Abstract 5339: Prevalence and distinct biological programs of small cell lung cancer molecular subtypes revealed by real-world clinico-genomics database.

Abstract Background: Molecular subtypes of Small Cell Lung Cancer (SCLC) highlight patient heterogeneity and continue to carry potential as predictive biomarkers. Resistance to Immune Checkpoint inhibitor (ICI) and impact of RW treatments remain outstanding. We aim to enhance understanding of SCLC subtype biology and therapeutic implications by leveraging the multi-modal RWD with curated outcome and treatment pattern. Methods: We analyzed a US-based SCLC cohort (381 samples; 376 patients) from Caris-ConcertAI linked clinico-genomic database. We performed Non-negative Matrix Factorization (NMF) on transcriptomic data from patients with treatment-naïve samples (n=230), and applied to treatment-experienced samples (n=66). Subsequent treatments included ICI+chemo (n=179) and chemo only (n=49). PD-L1 IHC results are based on the 22c3 assay using tumor percent cutoff of 1%. Gene signatures are obtained from MSigDB. Findings: Consistent with previous literature, expression of lineage-defining transcription factors varied across four NMF subtypes. Two groups (NMF1-SCLC-A/N, 33.5% treatment-naïve vs. 33.3% experienced; NMF2-SCLC-A-like, 15.2% vs. 19.7%) were neuroendocrine (NE)-enriched; two (NMF3-SCLC-I, 28.7% vs. 21.2%; NMF4-SCLC-P-like, 22.6% vs. 13.6%) displayed immune features, with prevalence remained in post-treated samples. NMF1 showed activated transcriptional programs of ASCL1/NEUROD1, high proliferation signatures, and low immune infiltration—consistent with classical NE subtypes. NMF2 exhibited ASCL1-associated activity, while its elevated metabolic pathways differ from canonical SCLC-A. NMF3 resembled the inflamed SCLC-I subtype, with broad immune signatures, confirmed by higher prevalence of PD-L1 IHC positivity. Yet the promise of an efficacious ICI therapy could be dampened by concurrently heightened immunosuppressive TGF-β/fibrosis pathways. REST/NOTCH activation and low NE features in NMF3 indicated YAP1-driven NE dedifferentiation, accompanied by hypoxia-related programs (angiogenesis, KRAS signaling) that may inform targeted therapy. NMF4 matched canonical SCLC-P with POU2F3 activation and MYC overexpression. Though atypical, the concurrent expression between POU2F3 and NEUROD1 may point to MYC-driven subtype mixing. Preliminary analysis revealed NMF3-SCLC-I subtype with the best prognosis. Survival outcomes vary by treatment received within NMF subtypes, warranting further validations on their predictive values. Conclusion: SCLC molecular subtypes from this US RW cohort align and extend those from global clinical trial cohorts, revealing distinct clusters enriched with metabolic pathway and subtype mixing. Current SoC regimen has limited impact on the subtype prevalence. The analysis of RW outcome associated with subtypes’ unique pathway activation will inform novel treatment strategies. Citation Format: Yige Luo, Lujia Wang, Nadine Jahchan, Peter Ansell, Xi Zhao, Weilong Zhao. Prevalence and distinct biological programs of small cell lung cancer molecular subtypes revealed by real-world clinico-genomics database abstract. In: Proceedings of the American Association for Cancer Research Annual Meeting 2026; Part 1 (Regular Abstracts); 2026 Apr 17-22; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2026;86(7 Suppl):Abstract nr 5339.