Comparison of QuickNAT and FSL-FIRST for segmentation of subcortical structures in patients with epilepsy

Abstract

Abstract Epilepsy is a neurological disorder that causes seizures and involves widespread structural alteration. Magnetic resonance (MR) is the preferred imaging tool for investigating patients with epilepsy and is also used for segmentation. We used to compare QuickNAT and FSL-FIRST software for the segmentation of subcortical structures in patients with temporal lobe epilepsy (TLE-R and TLE-L) and healthy control. We found that there were statistically significant differences among the automated methods in all groups (TLE-R, TLE-L, and control; mean ± SD) at the left pallidum (16.397 ± 9.326; 18.333 ± 11.062;39.322 ± 23.711) left putamen (29.561 ± 13.642;27.713 ± 13.696;22.499 ± 7.994), right amygdala (26.173 ± 19.743;22.822 ± 12.787;19.429 ± 11.617), right pallidum (24.205 ± 11.674;24.706 ± 10.647;38.976 ± 20.405), and right putamen (37.223 ± 19.498;31.143 ± 18.714;20.914 ± 9.885). We found the superiority of FSL-FIRST software over QuickNAT in calculating both volumes (raw and normalized) of subcortical structures.