Abstract： Objective To demonstrate the validity of volumetric QCT and dual energy X-ray absorptiometry( DXA ) in bone mineral density ( BMD ) measurement and compare the difference in discriminating osteoporotic postmenopausal women with and without vertebral fracture. Methods One hundred and eighteen postmenopausal women [mean age (62.1 ±7.0) years] who received thoracolumbar&nbsp;radiographic examination were enrolled and divided into four groups ( normal, osteopenia, osteoporotic and osteoporotic fractured group) also based on their BMD value of lumbar vertebra( AP-SPINE) measured by DXA: >xˉ-1s,xˉ-1s—xˉ-2s,<xˉ-2s without and with osteoporotic vertebral fracture.The second to four lumbar vertebrae were performed axial scan ( for conventional 2D trabecular BMD measurement, 2D-TRAB by corresponding QCT software) and volumetric scan using MSCT.The volumetric data were transferred to the workstation ADW4.2 for VR and MPR processing, and volumetric BMD were obtained based on the histogram of CT value, including integral BMD (3D-INTGL), cortical BMD (3D-CORT) and trabecular BMD (3D-TRAB), with the unit of mg/cm3.Apparent bone volume to total volume ratio (App BV/TV%) was calculated on the base of trabecular bone whose CT values were among 60 HU, 80 HU, 100 HU, 120—400 HU, respectively.Analysis of covariance ( ANCOVA) and calculation of coefficient of determination ( R2 ) were performed for each parameter among the 4 groups.Results The values of 2D-TRAB, 3D-INT, 3D-TRAB, App60 BV/TV%, App80 BV/TV%, App100 BV/TV% and App120 BV/TV% in osteoporotic fractured group [(48.8 ±24.9) mg/cm3, (94.4 ±20.2) mg/cm3, (59.3 ±28.0) mg/cm3, (56.1 ± 22.8)%, (43.2 ±22.2)%, (31.3 ±19.4)%, (21.3 ±15.6)%] were significantly lower than those in osteoporotic group [(74.9 ±21.0) mg/cm3, (115.0 ±14.3) mg/cm3, (82.0 ±23.7) mg/cm3, (75.2 ± 16.8)%, (62.6 ±20.5)%, (48.8 ±21.7)%, (35.5 ±20.1)%], osteopenia group [(89.2 ± 23.8) mg/cm3,(126.9 ±12.9)mg/cm3, (97.8 ±25.2) mg/cm3, (85.1 ±13.7)%, (75.1 ±17.9)%, (62.8 ±20.9)%, (49.2 ±21.9)%], and normal group [(120.6 ±19.4) mg/cm3, (154.0 ± 16.3) mg/cm3, (131.1 ±21.1) mg/cm3, (95.6 ±5.3)%, (91.4 ±8.7)%, (84.7 ±12.4)%, (75.2 ±15.5)%], P<0.01, respectively.For DXA parameter, there was no significant difference of AP-SPINE values found between osteoporotic fractured group [(0.84 ±0.16) g/cm2 ] and osteoporotic groups [ (0.85 ±0.06) g/cm2 ,P>0.05].In osteoporotic groups, AP-SPINE was not correlated significantly with other variables except 3D-CORT ( R2 =0.189, P <0.01 );parameters of App60, 80, 100, 120 BV/TV% were correlated significantly with 3D-TRAB(R2 =0.955, 0.951, 0.941, 0.912,P<0.01,respectively)and 2D-TRAB( R2 =0.912, 0.910, 0.878, 0.821, P <0.01, respectively ) . The precision of 3D-BMD measurement was between 0.70%—2.25%.Conclusions Parameters derived from 3D-vQCT technique can discriminate osteoporotic postmenopausal women with fractured vertebrae from those without fractured vertebrae, and have better the capability than DXA.Among them, 3D-INTGL was the best parameter for diagnosing serious osteoporosis.App BV/TV%could be used to reflect the lost of trabecular bone precisely and to clinically predict fracture risk.