Statistical Parametric Mapping of HR-pQCT Images: A Tool for Population-Based Local Comparisons of Micro-Scale Bone Features

Cortical thickness mapping at segmented regions in the distal radius using HR-pQCT

INTRODUCTION

An advanced method of analyzing the cortical bone microarchitecture of the distal radius using high-resolution peripheral quantitative computed tomography (HR-pQCT) was developed.

MATERIALS AND METHODS

The subjects were 60 women (20: aged 30-49, 20: aged 50-69, and 20: aged 70-89 years). The distal radius was scanned by HR-pQCT, and its cortical volumetric bone mineral density (Ct.vBMD), cortical porosity (Ct.Po), and cortical thickness (Ct.Th) were measured. The cortical bone was also divided into three areas according to whether its thickness was < 0.5 mm, 0.5-1.0 mm, or > 1.0 mm, and the percentage of each surface area in the total surface area of cortical bone was calculated (Ct.Th (<0.5), Ct.Th (0.5-1.0), Ct.Th (>1.0), respectively). The cortical bone at the distal radius was further segmented into dorsal, palmar, radial, and ulnar sides, and the above-described parameters were measured in these regions.

RESULTS

Integral analysis showed that Ct.vBMD and Ct.Th decreased and Ct.Po increased with age (R = - 0.62, - 0.55, and 0.54). Ct.Th (< 0.5) expanded with age (R = 0.49), with the rate of change between those aged 30-49 years and those aged 50-69 years being 106.7%. On regional analysis, the expansion of Ct.Th (< 0.5) with age was particularly marked on the dorsal and palmar side (R = 0.51 and 0.49), where the rate of change between those aged 30-49 years and those aged 50-69 years was the highest, at 196.1 and 149.6%.

CONCLUSION

The method to identify areas of cortical bone thinning in the segmented regions of the dorsal, palmar, radial, and ulnar sides of the distal radius using HR-pQCT may offer a sensitive assessment of age-related deterioration of cortical bone.

Augmenting Osteoporosis Imaging with Machine Learning

PURPOSE OF REVIEW

In this paper, we discuss how recent advancements in image processing and machine learning (ML) are shaping a new and exciting era for the osteoporosis imaging field. With this paper, we want to give the reader a basic exposure to the ML concepts that are necessary to build effective solutions for image processing and interpretation, while presenting an overview of the state of the art in the application of machine learning techniques for the assessment of bone structure, osteoporosis diagnosis, fracture detection, and risk prediction.

RECENT FINDINGS

ML effort in the osteoporosis imaging field is largely characterized by "low-cost" bone quality estimation and osteoporosis diagnosis, fracture detection, and risk prediction, but also automatized and standardized large-scale data analysis and data-driven imaging biomarker discovery. Our effort is not intended to be a systematic review, but an opportunity to review key studies in the recent osteoporosis imaging research landscape with the ultimate goal of discussing specific design choices, giving the reader pointers to possible solutions of regression, segmentation, and classification tasks as well as discussing common mistakes.

Global and Spatial Compartmental Interrelationships of Bone Density, Microstructure, Geometry and Biomechanics in the Distal Radius in a Colles' Fracture Study Using HR-pQCT

BACKGROUND

Bone parameters derived from HR-pQCT have been investigated on a parameter-by-parameter basis for different clinical conditions. However, little is known regarding the interrelationships of bone parameters and the spatial distribution of these interrelationships. In this work: 1) we investigate compartmental interrelationships of bone parameters; 2) assess the spatial distribution of interrelationships of bone parameters; and 3) compare interrelationships of bone parameters between postmenopausal women with and without a recent Colles' fracture.

METHODS

Images from the unaffected radius in fracture cases (n=84), and from the non-dominant radius of controls (n=98) were obtained using HR-pQCT. Trabecular voxel-based maps of local bone volume fraction (L.Tb.BV/TV), homogenized volumetric bone mineral density (H.Tb.BMD), homogenized μFEA-derived strain energy density (H.Tb.SED), and homogenized inter-trabecular distances (H.Tb.1/N) were generated; as well as surface-based maps of apparent cortical bone thickness (Surf.app.Ct.Th), porosity-weighted cortical bone thickness (Surf.Ct.SIT), mean cortical BMD (Surf.Ct.BMD), and mean cortical SED (Surf.Ct.SED). Anatomical correspondences across the parametric maps in the study were established via spatial normalization to a common template. Mean values of the parametric maps before spatial normalization were used to assess compartmental Spearman's rank partial correlations of bone parameters (e.g., between H.Tb.BMD and L.Tb.BV/TV or between Surf.Ct.BMD and Surf.app.Ct.Th). Spearman's rank partial correlations were also assessed for each voxel and vertex of the spatially normalized parametric maps, thus generating maps of Spearman's rank partial correlation coefficients. Correlations were performed independently within each group, and compared between groups using the Fisher's Z transformation.

RESULTS

All within-group global trabecular and cortical Spearman's rank partial correlations were significant; and the correlations of H.Tb.BMD-L.Tb.BV/TV, H.Tb.BMD-H.Tb.1/N, L.Tb.BV/TV-H.Tb.1/N, Surf.Ct.BMD-Surf.Ct.SED and Surf.Ct.SIT-Surf.Ct.SED were significantly different between controls and fracture cases. The spatial analyses revealed significant heterogeneous voxel- and surface-based correlation coefficient maps across the distal radius for both groups; and the correlation maps of H.Tb.BMD-L.Tb.BV/TV, H.Tb.BMD-H.Tb.1/N, L.Tb.BV/TV-H.Tb.1/N, H.Tb.1/N-H.Tb.SED and Surf.app.Ct.Th - Surf.Ct.SIT yielded small clusters of significant correlation differences between groups.

DISCUSSION

The heterogeneous spatial distribution of interrelationships of bone parameters assessing density, microstructure, geometry and biomechanics, along with their global and local differences between controls and fracture cases, may help us further understand different bone mechanisms of bone fracture.

Automated MRI-Based Deep Learning Model for Detection of Alzheimer’s Disease Process

In the context of neuro-pathological disorders, neuroimaging has been widely accepted as a clinical tool for diagnosing patients with Alzheimer’s disease (AD) and mild cognitive impairment (MCI). The advanced deep learning method, a novel brain imaging technique, was applied in this study to evaluate its contribution to improving the diagnostic accuracy of AD. Three-dimensional convolutional neural networks (3D-CNNs) were applied with magnetic resonance imaging (MRI) to execute binary and ternary disease classification models. The dataset from the Alzheimer’s disease neuroimaging initiative (ADNI) was used to compare the deep learning performances across 3D-CNN, 3D-CNN-support vector machine (SVM) and two-dimensional (2D)-CNN models. The outcomes of accuracy with ternary classification for 2D-CNN, 3D-CNN and 3D-CNN-SVM were [Formula: see text]%, [Formula: see text]% and [Formula: see text]% respectively. The 3D-CNN-SVM yielded a ternary classification accuracy of 93.71%, 96.82% and 96.73% for NC, MCI and AD diagnoses, respectively. Furthermore, 3D-CNN-SVM showed the best performance for binary classification. Our study indicated that ‘NC versus MCI’ showed accuracy, sensitivity and specificity of 98.90%, 98.90% and 98.80%; ‘NC versus AD’ showed accuracy, sensitivity and specificity of 99.10%, 99.80% and 98.40%; and ‘MCI versus AD’ showed accuracy, sensitivity and specificity of 89.40%, 86.70% and 84.00%, respectively. This study clearly demonstrates that 3D-CNN-SVM yields better performance with MRI compared to currently utilized deep learning methods. In addition, 3D-CNN-SVM proved to be efficient without having to manually perform any prior feature extraction and is totally independent of the variability of imaging protocols and scanners. This suggests that it can potentially be exploited by untrained operators and extended to virtual patient imaging data. Furthermore, owing to the safety, noninvasiveness and nonirradiative properties of the MRI modality, 3D-CNN-SMV may serve as an effective screening option for AD in the general population. This study holds value in distinguishing AD and MCI subjects from normal controls and to improve value-based care of patients in clinical practice.

Update on Imaging-Based Measurement of Bone Mineral Density and Quality

PURPOSE OF REVIEW

Patients with inflammatory arthropathies have a high rate of fragility fractures. Diagnostic assessment and monitoring of bone density and quality are therefore critically important. Here, we review standard and advanced techniques to measure bone density and quality, specifically focusing on patients with inflammatory arthropathies.

RECENT FINDINGS

Current standard procedures are dual-energy X-ray absorptiometry (DXA) and quantitative computed tomography (QCT). DXA-based newer methods include trabecular bone score (TBS) and vertebral fracture assessment (VFA). More advanced imaging methods to measure bone quality include high-resolution peripheral quantitative computed tomography (HR-pQCT) as well as multi-detector CT (MD-CT) and magnetic resonance imaging (MRI). Quantitative ultrasound has shown promise but is not standard to assess bone fragility. While there are limitations, DXA remains the standard technique to measure density in patients with rheumatological disorders. Newer modalities to measure bone quality may allow better characterization of bone fragility but currently are not standard of care procedures.

Development of a simple numerical model for trabecular bone structures

PURPOSE

Advances in additive manufacturing processes are enabling the fabrication of surrogate bone structures for applications including use in high-resolution anthropomorphic phantoms. In this research, a simple numerical model is proposed that enables the generation of microarchitecture with similar statistical distribution to trabecular bone.

METHODS

A human humerus, radius, ulna, and several vertebrae were scanned on the Imaging and Medical beamline at the Australian Synchrotron and the proposed numerical model was developed through the definition of two complex functions that encode the trabecular thickness and position-dependant spacing to generate volumetric surrogate trabecular structures. The structures reproduced those observed at 19 separate axial locations through the experimental bone volumes. The applicability of the model when incorporating a two-material approximation to absorption- and phase-contrast CT was also investigated through simulation.

RESULTS

The synthetic structures, when compared with the real trabecular microarchitecture, yielded an average mean thickness error of 2 μm, and a mean difference in standard deviation of 33 μm for the humerus, 24 μm for the ulna and radius, and 15 μm for the vertebrae. Simulated absorption- and propagation-based phase contrast CT projection data were generated and reconstructed using the derived mathematical simplifications from the two-material approximation, and the phase-contrast effects were successfully demonstrated.

CONCLUSIONS

The presented model reproduced trabecular distributions that could be used to generate phantoms for quality assurance and validation processes. The implication of utilizing a two-material approximation results in simplification of the additive manufacturing process and the generation of synthetic data that could be used for training of machine learning applications.

Cortical Bone Mapping: Measurement and Statistical Analysis of Localised Skeletal Changes

PURPOSE OF REVIEW

Cortical bone mapping (CBM) is a technique for measuring localised skeletal changes from computed tomography (CT) images. It can provide measurements with accuracy surpassing the underlying imaging resolution. CBM can detect changes in several properties of the cortex, with no prior assumptions about the likely location of said changes. This paper summarises the theory behind CBM, discusses its strengths and limitations, and reviews some studies in which it has been applied.

RECENT FINDINGS

CBM has revealed associations between fracture risk and cortical properties in specific regions of the proximal femur which present feasible therapeutic targets. Analyses of several pharmaceutical and exercise interventions quantify effects that are distinct both in location and in the nature of the micro-architectural changes. CBM has illuminated age-related changes in the proximal femur and has recently been applied to other bones, as well as to the assessment of cartilage. The CBM processing pipeline is designed primarily for large cohort studies. Its main impact thus far has not been in the realm of clinical practice, but rather to improve our fundamental understanding of localised bone structure and changes.

Spatial Differences in the Distribution of Bone Between Femoral Neck and Trochanteric Fractures

There is little knowledge about the spatial distribution differences in volumetric bone mineral density and cortical bone structure at the proximal femur between femoral neck fractures and trochanteric fractures. In this case-control study, a total of 93 women with fragility hip fractures, 72 with femoral neck fractures (mean ± SD age: 70.6 ± 12.7 years) and 21 with trochanteric fractures (75.6 ± 9.3 years), and 50 control subjects (63.7 ± 7.0 years) were included for the comparisons. Differences in the spatial distributions of volumetric bone mineral density, cortical bone thickness, cortical volumetric bone mineral density, and volumetric bone mineral density in a layer adjacent to the endosteal surface were investigated using voxel-based morphometry (VBM) and surface-based statistical parametric mapping (SPM). We compared these spatial distributions between controls and both types of fracture, and between the two types of fracture. Using VBM, we found spatially heterogeneous volumetric bone mineral density differences between control subjects and subjects with hip fracture that varied by fracture type. Interestingly, femoral neck fracture subjects, but not subjects with trochanteric fracture, showed significantly lower volumetric bone mineral density in the superior aspect of the femoral neck compared with controls. Using surface-based SPM, we found that compared with controls, both fracture types showed thinner cortices in regions in agreement with the type of fracture. Most outcomes of cortical and endocortical volumetric bone mineral density comparisons were consistent with VBM results. Our results suggest: 1) that the spatial distribution of trabecular volumetric bone mineral density might play a significant role in hip fracture; 2) that focal cortical bone thinning might be more relevant in femoral neck fractures; and 3) that areas of reduced cortical and endocortical volumetric bone mineral density might be more relevant for trochanteric fractures in Chinese women. © 2017 American Society for Bone and Mineral Research.