Genomic and Transcriptomic Correlates of Response to Tarlatamab in Small Cell Lung Cancer

PurposeTarlatamab is a DLL3-directed bispecific T-cell engager demonstrating clinically meaningful activity in relapsed small cell lung cancer (SCLC) in the phase II DeLLphi-301 trial. Determinants of tarlatamab sensitivity and resistance are incompletely understood, and thus we sought to identify genomic and transcriptional correlates of tarlatamab sensitivity using a clinical sequencing pipeline at a single comprehensive cancer center. Experimental DesignWe performed a retrospective, single-institution analysis of 12 patients with SCLC treated with tarlatamab. Whole-exome sequencing (WES) and exome-capture whole-transcriptome sequencing (WTS) were performed on 12 samples, and two matched samples after treatment with tarlatamab. Integrative analysis examined correlation between molecular features and clinical outcomes. ResultsThe overall response rate was 50%, which was consistent with outcomes reported in the DeLLphi-301 trial. Differences between SCLC driver alterations and tumor mutational burden were not significant between responders and non-responders, but homologous recombination deficiency scores were higher in responsive tumors. DLL3 expression was significantly greater in responders and demonstrated predictive discrimination for clinical response (AUC 0.83). Tumors responsive to tarlatamab were predominantly ASCL1-driven (SCLC-A) and demonstrated increased immune activation, such as enrichment of cytotoxic T-cell, NK-cell, and T cell transcriptional programs. Transcriptional subtype and a composite metric consisting of DLL3 expression and immune activity (DLI score) further discriminated between responders and non-responders (sensitivity 0.83, specificity 1). Paired post-treatment sample analysis identified loss of ASCL1 lineage and emergence of YAP1 expression and downregulation of DLL3, consistent with lineage plasticity as a mechanism of acquired resistance. ConclusionsSensitivity to tarlatamab is correlated with a combination of increased DLL3 expression, ASCL1-driven lineage, and an increased immune activation. Lineage state reprogramming and decrease in DLL3 expression accompany acquired resistance to tarlatamab. These findings highlight the utility of RNA based biomarkers which integrate target expression, lineage state, and immune context to guide tarlatamab therapy in SCLC. Prospective validation of the whole-transcriptome DLI score and transcriptional subtype will inform tarlatamab response prediction.