• Suraj D Serai, Jonathan Dudley, and James L Leach.
• Department of Radiology, Children's Hospital of Philadelphia, Philadelphia, PA, United States of America; Department of Radiology, Cincinnati Children's Hospital and Medical Center, Cincinnati, OH, United States of America. Electronic address: serais@email.chop.edu.
• Clin Imaging. 2019 Sep 1; 57: 77-82.

BackgroundMRI brain segmentation and volume estimation of gray matter (GM), white matter (WM) and cerebrospinal fluid (CSF) are important for many neurological applications. Signal intensity based measurements, such as the current statistical parametric mapping (SPM) based volume estimation techniques rely on T1W images that involve a series of pre-processing steps, making it impractical for clinical use.PurposeIn this study, we compared Synthetic MRI (SyMRI) generated relaxometry maps based brain segmentation and estimation of brain volumes with SPM image intensity based segmentation and volume estimation in children and young adults.Subjects176 studies were included for analysis with mean age of 10.9 ± 5.5 years.MethodsIncluded studies were quantitatively analyzed and segmented using SyMRI® software. In SPM, the segmentation routine segments brain T1W images into GM, WM, and CSF based on image intensity values. SPM and SyMRI segmentation based volume estimates were plotted. Scatter plots comparing the two methods were generated and agreement was assessed using correlation coefficients.ResultsCorrelation coefficient, r, of agreement between the 2 methods was 0.85 for GM, 0.91 for WM, and 0.38 for CSF (P < 0.0001 for all three volumes).ConclusionBrain imaging in children using SyMRI can identify and calculate estimates of GM, WM, CSF volumes. With our work, we have shown high similarity of volume estimates in GM and WM using SyMRI with a systematic bias for CSF values. The ease of use of this software can make this quantitative data to be used clinically along with the routine anatomical images.Copyright © 2019 Elsevier Inc. All rights reserved.

26 keywords...

hide…