Aller au contenu principal
Accès ouvert déclaré 2026 preprint

Artificial Intelligence in Thoracic Surgery: A Meta-research of the Evidence on Trends, Use, Benefits, Limitations, and Barriers to Clinical Implementation

0Citations signalées — pas une note de qualité
0Institutions déclarées
0Pays d’affiliation déclarés

Résumé fourni par la source

Artificial intelligence (AI) has been progressively incorporated across the thoracic surgery care pathway — from image-based diagnosis and surgical planning to intraoperative navigation, robotic surgery, and prediction of postoperative outcomes. Although recent studies report high discriminative performance, the methodological robustness, degree of external validation, reproducibility, and real clinical impact of these models remain uncertain. This project is a meta-research (a meta-epidemiological study with a bibliometric component) that critically appraises the scientific literature on AI in thoracic surgery. Its objectives are to map temporal publication trends; identify and classify the main clinical applications; assess methodological quality and risk of bias; determine the proportion of models with adequate external validation; analyze the reporting of calibration and clinical utility; evaluate transparency and reproducibility practices, including code and data availability; identify ethical, regulatory, and structural barriers to clinical implementation; and classify studies according to their translational readiness. A systematic search will be conducted in PubMed/MEDLINE, Embase, Scopus, Web of Science, Cochrane Library, and IEEE Xplore, covering January 2010 to March 2026, with study identification and selection based on the PRISMA 2020 guidelines. Original studies applying AI clinically in thoracic surgery will be included; narrative reviews, editorials, letters, and purely technical studies without clinical application will be excluded. Risk of bias and reporting quality will be appraised using PROBAST+AI, QUADAS-2 (with the QUADAS-AI extension), and TRIPOD+AI, and translational maturity will be classified using an adapted Clinical Readiness Levels framework. The study is based exclusively on secondary, already-published data.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.

Contrôle bibliographique ouvert

La source scientifique ouverte est momentanément indisponible.

Sujets associés

Artificial Intelligence in Healthcare and EducationLung Cancer Diagnosis and TreatmentRadiomics and Machine Learning in Medical Imaging

BNTIC News n’est pas le producteur de ces données. Recherche à la demande dans Crossref, OpenAIRE, DOAJ, Europe PMC, HAL, DataCite, ROR et la Banque mondiale, sans clé ; OpenAlex reste optionnel. Aucun service payant requis, aucune donnée externe enregistrée en base. Sources et limites.