Amplicon-based DNA and RNA NGS co-detection enhances ROS1 fusion detection and partner diversity compared with hybrid-capture DNA-NGS in Chinese NSCLC

Background: ROS1 rearrangements are important actionable drivers in non-small cell lung cancer (NSCLC), yet traditional DNA-based assays show limited sensitivity for detecting these fusions. RNA-NGS has been demonstrated to outperform DNA-NGS in fusion detection; however, the relative performance of amplicon-based DNA+RNA NGS co-detection versus hybrid-capture DNA-NGS for identifying ROS1 fusions in large real-world NSCLC populations remains insufficiently characterized. Methods: We retrospectively analyzed ROS1 fusion detection in NSCLC samples using two next-generation sequencing strategies: a 35-gene amplicon-based DNA+RNA NGS co-detection assay (n = 11,242) and a hybrid-capture DNA-NGS assay (n = 2,504). For both approaches, ROS1 fusion-positive cases were identified, and all confirmed rearrangements were annotated for fusion partners. Detection rates and partner diversity were compared to evaluate differences in analytical performance between the two platforms. Results: Amplicon-based DNA+RNA NGS co-detection demonstrated a higher ROS1 fusion detection rate (1.82%, 205/11,242) than hybrid-capture DNA-NGS (1.56%, 39/2,504). Amplicon-based DNA+RNA NGS co-detection identified a wide variety of classical 5′-3′ ROS1 fusion partners, with CD74-ROS1 (46.34%), EZR-ROS1 (22.93%), and SDC4-ROS1 (16.10%) being the most prevalent, followed by less frequent partners such as TPM3-ROS1(10/205, 4.88%), SLC34A2-ROS1(7/205,3.41%), CCDC6-ROS1(5/205,2.44%), GOPC-ROS1(2/205,0.98%), TPR-ROS1(2/205,0.98%), and several rare events including EML4-ROS1(1/205,0.49%), LRIG3-ROS1(1/205,0.49%), PPFIBPI-ROS1(1/205,0.49%), and ZCCHC8-ROS1(1/205,0.49%). In contrast, hybrid-capture DNA-NGS detected a narrower spectrum of partners, dominated by CD74-ROS1 (58.97%), SDC4-ROS1 (20.51%), and EZR-ROS1 (12.82%), with only isolated occurrences of GOPC-ROS1, TPM3-ROS1, and a single ROS1-SLC34A2 event. All rearrangements detected by either method were classical 5′-3′ fusions. Conclusion: In this large real-world cohort, amplicon-based DNA+RNA NGS co-detection outperformed hybrid-capture DNA-NGS by yielding a higher ROS1 fusion detection rate and uncovering a more diverse spectrum of ROS1 fusion partners. These findings highlight the importance of integrating RNA-based analysis into routine NGS workflows to enhance the detection accuracy of actionable gene fusions and support more precise therapeutic decision-making for patients with NSCLC.