Automated opportunistic screening for osteoporosis using deep learning-based automatic segmentation and radiomics on proximal femur images from low-dose abdominal CT

Table 2 Detailed results of automatic segmentation accuracy in all cohorts

Cohort	Group	Dice ^a	VD (cm³) ^b
Validation cohort	All patients	0.975 ± 0.012	0.057(0.022,0.129)
	Normal	0.979 ± 0.009	0.058(0.022,0.132)
	Osteopenia	0.973 ± 0.012	0.058(0.022,0.142)
	Osteoporosis	0.966 ± 0.017	0.058(0.023,0.132)
Test cohort	All patients	0.955 ± 0.137	0.041(0.021,0.068)
	Normal	0.956 ± 0.135	0.042(0.021,0.068)
	Osteopenia	0.944 ± 0.178	0.041(0.021,0.068)
	Osteoporosis	0.967 ± 0.011	0.047(0.023,0.068)

ISSN: 1471-2474