High-Resolution Peripheral Quantitative Computed Tomography in Rheumatic Diseases: A New Option for Knee Osteoarthritis.

High-resolution peripheral quantitative computed tomography (HR-pQCT) provides a new opportunity for the noninvasive measurement of bone microarchitecture in patient research of rheumatic disease in joints of the extremities. Rheumatoid arthritis of the hand and osteoarthritis of the knee are highly prevalent and there are no cures, so an improved understanding of the etiology of these diseases, particularly when combining HR-pQCT with other imaging modalities, is important to develop new disease management strategies. Analysis of HR-pQCT data is challenging due to the large size of the datasets, but the field continues to rapidly develop, and there is excellent opportunity to integrate sophisticated analyses for this novel research tool.

Statistical investigation of interdependence of trabecular microarchitectural parameters from micro-computed tomography.

Standard microarchitectural analysis of bone using micro-computed tomography produces a large number of parameters that quantify the structure of the trabecular network. Analyses that perform statistical tests on many parameters are at elevated risk of making Type I errors. However, when multiple testing correction procedures are applied, the risk of Type II errors is elevated if the parameters being tested are strongly correlated. In this article, we argue that four commonly used trabecular microarchitectural parameters (thickness, separation, number, and bone volume fraction) are interdependent and describe only two independent properties of the trabecular network. We first derive theoretical relationships between the parameters based on their geometric definitions. Then, we analyze these relationships with an aggregated in vivo dataset with 2987 images from 1434 participants and a synthetically generated dataset with 144 images using principal component analysis (PCA) and linear regression analysis. With PCA, when trabecular thickness, separation, number, and bone volume fraction are combined, we find that 92 % to 97 % of the total variance in the data is explained by the first two principal components. With linear regressions, we find high coefficients of determination (0.827-0.994) and fitted coefficients within expected ranges. These findings suggest that to maximize statistical power in future studies, only two of trabecular thickness, separation, number and bone volume fraction should be used for statistical testing.

Deep learning-based automated detection and segmentation of bone and traumatic bone marrow lesions from MRI following an acute ACL tear.

INTRODUCTION: Traumatic bone marrow lesions (BML) are frequently identified on knee MRI scans in patients following an acute full-thickness, complete ACL tear. BMLs coincide with regions of elevated localized bone loss, and studies suggest these may act as a precursor to the development of post-traumatic osteoarthritis. This study addresses the labour-intensive manual assessment of BMLs by using a 3D U-Net for automated identification and segmentation from MRI scans. METHODS: A multi-task learning approach was used to segment both bone and BML from T2 fat-suppressed (FS) fast spin echo (FSE) MRI sequences for BML assessment. Training and testing utilized datasets from individuals with complete ACL tears, employing a five-fold cross-validation approach and pre-processing involved image intensity normalization and data augmentation. A post-processing algorithm was developed to improve segmentation and remove outliers. Training and testing datasets were acquired from different studies with similar imaging protocol to assess the model's performance robustness across different populations and acquisition conditions. RESULTS: The 3D U-Net model exhibited effectiveness in semantic segmentation, while post-processing enhanced segmentation accuracy and precision through morphological operations. The trained model with post-processing achieved a Dice similarity coefficient (DSC) of 0.75 ± 0.08 (mean ± std) and a precision of 0.87 ± 0.07 for BML segmentation on testing data. Additionally, the trained model with post-processing achieved a DSC of 0.93 ± 0.02 and a precision of 0.92 ± 0.02 for bone segmentation on testing data. This demonstrates the approach's high accuracy for capturing true positives and effectively minimizing false positives in the identification and segmentation of bone structures. CONCLUSION: Automated segmentation methods are a valuable tool for clinicians and researchers, streamlining the assessment of BMLs and allowing for longitudinal assessments. This study presents a model with promising clinical efficacy and provides a quantitative approach for bone-related pathology research and diagnostics.

Accuracy of Volumetric Bone Mineral Density Measurement in Weight Bearing, Cone Beam Computed Tomography.

BACKGROUND: Weight bearing computed tomography (WBCT) utilizes cone beam CT technology to provide assessments of lower limb joint structures while they are functionally loaded. Grey-scale values indicative of X-ray attenuation that are output from cone beam CT are challenging to calibrate, and their use for bone mineral density (BMD) measurement remains debatable. To determine whether WBCT can be reliably used for cortical and trabecular BMD assessment, we sought to establish the accuracy of BMD measurements at the knee using modern WBCT by comparing them to measurements from conventional CT. METHODS: A hydroxyapatite phantom with three inserts of varying densities was used to systematically quantify signal uniformity and BMD accuracy across the acquisition volume. We evaluated BMD in vivo (n = 5, female) using synchronous and asynchronous calibration techniques in WBCT and CT. To account for variation in attenuation along the height (z-axis) of acquisition volumes, we tested a height-dependent calibration approach for both WBCT and CT images. RESULTS: Phantom BMD measurement error in WBCT was as high as 15.3% and consistently larger than CT (up to 5.6%). Phantom BMD measures made under synchronous conditions in WBCT improved measurement accuracy by up to 3% but introduced more variability in measured BMD. We found strong correlations (R = 0.96) as well as wide limits of agreement (-324 mgHA/cm3 to 183 mgHA/cm3) from Bland-Altman analysis between WBCT and CT measures in vivo that were not improved by height-dependent calibration. CONCLUSION: Whilst BMD accuracy from WBCT was found to be dependent on apparent density, accuracy was independent of the calibration technique (synchronous or asynchronous) and the location of the measurement site within the field of view. Overall, we found strong correlations between BMD measures from WBCT and CT and in vivo measures to be more accurate in trabecular bone regions. Importantly, WBCT can be used to distinguish between anatomically relevant differences in BMD, however future work is necessary to determine the repeatability and sensitivity of BMD measures in WBCT.

Rapid bone microarchitecture decline in older men with high bone turnover-the prospective STRAMBO study.

Older men with high bone turnover have faster bone loss. We assessed the link between the baseline levels of bone turnover markers (BTMs) and the prospectively assessed bone microarchitecture decline in men. In 825 men aged 60-87 yr, we measured the serum osteocalcin (OC), bone alkaline phosphatase (BAP), N-terminal propeptide of type I procollagen (PINP), and C-terminal telopeptide of type I collagen (CTX-I), and urinary total deoxypyridinoline (tDPD). Bone microarchitecture and strength (distal radius and distal tibia) were estimated by high-resolution pQCT (XtremeCT, Scanco Medical) at baseline and then after 4 and 8 yr. Thirty-seven men took medications affecting bone metabolism. Statistical models were adjusted for age and BMI. At the distal radius, the decrease in the total bone mineral density (Tt.BMD), cortical BMD (Ct.BMD), cortical thickness (Ct.Thd), and cortical area (Ct.Ar) and failure load was faster in the highest vs the lowest CTX-I quartile (failure load: -0.94 vs -0.31% yr-1, P < .001). Patterns were similar for distal tibia. At the distal tibia, bone decline (Tt.BMD, Ct.Thd, Ct.Ar, Ct.BMD, and failure load) was faster in the highest vs the lowest tDPD quartile. At each skeletal site, the rate of decrease in Tb.BMD differed between the extreme OC quartiles (P < .001). Men in the highest BAP quartile had a faster loss of Tt.BMD, Tb.BMD, reaction force, and failure load vs the lowest quartile. The link between PINP and bone decline was poor. The BTM score is the sum of the nos. of the quartiles for each BTM. Men in the highest quartile of the score had a faster loss of cortical bone and bone strength vs the lowest quartile. Thus, in the older men followed prospectively for 8 yr, the rate of decline in bone microarchitecture and estimated bone strength was 50%-215% greater in men with high bone turnover (highest quartile, CTX-I above the median) compared to the men with low bone turnover (lowest quartile, CTX-I below the median).

Fracture risk based on high-resolution peripheral quantitative computed tomography measures does not vary with age in older adults-the bone microarchitecture international consortium prospective cohort study.

Fracture risk increases with lower areal bone mineral density (aBMD); however, aBMD-related estimate of risk may decrease with age. This may depend on technical limitations of 2-dimensional (2D) dual energy X-ray absorptiometry (DXA) which are reduced with 3D high-resolution peripheral quantitative computed tomography (HR-pQCT). Our aim was to examine whether the predictive utility of HR-pQCT measures with fracture varies with age. We analyzed associations of HR-pQCT measures at the distal radius and distal tibia with two outcomes: incident fractures and major osteoporotic fractures. We censored follow-up time at first fracture, death, last contact or 8 years after baseline. We estimated hazard ratios (HR) and 95%CI for the association between bone traits and fracture incidence across age quintiles. Among 6835 men and women (ages 40-96) with at least one valid baseline HR-pQCT scan who were followed prospectively for a median of 48.3 months, 681 sustained fractures. After adjustment for confounders, bone parameters at both the radius and tibia were associated with higher fracture risk. The estimated HRs for fracture did not vary significantly across age quintiles for any HR-pQCT parameter measured at either the radius or tibia. In this large cohort, the homogeneity of the associations between the HR-pQCT measures and fracture risk across age groups persisted for all fractures and for major osteoporotic fractures. The patterns were similar regardless of the HR-pQCT measure, the type of fracture, or the statistical models. The stability of the associations between HR-pQCT measures and fracture over a broad age range shows that bone deficits or low volumetric density remain major determinants of fracture risk regardless of age group. The lower risk for fractures across measures of aBMD in older adults in other studies may be related to factors which interfere with DXA but not with HR-pQCT measures.

Transferability of bone phenotyping and fracture risk assessment by µFRAC from first-generation high-resolution peripheral quantitative computed tomography to second-generation scan data.

INTRODUCTION: The continued development of high-resolution peripheral quantitative computed tomography (HR-pQCT) has led to a second-generation scanner with higher resolution and longer scan region. However, large multicenter prospective cohorts were collected with first-generation HR-pQCT and have been used to develop bone phenotyping and fracture risk prediction (µFRAC) models. This study establishes whether there is sufficient universality of these first-generation trained models for use with second-generation scan data. METHODS: HR-pQCT data were collected for a cohort of 60 individuals, who had been scanned on both first- and second-generation scanners on the same day to establish the universality of the HR-pQCT models. These data were each used as input to first-generation trained bone microarchitecture models for bone phenotyping and fracture risk prediction, and their outputs were compared for each study participant. Reproducibility of the models were assessed using same-day repeat scans obtained from first-generation (n = 37) and second-generation (n = 74) scanners. RESULTS: Across scanner generations, the bone phenotyping model performed with an accuracy of 93.1%. Similarly, the 5-year fracture risk assessment by µFRAC was well correlated with a Pearson's (r) correlation coefficient of r > 0.83 for the three variations of µFRAC (varying inclusion of clinical risk factors, finite element analysis, and dual X-ray absorptiometry). The first-generation reproducibility cohort performed with an accuracy for categorical assignment of 100% (bone phenotyping) and a correlation coefficient of 0.99 (µFRAC), whereas the second-generation reproducibility cohort performed with an accuracy of 96.4% (bone phenotyping) and a correlation coefficient of 0.99 (µFRAC). CONCLUSION: We demonstrated that bone microarchitecture models trained using first-generation scan data generalize well to second-generation scans, performing with a high level of accuracy and reproducibility. Less than 4% of individuals' estimated fracture risk led to a change in treatment threshold, and in general, these dissimilar outcomes using second-generation data tended to be more conservative.

Establishing the universality of first-generation-trained HR-pQCT prediction models on second-generation scan data is important to move the bone microarchitecture field forward. We found that despite the difference in resolutions between the two HR-pQCT generations, models developed with first-generation data generalized well to second-generation systems. This avoids unnecessarily repeating complex studies.

Establishing error bounds for internal calibration of quantitative computed tomography.

Opportunistic computed tomography (CT) scans, which can assess relevant osteoporotic bones of interest, offer a potential solution for identifying osteoporotic individuals. CT scans usually do not contain calibration phantoms, so internal calibration methods have been developed to create a voxel-specific density calibration that can be used in opportunistic CT. It remains a challenge, however, to account for potential sources of error in internal calibration, such as beam hardening or heterogeneous internal reference tissues. The purpose of this work was to introduce our internal calibration method that accounts for these variations and to estimate error bounds for the bone mineral density (BMD) measurements taken from internally calibrated scans. The error bounds are derived by incorporating a combination of a Monte Carlo simulation and standard error propagation into our previously established internal calibration method. A cohort of 138 clinical abdominal CT scans were calibrated for BMD assessment with a phantom placed in the field of view and used as the ground truth. Our modified internal calibration method provided error bounds on the same images and was tested to contain the ground truth phantom-calibrated BMD. This was repeated using 10 different internal reference tissue combinations to explore how error bounds are affected by the choice of internal tissue referents. We found that the tissue combination of air, skeletal muscle, and cortical bone yielded the most accurate BMD estimates while maintaining error bounds that were sufficiently conservative to account for sources of error such as beam hardening and heterogeneous tissue samples. The mean difference between the phantom BMD and the BMD resulting from the tissue combination of air, skeletal muscle and cortical bone was 2.12 mg/cc (0.06% BMD error) and 1.13 mg/cc (0.02 % BMD error) for the left and right femur, respectively. Providing error bounds for internal calibration provides a method to explore the influence of internal reference tissues and confidence for BMD estimates.

Fixed and Relative Positioning of Scans for High Resolution Peripheral Quantitative Computed Tomography.

INTRODUCTION: High resolution peripheral quantitative computed tomography (HR-pQCT) imaging protocol requires defining where to position the ∼1 cm thick scan along the bone length. Discrepancies between the use of two positioning methods, the relative and fixed offset, may be problematic in the comparison between studies and participants. This study investigated how bone landmarks scale linearly with length and how this scaling affects both positioning methods aimed at providing a consistent anatomical location for scan acquisition. METHODS: Using CT images of the radius (N = 25) and tibia (N = 42), 10 anatomical landmarks were selected along the bone length. The location of these landmarks was converted to a percent length along the bone, and the variation in their location was evaluated across the dataset. The absolute location of the HR-pQCT scan position using both offset methods was identified for all bones and converted to a percent length position relative to the HR-pQCT reference line for comparison. A secondary analysis of the location of the scan region specifically within the metaphysis was explored at the tibia. RESULTS: The location of landmarks deviated from a linear relationship across the dataset, with a range of 3.6 % at the radius sites, and 4.5 % at the tibia sites. The consequent variation of the position of the scan at the radius was 0.6 % and 0.3 %, and at the tibia 2.4 % and 0.5 %, for the fixed and relative offset, respectively. The position of the metaphyseal junction with the epiphysis relative to the scan position was poorly correlated to bone length, with R2 = 0.06 and 0.37, for the fixed and relative offset respectively. CONCLUSION: The variation of the scan position by either method is negated by the intrinsic variation of the bone anatomy with respect both to total bone length as well as the metaphyseal region. Therefore, there is no clear benefit of either offset method. However, the lack of difference due to the inherent variation in the underlying anatomy implies that it is reasonable to compare studies even if they are using different positioning methods.