Molecular characteristics of primary pulmonary lymphoepithelioma-like carcinoma based on integrated genomic analyses

Primary pulmonary lymphoepithelioma-like carcinoma (pLELC) is a rare non-small cell lung cancer (NSCLC) subtype. Clinical features have been described in our previous report, but molecular characteristics remain unclear. Herein, pLELC genomic features were explored. Among 41,574 lung cancers, 128 pLELCs and 162 non-pLELC NSCLCs were enrolled. Programmed cell death ligand 1 (PD-L1) and protein 53 (p53) expression was detected in 47 surgically resected pLELC samples by immunohistochemical assays. Multiomics genomic analyses, including whole-genome sequencing (WGS), RNA whole-transcriptome sequencing (RNA-seq), and Epstein-Barr virus (EBV) integration analyses, were performed on eight frozen pLELC tissues and compared with 50 lung adenocarcinomas (LUADs) and 50 lung squamous cell carcinomas (LUSCs) from The Cancer Genome Atlas (TCGA) and another 26 EBV-positive nasopharynx cancers (EBV(+)-NPCs). Progression-free survival (PFS) and overall survival (OS) of pLELC patients were better than those of non-pLELC patients. High PD-L1 or p53 expression was associated with extended disease-free survival (DFS). pLELC had 14 frequently mutated genes (FMGs). Somatically mutated genes and enrichment of genetic lesions were found, which differed from observations in LUAD, LUSC, and EBV(+)-nasopharyngeal carcinoma (NPC). Three tumor-associated genes, zinc finger and BTB domain-containing 16 (ZBTB16), peroxisome proliferator activated receptor gamma (PPARG), and transforming growth factor beta receptor 2 (TGFBR2), were downregulated with copy number variation (CNV) loss. EBV was prone to integrating into intergenic and intronic regions with two upregulated miR-BamH1-A rightward transcripts (BARTs), BART5-3P and BART20-3P. Our findings reveal that pLELC has a distinct genomic signature. Three tumor-associated genes with CNV loss and two miR-BARTs might be involved in pLELC tumorigenesis.