Table 2.

Differentiation between patients with and without osteoporotic vertebral fractures including texture analysis – analysis on vertebral level.

Term	Description	β coefficient	95%-CI	p-value
CT_Correlation	Second-order texture feature, representing the linear spatial relationships between texture elements	-0.639	-1.308;-0.938	<0.001
CT_SRLGLE	Higher-order texture feature, representing the joint distribution of short runs and low gray-level values	0.173	-44.109;2028.023	0.060
PDFF_SumAverage	Second-order texture feature, representing the spread of the mean voxel co-occurrence distribution	-0.183	-134.472;-30.952	0.002
CT_Varianceglobal	Global texture feature, representing the spread of gray-level distribution	-0.435	-0.020;-0.005	0.001
CT_LRHGLE	Higher-order texture feature, representing the joint distribution of long runs and high gray-level values	-0.724	0.000;0.000	<0.001
CT_Contrast	Second-order texture feature, representing the local intensity variation	0.551	0.000;0.000	<0.001
PDFF_Energy	Second-order texture feature, representing uniformity	-0.201	-227.524;-36.375	0.007

This table shows the variables kept in the final linear regression model (adjusted R² [R² _a] = 0.66, (F(10, 160) = 34.7, p < 0.001) after a stepwise approach using the binary fracture status (at least one osteoporotic vertebral fracture present/no osteoporotic vertebral fracture present) as the dependent variable (vertebral level-wise analyses). Specifically, it included the texture features CT_Correlation, CT_SRLGLE, PDFF_SumAverage, CT_Varianceglobal, CT_LRHGLE, CT_Contrast, and PDFF_Energy (β coefficients, 95%-confidence intervals [CIs], and p-values shown per texture feature). Patient age, sex, the number of independent variables, and the vertebral level (T1-L5) were considered for adjustment. For vertebral level-wise analyses, the data from each vertebral body were considered as a separate data point.