Performing Best When Needed Least: Reader Experience Shapes Accuracy Gains in Large Language Model–assisted Brain MRI Differential Diagnosis
Authors:
- Severin SchrammSSSeverin Schramm
- Bastien Le GuellecBGBastien Le Guellec
- Mortimer SvecMSMortimer Svec
- Viktor Maria EisenkolbVEViktor Maria Eisenkolb
- Evamaria O. RiedelEREvamaria O. Riedel
- Mirjam BeyrleMBMirjam Beyrle
- Constanze RamschützCRConstanze Ramschütz
- Karolin J. PaprottkaKPKarolin J. Paprottka
- Martin RenzMRMartin Renz
- 1 May 2026
- journal article
- Published by Radiological Society of North America (RSNA) in Radiology
- Vol. 319 (2) , e253477
Abstract
With increasing reader experience, large language model accuracy improved, whereas diagnostic gains from model assistance paradoxically diminished, highlighting the divergence between model performance and clinically relevant benefits for readers.Keywords
This publication has 33 references indexed in Scilit:
- Endoscopist deskilling risk after exposure to artificial intelligence in colonoscopy: a multicentre, observational studyThe Lancet Gastroenterology & Hepatology, 2025
- Automated MRI protocoling in neuroradiology in the era of large language modelsLa radiologia medica, 2025
- Boosting LLM-assisted diagnosis: 10-minute LLM tutorial elevates radiology residents’ performance in brain MRI interpretationNeuroradiology, 2025
- Comparison between multimodal foundation models and radiologists for the diagnosis of challenging neuroradiology cases with text and imagesDiagnostic and Interventional Imaging, 2025
- Human-AI collaboration in large language model-assisted brain MRI differential diagnosis: a usability studyEuropean Radiology, 2025
- Automation bias in AI-assisted detection of cerebral aneurysms on time-of-flight MR angiographyLa radiologia medica, 2025
- Bias Sensitivity in Diagnostic Decision-Making: Comparing ChatGPT with ResidentsJournal of General Internal Medicine, 2024
- ChatGPT’s diagnostic performance based on textual vs. visual information compared to radiologists’ diagnostic performance in musculoskeletal radiologyEuropean Radiology, 2024
- The virtual reference radiologist: comprehensive AI assistance for clinical image reading and interpretationEuropean Radiology, 2024
- What influences algorithmic decision-making? A systematic literature review on algorithm aversionTechnological Forecasting and Social Change, 2022