MetaMind: A multi-agent transformer-driven framework for automated network meta-analyses
PLOS One · 13.02.2026 · 21 (2) · e0342895
DOI: 10.1371/journal.pone.0342895 ↗Kurz gefasst
Netzwerk-Metaanalyse in der Zeitschrift PLOS One, online erschienen am 13.02.2026. Verfasst von Achilleas Livieratos und 10 weiteren Autor:innen. Beteiligte Einrichtung: Independent Researcher, Athens, Greece. Hauptthemen laut MEDLINE: Software. Die Arbeit wurde bisher 1-mal zitiert (Crossref) und nennt 33 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: NLM NIH HHS.
Abstract (Originalsprache)
Background
Network 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.
Objective
To 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.
Methods
MetaMind 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).
Results
Promptriever 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.
Conclusions
In 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.
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Autor:innen und Institutionen
- Independent Researcher, Athens, Greece.
- Pfizer Inc., Collegeville, Pennsylvania, United States of America.
- Takeda Pharmaceuticals U.S.A., Inc., Cambridge, Massachusetts, United States of America.
- Teva Pharmaceuticals USA, Inc., West Chester, Pennsylvania, United States of America.
- Sarepta Therapeutics, Inc., Cambridge, Massachusetts, United States of America.
- 4th Department of Internal Medicine, Attikon University Hospital, Athens, Greece.
ORCID-Profile: Achilleas Livieratos, Vivek Rudrapatna
Förderung
- NLM NIH HHS
Zitieren
APA
Livieratos, A., Kudela, M., Zhao, Y., Chen, A. S., Luo, X., Lin, J., Zhang, D., Dharmarajan, S., Tsiodras, S., Rudrapatna, V., & Gamalo, M. (2026). MetaMind: A multi-agent transformer-driven framework for automated network meta-analyses. PLOS One, 21(2), e0342895. https://doi.org/10.1371/journal.pone.0342895
Vancouver
Livieratos A, Kudela M, Zhao Y, Chen AS, Luo X, Lin J, et al. MetaMind: A multi-agent transformer-driven framework for automated network meta-analyses. PLOS One. 2026;21(2):e0342895. doi:10.1371/journal.pone.0342895
BibTeX
@article{Livieratos2026,
author = {Livieratos, Achilleas and Kudela, Maria and Zhao, Yuxi and Chen, All-Shine and Luo, Xin and Lin, Junjing and Zhang, Di and Dharmarajan, Sai and Tsiodras, Sotirios and Rudrapatna, Vivek and Gamalo, Margaret},
title = {MetaMind: A multi-agent transformer-driven framework for automated network meta-analyses},
journal = {PLOS One},
year = {2026},
volume = {21},
number = {2},
pages = {e0342895},
doi = {10.1371/journal.pone.0342895}
}
EUVIMED