3D Pediatric Cranial Bone Imaging Using High Resolution MRI
kamlesh patel, MD, MSc, Cihat Eldeniz, PhD, Gary Skolnick, BS, Udayabhanu Jammalamadaka, PhD, Paul Commean, BEE, Manu Goyal, MD, Matthew Smyth, MD, An Hongyu, PhD.
Washington University in St. Louis, saint louis, MO, USA.
There is an unmet need to safely perform head imaging in young children and obtain CT-equivalent cranial bone images without sedation and radiation. Eley et al. developed a “Black Bone” (BB) MR protocol as an alternative to computed tomography (CT); however, challenges in obtaining reproducible 3D reconstructions have prevented clinical application. In this study, we propose a high-resolution Fast Low-Angle Shot (FLASH) Golden-Angle 3D stack-of-stars radial VIBE sequence (GA-VIBE) MR sequence that is intrinsically robust to motion and has enhanced bone vs. soft-tissue contrast.
Pediatric patients under 11 years old presenting for craniosynostosis or other craniofacial abnormalities were eligible for the study. Participants all had undergone a head CT scan as part of routine clinical care. Each patient also underwent the 5-minute GA-VIBE MRI sequence, performed as a stand-alone scan or as an add-on to a clinically indicated MR scan. The imaging data was processed to develop 3D-reconstructed images. The 3D images created from MRI and the gold standard CT 3D reconstructions were randomized and presented to three blinded reviewers for assessment (pediatric neurosurgeon, craniofacial plastic surgeon, neuroradiologist). For each image set, the reviewers individually noted whether they recommended that a second scan should be performed “to allow for clinical diagnosis” and/or “pre-surgical planning” on five-point Likert scales. 54 reviews were returned for analysis (9 patients x 2 imaging methods x 3 reviewers).
The presence or absence of the six primary cranial sutures was recorded. For each imaging method (CT and MRI), there were 162 assessments of suture closure and imaging quality in the regions of interest (9 patients x 6 sutures x 3 reviewers).
Following IRB approval, nine patients underwent an MRI after a clinical head CT scan. Subject age ranged 3 weeks to 9 years (median age 1.6 years). Median time after CT was 47 days (range: 5 - 193). None of the patients were sedated for CT whereas 5 of the 9 MR scans were performed under sedation per routine clinical care. In 51 of the 54 cases, the reviewer stated that both MRI and CT were acceptable for diagnosis and surgical planning. In one review (Patient 5), the CT was considered unacceptable and the corresponding MRI was acceptable; and in two subjects (Patient 4 and Patient 5) the MRI was deemed unacceptable by one reviewer and the corresponding CT was acceptable.
Reviewers reported clear imaging of the cranial sutures on 98% of CT reviews (159/162 independent assessments) and 97% of MRI reviews (157/162). Sensitivity to detect suture closure from MRI was 100% and specificity was 95%. Figure 1 depicts sample images.
The 3D reconstructed images using the GA-VIBE sequence in comparison to CT created clinically acceptable cranial images with ability to detect suture patency. Future directions include reducing the scan time, improve and automate motion correction and post processing for clinical utility.
Back to 2020 Abstracts