Bone morphology of the femur and tibia captured by statistical shape modelling predicts rapid bone loss in acute spinal cord injury patients

Regional and temporal variation in bone loss during the first year following spinal cord injury

Osteoporosis is a consequence of spinal cord injury (SCI) that leads to fragility fractures. Visual assessment of bone scans suggests regional variation in bone loss, but this has not been objectively characterised. In addition, substantial inter-individual variation in bone loss following SCI has been reported but it is unclear how to identify fast bone losers. Therefore, to examine regional bone loss, tibial bone parameters were assessed in 13 individuals with SCI (aged 16-76 years). Peripheral quantitative computed tomography scans at 4 % and 66 % tibia length were acquired within 5 weeks, 4 months and 12 months postinjury. Changes in total bone mineral content (BMC), and bone mineral density (BMD) were assessed in ten concentric sectors at the 4 % site. Regional changes in BMC and cortical BMD were analysed in thirty-six polar sectors at the 66 % site using linear mixed effects models. Relationships between regional and total loss at 4 months and 12 months timepoints were assessed using Pearson correlation. At the 4 % site, total BMC (P = 0.001) decreased with time. Relative losses were equal across the sectors (all P > 0.1). At the 66 % site, BMC and cortical BMD absolute losses were similar (all P > 0.3 and P > 0.05, respectively) across polar sectors, but relative loss was greatest in the posterior region (all P < 0.01). At both sites, total BMC loss at 4 months was strongly positively associated with the total loss at 12 months (r = 0.84 and r = 0.82 respectively, both P < 0.001). This correlation was stronger than those observed with 4-month BMD loss in several radial and polar sectors (r = 0.56-0.77, P < 0.05). These results confirm that SCI-induced bone loss varies regionally in the tibial diaphysis. Moreover, bone loss at 4 months is a strong predictor of total loss 12 months postinjury. More studies on larger populations are required to confirm these findings.

Statistical Shape and Appearance Models: Development Towards Improved Osteoporosis Care

PURPOSE OF REVIEW

Statistical models of shape and appearance have increased their popularity since the 1990s and are today highly prevalent in the field of medical image analysis. In this article, we review the recent literature about how statistical models have been applied in the context of osteoporosis and fracture risk estimation.

RECENT FINDINGS

Recent developments have increased their ability to accurately segment bones, as well as to perform 3D reconstruction and classify bone anatomies, all features of high interest in the field of osteoporosis and fragility fractures diagnosis, prevention, and treatment. An increasing number of studies used statistical models to estimate fracture risk in retrospective case-control cohorts, which is a promising step towards future clinical application. All the reviewed application areas made considerable steps forward in the past 5-6 years. Heterogeneities in validation hinder a thorough comparison between the different methods and represent one of the future challenges to be addressed to reach clinical implementation.

Statistical shape modelling provides a responsive measure of morphological change in knee osteoarthritis over 12 months

OBJECTIVES

Responsive biomarkers are needed to assess the progression of OA and their lack has hampered previous clinical trials. Statistical shape modelling (SSM) from radiographic images identifies those at greatest risk of fast-progression or joint replacement, but its sensitivity to change has not previously been measured. This study evaluates the responsiveness of SSM in knee OA in a 12-month observational study.

METHODS

A total of 109 people were recruited who had undergone knee radiographs in the previous 12 months, and were grouped based on severity of radiographic OA (Kellgren-Lawrence grading). An SSM was built from three dual-energy X-ray absorptiometry scans at 6-month intervals. Change-over-time and OA were assessed using generalized estimating equations, standardized response means (SRM) and reliable change indices.

RESULTS

Mode 1 showed typical features of radiographic OA and had a strong link with Kellgren-Lawrence grading but did not change significantly during the study. Mode 3 showed asymmetrical changes consistent with medial cartilage loss, osteophytes and joint malalignment, and was responsive to change, with a 12-month SRM of 0.63. The greatest change was observed in the moderate radiographic OA group (SRM 0.92) compared with the controls (SRM 0.21), and the reliable change index identified 14% of this group whose progression was clinically significant.

CONCLUSION

Shape changes linked the progression of osteophytosis with increasing malalignment within the joint. Modelling of the whole joint enabled quantification of change beyond the point where bone-to-bone contact has been made. The knee SSM is, therefore, a responsive biomarker for radiographic change in knees over 12 months.

Perinatal factors associate with vertebral size and shape but not lumbar lordosis in 10-year-old children

The intrauterine environment is known to influence foetal development and future health. Low birthweight has been linked to smaller vertebral canals in children and decreased adulthood spine bone mineral content. Perinatal factors affecting lumbar spine curvature have not yet been considered but could be important for adult spinal health, as lumbar movement during lifting, a risk factor for backpain, is associated with lordosis. To investigate this, lumbar spine magnetic resonance images at age 10 years and perinatal and maternal data (birthweight, placental weight, gestation length, crown-heel length, maternal age, height, weight and smoking status) from 161 children born in Aberdeen in 1988-1989 were acquired. Statistical shape modelling, using principal component analysis, quantified variations in lumbar spine shape and resulting modes of variation were assessed in combination with perinatal data using correlations and analyses of covariance, adjusted for potential confounders. Spine modes 1-3 (SM1-SM3) captured 75% of the variation in lumbar spine shape. The first and third modes described the total amount (SM1) and evenness of curvature distribution (SM3). SM2 accounted for variations in antero-posterior vertebral diameter relative to vertebral height, increasing positive scores representing a larger relative diameter. Adjusting for gestation length and sex, SM2 positively correlated with birthweight (r = 0.25, P < 0.01), placental weight (r = 0.20, P = 0.04), crown-heel length (r = 0.36, P < 0.001) and maternal weight (r = 0.19, P = 0.04), and negatively with maternal age (r = -0.22, P = 0.02). SM2 scores were lower in girls (P < 0.001) and in the low birthweight group (P = 0.02). There were no significant differences in SM1 and SM3 scores between birthweight groups, boys and girls or children of smokers (31%) and non-smokers (69%). In conclusion, some perinatal factors were associated with vertebral body morphology but had little effect on lumbar curvature.

Effects of drugs on bone metabolism in a cohort of individuals with traumatic spinal cord injury

Study Design

This study is a retrospective review examining the prevalence of drugs commonly used in the management of spinal cord injury (SCI) which may influence bone health.

Objective

The aim of our study was to examine the role commonly prescribed medications play in post-SCI bone health.

Setting

We included all males 21 years of age and older who were evaluated over a 10-year period at an SCI-specialized center for a trauma-induced SCI.

Method

We compared characteristics of individuals with normal bone mass to those with low bone mass according to their dual-energy X-ray absorptiometry (DXA) scan. Medication lists were reviewed for the presence of drugs considered to either positively or negatively affect bone metabolism.

Results

Comparing individuals with normal bone mass (n = 68) to those with low bone mass (n = 211), only "Time after Injury" and "Level of Injury" were found to influence the likelihood of having low bone mass. Multivariate analysis failed to demonstrate significant associations between bone mass and the sum of drugs which either positively or negatively affect bone metabolism. When medications were reviewed individually, only bisphosphonates and anticonvulsants were found to be significantly associated with bone mass.

Conclusions

Although 76% of our cohort was found to have low bone mass, the only major risk factors were "Time after Injury" and "Level of Injury". Anticonvulsant use was more common in individuals with low bone mass compared to those with normal bone mass. Given the retrospective methodology of this work, our findings underline associations that warrant further investigation.

Prenatal growth map of the mouse knee joint by means of deformable registration technique

Joint morphogenesis is the process during which distinct and functional joint shapes emerge during pre- and post-natal joint development. In this study, a repeatable semi-automatic protocol capable of providing a 3D realistic developmental map of the prenatal mouse knee joint was designed by combining Optical Projection Tomography imaging (OPT) and a deformable registration algorithm (Sheffield Image Registration toolkit, ShIRT). Eleven left limbs of healthy murine embryos were scanned with OPT (voxel size: 14.63μm) at two different stages of development: Theiler stage (TS) 23 (approximately 14.5 embryonic days) and 24 (approximately 15.5 embryonic days). One TS23 limb was used to evaluate the precision of the displacement predictions for this specific case. The remaining limbs were then used to estimate Developmental Tibia and Femur Maps. Acceptable uncertainties of the displacement predictions computed from repeated images were found for both epiphyses (between 1.3μm and 1.4μm for the proximal tibia and between 0.7μm and 1.0μm for the femur, along all directions). The protocol was found to be reproducible with maximum Modified Housdorff Distance (MHD) differences equal to 1.9 μm and 1.5 μm for the tibial and femoral epiphyses respectively. The effect of the initial shape of the rudiment affected the developmental maps with MHD of 21.7 μm and 21.9 μm for the tibial and femoral epiphyses respectively, which correspond to 1.4 and 1.5 times the voxel size. To conclude, this study proposes a repeatable semi-automatic protocol capable of providing mean 3D realistic developmental map of a developing rudiment allowing researchers to study how growth and adaptation are directed by biological and mechanobiological factors.

Peripheral Quantitative Computed Tomography: Review of Evidence and Recommendations for Image Acquisition, Analysis, and Reporting, Among Individuals With Neurological Impairment

In 2015, the International Society for Clinical Densitometry (ISCD) position statement regarding peripheral quantitative computed tomography (pQCT) did not recommend routine use of pQCT, in clinical settings until consistency in image acquisition and analysis protocols are reached, normative studies conducted, and treatment thresholds identified. To date, the lack of consensus-derived recommendations regarding pQCT implementation remains a barrier to implementation of pQCT technology. Thus, based on description of available evidence and literature synthesis, this review recommends the most appropriate pQCT acquisition and analysis protocols for clinical care and research purposes, and recommends specific measures for diagnosis of osteoporosis, assigning fracture risk, and monitoring osteoporosis treatment effectiveness, among patients with neurological impairment. A systematic literature search of MEDLINE, EMBASE^©, CINAHL, and PubMed for available pQCT studies assessing bone health was carried out from inception to August 8th, 2017. The search was limited to individuals with neurological impairment (spinal cord injury, stroke, and multiple sclerosis) as these groups have rapid and severe regional declines in bone mass. Of 923 references, we identified 69 that met review inclusion criteria. The majority of studies (n = 60) used the Stratec XCT 2000/3000 pQCT scanners as reflected in our evaluation of acquisition and analysis protocols. Overall congruence with the ISCD Official Positions was poor. Only 11% (n = 6) studies met quality reporting criteria for image acquisition and 32% (n = 19) reported their data analysis in a format suitable for reproduction. Therefore, based on current literature synthesis, ISCD position statement standards and the authors' expertise, we propose acquisition and analysis protocols at the radius, tibia, and femur sites using Stratec XCT 2000/3000 pQCT scanners among patients with neurological impairment for clinical and research purposes in order to drive practice change, develop normative datasets and complete future meta-analysis to inform fracture risk and treatment efficacy evaluation.

Significant morphological change in osteoarthritic hips identified over 6-12 months using statistical shape modelling

OBJECTIVE

Predicting who will develop osteoarthritis, assessing how rapidly their disease will progress and monitoring early responses to treatment are key to the development of therapeutic agents able to treat this crippling disease and to their future clinical use. Statistical Shape Modelling (SSM) enables quantification of variations in multiple geometric measures describing the whole hip joint to be considered in concert. This prospective study evaluates the responsiveness of SSM to changes in hip-shape within 1 year.

METHODS

Sixty-two people, mean age 67.1 yrs, were recruited. Dual-energy X-ray Absorptiometry images were taken at three timepoints (baseline, 6 and 12 months). Based on Kellgren-Lawrence grading (KLG) of their baseline images, subjects were classified into control/doubtful OA: KLG < 1 in both hips; moderate OA: KLG = 2; and severe OA: KLG ≥ 3 in their most severe hip. Morphology was quantified using SSM and changes in shape were assessed using generalised estimating equations. Standardized response means (SRMs) were calculated for the first and second 6 month periods, then the full 12 months.

RESULTS

Disease severity ranged from KLG0-KLG4 in the 124 hips assessed at baseline. Three SSM modes (Modes 1, 3 and 4) were associated with OA severity. Across the whole cohort, SRM magnitudes ranged from 0.16 to 0.63. The greatest subgroup SRM (magnitude 0.91) was observed over 12 months in those subjects with moderate OA (KLG2).

CONCLUSIONS

We have demonstrated that SSM can capture changes in hip shape over 6 and 12 months across the entire hip joint providing a sensitive measure of hip OA progression.

Statistical shape modelling of hip and lumbar spine morphology and their relationship in the MRC National Survey of Health and Development

The anatomical shape of bones and joints is important for their proper function but quantifying this, and detecting pathological variations, is difficult to do. Numerical descriptions would also enable correlations between joint shapes to be explored. Statistical shape modelling (SSM) is a method of image analysis employing pattern recognition statistics to describe and quantify such shapes from images; it uses principal components analysis to generate modes of variation describing each image in terms of a set of numerical scores after removing global size variation. We used SSM to quantify the shapes of the hip and the lumbar spine in dual-energy x-ray absorptiometry (DXA) images from 1511 individuals in the MRC National Survey of Health and Development at ages 60-64 years. We compared shapes of both joints in men and women and hypothesised that hip and spine shape would be strongly correlated. We also investigated associations with height, weight, body mass index (BMI) and local (hip or lumber spine) bone mineral density. In the hip, all except one of the first 10 modes differed between men and women. Men had a wider femoral neck, smaller neck-shaft angle, increased presence of osteophytes and a loss of the femoral head/neck curvature compared with women. Women presented with a flattening of the femoral head and greater acetabular coverage of the femoral head. Greater weight was associated with a shorter, wider femoral neck and larger greater and lesser trochanters. Taller height was accompanied by a flattening of the curve between superior head and neck and a larger lesser trochanter. Four of the first eight modes describing lumbar spine shape differed between men and women. Women tended to have a more lordotic spine than men with relatively smaller but caudally increasing anterior-posterior (a-p) vertebral diameters. Men were more likely to have a straighter spine with larger vertebral a-p diameters relative to vertebral height than women, increasing cranially. A weak correlation was found between body weight and a-p vertebral diameter. No correlations were found between shape modes and height in men, whereas in women there was a weak positive correlation between height and evenness of spinal curvature. Linear relationships between hip and spine shapes were weak and inconsistent in both sexes, thereby offering little support for our hypothesis. In conclusion, men and women entering their seventh decade have small but statistically significant differences in the shapes of their hips and their spines. Associations with height, weight, BMI and BMD are small and correspond to subtle variations whose anatomical significance is not yet clear. Correlations between hip and spine shapes are small.

Patient-specific bone mineral density distribution in the tibia of individuals with chronic spinal cord injury, derived from multi-slice peripheral Quantitative Computed Tomography (pQCT) - A cross-sectional study

BACKGROUND

The high risk of fracture associated with chronic spinal cord injury (SCI) is attributed to extensive disuse-related bone loss in previously weight-bearing long bones. Changes in bone mineral density (BMD) after SCI have been documented extensively for the epiphyses of the tibia and femur, fracture-prone sites in this patient group. Less attention has been given to patterns of cortical bone loss in the diaphyses, but variability in BMD distributions throughout the long bones may contribute to some patients' increased susceptibility to shaft fractures in chronic SCI.

AIM

A cross-sectional study was carried out to determine whether BMD distributions along the tibia differ between individuals with chronic SCI and healthy able-bodied (AB) controls, in both the trabecular and cortical bone compartments. The effects of time post-injury and gender on BMD distribution were also explored.

METHODS

Individuals with chronic (≥6months post-injury) motor-complete SCI were recruited from the Queen Elizabeth National Spinal Injuries Unit (Glasgow, UK). AB control subjects were recruited to achieve similar age and gender profiles for the SCI and control groups. Multi-slice pQCT (XCT3000, Stratec) was performed along the length of the tibia (2mm thickness, 0.5mm voxel size), at 1% intervals in the epiphyses and 5% intervals in the diaphysis (34 slices in total). These were used to reconstruct full 3-D subject-specific models (Mimics, Materialise) of BMD distribution, by interpolating between slices. Subjects with chronic SCI were subdivided into 'early' (<4years post-injury) and 'established' SCI (≥4years post-injury). Subject-specific BMD distribution was described according to new parameters determined from the 3-D patient-specific models, quantifying descriptors of the trabecular and cortical BMD regions separately (volume, peak BMD, half-peak width, area under the curve). These were compared between sub-groups (using independent-samples t-tests or Mann-Whitney tests, significance level of 5%).

RESULTS

11 men (age range 17-59years old; mean 35.7±10.6) and 3 post-menopausal women (age range 56-58years old; mean 56.7±1.2years) with motor-complete SCI (ranging from 6months to 27years post-injury) were recruited; 6 men (age range 20-56years old; 33.0±12.7years) and 1 post-menopausal woman (56years) formed the AB control group. Overall, SCI resulted in lower BMD at both trabecular and cortical regions of the tibia. In men, longer time since injury resulted in greater BMD differences when compared to AB, throughout the tibia. For the post-menopausal women, differences in BMD between SCI and AB were greater in cortical bone than in trabecular bone. From the models, individual BMD distribution curves showed healthy double-peaks in AB subjects: one trabecular peak (around 200-300mg/cm³) and the other cortical (around 1000-1100mg/cm³). In most subjects with established SCI, trabecular peaks were exaggerated whilst the cortical peaks were barely discernible, with crucially some individuals already exhibiting a diminishing cortical BMD peak even <4years post-injury.

CONCLUSIONS

These findings may have implications for determining the fracture susceptibility of the long bones in individual patients with SCI. Epiphyseal fractures associated with low trabecular BMD are well characterised, but our data show that some individuals with SCI may also be at higher risk of shaft fractures. The proposed BMD distribution description parameters, determined from patient-specific models, could be used to identify patients with a weakened diaphysis who may be susceptible to fractures of the tibial shaft, but this requires validation.

Analysis of the three-dimensional anatomical variance of the distal radius using 3D shape models

BACKGROUND

Various medical fields rely on detailed anatomical knowledge of the distal radius. Current studies are limited to two-dimensional analysis and biased by varying measurement locations. The aims were to 1) generate 3D shape models of the distal radius and investigate variations in the 3D shape, 2) generate and assess morphometrics in standardized cut planes, and 3) test the model's classification accuracy.

METHODS

The local radiographic database was screened for CT-scans of intact radii. 1) The data sets were segmented and 3D surface models generated. Statistical 3D shape models were computed (overall, gender and side separate) and the 3D shape variation assessed by evaluating the number of modes. 2) Anatomical landmarks were assigned and used to define three standardized cross-sectional cut planes perpendicular to the main axis. Cut planes were generated for the mean shape models and each individual radius. For each cut plane, the following morphometric parameters were calculated and compared: maximum width and depth, perimeter and area. 3) The overall shape model was utilized to evaluate the predictive value (leave one out cross validation) for gender and side identification within the study population.

RESULTS

Eighty-six radii (45 left, 44% female, 40 ± 18 years) were included. 1) Overall, side and gender specific statistical 3D models were successfully generated. The first mode explained 37% of the overall variance. Left radii had a higher shape variance (number of modes: 20 female / 23 male) compared to right radii (number of modes: 6 female / 6 male). 2) Standardized cut planes could be defined using anatomical landmarks. All morphometric parameters decreased from distal to proximal. Male radii were larger than female radii with no significant side difference. 3) The overall shape model had a combined median classification probability for side and gender of 80%.

CONCLUSIONS

Statistical 3D shape models of the distal radius can be generated using clinical CT-data sets. These models can be used to assess overall bone variance, define and analyze standardized cut-planes, and identify the gender of an unknown sample. These data highlight the potential of shape models to assess the 3D anatomy and anatomical variance of human bones.

Evidence of the Role of R-Spondin 1 and Its Receptor Lgr4 in the Transmission of Mechanical Stimuli to Biological Signals for Bone Formation

The bone can adjust its mass and architecture to mechanical stimuli via a series of molecular cascades, which have been not yet fully elucidated. Emerging evidence indicated that R-spondins (Rspos), a family of secreted agonists of the Wnt/β-catenin signaling pathway, had important roles in osteoblastic differentiation and bone formation. However, the role of Rspo proteins in mechanical loading-influenced bone metabolism has never been investigated. In this study, we found that Rspo1 was a mechanosensitive protein for bone formation. Continuous cyclic mechanical stretch (CMS) upregulated the expression of Rspo1 in mouse bone marrow mesenchymal stem cells (BMSCs), while the expression of Rspo1 in BMSCs in vivo was downregulated in the bones of a mechanical unloading mouse model (tail suspension (TS)). On the other hand, Rspo1 could promote osteogenesis of BMSCs under CMS through activating the Wnt/β-catenin signaling pathway and could rescue the bone loss induced by mechanical unloading in the TS mice. Specifically, our results suggested that Rspo1 and its receptor of leucine-rich repeat containing G-protein-coupled receptor 4 (Lgr4) should be a novel molecular signal in the transmission of mechanical stimuli to biological signal in the bone, and this signal should be in the upstream of Wnt/β-catenin signaling for bone formation. Rspo1/Lgr4 could be a new potential target for the prevention and treatment of disuse osteoporosis in the future.

Treatment with hydrogen sulfide attenuates sublesional skeletal deterioration following motor complete spinal cord injury in rats

Treatment with hydrogen sulfide mitigates spinal cord injury-induced sublesional bone loss, possibly through abating oxidative stress, suppressing MMP activity, and activating Wnt/β-catenin signaling.

INTRODUCTION

Spinal cord injury (SCI)-induced sublesional bone loss represents the most severe osteoporosis and is resistant to available treatments to data. The present study was undertaken to explore the therapeutic potential of hydrogen sulfide (H₂S) against osteoporosis in a rodent model of motor complete SCI.

METHODS

SCI was generated by surgical transaction of the cord at the T3-T4 levels in rats. Treatment with NaHS was initiated through intraperitoneal injection of 0.1 ml/kg/day of 0.28 mol/l NaHS from 12 h following the surgery and over 14 subsequent days.

RESULTS

H₂S levels in plasma of SCI rats were lower, which was restored by treatment with exogenous H₂S. Treatment of SCI rats with exogenous H₂S had no significant effect on body mass but increased bone mineral density in femurs and tibiae, increased BV/TV, Tb.Th, and Tb.N and reduced Tb.Sp in proximal tibiae, and increased mineral apposition rate (MAR), bone formation rate (BFR), and osteoblast surface and reduced eroded surface and osteoclast surface in proximal tibiae. More importantly, H₂S treatment led to a significant enhancement in ultimate load, stiffness, and energy to max force of femoral diaphysis. Treatment of SCI rats with exogenous H₂S reduced malondialdehyde (MDA) levels in serum and femurs, decreased hydroxyproline levels, suppressed activities of matrix metallopeptidase 9 (MMP9), and upregulated Wnt3a, Wnt6, Wnt10, and ctnnb1 expression in femurs.

CONCLUSION

Treatment with H₂S mitigates SCI-induced sublesional bone loss, possibly through abating oxidative stress, suppressing MMP activity, and activating Wnt/β-catenin signaling.

Measuring muscle and bone in individuals with neurologic impairment; lessons learned about participant selection and pQCT scan acquisition and analysis

Peripheral quantitative computed tomography (pQCT) can be used to examine bone strength outcomes and muscle size and fatty infiltration. Our research team and others have used it to examine bone loss after spinal cord injury (SCI). However, the high prevalence of restricted lower extremity range of motion, spasticity, edema, excessive muscle atrophy, or severe osteoporosis necessitates changes to standard protocols for screening, positioning during scan acquisition, and analysis methods. This manuscript outlines the challenges that we experienced using pQCT in individuals with SCI, and provides solutions, ones that may also be applicable when using pQCT in individuals with other chronic conditions or in older adults. Suggestions for participant screening, positioning individuals for scanning while in a wheelchair, scan site selection, need for attendant assistance, and considerations in the presence of secondary complications, such as contracture, spasticity, and paralysis, are presented. In the presence of very low bone mineral density or severe muscle atrophy, the default analysis modes provided by the manufacturer may not provide valid estimates of bone or muscle indices; we propose alternates. We have used watershed segmentation methods to determine muscle size and density based on lower precision error compared to threshold-based edge-detection segmentation, particularly for adults with SCI, where more fatty infiltration was present. By presenting our "lessons learned," we hope to reduce the learning curve for researchers using pQCT in the future.