Deep learning predicts brain tumor enhancement from non-contrast MRI, aiming to reduce gadolinium use

A deep learning model predicted brain tumor contrast enhancement from non-contrast MRI alone: 83% accuracy, 92% sensitivity, 74% specificity on over 1,100 scans, but authors say it is not yet accurate enough to replace gadolinium-enhanced MRI.

• Deep learning model trained on 11,000 MRI scans from more than 8,500 patients across four countries predicted tumor contrast enhancement from non-contrast sequences.
• Performance was high for meningiomas (93 of 100 cases) but low in children (nine of 20 cases), attributed to fewer pediatric exams in training data.
• Authors caution the model is not currently accurate enough to replace contrast-enhanced MRI; the goal is individualized contrast decisions.

Automated summary

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