Thursday, March 13, 2025
3/13/2025
Project Summary: Brain MRIs are widely used in children for diagnosis and treatment monitoring. A typical exam comprises multiple sequences with different contrast preparations, and often requires 20+ minutes to complete. Potential motion during such lengthy acquisitions necessitates sedation or anesthesia. However, repeated sedation or anesthesia in children increase the risk of long-term detrimental effects on cognitive development. 40% of these pediatric scans also require gadolinium (Gd) based contrast agent (GBCA) administration. It has been recently shown that Gd is retained in the brain and body, which may be particularly harmful to children since their developing brains are more susceptible to heavy metal exposure, and free lanthanides are known to be neurotoxic. Pediatric brain exams are excessively long because of inefficient use of parallel imaging technology that only provides R=2–3-fold acceleration. Acceleration in two dimensions, including the slice/partition axis, through controlled aliasing in parallel imaging (CAIPI) has enabled R=4–6-fold speed-up, and has become popular in functional/diffusion imaging using simultaneous multislice (SMS) encoding. Unfortunately, adoption of SMS in clinical sequences has been extremely slow. Acceleration via Compressed Sensing (CS) is a promising solution, but its availability as a product solution is variable among MR vendors, and it often comes at the cost of low-contrast image features. Deep learning (DL) has emerged as a powerful reconstruction and image enhancement tool. Vendors’ DL solutions include denoising and super- resolution enhancement, but these are limited to the newest software versions and host computers with GPUs. While promising a better trade-off between image quality and scan time, they are currently implemented for a small number of sequences on different vendors, and are therefore constrained by poor availability and scalability. Thus, current clinical technologies have been hampered by limited availability and faster has often meant suboptimal quality. Lastly, no vendor has DL-based solutions for Gd dose reduction for either adults or children. Given the need for repeated injections of sedatives and GBCA in children who are scanned periodically for treatment monitoring, there is an unmet need in imaging technology that makes these young populations vulnerable to severe and long-term health risks. We propose data acquisition, reconstruction and contrast enhancement strategies to address this unmet need. In Aim1, we will develop a rapid, comprehensive brain exam by combining our advanced controlled aliasing strategy, wave-CAIPI, and extend this to SMS encoding for rapid FLAIR/TSE imaging. Combining this with DL super-resolution reconstruction will enable R=9-fold acceleration with high fidelity to create a 6-min protocol. In Aim2, we will develop and validate a DL- based contrast enhancement algorithm to synthesize full dose images from 5× reduced Gd dose in pediatric exams using our rapid protocol from Aim1. To that end, our novel technologies would speed up the clinical MR exams and minimize both the amount of sedation and the injected contrast agents dose in children.
HHS’ Tracking Accountability in Government Grants System (TAGGS) website provides access to detailed descriptions of grants, loans, aggregated direct payments and other types of financial assistance awarded by the HHS Operating Divisions (OPDIVs) and the Office of the Secretary Staff Divisions (STAFFDIVs).