Cultural bias in large language models’ ability to follow neuroradiology guidelines - Takeaways - MDSpire
Coming Soon: Introducing MDSpire News. Learn more
Conexiant’s news site is now MDSpire News. Learn more

Cultural Influences on the Compliance of Large Language Models with Neuroradiology Guidelines

  • By

  • Naël Bazerbachi

  • Raphaël Bentegeac

  • Giada Pistilli

  • Su-Hwan Kim

  • Philippe Amouyel

  • Jean-Pierre Pruvo

  • Lotfi Hacein-Bey

  • Aghiles Hamroun

  • Grégory Kuchcinski

  • Bastien Le Guellec

  • May 26, 2026

Share

  • 1

    Medical imaging complexity leads to clinician challenges in ordering appropriate studies and frequent over-utilization of imaging modalities.

  • 2

    A multi-country evaluation showed significant variations in guideline adherence, ranging from 58% to 86%, highlighting inconsistencies in neuroradiology.

  • 3

    Large language models (LLMs) have potential as decision-support tools in neuroradiology but exhibit biases, including demographic and geographic biases.

  • 4

    Geographic bias in LLMs may misalign clinical workflows with local practice standards, impacting integrated care pathways and treatment philosophies.

  • 5

    The study investigates whether LLMs disproportionately adhere to U.S. neuroradiology guidelines over international ones, using fictitious clinical vignettes.

Original Source(s)

Related Content