Bone Strength Evaluation Of Lumber Vertebra In Postmenopausal Women By VQCT And FEA

[Objective] To compare the ability of vQCT and DXA in discriminatingosteoporotic fracture through measuring lumbar vertebral volumetric BMD ofpostmenopausal women respectively. To determine the load distribution of lumberspine under physical status by 3D-finite element model of the functional unit ofmultiple segments of lumber spine(L2～L4), and analyze the influence of BMD andthe status of intervertebral disc on load transfer; to calculate the difference of fracturerisk between healthy, osteoporotic and fractured vertebra; to evaluate the usage ofFEA on bone strength; to compare the ability of vQCT and FEA in discriminatingosteoporotic fracture.[Material and Method]1. 118 postmenopausal women were collected and divided into four groupsbased on the BMD value of lumber vertebra measured by DXA:＞M-1SD, M-1SD～M-2SD,＜M-2SD without and with osteoporotic fracture. The second to four lumbervertebra were scanned using MSCT (GE Healthcare, LightSpeedl6) and thetransversal mid-slice images were obtained axially. The following parameters werecalculated: (1) QCT of the L2～L4 lumber spine for trabecular BMD(2D-BMD); (2)The volumetric data were transferred to the workstation ADW4.2 for VR and MPR, and volumetric BMD were obtained based on the histogram of CT value, includingintegral BMD(3D-INT), cortical BMD(3D-COR) and trabecular BMD(3D-TRAB), with the unit mg/cm³. Apparent bone volume ratio to total volume was presumed asthe trabecular whose CT values were among 60Hu, 80Hu, 100Hu, 120Hu to 400Hu respectively(App BV/TVï¼…).2. The volumetric data of one normal case was reconstructed into 1.25mm slices.These slices were reconstructed into a 3D finite element model of the L2～4, usingthe software of MATLAB and ANSYS.3. 30 cases were analyzed using the FE model, including healthy, osteoporosisand osteoporotic fractured cases. The vBMD of the lumber vertebra were transferredto elastic modulus for finite element analysis. We investigated the load distribution oflumbor spine under physical status, and analyzed the influence of osteoporosis and thestatus of disc on the load transfer. Calculated the ratio of trabecular volume to totalvolume whose strains exceeded 5000μstrain, which was the ratio of fracture risk (Fï¼…).4. Statistical analysis: Every BMD parameter of every group was comparedusing ANOVA and covariance analysis of variance. The comparison of the differenceof fracture ratio of trabecular with normal and degenerated disc was conducted using apaired t test. P value less than 0.05 was considered as statistical significance. TheBMD values of the vertebra were predicted using linear correlation and regressionwith the following variables as explanatory variables: AP-SPINE, 2D-TRAB, vBMD, App BV/TVï¼…and Fï¼…. The ability of discriminating osteoporotic fracture wasevaluated using ROC curve with the following parameter: 2D-TRAB, 3D-INT, 3D-TRAB and App BV/TVï¼….[Results]Partâ… 1. The value of 2D-TRAB, 3D-INT, 3D-TRAB, App60 BV/TVï¼…, App₈₀BV/TVï¼…, App₁₀₀ BV/TVï¼…and App₁₂₀ BV/TVï¼…were significantly lower inosteoporotic fractured group than in osteoporotic group(P＜0.01). There was nosignificant difference of 3D-COR, AP-SPINE, NECK, TROCH, WARD between twogroups above(Pvalue0.794, 0.863, 0.148, 0.167, 0.144, respectively).2. In general, AP-SPINE was correlated significantly with 2D-TRAB, 3D-INT, 3D-TRAB and 3D-COR (P=0.0000), and the strongest correlation was betweenAP-SPINE and 3D-INT(R²=0.618); App BV/TVï¼…was correlated significantly with2D-TRAB and 3D-TRAB, and the strongest correlation was between App120 BV/TVï¼…and 3D-TRAB(R²=0.958).In OP groups, AP-SPINE was not correlated significantly with other variablesexcept 3D-COR (R²=0.189).App BV/TVï¼…was correlated significantly with2D-TRAB and 3D-TRAB, and the strongest correlation was between App60 BV/TVï¼…and 3D-TRAB (R²=0.955).3. The area under the ROC curve for 2D-TRAB, 3D-INT and 3D-TRAB was0.885, 0.902, and 0.85, respectively. The threshold of 3D-INT for diagnosingosteoporotic fracture was 116.48 (sensitivity 85.7ï¼…, specificity 75.6ï¼…).The area under the ROC curve for App60 BV/TVï¼…, App₈₀ BV/TV), App₁₀₀BV/TVï¼…and App₁₂₀ BV/TVï¼…was 0.864, 0.862, 0.854, and 0.849, respectively. Thethreshold of App60 BV/TVï¼…for diagnosing osteoporotic fracture was 71.5 (sensitivity78.6ï¼…, specificity 82.2ï¼…).4. App60 BV/TVï¼…showed the highest precision, 0.25ï¼…between twicemeasures, and 0.48ï¼…between observers. The precision of 3D-BMD was between0.70％～2.25ï¼…. All of the above parameters were useful for monitoring therapyefficacy.Partâ…¡1. Under normal intervertebral disc status, the stress distribution in L3 was likea three dimensional"å·¥",that was a high stress region near the endplates, and then themidtransverse region following. With increasing BMD, the percentage of trabecularunder high stress increased, and then trabecular at risk of fracture increased, whichlocated in the center of the trabecular core. For degenerated discs, load wastransferred from the trabecular core to the cortical shell, the stress in the trabecularcore decreased, then the percentage of trabecular at risk of fracture decreased (83.6ï¼…in fractured group). 2. The amount of trabecular bone at risk of fracture was about 0ï¼…for healthyvertebra, about 0.235ï¼…for osteoporotic vertebra, and about 21.90ï¼…for fracturedvertebra. There was significant difference between fractured and nonfractured groups(P=0.000).3. The variable Fï¼…measured in FEA was correlated significantly with thevBMD parameters and App BV/TVï¼…(r -0.679～-0.861), and the strongestcorrelation was between Fï¼…and App60 BV/TVï¼….[Conclusion]1. Parameters derived from QCT could different osteoporotic fractured vertebrafrom unfractured vertebra, so they could predict fracture risk clinically. 3D-INT wasthe best parameter for diagnosing fracture.2. App BV/TVï¼…could predict fracture risk because it could reflect the lost oftrabecular bone precisely, among which the strongest one was App60 BWTVï¼….3. The precision of 3D-BMD was high enough for monitoring therapy efficacy, and App₆₀ BV/TVï¼…showed the highest precision.4. A functional lumber spine model of L2～4 was established successfully usingthe volumetric data from MSCT, which could be used for biomechanical analyze ofthe lumbar spine. It will offer great promise for improvement of clinical osteoporosisdiagnosis, predicting fracture risk and bone strength assessment.5. The strength of trabecular core decreased after osteoporosis, which increasedthe amount of trabecular at risk of fracture. The dangerous region located at the centerof the core. For a degenerated disc, the stress was transferred to the peripheral region.The trabecular core carried fewer loads, which decreased the amount of trabecular atrisk of fracture.6. The trabecular bone at risk of fracture measured in FEA (Fï¼…) was correlatedinversely with the parameters derived from vQCT. FEA was valuable clinically forimprovement of bone strength assessment in osteoporotic vertebra.[Innovative point] 1. To measure the vBMD of lumbar spine of the postmenopausal women usingvQCT, and analyze the difference of the volume of trabecular bone to assess it's valuefor bone strength assessment comparing to DXA.2. We established a 3D-FE model of the functional lumbor spine of L2～4 using16-MSCT for the first time. The model resembled geometrically the actuality perfectly, and conformed to the physiologic characteristic of the lumbar spine.3. Investigate the influence of osteoporosis and the state of disc on the loaddistribution of the lumbar spine to provide the basis for predicting bone strength andthe dangerous spot of fracture risk.4. Assess the bone strength and fracture risk of osteoporotic vertebra clinicallyusing vQCT and FEA respectively in order to provide a reliable basis for the bonestrength appraisal using imaging modality.