Role of large language models for etiological classification of brain stroke based on MRI brain reports: a feasibility study.
Gespeichert in:
| Titel: | Role of large language models for etiological classification of brain stroke based on MRI brain reports: a feasibility study. |
|---|---|
| Autoren: | Escartín J; Neurorradiología Diagnostica e Intervencionista, HT Médica Ávila, Spain. Electronic address: j.escartin@htmedica.com., López-Úbeda P; NLP department, HT medica, Jaén, Spain. Electronic address: p.lopez@htmedica.com., Martín-Noguerol T; MRI unit, Radiology department, HT medica, Jaén, Spain. Electronic address: t.martin.f@htime.org., Luna A; MRI unit, Radiology department, HT medica, Jaén, Spain. Electronic address: aluna70@htmedica.com. |
| Quelle: | Magnetic resonance imaging [Magn Reson Imaging] 2025 Dec; Vol. 124, pp. 110538. Date of Electronic Publication: 2025 Oct 06. |
| Publikationsart: | Journal Article |
| Sprache: | English |
| Info zur Zeitschrift: | Publisher: Elsevier Country of Publication: Netherlands NLM ID: 8214883 Publication Model: Print-Electronic Cited Medium: Internet ISSN: 1873-5894 (Electronic) Linking ISSN: 0730725X NLM ISO Abbreviation: Magn Reson Imaging Subsets: MEDLINE |
| Imprint Name(s): | Publication: <2008->: Amsterdam : Elsevier Original Publication: New York : Pergamon, c1982- |
| MeSH-Schlagworte: | Magnetic Resonance Imaging*/methods , Natural Language Processing* , Brain*/diagnostic imaging , Stroke*/diagnostic imaging , Stroke*/classification , Stroke*/etiology, Humans ; Feasibility Studies ; Algorithms ; Image Interpretation, Computer-Assisted/methods ; Large Language Models |
| Abstract: | Purpose: Ischemic stroke, a leading cause of global disability and mortality, demands precise etiological classification for effective management. The variability in the use of existing stroke classification systems, along with the challenges in manual etiological labeling from brain MRI radiological reports, calls for an innovative approach. This study aims to develop and evaluate a Natural Language Processing (NLP) algorithm using transformer-based models for the extraction and classification of ischemic stroke types from MRI reports, enhancing diagnostic efficiency and stroke management. Methods: We built a dataset comprising 635 brain MRI reports, annotated for four distinct ischemic stroke types. All were clinically consistent with focal neurologic impairment due to stroke. The study involved evaluating two pre-trained models BERT (Bert clinical and Beto) and two models RoBERTa (Roberta clinical trials and Roberta biomedical), focusing on their ability to accurately classify stroke subtypes. Results: The Roberta biomedical model emerged as the most effective, demonstrating superior performance with an accuracy of 76.7 % with statistically significant results. This model also achieved the highest precision, recall, and F1 scores across all stroke types, indicating its robustness in stroke subtype classification. Conclusion: The study highlights the potential of NLP algorithms in automating stroke classification from MRI reports, which could significantly aid in diagnostic processes and streamline stroke management strategies. (Copyright © 2025 Elsevier Inc. All rights reserved.) |
| Contributed Indexing: | Keywords: Artificial intelligence; Large language model; MRI; Natural language processing; Stroke |
| Entry Date(s): | Date Created: 20251007 Date Completed: 20251025 Latest Revision: 20251025 |
| Update Code: | 20251026 |
| DOI: | 10.1016/j.mri.2025.110538 |
| PMID: | 41057116 |
| Datenbank: | MEDLINE |
Full Text Finder 
Nájsť tento článok vo Web of Science | Abstract: | Purpose: Ischemic stroke, a leading cause of global disability and mortality, demands precise etiological classification for effective management. The variability in the use of existing stroke classification systems, along with the challenges in manual etiological labeling from brain MRI radiological reports, calls for an innovative approach. This study aims to develop and evaluate a Natural Language Processing (NLP) algorithm using transformer-based models for the extraction and classification of ischemic stroke types from MRI reports, enhancing diagnostic efficiency and stroke management.<br />Methods: We built a dataset comprising 635 brain MRI reports, annotated for four distinct ischemic stroke types. All were clinically consistent with focal neurologic impairment due to stroke. The study involved evaluating two pre-trained models BERT (Bert clinical and Beto) and two models RoBERTa (Roberta clinical trials and Roberta biomedical), focusing on their ability to accurately classify stroke subtypes.<br />Results: The Roberta biomedical model emerged as the most effective, demonstrating superior performance with an accuracy of 76.7 % with statistically significant results. This model also achieved the highest precision, recall, and F1 scores across all stroke types, indicating its robustness in stroke subtype classification.<br />Conclusion: The study highlights the potential of NLP algorithms in automating stroke classification from MRI reports, which could significantly aid in diagnostic processes and streamline stroke management strategies.<br /> (Copyright © 2025 Elsevier Inc. All rights reserved.) |
|---|---|
| ISSN: | 1873-5894 |
| DOI: | 10.1016/j.mri.2025.110538 |