1540 Background: Biomarker actionability and therapy matching are two central concepts in precision oncology. However, the best strategy to align therapy with genomic alterations remains unclear. Here we report the updated results from a cohort study conducted in an Australian precision oncology program (MoST, ACTRN12616000908437), examining a therapy matching strategy based on comprehensive genomic profiling (CGP) results. Methods: All patients (pts) with rare or advanced solid tumours undergoing CGP from 2016 to 2021 after exhausting standard treatments were included. The primary outcome was overall survival (OS) from the date of CGP result, estimated using the Kaplan-Meier method. Pts were grouped according to tiers of actionability determined by matching against the TOPOGRAPH knowledge base ( https://topograph.info ) and stratified into clinically active (Tier 1-3, therapies with evidence from prospective trials), investigational (Tier 3B and 4, therapies with evidence from another cancer type, preclinical, or retrospective studies), or unmatched tier groups. To assess between-group differences in OS, hazard ratios (HR) were estimated in a time-varying Cox regression model adjusting for time to initiation of subsequent therapy. Results: This updated analysis included 3,383 pts (79% rare and less common). The median follow-up was 22.6 months (mo). For 2,065 pts who did not receive treatment after CGP, the median OS (mOS) was 8.2 mo (95% CI 7.4 to 9.0). For 1,318 pts who received ≥1 line of therapy after CGP, the mOS was 14.1 mo (13.4 to 15.2). 1,270 pts (38%) carried a genomic alteration linked to clinically active therapies (Tier 1-3). Of these, 116 (3.4%) received matched treatment after CGP, experiencing longer survival compared with 410 (32.2%) that received only unmatched therapy (mOS 21.2 v 12.8 mo, HR 0.58, 0.45 to 0.76, P<0.001). Whilst receiving a matched therapy in the investigational tier group (n=133, T3B/4) did not show significant differences in outcomes (mOS 14.5 v unmatched 12.8 mo, n=536, HR 0.88, 0.72 to 1.08, p=0.24), exploratory analyses identified differences in survival for a subset of 41 pts with Switch/Sucrose Non-Fermentable complex mutations receiving matched Tier 4 therapies (mOS 30.1 v 9.6 mo, HR 0.29, 0.17 to 0.51, P<0.0001). Pts receiving repurposed drugs that matched solely on biomarkers in non-cognate cancer types (Tier 3B) showed no survival difference over unmatched therapy (n=35 v 469, mOS 13.6 v 12.5 mo, HR 0.96, 0.65 to 1.41, p=0.83). Conclusions: This study provides insight into how biomarker-linked therapies can be rationally prioritised in rare and advanced cancer populations. Addressing the barriers of access to Tier 1-3 therapies may broaden the utility of genomic biomarker testing. Off-label drug repurposing without direct supporting evidence should only be undertaken in a clinical trial setting.