GPT-5 for endolymphatic hydrops MRI grading: human-AI collaboration boosts accuracy

With enhanced prompting, GPT-5 graded cochlear hydrops on MRI with 87% accuracy and vestibular with 80%, but human-AI collaboration improved junior physicians' accuracy from 36% to 59%.

• Retrospective study of 436 patients (872 ears) with Ménière's disease.
• With enhanced prompting (structured image descriptions, clinical context, examples), GPT-5 achieved 87% accuracy for cochlear and 80% for vestibular endolymphatic hydrops grading.
• Human-AI collaboration (Human-first mode) improved junior physicians' accuracy from 36% to 59% (cochlear) and 42% to 53% (vestibular).

Automated summary

RadPigeon summaries are original and for information only. They are not clinical advice.