Effect of slice thickness on brain magnetic resonance image texture analysis

Table 2 The accuracy of LDA method for T1 and T1C

ROIs	1 mm -> 1 mm		1 mm -> 3 mm		3 mm -> 1 mm		3 mm -> 3 mm
	T1	T1C	T1	T1C	T1	T1C	T1	T1C
WM vs. NAWM	60 [59, 61]	55 [53 56]	52 [50, 54]	50 [48, 53]	57 [55, 58]	55 [53, 56]	60 [59, 61]	62 [60, 63]
WM vs. MSi	92 [91, 95]	98 [97, 98]	93 [93, 94]	93 [92, 93]	92 [91, 95]	89 [89, 90]	96 [95, 96]	95 [94, 95]
WM vs. MSr	93 [93, 94]	93 [92, 94]	88 [88, 89]	91 [91, 93]	90 [90, 91]	84 [83, 85]	89 [88, 89]	92 [92, 93]
NAWM vs. MSi	91 [90, 92]	98 [97, 98]	92 [92, 93]	98 [98, 98]	95 [94, 96]	98 [98, 100]	95 [94, 96]	100 [98, 100]
NAWM vs. MSr	91 [90, 92]	90 [89, 91]	88 [86, 88]	90 [89, 90]	90 [90, 91]	91 [90, 92]	88 [86, 88]	94 [93, 94]

The accuracy of the LDA method in percentages for T1 and T1C (with contrast agent) images, in comparisons of white matter (WM), normal appearing white matter (NAWM), MS plaques with irregular ROI (MSi) and MS plaques with regular ROI (MSr). Each comparison includes three of the original texture parameters. Texture parameters have been chosen for LDA based on the Fisher coefficient. In the top row, "X -> Y" refers to the case in which the training set consists of slices with thickness X (1 mm/3 mm), whereas the test set comprises of slices with thickness Y (1 mm/3 mm). The 95% confidence interval for the median is shown in square brackets.

ISSN: 1475-925X