Room: Track 1
Purpose: Shorter brain protocols can increase patient throughput, reduce the risk of motion, and reduce the need for sedation for claustrophobic or pediatric patients. However, reducing scan time for fast brain protocols typically results in compromised image quality. Deep learning based image reconstruction (DL Recon) offers the potential for increased SNR, reduced artifacts, and enhanced resolution. To gain acceptance these improvements must be verified while ensuring that image contrast maintained. In this work, we quantify the effect of a DL Recon on image quality on an accelerated brain protocol with gadolinium contrast.
Methods: The DL Recon used in our study is a deep convolutional residual encoder network trained to reconstruct images from 2D MR data with reduced noise, reduced Gibbs ringing, and enhanced resolution. The network has an adjustable parameter that ranges between 0 and 100% to control the noise level of the final reconstructed images.
Raw data for twelve patients receiving our standard of care accelerated brain protocol were saved for this study. In addition to the standard images generated with conventional reconstruction on the scanner, DL Recon was retroactively applied at 75% noise reduction level to the post contrast T1 weighted gradient echo sequence in the protocol.
ROI’s were drawn in the CSF, white matter, and background air, and mean and standard deviation of the ROI’s were recorded. These values were used to calculate SNR in CSF and white matter-CSF contrast. Difference images were generated to observe any change in noise and image features.
Results: DL recon improved SNR in all subjects. Mean background signal was reduced while tissue contrast was maintained. Difference images demonstrate reduction in ringing artifacts as well as noise in both the subject and background.
Conclusion: DL Recon is a promising technique for restoring image quality that is lost from acceleration of brain sequences.
Funding Support, Disclosures, and Conflict of Interest: Ken Hwang receives research funding from GE Healthcare.