A real-world study of ALK fusion detection, treatment patterns, and survival outcomes of patients with advanced non-small-cell lung cancer (aNSCLC) in Brazil

e18655 Background: Clinical trials have firmly established that ALK tyrosine kinase inhibitors (TKIs) are safe and effective in aNSCLC, with substantial anti-tumor activity of multiple agents across different treatment lines. Real-world ALK fusion diagnosis and targeted therapy effectiveness have not been described in Brazilian private healthcare setting, with ALK TKIs approved since 2016. Methods: This study retrospectively explored the patterns of use and outcome after initial exposure to ALK TKIs in patients with ALK positive aNSCLC in a real-world clinical cohort from Oncoclínicas Group, Brazil's largest network of community oncology practices, between 2016 and 2022. For data collection we used a platform where structured EHR variables are integrated with unstructured elements from physician notes using technology-based abstraction and expert curation. We also describe molecular epidemiology and ALK fusion variant prevalence for a subset of patients tested with an in-house RNA sequencing assay (GS180, ArcherDx) at Oncoclínicas Precision Medicine Lab from 2020 to 2022. Results: Out of 100 ALK-positive aNSCLC patients treated with ALK TKIs, median age was 60 years and 62% were female. Overall, median number of treatment lines was 2, with 29% receiving 3 or more regimens. In first line, 64% of the patients received an ALK TKI (58% crizotinib, 33% alectinib, and 9% brigatinib, ceritinib or lorlatinib). The remaining 36% were treated with chemo +/- immunotherapy first, and 28% switched to an ALK TKI right after driver diagnosis. When comparing first line treatment practice in 2020-2022 versus 2016-2019 we detected major reduction in crizotinib as first ALK TKI (83% versus 35%). Overall, 14% received 2 or more sequential ALK TKIs. With a median follow-up of 12 months, median Overall Survival (mOS) from start of first ALK TKI was 28 months (CI95% 18-NR). In first line, exposure to alectinib was associated with numerically higher mOS (36 months) when compared to crizotinib (18 months, HR = 0.58, CI95% 0.2-1.6). From 592 cases of non-squamous NSCLC cases with in-house RNA sequencing, ALK fusions were detected in 5.2%. EML4-ALK fusion was found in 84%, with most common variants being E5:A20 (non-canonical) in 41% and E13:E20 (v1) in 38%. We also found novel fusions partners HIP1, KLC1, LTK and STRN in 16%. Conclusions: The largest clinico-genomics cohort in Latin America reveals that the impact of ALK TKIs in ALK fusion aNSCLC may not be as profound in the real world as in clinical trials. Fast access to molecular diagnosis and exposure to ALK TKI in first line are being incentivized. EML4-ALK variant prevalence in Brazilian patients with aNSCLC is different from published cohorts, and the effect that non-canonical variants may have on ALK TKI activity must be studied further.