Conectar
by Achilleas Livieratos, Maria Kudela, Yuxi Zhao, All-shine Chen, Xin Luo, Junjing Lin, Di Zhang, Sai Dharmarajan, Sotirios Tsiodras, Vivek Rudrapatna, Margaret Gamalo
BackgroundNetwork meta-analysis (NMA) can compare several interventions at once by combining head-to-head and indirect trial evidence. However, identifying, extracting, and modelling these often takes months, delaying updates in many therapeutic areas.
ObjectiveTo develop and validate MetaMind, an end-to-end, transformer-driven framework that automates NMA processes—including study retrieval, structured data extraction, and meta-analysis execution—while minimizing human input.
MethodsMetaMind integrates Promptriever, a fine-tuned retrieval model, to semantically retrieve high-impact clinical trials from PubMed; a multi-agent LLM architecture--Mixture of Agents (MoA)-- pipeline to extract PICO-structured (Population, Intervention, Comparison, Outcome) endpoints; and GPT-4o–generated Python and R scripts to perform Bayesian random-effects NMA and other NMA designs within a unified workflow. Validation was conducted by comparing MetaMind’s outputs against manually performed NMAs in ulcerative colitis (UC) and Crohn’s disease (CD).
ResultsPromptriever outperformed baseline SentenceTransformer with higher similarity scores (0.7403 vs. 0.7049 for UC; 0.7142 vs. 0.7049 for CD) and narrower relevance ranges. Promptriever performance achieved 82.1% recall, 91.1% precision and an F1 score of 86.4% when compared to a previously published NMA. MetaMind achieved 100% accuracy on a limited set of remission endpoints regarding PICO (Population, Intervention, Comparator, Outcome) element extraction and produced comparative effect estimates and credible intervals closely matching manual analyses.
ConclusionsIn our validation studies, MetaMind reduced the end-to-end NMA process to less than a week, compared with the several months typically needed for manual workflows, while preserving statistical rigor. This suggests its potential for future scaling of evidence synthesis to additional therapeutic areas.
Etrasimod is an oral, once-daily, selective sphingosine 1-phosphate1,4,5 receptor modulator for the treatment of moderately to severely active ulcerative colitis (UC). While etrasimod demonstrated efficacy in randomised controlled trials, understanding its effectiveness in an observational setting is crucial.
EFFECT-UC is a prospective, multinational, non-interventional study to evaluate the real-world effectiveness of etrasimod in adults with moderately to severely active UC. The study consists of a 52-week treatment period and a 28-day safety follow-up period and aims to enrol ~300 patients per cohort. Eligible patients (18–64 years) are advanced therapy naïve or experienced and are initiating etrasimod in a real-world clinical setting. Treatment will be guided independently by the clinician’s judgement. Patient-reported outcomes will be collected electronically throughout the study and daily for the first 2 weeks. Exploratory data, including faecal calprotectin, endoscopy and intestinal ultrasound, will be collected at predefined visits or during standard care. Primary endpoints are symptomatic remission at week 12 and week 52. Secondary endpoints include patient-reported outcome 2 (combined rectal bleeding and stool frequency subscores) response at week 12 and week 52 and corticosteroid-free symptomatic remission at week 52.
Ethics approval was obtained for all sites. Recruitment is underway for cohort 1, comprising patients from the UK, Germany and Canada. Interim results for this cohort are expected in 2026 and final results in 2028; these will be submitted for publication in peer-reviewed journals and presented at appropriate congresses.
FreshRSS 