Introduction
Materials and methods
Cohort characteristics
Marseille database
iDB-MRXFDG database
Patient group
Image processing
Processing of the Marseille database
Visual inspection
SPM analysis (Fig. 1)
Reorientation and coregistration
Spatial normalisation
Intensity normalisation
Leave-one-out SPM analysis
Regional analysis
Processing of MRXFDG database
Processing of patient group
Merging of control databases
Image resolution
Age correlation analysis
Voxel-based differences in [18F]FDG maps
Detection of FCD-related decreases in [18F]FDG uptake
Analysis against each control database separately
Analysis against combined control databases
Two-sample t test
Full factorial design
Statistical assessment of patient results
Influence of scanner type and patient characteristics
Results
Marseille database
Leave-one-out SPM analysis
Contrast | Subject-level (%) | Cluster-level (%) | Voxel-level (%) |
---|---|---|---|
N = 58 subjects | |||
Hypermetabolism | 12.1 (n = 7) | 7.4 | 0.98 |
Hypometabolism | 12.1 (n = 7) | 7.4 | 1.10 |
Removing the aforementioned subjects from this analysis (n = 54) | |||
Hypermetabolism | 8.6 (n = 5) | 3.7 | 0.45 |
Hypometabolism | 10.3 (n = 6) | 4.2 | 0.41 |
Demographics
Regional analysis
Database comparability for healthy controls
PET image resolution and smoothing level
Gaussian filter used | Gradient (mean ± SD) | Coefficient of variation (%) |
---|---|---|
Marseille database | ||
No filter | 21.4 ± 3.1 | 14.5 |
8 × 8 × 8 mm | 15.9 ± 1.5 | 9.4 |
iDB-MRXFDG | ||
No filter | 22.2 ± 4.0 | 18.0 |
8 × 8 × 8 mm | 17.2 ± 2.5 | 14.5 |
10 × 10 × 10 mm | 15.0 ± 2.1 | 14.0 |
Regional SUVs
Age correlation
Marseille database
iDB-MRXFDG
Voxel-wise two-sample t test between databases of healthy controls
Detection of FCD-related hypometabolism
FCD detection as a function of statistical model
Marseille | MRXFDG | Both DB (2STT) | Both DB (FF—under 60 years) | Both DB (FF—all ages, linear age covariate) | Both DB (FF—all ages, quadratic age covariate) | |
---|---|---|---|---|---|---|
Datasets in database(s) used | 54 | 37 | 91 | 70 | 91 | 91 |
Lesions found (any p value) | ||||||
n | 17 | 18 | 14 | 24 | 22 | 19 |
% | 50 | 53 | 41 | 71 | 65 | 56 |
Lesions found (significant p value) | ||||||
n | 14 | 18 | 12 | 22 | 19 | 18 |
% | 41 | 53 | 41 | 65 | 56 | 53 |