Presentation (CTI)

OBJECTIVES: Joint Clinical Assessments (JCAs) require rapid generation of comparative evidence. Multiple anticipated PICOs (Population, Intervention, Comparator, Outcome) drive extensive literature searches, increasing the volume of studies in systematic literature reviews (SLRs) and the burden of screening and data extraction. Study heterogeneity further complicates indirect treatment comparisons (ITCs). To address these challenges, we propose a structured, scalable framework to efficiently identify, prioritize, and select trials for ITC, illustrated in treatment‑naïve ALK‑positive advanced or metastatic non-small cell lung cancer (a/m NSCLC).
METHODS: Systematic searches were conducted in EMBASE, MEDLINE, and CENTRAL. Following deduplication and title/abstract screening, a tailored prioritization strategy was applied, deprioritizing non-systemic interventions, phase I or dose-finding studies, small trials (≤20 patients), and non-English publications. Topline data extraction identified studies aligned with predicted JCA PICOs, capturing key study characteristics (randomized/single-arm), interventions, population characteristics (ALK status, co-mutations, histology, metastases, central nervous system involvement), and availability of overall and progression-free survival data.
RESULTS: The SLR identified 36,748 records, of which 2,985 advanced to full‑text review and 641 were initially eligible for data extraction. Following prioritization, full‑text records were reduced to 1,946 and extracted publications to 180. Based on topline data extraction for the 180 publications, studies were categorized into three evidence tiers aligned with the anticipated JCA PICOs. Tier 1 comprised 18 trials representing the highest‑priority evidence, evaluating relevant treatments in the target population according to predicted JCA PICOs. Tier 2 included 10 trials assessing relevant interventions in a broader a/m NSCLC setting consistent with the anticipated PICOs. Tier 3 encompassed trials evaluating other treatments within the a/m NSCLC population.
CONCLUSIONS: Application of the proposed framework in a case study of ALK‑positive a/m NSCLC streamlined evidence identification and selection of relevant studies for evidence synthesis. It offers a scalable and pragmatic approach to generating high‑quality comparative evidence within the constraints of the JCA process.