贵州省肺癌的基因组分析以及DNA-NGS与DNA+RNA-NGS在融合检测中的性能比较

Background: Targeted therapy plays a crucial role in non-small cell lung cancer (NSCLC), making accurate genomic profiling essential for treatment selection. RNA-based fusion detection has been increasingly recommended by major guidelines due to its superior analytical performance, yet real-world evidence remains limited. This study characterized the genomic landscape of NSCLC in Guizhou Province and compared DNA-based next-generation sequencing (DNA-NGS) with combined DNA+RNA sequencing (D+R-NGS), particularly for fusion detection. Methods: From 2017 to 2024, 880 tumor samples from 571 NSCLC patients underwent NGS testing, including 750 samples analyzed by DNA-NGS and 130 by D+R-NGS. Somatic SNVs/indels, copy-number variations (CNVs), and gene fusions were comprehensively assessed and compared across platforms. Results: Across all samples, 2,051 SNVs/indels, 198 CNVs, and 73 gene fusions were identified. EGFR (53%) and TP53 (45%) were the most common mutations, consistent with previously reported Chinese NSCLC data. Except for one large-cell neuroendocrine carcinoma, all fusion-positive cases were adenocarcinomas, with female patients representing a higher proportion (60.2%).Fusion detection rates differed substantially between platforms. The D+R-NGS group showed a significantly higher fusion detection rate than DNA-NGS (15.4% vs. 7.1%, p = 0.003). And the D+R-NGS also had a numerically higher detection rate (13.5% VS 8.1%, p=0.095), even when significant differences were lost after filtering multiple samples from the same patient. Fusion subtype analysis indicated similar overall patterns, with ALK rearrangements most common in both groups (45% in D+R-NGS; 64% in DNA-NGS). Notably, the detection of MET exon 14 skipping was substantially higher in the D+R-NGS cohort (20%) compared with DNA-NGS (2.4%).This difference may be related to the fact that the MET exon 14 skipping on DNA is usually caused by the variant drive of intron 13 or 14, regions that are challenging to assess using DNA-based assays, whereas RNA sequencing provides inherent structural advantages for detecting such events.Importantly, the improvement in fusion detection with D+R-NGS did not reduce sensitivity for other mutation types. Instead, actionable alterations such as EGFR (64.7% vs. 46.1%) and KRAS (12.9% vs. 6.53%) were identified at higher rates in the D+R-NGS group, suggesting broader analytical benefits. Conclusions: The genomic profile of NSCLC in Guizhou Province aligns with broader Chinese cohorts. Combined DNA+RNA sequencing significantly enhances fusion detection while maintaining strong performance for SNVs/indels and CNVs. These findings support integrating RNA into routine NGS workflows to improve diagnostic accuracy and guide targeted therapy more effectively in NSCLC.