What question did this study set out to answer?

This study evaluates how well large language models can replicate neurologists’ stroke scoring metrics like NIHSS and ASPECTS.

May 8, 2026European Stroke Journal

Abstract Number: Esoc2026a1487 Automated Stroke Scoring: Agreement Between Large Language Models and Neurologists in Nihss and Aspects

View Full Paper

Ask AI

Bookmark

Authors

RLR Domingos Da Costa LopesUniversidade Federal do Rio de Janeiro MPMaria Carlos PereiraAdministração Regional de Saúde de Lisboa e Vale do Tejo CGCarolina GonçalvesAdministração Regional de Saúde de Lisboa e Vale do Tejo

Discussion

Member takes

Implication

Randomized trial evaluates agreement of LLMs with neurologists on NIHSS and ASPECTS scores in stroke assessment, suggesting efficient data extraction methods.

Key Points

This study evaluates how well large language models can replicate neurologists’ stroke scoring metrics like NIHSS and ASPECTS.
Processed 487 neurological examination records and CT scans using five LLMs.
Assessed agreement with neurologist ratings using various statistical methods including ICC and kappa.
Measured time reduction for assessment tasks with LLM support.
Mean NIHSS score was 10.9; LLMs showed good agreement with neurologists, particularly GeminiTM with an ICC of 0.964.
Kappa for NIHSS categorization was 0.827, indicating almost perfect agreement.
LLM-based scoring reduced assessment time by up to 16.4 seconds per patient.

Cite This Study

Lopes et al. (2026) studied this question.

synapsesocial.com/papers/69fd7e00bfa21ec5bbf06320 https://doi.org/10.1093/esj/aakag023.719

View Full Paper

Ask AI

Bookmark

Implication

Randomized trial evaluates agreement of LLMs with neurologists on NIHSS and ASPECTS scores in stroke assessment, suggesting efficient data extraction methods.

Abstract Number: Esoc2026a1487 Automated Stroke Scoring: Agreement Between Large Language Models and Neurologists in Nihss and Aspects

Discussion

Member takes

Implication

Key Points

Cite This Study

Abstract Number: Esoc2026a1487 Automated Stroke Scoring: Agreement Between Large Language Models and Neurologists in Nihss and Aspects

Discussion

Member takes

Implication

Key Points

Cite This Study

Also Consider

Also Consider