通过在基于DNA的检测中整合复杂基因组重排来增强EML4-ALK融合的检出

Introduction: ALK fusions, particularly EML4-ALK , are well-established therapeutic biomarkers in non-small cell lung cancer (NSCLC) and other malignancies. In DNA-based fusion detection, genomic rearrangements at ALK loci can be complex, often involving non-canonical or fragmented configurations. These include events where EML4 and ALK breakpoints are both detected, but occur at separated genomic positions, sometimes bridged by intergenic sequences or intervening genes. Such configurations may still yield functional EML4 - ALK transcripts yet fall outside canonical DNA-based fusion-calling rules. In this study, we explore the potential impact of broadening reporting criteria to include these distanced non-canonical EML4 - ALK rearrangements and assess their clinical relevance. Methods: We analyzed paired DNA and RNA sequencing data from 7,810 tissue samples (Guardant360 Tissue; Guardant Health, Inc. Palo Alto, CA) and DNA sequencing data from 151,517 liquid samples (Guardant 360 Liquid; Guardant Health, Inc. Palo Alto, CA) from advanced stage patients across cancer types to determine the prevalence of EML4 - ALK fusions. DNA-based fusions were classified as canonical ( EML4 - ALK with direct adjacency) or non-canonical ( EML4 - ALK with separated intra-chromosomal breakpoints). For tissue samples with paired RNA data, we compared results across the two assays to evaluate concordance and to determine whether RNA evidence supported transcriptionally active EML4 - ALK fusions from these complex, non-canonical DNA events. Results: Among pan-cancer tissue samples, 45 (0.57%) harbored canonical EML4 - ALK fusions, while 5 (0.06%) contained non-canonical EML4- ALK rearrangements with preserved ALK kinase domains. All five were confirmed by RNA and found exclusively in lung cancers. In liquid, 612 (0.40%) cases with a canonical event and 55 (0.04%) with non-canonical events were observed; 91% retained a complete kinase domain, and 73% were found in lung cancer patients. The genomic distance between EML4 and ALK breakpoints typically fell within a few kilobases to several megabases but within the same chromosomal context, suggesting structural complexity. Including non-canonical events improved fusion detection sensitivity by ~13% in tissue and ~8% in liquid while maintaining high cancer-type specificity. Conclusions: A subset of EML4 - ALK fusions exhibit non-canonical, intra-chromosomal breakpoints detectable at the DNA level yet yield functional, targetable, EML4 - ALK fusion products at the RNA level. These rearrangements preserve the ALK kinase domain and are highly enriched in NSCLC, supporting their biological and therapeutic relevance. Expanding DNA fusion criteria to capture these complex, functional events could enable the identification of additional ALK -positive NSCLC patients eligible for targeted therapies by NGS, improving precision oncology outcomes.