Deep-learning-based assessment of tumor response in patients with glioblastoma

Takeaway

  • A deep convolutional neural network (dCNN) accurately assessed glioblastoma patient response to treatment, with no difference when compared with true post-contrast T1-weighted magnetic resonance imaging (MRI) sequences.

Why this matters

  • Gadolinium-based contrast agents (GBCAs) are commonly used to enhance tissue contrast during MRI. Although they appear to be safe, previous studies have shown gadolinium deposition in the brain after repeated administration, with unknown significance.

  • This study supports the potential utilization of dCNN in patients undergoing frequent MRI scans in order to reduce GBCA exposure.