Real-World Insights on Treatment Patterns and Outcomes in ROS1 NSCLC: An Australian Multicenter Study (AURORA).

Abstract

Introduction: ROS1-rearranged non-small cell lung cancer (ROS1 NSCLC) is a rare, yet clinically significant subset driven by oncogenic ROS1 fusions. While targeted therapies have transformed outcomes, real-world data on treatment patterns and survival remain limited. This consumer-informed analysis leverages the AUstralasian thoRacic cancers lOngitudinal cohoRt study and biobAnk (AURORA) to present one of the largest real-world ROS1 NSCLC cohorts, providing a comprehensive evaluation of patient characteristics, treatment ap proaches, and survival across multiple Australian hospitals. Method(s): Patients with ROS1 NSCLC diagnosed by fluorescence in situ hybridization (FISH) or next-generation sequencing (NGS) enrolled in AURORA (ACTRN12625000038493) were included. Key outcomes included demographics, diagnostics, treatment patterns, and overall survival (OS). OS was estimated using Kaplan-Meier analysis and measured from ROS1 NSCLC diagnosis. The cohort continues to enrol as part of ongoing AURORA enrolment, with a final data cut on this analysis in June 2025 to enable multivariate modelling on clinical and molecular markers of treatment performance. This study is supported by the Kha Kiang Khoo Memorial Award, Lung Foundation Australia. Result(s): Among 104 patients enrolled across 15 hospitals with ROS1 NSCLC diagnosed 2009-2024, median age was 56 years (range: 24 79), and 66% (69/104) were female. At diagnosis, 24% (24/104) had early-stage disease treated with curative intent, while 76% (80/104) presented with advanced/metastatic disease, including 14% (11/80) with brain metastases. Among early-stage patients, 54% (13/24) subsequently relapsed and received treatment for advanced disease. In the advanced treatment setting (93/104, 89%), the median number of treatment lines was 2 (range: 1-8). Most received one (32/93, 34%) or two (29/93, 31%) therapy lines, while 34% (32/93) underwent three-lines or more, including one patient with eight lines. ROS1- directed therapy was given first-line in 72% (67/93) and at any line in 96% (89/93), accounting for total 239 lines of ROS-1 therapies. Entrectinib (31/93, 33%) and crizotinib (23/93, 25%) were the most common first-line targeted agents. Across all lines, entrectinib (35/93, 38%) and crizotinib (33/93, 35%) were frequently used, followed by lorlatinib (26/93, 28%), NVL-520 (24/93, 26%), repotrectinib (17/93, 18%). Clinical trial enrolment was 17% (16/93) at first line and 59% (55/93) overall. Median OS was 92 months (95%CI: 54-NR) in earlystage patients treated with curative intent (N=24). For those diagnosed with de-novo advanced/metastatic disease (N=80), median OS was 56 months (95%CI: 49-82). In the subgroup receiving first-line ROS1-directed therapy in the advanced disease setting (N=67), me dian OS was 56 months (95%CI: 49-92). Conclusion(s): In this large real-world ROS1 NSCLC cohort, 72% received ROS1-directed therapy in the first line, and 96% at any line, with 59% enrolled in trials. Firstline ROS1 therapy demonstrated a median survival of 56 months, comparable to PROFILE 1001 (51 months) and exceeding other realworld cohorts (24 months), likely due to reflex molecular testing, targeted therapy access, and trial prioritisation. This real-world data provides valuable insights into the longitudinal patient journey, outcomes with serial therapies, and the potential impact of treatment sequencing.