Can GPT-4 suggest the optimal sequence for brain magnetic resonance imaging?
Suzuki, K.; Abe, K.; Sakai, S.
Show abstract
PurposeThis study aimed to evaluate the potential of GPT-4, a large language model, in assisting radiologists to determine brain magnetic resonance imaging (MRI) protocols. MethodsWe used brain MRI protocols from a specific hospital, covering 20 diseases or examination purposes, excluding brain tumor protocols. GPT-4 was given system prompts to add one MRI sequence for the basic brain MRI protocol and disease names were input as user prompts. The models suggestions were evaluated by two radiologists with over 20 years of relevant experience. Suggestions were scored based on their alignment with the hospitals protocol as follows: 0 for inappropriate, 1 for acceptable but nonmatching, and 2 for matching the protocol. The experiment was conducted in both Japanese and English to compare GPT-4s performance in different languages. ResultsGPT-4 scored 27/40 points in English and 28/40 points in Japanese. GPT-4 gave inappropriate suggestions for Moyamoya disease and neuromyelitis optica in both languages and cerebral infarction in Japanese. For the other protocols, the suggested sequences were either appropriate or better. The suggestions in English differed from those in Japanese for seven protocols. ConclusionGPT-4 can suggest appropriate MRI sequences for each disease in addition to the standard brain MRI protocol. GPT-4s output is language-dependent and suggests brain MRI protocols tailored to specific regions and domains.
Matching journals
The top 9 journals account for 50% of the predicted probability mass.
Similar papers in this journal
- PRCnet: An Efficient Model for Automatic Detection of Brain Tumor in MRI Images 94%
- Cerebro-spinal Flow Pattern in the Cervical Subarachnoid Space of Healthy Volunteers: Influence of the Spinal Cord morphology 93%
- tbiExtractor: A framework for Extracting Traumatic Brain Injury Common Data Elements from Radiology Reports 93%
Similar papers in this journal
- Evaluating Large Language Model-Generated Brain MRI Protocols: Performance of GPT4o, o3-mini, DeepSeek-R1 and Qwen2.5-72B 96%
- Assessing GPT-4 Multimodal Performance in Radiological Image Analysis 94%
- Impact of Non-Contrast Enhanced Imaging Input Sequences on the Generation of Virtual Contrast-Enhanced Breast MRI Scans using Neural Networks 90%
Similar papers in this journal
- “This is a quiz” Premise Input: A Key to Unlocking Higher Diagnostic Accuracy in Large Language Models 95%
- Effects of contrast-medium and vertebral measurement level on computed tomography-based body composition parameters of skeletal muscle and adipose tissue 92%
- Benchmarking Deep Learning-based Image Retrieval of Oral Tumor Histology 90%
Similar papers in this journal
- Automatic quantification of brain lesion volume from post-trauma MR Images 93%
- LesionQuant for assessment of MRI in multiple sclerosis - a promising supplement to the visual scan inspection 91%
- Effects of variability in manually contoured spinal cord masks on fMRI co-registration and interpretation 90%
Similar papers in this journal
- Simulated Diagnostic Performance of Ultra-Low-Field MRI: Harnessing Open-Access Datasets to Evaluate Novel Devices 92%
- Increased Brain Volumetric Measurement Precision from Multi-Site 3D T1-weighted 3T Magnetic Resonance Imaging by Correcting Geometric Distortions 91%
- Anisotropy Measure from Three Diffusion-Encoding Gradient Directions 91%
"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.