Related Experiment Video
Updated: Apr 1, 2026
Automated Segmentation of Cortical Grey Matter from T1-Weighted MRI Images
Published on: January 7, 2019
Partial volume correction and image segmentation for accurate measurement of standardized uptake value of grey matter
Gonca Bural1, Drew Torigian, Sandip Basu
1aDepartment of Nuclear Medicine, İzmir Kâtip Çelebi University Atatürk Research and Training Hospital, İzmir,Turkey bDepartment of Radiology, Hospital of the University of Pennsylvania, Philadelphia, Pennsylvania, USA cRadiation Medicine Centre (BARC), Tata Memorial Centre Annexe, Mumbai, Maharashtra, India dDepartment of Radiology, National Liver Institute, Menofia University, Egypt eDepartment of Nuclear Medicine, PET/CT Centre, Radiology, NeuroRadiology, Medical Physics, 'Santa Maria della Misericordia' Hospital, Rovigo, Italy.
Objectives:
Our aim was to explore a novel quantitative method [based upon an MRI-based image segmentation that allows actual calculation of grey matter, white matter and cerebrospinal fluid (CSF) volumes] for overcoming the difficulties associated with conventional techniques for measuring actual metabolic activity of the grey matter.
Methods:
We included four patients with normal brain MRI and fluorine-18 fluorodeoxyglucose (F-FDG)-PET scans (two women and two men; mean age 46±14 years) in this analysis. The time interval between the two scans was 0-180 days. We calculated the volumes of grey matter, white matter and CSF by using a novel segmentation technique applied to the MRI images. We measured the mean standardized uptake value (SUV) representing the whole metabolic activity of the brain from the F-FDG-PET images. We also calculated the white matter SUV from the upper transaxial slices (centrum semiovale) of the F-FDG-PET images. The whole brain volume was calculated by summing up the volumes of the white matter, grey matter and CSF. The global cerebral metabolic activity was calculated by multiplying the mean SUV with total brain volume. The whole brain white matter metabolic activity was calculated by multiplying the mean SUV for the white matter by the white matter volume. The global cerebral metabolic activity only reflects those of the grey matter and the white matter, whereas that of the CSF is zero. We subtracted the global white matter metabolic activity from that of the whole brain, resulting in the global grey matter metabolism alone. We then divided the grey matter global metabolic activity by grey matter volume to accurately calculate the SUV for the grey matter alone.
Results:
The brain volumes ranged between 1546 and 1924 ml. The mean SUV for total brain was 4.8-7. Total metabolic burden of the brain ranged from 5565 to 9617. The mean SUV for white matter was 2.8-4.1. On the basis of these measurements we generated the grey matter SUV, which ranged from 8.1 to 11.3.
Conclusion:
The accurate metabolic activity of the grey matter can be calculated using the novel segmentation technique that we applied to MRI. By combining these quantitative data with those generated from F-FDG-PET images we were able to calculate the accurate metabolic activity of the grey matter. These types of measurements will be of great value in accurate analysis of the data from patients with neuropsychiatric disorders.
