Transcription factor‐based subtype assignment in pulmonary large cell neuroendocrine carcinoma

AIMS: Pulmonary large cell neuroendocrine carcinoma (LCNEC) can be difficult to diagnose due to histological overlap with small cell lung carcinoma (SCLC). SCLC comprises multiple transcription factor-based subtypes. We aimed to evaluate the clinicopathological significance of transcription factor-based subtyping in pulmonary LCNEC. METHODS AND RESULTS: We identified a consecutive series of 117 patients in 2010-2024 with samples diagnosed as pulmonary LCNEC (n = 70) or high-grade neuroendocrine carcinoma with combined or intermediate morphology (n = 47). Cytomorphological score was assessed as a weighted average from two areas. Immunohistochemistry (IHC) for ASCL1, NeuroD1, POU2F3, YAP1, and HNF4A was evaluated using H-scores, with subtype assignment based on the highest H-score. Next-generation sequencing (NGS) was performed in selected cases. IHC subtyping identified 73 (62%) ASCL1-dominant, 20 (17%) YAP1-dominant, 9 (8%) NeuroD1-dominant, 7 (6%) POU2F3-dominant, 2 (2%) HNF4A-dominant, and 6 (5%) quintuple-negative samples. While YAP1 was often co-expressed with other subtypes and HNF4A was frequently co-expressed with ASCL1, POU2F3 was mutually exclusive from ASCL1/NeuroD1/HNF4A. Unlike ASCL1/NeuroD1/POU2F3, YAP1 and HNF4A H-scores each correlated with large-cell morphology-both across the entire cohort and in the lung resection subgroup. NeuroD1 dominance was more common in tumours with combined/intermediate morphology than LCNEC. Some of the tumours with intermediate morphology straddling between prototypical LCNEC and SCLC harboured POU2F3 dominance, or EGFR or other non-KRAS driver mutations. In 19 patients with multiple samples (including nine with paired pre- and post-treatment samples), all showed concordant subtypes after accounting for codominance. CONCLUSION: YAP1 and HNF4A expression correlated significantly with large-cell morphology.