Hip Fractures in Older Adults Are Associated With the Low Density Bone Phenotype and Heterogeneous Deterioration of Bone Microarchitecture

AI's ongoing impact: Implications of AI's effects on health equity for women's healthcare providers

Objective

To assess the effects of the current use of artificial intelligence (AI) in women's health on health equity, specifically in primary and secondary prevention efforts among women.

Methods

Two databases, Scopus and PubMed, were used to conduct this narrative review. The keywords included "artificial intelligence," "machine learning," "women's health," "screen," "risk factor," and "prevent," and papers were filtered only to include those about AI models that general practitioners may use.

Results

Of the 18 articles reviewed, 8 articles focused on risk factor modeling under primary prevention, and 10 articles focused on screening tools under secondary prevention. Gaps were found in the ability of AI models to train using large, diverse datasets that were reflective of the population it is intended for. Lack of these datasets was frequently identified as a limitation in the papers reviewed (n = 7).

Conclusions

Minority, low-income women have poor access to health care and are, therefore, not well represented in the datasets AI uses to train, which risks introducing bias in its output. To mitigate this, more datasets should be developed to validate AI models, and AI in women's health should expand to include conditions that affect men and women to provide a gendered lens on these conditions. Public health, medical, and technology entities need to collaborate to regulate the development and use of AI in health care at a standard that reduces bias.

Choose your mother wisely: the familial resemblance of bone adaptation

This study demonstrates how complex bone microarchitectural features can be summarized to describe bone adaptations seen with aging in women, which are consistent with the stages of osteoporosis. Additionally, we showed familial resemblance in these bone microarchitectural traits between mothers and daughters that can be used to predict bone adaptations.

INTRODUCTION

Patient-specific characterization of bone quality can reduce complex microarchitectural features to common combinations of bone characteristics, known as bone phenotypes. This study investigated whether there is a general trend in bone phenotype change over time seen with aging in females and whether there is a familial resemblance to phenotype membership between mothers and daughters.

METHODS

Bone phenotype membership was calculated on biological mother and daughter pairs (Participants = 101), scanned using high resolution peripheral quantitative computed tomography, to the three pre-defined phenotypes (healthy, low volume, and low density). The trajectory of bone phenotype with age was explored using all participant's data. Linear regression models were used to assess the familial resemblance of phenotyping in the mother-daughter pairs.

RESULTS

When stratified for age, the trajectory of the phenotype membership transitioned from healthy (20-40 years), to low volume (40-60 years), to low density (60-80 years), which similarly aligns with the stages of osteoporosis observed in females. Familial resemblance (½h2) was observed in the healthy phenotype (β = 0.432, p < 0.01). Predictive modelling showed a significant association in phenotype membership between mothers and daughters in the healthy (R2 = 0.347, p = 0.04) and low volume (R2 = 0.416, p < 0.01) phenotypes, adjusted for age, height, and weight.

CONCLUSION

Our results suggest that phenotype membership in females changes with age in a pattern that is consistent with the stages of osteoporosis. Additionally, we showed familial resemblance in bone phenotype, which can be used to predict bone adaptations between mothers and daughters that are associated with bone loss with aging.

The Fracture Phenotypes in Women and Men of 50 Years and Older with a Recent Clinical Fracture

PURPOSE OF REVIEW

We review the literature about patients 50 years and older with a recent clinical fracture for the presence of skeletal and extra-skeletal risks, their perspectives of imminent subsequent fracture, falls, mortality, and other risks, and on the role of the fracture liaison service (FLS) for timely secondary fracture prevention.

RECENT FINDINGS

Patients with a recent clinical fracture present with heterogeneous patterns of bone-, fall-, and comorbidity-related risks. Short-term perspectives include bone loss, increased risk of fractures, falls, and mortality, and a decrease in physical performance and quality of life. Combined evaluation of bone, fall risk, and the presence of associated comorbidities contributes to treatment strategies. Since fractures are related to interactions of bone-, fall-, and comorbidity-related risks, there is no one-single-discipline-fits-all approach but a need for a multidisciplinary approach at the FLS to consider all phenotypes for evaluation and treatment in an individual patient.

Distinguishing risk of curve progression in adolescent idiopathic scoliosis with bone microarchitecture phenotyping: a 6-year longitudinal study

Low bone mineral density and impaired bone quality have been shown to be important prognostic factors for curve progression in adolescent idiopathic scoliosis (AIS). There is no evidence-based integrative interpretation method to analyze high-resolution peripheral quantitative computed tomography (HR-pQCT) data in AIS. This study aimed to (1) utilize unsupervised machine learning to cluster bone microarchitecture phenotypes on HR-pQCT parameters in girls with AIS, (2) assess the phenotypes' risk of curve progression and progression to surgical threshold at skeletal maturity (primary cohort), and (3) investigate risk of curve progression in a separate cohort of girls with mild AIS whose curve severity did not reach bracing threshold at recruitment (secondary cohort). Patients were followed up prospectively for 6.22 ± 0.33 years in the primary cohort (n = 101). Three bone microarchitecture phenotypes were clustered by fuzzy C-means at time of peripubertal peak height velocity (PHV). Phenotype 1 had normal bone characteristics. Phenotype 2 was characterized by low bone volume and high cortical bone density, and phenotype 3 had low cortical and trabecular bone density and impaired trabecular microarchitecture. The difference in bone quality among the phenotypes was significant at peripubertal PHV and continued to skeletal maturity. Phenotype 3 had significantly increased risk of curve progression to surgical threshold at skeletal maturity (odd ratio [OR] = 4.88; 95% CI, 1.03-28.63). In the secondary cohort (n = 106), both phenotype 2 (adjusted OR = 5.39; 95% CI, 1.47-22.76) and phenotype 3 (adjusted OR = 3.67; 95% CI, 1.05-14.29) had increased risk of curve progression ≥6° with mean follow-up of 3.03 ± 0.16 years. In conclusion, 3 distinct bone microarchitecture phenotypes could be clustered by unsupervised machine learning on HR-pQCT-generated bone parameters at peripubertal PHV in AIS. The bone quality reflected by these phenotypes was found to have significant differentiating risk of curve progression and progression to surgical threshold at skeletal maturity in AIS.

Evaluation of bone-related mechanical properties in female patients with long-term remission of Cushing's syndrome using quantitative computed tomography-based finite element analysis

BACKGROUND

Hypercortisolism in Cushing's syndrome (CS) is associated with bone loss, skeletal fragility, and altered bone quality. No studies evaluated bone geometric and strain-stress values in CS patients after remission thus far.

PATIENTS AND METHODS

Thirty-two women with CS in remission (mean age [±SD] 51 ± 11; body mass index [BMI], 27 ± 4 kg/m2; mean time of remission, 120 ± 90 months) and 32 age-, BMI-, and gonadal status-matched female controls. Quantitative computed tomography (QCT) was used to assess volumetric bone mineral density (vBMD) and buckling ratio, cross-sectional area, and average cortical thickness at the level of the proximal femur. Finite element (FE) models were generated from QCT to calculate strain and stress values (maximum principal strain [MPE], maximum strain energy density [SED], maximum Von Mises [VM], and maximum principal stress [MPS]). Areal BMD (aBMD) and trabecular bone score (TBS) were assessed by dual-energy X-ray absorptiometry (2D DXA).

RESULTS

Trabecular vBMD at total hip and trochanter were lower in CS as compared with controls (P < .05). Average cortical thickness was lower, and buckling ratio was greater in CS vs controls (P < .01). All strain and stress values were higher in CS patients vs controls (P < .05). 2D DXA-derived measures were similar between patients and controls (P > .05). Prior hypercortisolism predicted both VM (β .30, P = .014) and MPS (β .30, P = .015), after adjusting for age, BMI, menopause, delay to diagnosis, and duration of remission.

CONCLUSIONS

Women with prior hypercortisolism have reduced trabecular vBMD and impaired bone geometrical and mechanical properties, which may contribute to an elevated fracture risk despite long-term remission.

Characterizing Bone Phenotypes Related to Skeletal Fragility Using Advanced Medical Imaging

PURPOSE OF REVIEW

Summarize the recent literature that investigates how advanced medical imaging has contributed to our understanding of skeletal phenotypes and fracture risk across the lifespan.

RECENT FINDINGS

Characterization of bone phenotypes on the macro-scale using advanced imaging has shown that while wide bones are generally stronger than narrow bones, they may be more susceptible to age-related declines in bone strength. On the micro-scale, HR-pQCT has been used to identify bone microarchitecture phenotypes that improve stratification of fracture risk based on phenotype-specific risk factors. Adolescence is a key phase for bone development, with distinct sex-specific growth patterns and significant within-sex bone property variability. However, longitudinal studies are needed to evaluate how early skeletal growth impacts adult bone phenotypes and fracture risk. Metabolic and rare bone diseases amplify fracture risk, but the interplay between bone phenotypes and disease remains unclear. Although bone phenotyping is a promising approach to improve fracture risk assessment, the clinical availability of advanced imaging is still limited. Consequently, alternative strategies for assessing and managing fracture risk include vertebral fracture assessment from clinically available medical imaging modalities/techniques or from fracture risk assessment tools based on clinical risk factors. Bone fragility is not solely determined by its density but by a combination of bone geometry, distribution of bone mass, microarchitecture, and the intrinsic material properties of bone tissue. As such, different individuals can exhibit distinct bone phenotypes, which may predispose them to be more vulnerable or resilient to certain perturbations that influence bone strength.

Comparison of Motion Grading in 1,000 Patients by First- and Second-Generation HR-pQCT: A Propensity Score Matched Cohort Study

In-vivo bone microstructure measured by high-resolution peripheral quantitative computed tomography (HR-pQCT) is gaining importance in research and clinical practice. Second-generation HR-pQCT (XCT2) shows improved image quality and shorter measurement duration compared to the first generation (XCT1). Predicting and understanding the occurrence of motion artifacts is crucial for clinical practice. We retrospectively analyzed data from HR-pQCT measurements at the distal radius and tibia of 1,000 patients (aged 20 to 89) evenly distributed between both generations of HR-pQCT. Motion artifacts were graded between 1 (no motion) and 5 (severe motion), with grades greater 3 considered unusable. Additionally, baseline characteristics and patients' muscle performance and balance were measured. Various group comparisons between the two generations of HR-pQCT and regression analyses between patient characteristics and motion grading were performed. The study groups of XCT1 and XCT2 did not differ by age (XCT1: 64.9 vs. XCT2: 63.8 years, p = 0.136), sex (both 74.5% females, p > 0.999), or BMI (both 24.2 kg/m2, p = 0.911) after propensity score matching. XCT2 scans exhibited significantly lower motion grading in both extremities compared to XCT1 (Radius: p < 0.001; Tibia: p = 0.002). In XCT2 motion-corrupted scans were more than halved at the radius (XCT1: 35.3% vs. XCT2: 15.5%, p < 0.001), and at the tibia the frequency of best image quality scans was increased (XCT1: 50.2% vs. XCT2: 63.7%, p < 0.001). The strongest independent predictor for motion-corrupted images is the occurrence of high motion grading at the other scanning site during the same consultation. The association between high motion grading in one scan and a corresponding high motion grading in another scan within the same session suggests a non-resting patient. Additionally, aged, female, and patients with smaller stature tend towards higher motion grading, requiring special attention to a correct extremity fixation.

High-resolution peripheral quantitative computed tomography: research or clinical practice?

High-resolution peripheral quantitative CT (HR-pQCT) is a low-dose three-dimensional imaging technique, originally developed for in vivo assessment of bone microarchitecture at the distal radius and tibia in osteoporosis. HR-pQCT has the ability to discriminate trabecular and cortical bone compartments, providing densitometric and structural parameters. At present, HR-pQCT is mostly used in research settings, despite evidence showing that it may be a valuable tool in osteoporosis and other diseases. This review summarizes the main applications of HR-pQCT and addresses the limitations that currently prevent its integration into routine clinical practice. In particular, the focus is on the use of HR-pQCT in primary and secondary osteoporosis, chronic kidney disease (CKD), endocrine disorders affecting bone, and rare diseases. A section on novel potential applications of HR-pQCT is also present, including assessment of rheumatic diseases, knee osteoarthritis, distal radius/scaphoid fractures, vascular calcifications, effect of medications, and skeletal muscle. The reviewed literature seems to suggest that a more widespread implementation of HR-pQCT in clinical practice would offer notable opportunities. For instance, HR-pQCT can improve the prediction of incident fractures beyond areal bone mineral density provided by dual-energy X-ray absorptiometry. In addition, HR-pQCT may be used for the monitoring of anti-osteoporotic therapy or for the assessment of mineral and bone disorder associated with CKD. Nevertheless, several obstacles currently prevent a broader use of HR-pQCT and would need to be targeted, such as the small number of installed machines worldwide, the uncertain cost-effectiveness, the need for improved reproducibility, and the limited availability of reference normative data sets.

Automatic segmentation of trabecular and cortical compartments in HR-pQCT images using an embedding-predicting U-Net and morphological post-processing

High-resolution peripheral quantitative computed tomography (HR-pQCT) is an emerging in vivo imaging modality for quantification of bone microarchitecture. However, extraction of quantitative microarchitectural parameters from HR-pQCT images requires an accurate segmentation of the image. The current standard protocol using semi-automated contouring for HR-pQCT image segmentation is laborious, introduces inter-operator biases into research data, and poses a barrier to streamlined clinical implementation. In this work, we propose and validate a fully automated algorithm for segmentation of HR-pQCT radius and tibia images. A multi-slice 2D U-Net produces initial segmentation predictions, which are post-processed via a sequence of traditional morphological image filters. The U-Net was trained on a large dataset containing 1822 images from 896 unique participants. Predicted segmentations were compared to reference segmentations on a disjoint dataset containing 386 images from 190 unique participants, and 156 pairs of repeated images were used to compare the precision of the novel and current protocols. The agreement of morphological parameters obtained using the predicted segmentation relative to the reference standard was excellent (R² between 0.938 and > 0.999). Precision was significantly improved for several outputs, most notably cortical porosity. This novel and robust algorithm for automated segmentation will increase the feasibility of using HR-pQCT in research and clinical settings.