Performance of GPT-5 in Brain Tumor MRI Reasoning

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Safari, Mojtaba, Wang, Shansong, Hu, Mingzhe, Eidex, Zach, Li, Qiang, Yang, Xiaofeng
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866915446825943040
author Safari, Mojtaba
Wang, Shansong
Hu, Mingzhe
Eidex, Zach
Li, Qiang
Yang, Xiaofeng
author_facet Safari, Mojtaba
Wang, Shansong
Hu, Mingzhe
Eidex, Zach
Li, Qiang
Yang, Xiaofeng
contents Accurate differentiation of brain tumor types on magnetic resonance imaging (MRI) is critical for guiding treatment planning in neuro-oncology. Recent advances in large language models (LLMs) have enabled visual question answering (VQA) approaches that integrate image interpretation with natural language reasoning. In this study, we evaluated GPT-4o, GPT-5-nano, GPT-5-mini, and GPT-5 on a curated brain tumor VQA benchmark derived from 3 Brain Tumor Segmentation (BraTS) datasets - glioblastoma (GLI), meningioma (MEN), and brain metastases (MET). Each case included multi-sequence MRI triplanar mosaics and structured clinical features transformed into standardized VQA items. Models were assessed in a zero-shot chain-of-thought setting for accuracy on both visual and reasoning tasks. Results showed that GPT-5-mini achieved the highest macro-average accuracy (44.19%), followed by GPT-5 (43.71%), GPT-4o (41.49%), and GPT-5-nano (35.85%). Performance varied by tumor subtype, with no single model dominating across all cohorts. These findings suggest that GPT-5 family models can achieve moderate accuracy in structured neuro-oncological VQA tasks, but not at a level acceptable for clinical use.
format Preprint
id arxiv_https___arxiv_org_abs_2508_10865
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Performance of GPT-5 in Brain Tumor MRI Reasoning
Safari, Mojtaba
Wang, Shansong
Hu, Mingzhe
Eidex, Zach
Li, Qiang
Yang, Xiaofeng
Computer Vision and Pattern Recognition
Artificial Intelligence
Accurate differentiation of brain tumor types on magnetic resonance imaging (MRI) is critical for guiding treatment planning in neuro-oncology. Recent advances in large language models (LLMs) have enabled visual question answering (VQA) approaches that integrate image interpretation with natural language reasoning. In this study, we evaluated GPT-4o, GPT-5-nano, GPT-5-mini, and GPT-5 on a curated brain tumor VQA benchmark derived from 3 Brain Tumor Segmentation (BraTS) datasets - glioblastoma (GLI), meningioma (MEN), and brain metastases (MET). Each case included multi-sequence MRI triplanar mosaics and structured clinical features transformed into standardized VQA items. Models were assessed in a zero-shot chain-of-thought setting for accuracy on both visual and reasoning tasks. Results showed that GPT-5-mini achieved the highest macro-average accuracy (44.19%), followed by GPT-5 (43.71%), GPT-4o (41.49%), and GPT-5-nano (35.85%). Performance varied by tumor subtype, with no single model dominating across all cohorts. These findings suggest that GPT-5 family models can achieve moderate accuracy in structured neuro-oncological VQA tasks, but not at a level acceptable for clinical use.
title Performance of GPT-5 in Brain Tumor MRI Reasoning
topic Computer Vision and Pattern Recognition
Artificial Intelligence
url https://arxiv.org/abs/2508.10865