SS 610 - Bone health and osteoporosis
Dose reduction in MDCT-based bone mineral density and microstructure assessment: effects of low-dose simulation and sparse sampling
Purpose: To evaluate whether in-vivo quantitative bone mineral density (BMD) and microstructure assessment after simulated dose reduction and sparse sampling combined with statistical iterative reconstruction (SIR) is still feasible for differentiating subjects with and without vertebral fractures.
Methods and Materials: In 12 subjects with osteoporotic vertebral fractures and 12 controls pairwise-matched for age and sex undergoing routine thoracic/abdominal MDCT examinations, lower radiation doses were simulated by virtual lower tube current (10, 25 and 50% of the original current) and sparse sampling (10, 25 and 50% of the original raw data). All images were reconstructed with SIR. BMD and trabecular bone microstructure parameters (including bone fraction and trabecular thickness) were extracted from the reconstructed images in T10 to L5.
Results: Effective dose of the original scan was 10 mSv, and 5, 3, 1 mSv for the 50, 25, 10% sparse sampled/low-dose simulated data, respectively. All BMD measurements calculated from different virtual lower tube currents and different sparse samplings were significantly lower in subjects with fractures compared to controls (range: 89-110 vs. 125-187 mg/ml; P<0.003 for all). Analogously, bone fraction (0.31-0.47 vs. 0.50-0.52, P<0.004 for all), and trabecular thickness were consistently lower in subjects with fractures (1.1-1.3 vs. 1.7-2.4 mm; P<0.011 for all).
Conclusion: After simulated dose reduction and sparse sampling, BMD and microstructure parameters obtained from MDCT examinations and processed with statistical iterative reconstruction were still significantly different in subjects with and without vertebral fractures, suggesting osteoporosis diagnosis to be feasible in low-dose protocols.
Purpose: Combining bone structure and density measurement in 3D is required for site-specific bone quality assessment. Spectral molecular imaging can measure bone density in relation to bone structure by quantifying calcium hydroxyapatite (CaHA) in 3D space. This study aimed to optimise spectral CT methodology to measure bone density in excised human bone samples.
Methods and Materials: MARS CT with CdTe Medipix3RX detector was used with brass filter in four energy binds between 26 and 118 keV for following experiments. Two energy protocols were tested using a Perspex phantom containing four concentrations of CaHA (50, 200, 800 and 1200 mg/cm3) to assess the linear relationship between attenuation and density and to detect anomalies in measured mass attenuation curve. Using the chosen protocol, linear response and mass attenuation curves were compared with theoretical values for CaHA and CaCl2 phantoms. Two parts of the same human femoral head were prepared in two conditions (in air and in PBS) then scanned to assess the effect of intra-osseous air on trabecular bone density measurement.
Results: The protocol with energy bins [30-45, 45-60, 60-78 and 78-118 keV] displayed linear response and anticipated mass attenuation behaviour for both CaHA and CaCl2. Due to partial volume effect at bone-air interface, trabecular CaHA density measured 172 ± 46 mg/cm3 (air present) versus 319 ± 20 mg/cm3 (prepared in PBS).
Conclusion: We established a bone density protocol for quantifying bone density with spectral CT. In future, this method could be combined with bone structure measurement to study site-specific bone quality.
L1 vertebra CT density measurements are too variable with different scanning protocols to be used as a simple screening test for osteoporosis
Purpose: Osteoporosis is an increasing problem in the aging population, with insufficiency fractures causing significant morbidity and mortality. CT density of the L1 vertebral body has previously been found to correlate with bone mineral density. There has been recent interest in screening for osteoporosis in patients undergoing CT by measuring their L1 density. We aimed to determine the feasibility of a simple L1 density threshold to identify patients with osteoporosis. We investigated some of the parameters causing variability in L1 density measurements.
Methods and Materials: Retrospective study of 200 randomly selected patients from a previous audit of CT reporting of vertebral fractures. The L1 density was measured on 864 CT scans with a variety of scanners, kVp and iv contrast protocols.
Results: L1 densities for the same patient with multiple scans within 6 months varied by a mean of 27.5 HU (95%CI 23.8-31.1 HU). Iv contrast resulted in a mean increase of 24.5 HU (95%CI 21.6-27.4 HU). Changing the scan kVp from 100 kVp to 130 kVp lead to a mean decrease of 24.1 HU (95%CI 13.45 - 34.8 HU). Increasing age decreased the L1 density by approximately 1.7 HU per year.
Conclusion: There is too much variation in L1 CT density due to patient and scanning parameters to define a single threshold to identify those at risk of osteoporosis. Previous studies have been performed on more defined subsets of patients. Patient age and scanning protocol have significant effects on the measured CT density of the L1 vertebral body.
Purpose: Multi-Detector Computed Tomography (MDCT) image-based finite-element (FE) analysis has shown significant improvement in prediction of vertebral bone strength beyond bone mineral density (BMD). However, high radiation exposure limits its use in clinical routine. This study aimed to evaluate the effect of dose reduction, by means of tube exposure reduction, on bone strength prediction from FE analysis.
Methods and Materials: Fresh thoracic midvertebral specimens (n = 11) were imaged in a water bath to simulate the soft tissue environment using MDCT with different X-ray tube exposures (80, 150, 220 and 500 mAs). Differences in image quality and geometry of each specimen were measured. FE analysis was performed on all specimens to predict fracture load. Paired t-tests were used to compare the results obtained, using the highest MDCT dose (500 mAs) as reference.
Results: Dose reduction had no significant impact on FE-predicted fracture loads, with significant correlations obtained with reference to 500 mAs, for 80 mAs (R2 = 0.997, p < 0.001), 150 mAs (R2 = 0.998, p < 0.001) and 220 mAs (R2 = 0.987, p < 0.001). There were no significant differences in volume quantification between the different doses examined.
Conclusion: MDCT radiation dose could be reduced substantially with no impact on strength estimates obtained from FE analysis. This finding may enable early diagnosis and advanced monitoring of osteoporosis and associated fracture risk.
Femoral neck strength prediction in osteoporosis patients: trabecular bone analysis using tomosynthesis images
Purpose: Although bone mineral density (BMD) by dual x-ray absorptiometry (DXA) has been used for the diagnosis of osteoporosis, bone marrow quality also has important implications in bone strength prediction. The purpose of this study is to determine the value of trabecular bone analysis using tomosynthesis (TS) images to BMD in femoral neck strength prediction of osteoporosis patients.
Methods and Materials: Forty-seven consecutive osteoporosis patients were included in this study. They underwent DXA, TS, and CT covering the hip joints within a week. We extracted the trabecular patterns of TS images, and obtained the total strut length (TSL), the bone volume per tissue volume (BV/TV) and the textural features (HOM: homogeneity, ENT: entropy, COR: correlation, CON: contrast, VAR: variance) as the indices of TS images. Failure load of the femoral neck, determined by the CT-based finite-element method (FEM), was used as the gold standard for bone strength. Stepwise multiple regression analysis for evaluating the availability of the TS image indices was performed.
Results: Amongst the combinations of the BMD and each index of TS images, the combination of BMD with the TSL and the VAR showed the highest correlation to the failure load by CT-FEM. The correlation between the failure load and the BMD with the TSL and the VAR (r2=0.72) was significantly higher than that between the failure load and the BMD alone (r2=0.67; p=0.0397).
Conclusion: TS-based trabecular bone analysis in combination with BMD measurements can potentially be used in predicting bone strength in osteoporosis patients.
Validation of bone volume percentage and pore size measurements extracted from MR and MDCT against synthetic bone phantom
Purpose: The characterisation of the bone microarchitecture from MR or CT is significantly relevant to predict the risk of bone fracture. Our purpose was to evaluate the accuracy of the bone analysis methodology and compare it to the reference standard under different imaging modalities.
Methods and Materials: The reference is based on synthetic bones, consisting of 10 samples, 5 with 15% and 5 with 30% of material density that were acquired from SAWBONES® (Washington, USA) with known mass, density and pore size and were scanned using MDCT and MR. The voxel size in MDCT was of 0.234x0.236x0.67mm. The MR sequence was a T2-GRE with a voxel of 0.146x0.146x0.5mm. The image processing steps were different in MDCT and MR, according to a methodology developed in our group. Bone volume to total volume (BV/TV) was measured to define the bone percentage in the volume. Trabecular separation (Tb.Sp) was quantified by taking into account consecutive voxels corresponding to marrow cavities.
Results: Mean relative errors of 9% and 27% were obtained for MDCT and MR measurements of BV/TV, with a tendency to overestimate the parameter in 15% density samples but underestimate in 30% density samples with similar errors. Mean relative errors for Tb.Sp were of 4% and 34% for MDCT and MR, respectively.
Conclusion: MR and MDCT measurements of trabecular bone volume percentage and pore size were calibrated against the known ground truth from synthetic bones. These results add insight into the validation of bone microarchitecture imaging biomarkers.
Characterisation of sharp force trauma and thermal injuries on human bone samples: a forensic MicroCT study
Purpose: Identification of sharp force and thermal injuries is a main challenge in forensic medicine.Thus, aim of this study was to investigate the role of MicroCT for the detection of experimentally produced saw marks and the characterization of bone structural parameters after fire exposure.
Methods and Materials: Saw-marks were experimentally produced on eight human bone samples using either a 24-teeth-per-inch (TPI) with alternating-set or a 18-TPI with wavy-set rip-cut-saw (ie.,four samples each). Each sample was then examined by MicroCT before and after exposure to open-flame fire for two and ten minutes. For all investigated conditions, saw-marks’ detectability and structural cortical (bone volume (BVc), bone surface (BSc), bone surface/bone volume (BS/BVc)) and trabecular bone parameters (BVt, BSt, BS/BVt, structural model index, anisotropy degree, trabecular thickness, -spacing, -number, -pattern factor) were collected and compared using a one-way-repeated measures ANOVA with Greenhouse-Geisser correction and Bonferroni post-hoc tests.
Results: All saw-marks (n=8) were recognizable at each investigated condition (i.e., fresh, after two and ten minutes of fire exposure). BVt and BSt significantly decreased comparing two (BVt mean±SD=5.16±2.96 mm3; BSt=133.40±69.13 mm2) and ten (BVt=3.48±2.23 mm3;BSt=97.15±56.72 mm2) minutes exposure (p<0.05, each). BVc decreased whereas BS/BVc increased at all tested condition (BVc values: 139.95±67.18 mm3,131.66±71.64 mm3,119.89±68.91 mm3; BS/BVc values: 5.30±1.06%, 5.48±1.29%, 5.94±1.17%, respectively fresh, at two and at ten minutes fire exposure; p<0.05,each). All other cortical and trabecular parameters did not show any significant difference (p>0.05).
Conclusion: MicroCT demonstrated to be an accurate forensic tool to investigate bone saw marks, even after thermal injuries, providing information about heat-induced bone modifications. Future studies including prolonged fire exposure are necessary to further assess these evidences.
The effect of vertebral osteoarthrosis (VO) and vertebral fractures (VF) on trabecular bone score (TBS): preliminary results
Purpose: TBS is a textural score that provides an indirect index of trabecular microarchitecture from lumbar spine (LS) dual energy x-ray absorptiometry (DXA). We evaluated the effect of VO and VF on bone mineral density (BMD) and TBS.
Methods and Materials: We retrospectively identified all patients that performed an LS DXA in January-March 2015 together with an LS x-ray/MRI (maximum interval with DXA=6 months). Among them, we included patients with VO or VF diagnosed on LS x-ray/MRI with a maximum of two vertebrae involved. We calculated the BMD/TBS difference between vertebrae affected by VO/VF and the adjacent vertebrae with the greater values of BMD/TBS; between vertebrae affected by VO/VF and the average L1-L4 BMD/TBS (average value included the vertebrae with VO/VF).
Results: Out of 258, we included 20 patients (19 females, age=72±10y) with VO (n=13) or VF (n=7). Mean BMD (g/cm2): vertebrae with VO/VF=1.004±0.167; adjacent-vertebrae=0.935±0.154; L1-L4=0.946±0.127. Mean TBS: vertebrae with VO/VF=1.300±0.122; adjacent-vertebrae=1.252±0.122; L1-L4=1.261±0.091. Considering VO+VF together, all differences were significant (p<0.05), except for BMD difference between vertebrae with VO/VF and adjacent-vertebrae (p=0.082). Considering only VO vertebrae, all differences between vertebrae were significant (p<0.05). Considering only VF vertebrae, no differences were found with adjacent-vertebrae/L1-L4 for BMD and TBS (p>0.119). Correlations between BMD and TBS for vertebrae with VO/VF were all significant (p<0.05): R=0.5247 (VO+VF), R=0.6194 (only VO), R=0.9224 (only VF).
Conclusion: Compared to adjacent vertebrae and L1-L4, VO significantly impacts both on BMD and TBS, while no differences were found for VF. BMD and TBS are positively correlated, especially for VF.
Lumbar bone marrow perfusion in osteoporosis rabbit with dynamic contrast enhancement MRI: correlation with BMD and MVD
Purpose: To use dynamic contrast enhancement (DCE-MRI) to assess lumbar spine bone marrow blood perfusion in osteoporosis rabbit and correlate perfusion with bone mineral density (BMD) and microvessel density (MVD).
Methods and Materials: 22 New Zealand white rabbits were randomly divided into 2 groups, including experimental group (n=12) and control group (n=10). Bilateral ovariectomies (OVX) were applied in experimental group. DCE-MRI, DEXA and immunohistochemistry exam were respectively performed both groups in 3, 5, 7 moths after OVX. Vertebral maximum percentage of enhancement (Emax), enhancement slope (ES), BMD and MVD were measured from them.
Results: BMD of lumber spine in OVX group in postoperative 7m was obviously decreased compared to control group (P<0.05). In postoperative 3 and 5m lumber spine bone marrow Emax and ES had no difference both groups(P>0.05), however the significant difference were found in postoperative 7m (P<0.05), Emax and ES were significantly decreased in OVX group. MVD of vertebral marrow was 67.58±11.6 in control group and 39.32±9.54 in OVX group, which were significant difference between them (P<0.05).Vertebral marrow Emax and ES were significant positive correlation with BMD and MVD in 7m OVX group (r=0.714, 0.820, 0.866 and 0.771,P<0.05).
Conclusion: Lumbar bone marrow perfusion is significantly decreased in osteoporosis rabbit, similarly vertebral MVD obviously reduce too. DCE-MRI can provide more information about osteoporosis blood perfusion. Emax and ES are significant correlation with BMD and MVD. This result may suggest a vascular component in the pathogenesis of OP and is helpful to diagnosis and treatment of OP.
Prediction of fracture non-union healing using clinical scores, contrast-enhanced ultrasound (CEUS) and dynamic contrast-enhanced MRI (DCE-MRI)
Purpose: To assess whether dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) and contrast-enhanced ultrasound (CEUS) can predict fracture non-union consolidation after revision surgery.
Methods and Materials: Perfusion within non-unions of extremities was prospectively quantified in 205 patients (mean, 51.5 years; 76 female) before revision surgery and at 6, 12, 26, 52 and more weeks follow-up. 3-Tesla DCE-MRI derived signal-intensity-curves obtained from a region-of-interest analysis were normalized to adjacent muscle tissue and non-unions were classified as vascularised or not depending on the contrast uptake within the non-union. Potential infection was measured by positive microbiologic culture of the resected non-union tissue. In addition, CEUS was performed in 43 patients (mean, 46.8 years; 14 female). The ability to predict successful outcome with osseous consolidation at CT of DCE-MRI and CEUS parameters, body mass index (BMI), and the non-union scoring system (NUSS) was estimated by receiver-operating-characteristic analysis.
Results: 103/169 (61%) non-unions eventually healed and demonstrated higher perfusion than in failed consolidation at 6 (p=0.023), 12 (p=0.025) and 26 (p=0.009) weeks follow-up (sensitivity/specificity of DCE-MRI at 26 weeks for prediction of consolidation: 75%/87%, false classification rate: 19%). 50% of all non-unions consolidated within 14 months after revision surgery. Both a low BMI (p=0.041) and NUSS (p<0.0001) were associated with eventual consolidation. CEUS detected preoperatively increased perfusion kinetics in infected non-unions (n=16/43, rise time (p=0.024), time to peak (p=0.011), sensitivity/specificity: 93%/65%).
Conclusion: DCE-MRI at 26 weeks after revision surgery of non-unions predicts successful outcome. Preoperative vascularity in CEUS, NUSS and BMI are important prognostic factors concerning consolidation.
Purpose: To establish a primary osteoporosis rabbit model by ovariectomy (OVX) and measure lumbar spine T2* and R2* MR relaxation parameters in osteoporosis rabbit and compare them with bone mineral density (BMD) and histomorphology.
Methods and Materials: 22 New Zealand white rabbits were randomly divided into 2 groups, including experimental group (n=12) and control group (n=10). Bilateral ovariectomies were applized in experimental group. MRI, DEXA and pathohistology exam were respectively performed both groups in 3, 5, 7 moths after OVX. T2*, R2* and BMD of lumber spine were obtained.
Results: BMD of lumber spine in OVX group in postoperative 7m was obviously decreased compared to control group (P<0.05). In postoperative 3m lumber spine T2* and R2* value had no difference both groups; but the significant difference were found in postoperative 5m (P<0.05) and 7m (P<0.001), T2* was significantly increased and R2* was obviously decreased. Histomorphological characteristics showed the trabecular bone were thinner, interrupted and trabecular spaces enlarged and trabecular number lessened in OVX group. Lumber spine T2* value was significant inverse correlation with BMD in OVX group (r =-0.599, P< 0.001), and R2* was positive association with BMD (r =0.746, P< 0.001).
Conclusion: T2* and R2* MR relaxation time can sensitively detect trabecular bone micro-architecture changes in osteoporosis rabbit and are significant correlation with BMD. Combination with T2*, R2* and BMD can better reflect bone strength and bong quality variation in early OP.