed in duplicate. Data were acquired using the Bio-Plex 200 system and Bio-Plex Manager Software v6.1 (Bio-Rad) per manufacturer's instructions. Cytokine concentrations were determined using a standard curve from recombinant cytokine standards. ELISA tests followed the product protocols.

Case presentation
Case 1: A 65-year-old man was diagnosed with extensive-stage (ES) SCLC (cT2aN2M1c, stage IVB) in January 2024. He received four cycles of cisplatin, etoposide, and durvalumab as first-line therapy, followed by four cycles of durvalumab maintenance, after which his disease progressed. Amrubicin monotherapy began as second-line therapy in October 2024, but progression was noted by April 2025. Tarlatamab was introduced as a third-line therapy in May 2025. On day 2 post-administration, he developed fever (38 °C) and oxygen desaturation and was diagnosed with grade 2 CRS. Dexamethasone and tocilizumab were administered, leading to recovery by day 4. However, he subsequently developed interstitial pneumonia and upper gastrointestinal bleeding, precluding further treatment. He died 5 days after tarlatamab initiation.
Case 2: A 70-year-old man was diagnosed with limited-stage SCLC (pT2aN0M0, stage IB) and underwent surgery in June 2021. He received two cycles of adjuvant carboplatin and etoposide. In March 2023, he relapsed with intrapulmonary and left tenth rib metastases. He underwent 12 cycles of amrubicin as second-line therapy and 11 cycles of nogitecan as third-line therapy from September 2024. Following progression, tarlatamab was introduced in May 2025. He experienced CRS twice: first on day 2, presenting with fever and hypoxemia, and again on day 9 with fever (≥ 38.5 °C). Both episodes responded to dexamethasone and tocilizumab within 48 h. The tarlatamab treatment was continued. The only notable adverse event was grade 2 dysgeusia, which developed during the second course and has persisted thereafter. Following the third course, he achieved a partial response, and tumor markers (pro-gastrin-releasing peptide and neuron-specific enolase) markedly declined. After the sixth course evaluation, a computed tomography scan revealed progression of liver metastases, leading to a diagnosis of progressive disease with a progression-free survival of 4.0 months.
Case 3: A 64-year-old man was diagnosed with ES-SCLC (cT4N0M1c, stage IVB) in August 2024. He received four cycles of carboplatin, etoposide, and atezolizumab as first-line therapy. Maintenance atezolizumab was discontinued due to immune-related hepatitis. After high-dose corticosteroids, he remained on 8 mg/day methylprednisolone. With disease progression and worsening pleural effusion, amrubicin was initiated but was ineffective. Tarlatamab was started as third-line therapy in May 2025. No notable treatment-related adverse events, including CRS, occurred during the first treatment course. After the first course, his cancer rapidly progressed, and his general condition deteriorated due to worsening pain from bone metastases and the development of superior vena cava syndrome. He died on day 39 following the initiation of tarlatamab treatment.

Results
The characteristics of the three patients are summarized in Table 1. Blood samples were collected per protocol in all cases. Figure 1 outlines the clinical course, timing of serum collection, and CRS treatments.
Figure 2 presents inflammatory cytokine trajectories. Interleukin (IL)-6, IL-1 receptor antagonist (Ra), IL-10, and granulocyte colony-stimulating factor (G-CSF) rose from baseline at the initial CRS onset in cases 1 and 2. These decreased following steroid and tocilizumab treatment and remained suppressed during the second CRS in case 2. In case 3, which did not develop CRS, levels showed no significant increase (Fig. 2A). Monokine induced by interferon (IFN)-γ (MIG) and IFN-γ-induced protein 10 (IP-10) increased during CRS in case 1 and both CRS episodes in case 2, forming a bimodal pattern. In case 3, changes from baseline were minimal (Fig. 2B). Hepatocyte growth factor (HGF), IFN-α2, IFN-γ, IL-2Rα, and macrophage inflammatory protein (MIP)-1α increased only in case 1 during CRS onset. In contrast, cases 2 and 3 showed no substantial changes. At baseline, HGF, IFN-α2, and IFN-γ showed comparatively higher concentrations in case 1 than in the other two cases (Fig. 2C). Monocyte chemoattractant protein (MCP)-1 and TNF-related apoptosis-inducing ligand (TRAIL) were elevated during both CRS episodes in case 2 but remained stable in cases 1 and 3 (Fig. 2D). Regulated on activation, normal T-cell expressed and secreted (RANTES) decreased during CRS in case 1 yet rose and remained elevated in case 2. In case 3, it was higher at baseline and stayed elevated throughout the treatment (Fig. 2E).
All cytokine profiles are shown in the online supplemental figure.

Discussion
This preliminary study demonstrates early inflammatory cytokine kinetics following tarlatamab administration and explores their role in CRS pathophysiology and treatment response. A key strength is the comprehensive profiling of 48 cytokines before and after tarlatamab initiation. To our knowledge, this is the first exhaustive investigation of early cytokine trajectories after tarlatamab.
In cases 1 and 2, IL-6, a central proinflammatory cytokine, was significantly elevated during the first CRS episode, alongside immunosuppressive cytokines (IL-1Ra, IL-10) and G-CSF. This pattern likely reflects concurrent systemic inflammation and compensatory regulation. After corticosteroids and tocilizumab (an IL-6 receptor antagonist), cytokine levels declined. Notably, in case 2, these levels remained low during the second CRS, suggesting sustained anti-inflammatory effects from initial treatment. In contrast, the absence of CRS in case 3 may be attributed to ongoing methylprednisolone therapy for prior irAE hepatitis.
Despite low IL-6 levels, case 2 experienced a second CRS episode, indicating that cytokines beyond the IL-6 axis may drive recurrence. During this phase, MIG, IP-10, MCP-1, and TRAIL were elevated while minimal or absent increases were observed for IL-1Ra, IL-10, IFN-γ, and IFN-α2. MIG and IP-10 are IFN-γ-inducible chemokines involved in T-cell and NK-cell recruitment [6, 7]. MCP-1 recruits monocytes/macrophages, while TRAIL promotes apoptosis [6, 8]. Their elevation suggests alternative immune pathways, particularly cytotoxic cell recruitment and macrophage activation, may contribute to CRS, even with IL-6 suppression.
Distinct cytokine patterns were observed between cases 1 and 2. HGF, IFN-α2, IFN-γ, and IL-2Rα were elevated only in case 1, while MCP-1 and TRAIL increased exclusively in case 2. These cytokines have been reported in CRS associated with CAR-T or ICI therapies [9, 10]. However, the variation across cases suggests CRS is not a uniform entity but comprises multiple subtypes within a complex cytokine network. Recognizing this heterogeneity is essential for developing therapies beyond corticosteroids and IL-6 blockade, and for guiding personalized management.
RANTES dynamics also varied. Case 3, which did not develop CRS, had persistently high baseline RANTES. In case 1, RANTES decreased during CRS, while in case 2, it increased during the initial CRS and remained elevated, yet the patient still developed a second CRS. RANTES, a chemokine that promotes migration of T-cells, monocytes, and eosinophils, is generally proinflammatory [11]. However, it may also recruit regulatory T-cells to tumors, indirectly dampening inflammation [12]. These findings suggest that the dual inflammatory and immunosuppressive functions of RANTES and its interplay with other cytokines may contribute to differing CRS phenotypes. Prior studies have shown elevated RANTES in ICI-induced CRS with increasing severity [10], which differs from our findings.
This preliminary study is limited by its single-center design and small sample size. Patient selection bias and clinical heterogeneity may have influenced the results. Further prospective studies with larger cohorts are needed to validate and extend these findings.

Conclusion
Cytokine profiling in these three cases illustrates the complexity of CRS pathophysiology and the involvement of cytokine networks beyond the IL-6 axis. Case 2, where CRS occurred despite low IL-6 and elevated alternative cytokines, supports the presence of CRS subtypes. Differences in cytokine signatures between cases further underscore this heterogeneity. The unique RANTES dynamics suggest a possible suppressive mechanism in CRS progression. These findings may guide future biomarker development, disease classification, and therapeutic strategies beyond IL-6 inhibition, advancing personalized CRS management based on individual cytokine profiles.

Supplementary Information
Below is the link to the electronic supplementary material.