Quantitative ultrasound of the phalanges and DXA of the lumbar spine and proximal femur in evaluating the risk of osteoporotic vertebral fracture in postmenopausal women

Evaluation of fragility fracture risk using deep learning based on ultrasound radio frequency signal

BACKGROUND

It was essential to identify individuals at high risk of fragility fracture and prevented them due to the significant morbidity, mortality, and economic burden associated with fragility fracture. The quantitative ultrasound (QUS) showed promise in assessing bone structure characteristics and determining the risk of fragility fracture.

AIMS

To evaluate the performance of a multi-channel residual network (MResNet) based on ultrasonic radiofrequency (RF) signal to discriminate fragility fractures retrospectively in postmenopausal women, and compared it with the traditional parameter of QUS, speed of sound (SOS), and bone mineral density (BMD) acquired with dual X-ray absorptiometry (DXA).

METHODS

Using QUS, RF signal and SOS were acquired for 246 postmenopausal women. An MResNet was utilized, based on the RF signal, to categorize individuals with an elevated risk of fragility fracture. DXA was employed to obtain BMD at the lumbar, hip, and femoral neck. The fracture history of all adult subjects was gathered. Analyzing the odds ratios (OR) and the area under the receiver operator characteristic curves (AUC) was done to evaluate the effectiveness of various methods in discriminating fragility fracture.

RESULTS

Among the 246 postmenopausal women, 170 belonged to the non-fracture group, 50 to the vertebral group, and 26 to the non-vertebral fracture group. MResNet was competent to discriminate any fragility fracture (OR = 2.64; AUC = 0.74), Vertebral fracture (OR = 3.02; AUC = 0.77), and non-vertebral fracture (OR = 2.01; AUC = 0.69). After being modified by clinical covariates, the efficiency of MResNet was further improved to OR = 3.31-4.08, AUC = 0.81-0.83 among all fracture groups, which significantly surpassed QUS-SOS (OR = 1.32-1.36; AUC = 0.60) and DXA-BMD (OR = 1.23-2.94; AUC = 0.63-0.76).

CONCLUSIONS

This pilot cross-sectional study demonstrates that the MResNet model based on the ultrasonic RF signal shows promising performance in discriminating fragility fractures in postmenopausal women. When incorporating clinical covariates, the efficiency of the modified MResNet is further enhanced, surpassing the performance of QUS-SOS and DXA-BMD in terms of OR and AUC. These findings highlight the potential of the MResNet as a promising approach for fracture risk assessment. Future research should focus on larger and more diverse populations to validate these results and explore its clinical applications.

Exploring deep learning radiomics for classifying osteoporotic vertebral fractures in X-ray images

Purpose

To develop and validate a deep learning radiomics (DLR) model that uses X-ray images to predict the classification of osteoporotic vertebral fractures (OVFs).

Material and methods

The study encompassed a cohort of 942 patients, involving examinations of 1076 vertebrae through X-ray, CT, and MRI across three distinct hospitals. The OVFs were categorized as class 0, 1, or 2 based on the Assessment System of Thoracolumbar Osteoporotic Fracture. The dataset was divided randomly into four distinct subsets: a training set comprising 712 samples, an internal validation set with 178 samples, an external validation set containing 111 samples, and a prospective validation set consisting of 75 samples. The ResNet-50 architectural model was used to implement deep transfer learning (DTL), undergoing -pre-training separately on the RadImageNet and ImageNet datasets. Features from DTL and radiomics were extracted and integrated using X-ray images. The optimal fusion feature model was identified through least absolute shrinkage and selection operator logistic regression. Evaluation of the predictive capabilities for OVFs classification involved eight machine learning models, assessed through receiver operating characteristic curves employing the "One-vs-Rest" strategy. The Delong test was applied to compare the predictive performance of the superior RadImageNet model against the ImageNet model.

Results

Following pre-training separately on RadImageNet and ImageNet datasets, feature selection and fusion yielded 17 and 12 fusion features, respectively. Logistic regression emerged as the optimal machine learning algorithm for both DLR models. Across the training set, internal validation set, external validation set, and prospective validation set, the macro-average Area Under the Curve (AUC) based on the RadImageNet dataset surpassed those based on the ImageNet dataset, with statistically significant differences observed (P<0.05). Utilizing the binary "One-vs-Rest" strategy, the model based on the RadImageNet dataset demonstrated superior efficacy in predicting Class 0, achieving an AUC of 0.969 and accuracy of 0.863. Predicting Class 1 yielded an AUC of 0.945 and accuracy of 0.875, while for Class 2, the AUC and accuracy were 0.809 and 0.692, respectively.

Conclusion

The DLR model, based on the RadImageNet dataset, outperformed the ImageNet model in predicting the classification of OVFs, with generalizability confirmed in the prospective validation set.

MicroRNAs as potential biopredictors for premenopausal osteoporosis: a biochemical and molecular study

BACKGROUND

Circulating micro-RNAs have been proposed as a new type of biomarker in several diseases, particularly those related to bone health. They have shown great potential due to their feasibility and simplicity of measurement in all body fluids, especially urine, plasma, and serum.

AIM

This study aimed to evaluate the expression of a set of mRNAs, namely miR-21, miR-24, mir-100, miR-24a, miR-103-3p, and miR-142-3p. Their proposed roles in the progression of osteoporosis were identified using a real-time polymerase chain reaction (RT-PCR) analysis in premenopausal women. In addition, their correlations with osteocalcin (OC), bone-specific alkaline phosphatase (BAP), and deoxypyridinoline (DPD) bone markers were explored.

METHODS

A total of 85 healthy premenopausal women aged 25-50 years old were included in this study. Based on a DXA scan (Z-score) analysis and calcaneus broadband ultrasound attenuation scores (c-BUAs), measured via quantitative ultrasound (QUS), the subjects were classified into three groups: normal group (n = 25), osteopenia (n = 30), and osteoporosis (n = 30). Real-time-PCR and immunoassay analyses were performed to determine miRNA expression levels and serum OC, s-BAP, and DPD, respectively, as biomarkers of bone health.

RESULTS

Among the identified miRNAs, only miR-21, miR-24, and mir-100 were significantly upregulated and increased in the serum of patients with osteopenia and osteoporosis, and miR-24a, miR-103-3p, and miR-142-3p were downregulated and significantly decreased in osteoporosis. Both upregulated and downregulated miRNAs were significantly correlated with BMD, c-BUA, OC, s-BAP, and DPD.

CONCLUSION

A group of circulating miRNAs was shown to be closely correlated with the parameters BMD, c-BUA, OC, s-BAP, and DPD, which are traditionally used for bone-health measurements. They could be identified as non-invasive biomarkers in premenopausal patients with osteoporosis. More studies with large sample sizes are recommended to estimate the mechanistic role of miRNAs in osteoporosis pathogenesis and to provide evidence for the use of these miRNAs as a non-invasive method of diagnosing clinical osteoporosis, especially in premenopausal patients.

Assessment of Lumbar Vertebrae Morphology by Computed Tomography in Older Adults with Osteoporosis

BACKGROUND

Hounsfield units (HU) values derived from computerized tomography (CT) have been used in diagnosis for osteoporosis in the lumbar spine.

OBJECTIVE

This study aimed to identify anatomical dimensions of lumbar vertebrae on CT images, which were different between older normal, osteopenic and osteoporotic subjects.

METHODS

This prospective pilot study enrolled 79 older adults. Based on CT measurements of lumbar vertebrae in HU, participants were classified into three groups: normal (HU > 109), osteopenia (HU: 94-108), and osteoporosis (HU < 93). Altogether, 42 anatomical variables of lumbar vertebrae, L2, L3, L4, and L5, were measured in each participant by CT, including 24 parameters measurable by MRI or plain X-ray, and 18 parameters measurable by MRI only.

RESULTS

Among the morphological measurements also measurable by MRI and plain X-ray, the length upper curve, 50% and 75% of L5, length upper with cortex of L4, length center of cortex of L3, as well as width upper curve 75% of L2, were significantly different between the three groups (p= 0.008, 0.007, 0.035, 0.036, and 0.003 respectively). Among the morphological measurements also measurable by MRI, only width upper cortex 75% of L5 and width lower cortex 25% of L3 were significantly different between the three groups (p= 0.031 and 0.020, respectively).

CONCLUSION

Seven CT morphological measurements may be used as "reference standard" CT measurements for preliminarily diagnosing osteoporosis and osteopenia in older adults.

Diagnostic value of calcaneal quantitative ultrasound in the evaluation of osteoporosis in middle-aged and elderly patients

To study the correlation between calcaneal quantitative ultrasound (QUS) and dual-energy X-ray absorptiometry (DXA), and analyze the diagnostic value of calcaneal QUS in the evaluation of middle-aged and elderly osteoporosis.We assessed bone mineral density (BMD) at the femoral neck and intertrochanteric of left hip and lumbar spine (L1-L4) sites with DXA and QUS parameters of the right and left calcanei in a cohort of 82 patients over the age of 50 years. Using DXA parameters as the gold standard for the diagnosis of osteoporosis, the correlation coefficient between BMD and QUS parameters was calculated. Receiver operating characteristic curve was generated and areas under the curves were evaluated. Cut-off values for QUS were defined.In men, there was a moderate correlation between calcaneal QUS and proximal femoral BMD (P < .05), but no significant correlation between calcaneal QUS and lumbar BMD (P > .05). In women, calcaneal QUS were moderately correlated with lumbar spine and proximal femoral BMD (P < .05). Using DXA as the gold standard, the accuracy, sensitivity, specificity, positive predictive value, and negative predictive value of calcaneal QUS in the diagnosis of osteoporosis were 90.2%, 89.2%, 100%, 100%, and 50.0%, respectively. According to the receiver operating characteristic curve, when the QUS T-score of calcaneum was -1.8, the area under the curve was 0.888, the sensitivity was 73.21%, and the specificity was 92.31% (P < .05). When the QUS T-score of calcaneum was -2.35, the sensitivity was 37.2% and the specificity was 100%.Calcaneal QUS can be used to predict proximal femoral BMD in middle-aged and elderly people, as well as lumbar BMD in women. As a screening method for osteoporosis, calcaneal QUS has good specificity, so it can be recommended to use it as a pre-screening tool to reduce the number of DXA screening. When the QUS T-score of calcaneum is -1.8, it has the greatest diagnostic efficiency for osteoporosis; when the QUS T-score of calcaneum is ≤-2.35, it can be diagnosed as osteoporosis.

Quantitative Ultrasound of Phalanx in Primary and Secondary Osteoporosis: Mini-review and Practical Experience

Objective:

Dual x-ray absorptiometry (DXA) is gold standard of bone densitometry, but quantitative ultrasound (QUS) of bone is less expensive and portable. This study was designed to assess its usefulness in secondary osteoporosis diagnosis.

Materials and Methods:

There were 200 secondary osteoporosis cases (rheumatoid arthritis, hemodialysis, kidney transplant patients, and levothyroxine users) and of those, their phalanx QUS results were compared with normal controls. Also, the QUS and DXA results were compared to find any correlation of these methods for diagnosing osteoporosis.

Results:

There was not significantly different results compared with normal controls, except for those of hemodialysis patients ( P = .00). Also, the comparison of QUS with DXA results showed no significant correlation except in hemodialysis patients, in both spinal and femoral regions ( P = .023 and .21, respectively), as well as the levothyroxine group’s spinal region ( P = .005).

Conclusion:

These results suggest that QUS of phalanx may be useful in screening secondary osteoporosis but for establishment of diagnosis, DXA measurements are still needed.

Wavelet transform-based photoacoustic time-frequency spectral analysis for bone assessment

In this study, we investigated the feasibility of using photoacoustic time-frequency spectral analysis (PA-TFSA) for evaluating the bone mineral density (BMD) and bone structure. Simulations and ex vivo experiments on bone samples with different BMDs and mean trabecular thickness (MTT) were conducted. All photoacoustic signals were processed using the wavelet transform-based PA-TFSA. The power-weighted mean frequency (PWMF) was evaluated to obtain the main frequency component at different times. The y-intercept, midband-fit, and slope of the linearly fitted curve of the PWMF over time were also quantified. The results show that the osteoporotic bone samples with lower BMD and thinner MTT have higher frequency components and lower acoustic frequency attenuation over time, thus higher y-intercept, midband-fit, and slope. The midband-fit and slope were found to be sensitive to the BMD; therefore, both parameters could be used to distinguish between osteoporotic and normal bones (p < 0.05).

Association of serum leptin with bone mineral density in postmenopausal osteoporotic females

OBJECTIVE

Present study was designed to find out whether leptin is a predictor of bone mass density (BMD) in premenopausal women (PMW) and postmenopausal osteoporotic women (PMOPW) or it has no association with BMD.

METHODS

One hundred and ninety two women (98 PMOPW and 94 PMW) were recruited for this study. The control group was BMI matched with osteoporotic subjects. BMD assessment was done on calcaneus by peripheral ultrasound bone densitometry and T scores were determined. Serum leptin levels were measured by enzyme-linked immunosorbent assay (ELISA).

RESULTS

Serum leptin and BMD values were significantly different in both groups (leptin, 18.56 ± 8.65 ng/ml versus 21.64 ± 9.80 ng/ml, p = 0.02) and (BMD, -0.70 ± 0.19 versus -3.17 ± 0.59, p = 0.000), respectively. In PMOPW serum leptin and BMD were considerably correlated with weight (lep, r = 0.53, p = <0.001; BMD, r = -0.21, p = 0.02), BMI (lep, r = 0.52, p = <0.001; BMD, r = -0.27, p = 0.005), waist circumference (lep, r = 0.61, p = <0.001; BMD, r = 0.18, p = 0.04), hip circumference (lep, r = 0.58, p = <0.001). Multivariate linear stepwise regression analysis showed that weight and BMI in PMW and PMOPW were independent predictors of BMD. Serum leptin level was not found to be the predictor of BMD in both groups.

CONCLUSION

The present results indicate that body weight and BMI have an impact on BMD while serum leptin is not associated with BMD in PMW and PMOPW.

Falls, fractures and bone density in Parkinson's disease - a cross-sectional study

AIM

Evidence suggests that falls and associated bone fractures are more frequent in patients suffering from Parkinson's disease (PD) than in the general population. In this cross-sectional study we evaluated the clinical and biochemical characteristics that are associated to falls, fractures and bone health in a population of PD subjects.

MATERIALS AND METHODS

Forty-two consecutive subjects suffering from idiopathic PD (mild-to-moderate severity) with/without falls in the previous year were included. They were characterized as regards functional independence, balance, fear of falling, bone density (ultrasound densitometry) and plasma levels of vitamin D. Twenty-one age- and sex-matched healthy subjects were evaluated as controls.

RESULTS

We detected a greater degree of osteoporosis in PD subjects as compared to controls, more pronounced in males than in females (Z-score: M -3.8 ± 1.6, F -2.28 ± 0.92, p = 0.0006). A positive correlation was found between independence levels and bone density or vitamin D levels. Twenty seven patients (64%) reported falls in the previous year. These were associated to post-traumatic fractures in 16 subjects (59% of fallers). Women fell more than men (fallers: 20 F/7 M; non fallers: 4 F/11 M, χ² test p = 0.02), although the occurrence of post-traumatic fractures among fallers did not differ between sexes (F 11/9, M 5/2, χ² test p > 0.05). Fallers with post-traumatic fractures showed higher degrees of motor impairment.

CONCLUSIONS

These findings confirm that falls and osteoporosis represent major health issues in PD, already in the middle stages of disease.

An Advanced Quantitative Echosound Methodology for Femoral Neck Densitometry

The aim of this paper was to investigate the clinical feasibility and the accuracy in femoral neck densitometry of the Osteoporosis Score (O.S.), an ultrasound (US) parameter for osteoporosis diagnosis that has been recently introduced for lumbar spine applications. A total of 377 female patients (aged 61-70 y) underwent both a femoral dual X-ray absorptiometry (DXA) and an echographic scan of the proximal femur. Recruited patients were sub-divided into a reference database used for ultrasound spectral model construction and a study population for repeatability assessments and accuracy evaluations. Echographic images and radiofrequency signals were analyzed through a fully automatic algorithm that performed a series of combined spectral and statistical analyses, providing as a final output the O.S. value of the femoral neck. Assuming DXA as a gold standard reference, the accuracy of O.S.-based diagnoses resulted 94.7%, with k = 0.898 (p < 0.0001). Significant correlations were also found between O.S.-estimated bone mineral density and corresponding DXA values, with r(2) up to 0.79 and root mean square error = 5.9-7.4%. The reported accuracy levels, combined with the proven ease of use and very good measurement repeatability, provide the adopted method with a potential for clinical routine application in osteoporosis diagnosis.

New perspectives in echographic diagnosis of osteoporosis on hip and spine

Currently, the accepted "gold standard" method for bone mineral density (BMD) measurement and osteoporosis diagnosis is dual-energy X-ray absorptiometry (DXA). However, actual DXA effectiveness is limited by several factors, including intrinsic accuracy uncertainties and possible errors in patient positioning and/or post-acquisition data analysis. DXA employment is also restricted by the typical issues related to ionizing radiation employment (high costs, need of dedicated structures and certified operators, unsuitability for population screenings). The only commercially-available alternative to DXA is represented by "quantitative ultrasound" (QUS) approaches, which are radiation-free, cheaper and portable, but they cannot be applied on the reference anatomical sites (lumbar spine and proximal femur). Therefore, their documented clinical usefulness is restricted to calcaneal applications on elderly patients (aged over 65 y), in combination with clinical risk factors and only for the identification of healthy subjects at low fracture risk. Literature-reported studies performed some QUS measurements on proximal femur, but their clinical translation is mostly hindered by intrinsic factors (e.g., device bulkiness). An innovative ultrasound methodology has been recently introduced, which performs a combined analysis of B-mode images and corresponding "raw" radiofrequency signals acquired during an echographic scan of the target reference anatomical site, providing two novel parameters: Osteoporosis Score and Fragility Score, indicative of BMD level and bone strength, respectively. This article will provide a brief review of the available systems for osteoporosis diagnosis in clinical routine contexts, followed by a synthesis of the most promising research results on the latest ultrasound developments for early osteoporosis diagnosis and fracture prevention.

Characterization of bone microstructure using photoacoustic spectrum analysis

Osteoporosis is a progressive bone disease that is characterized by a decrease in bone mass and the deterioration in bone microarchitecture. This study investigates the feasibility of characterizing bone microstructure by analyzing the frequency spectrum of the photoacoustic (PA) signal from the bone. Modeling and numerical simulation of PA signal were performed on trabecular bone simulations and CT scans with different trabecular thicknesses. The resulting quasi-linear photoacoustic spectra were fittted by linear regression, from which the spectral parameter slope was quantified. The simulation based on two different models both demonstrate that bone specimens with thinner trabecular thicknesses have higher slope. Experiment on osteoporotic rat femoral heads with different mineral content was conducted. The finding from the experiment was in good agreement with the simulation, demonstrating that the frequency-domain analysis of PA signals can provide an objective assessment of bone microstructure and deterioration. Considering that PA measurement is non-ionizing, non-invasive, and has sufficient penetration in both calcified and non-calcified tissues, this new bone evaluation method based on photoacoustic spectral analysis holds potential for clinical management of osteoporosis and other bone diseases.

Epidemiological burden of postmenopausal osteoporosis in Italy from 2010 to 2020: estimations from a disease model

The article describes the adaptation of a model to estimate the burden of postmenopausal osteoporosis in women aged 50 years and over in Italy between 2010 and 2020. For this purpose, a validated postmenopausal osteoporosis disease model developed for Sweden was adapted to Italy. For each year of the study, the 'incident cohort' (women experiencing a first osteoporotic fracture) was identified and run through a Markov model using 1-year cycles until 2020. Health states were based on the number of fractures and deaths. Fracture by site (hip, clinical vertebral, non-hip non-vertebral) was tracked for each health state. Transition probabilities reflected fracture site-specific risk of death and subsequent fractures. Model inputs specific to Italy included population size and life tables from 1970 to 2020, incidence of hip fracture and BMD by age in the general population (mean and standard deviation). The model estimated that the number of postmenopausal osteoporotic women would increase from 3.3 million to 3.7 million between 2010 and 2020 (+14.3%). Assuming unchanged incidence rates by age group over time, the model predicted the overall number of osteoporotic fractures to increase from 285.0 to 335.8 thousand fractures between 2010 and 2020 (+17.8%). The estimated expected increases in hip, vertebral and non-hip non-vertebral fractures were 22.3, 17.2 and 16.3%, respectively. Due to demographic changes, the burden of fractures is expected to increase markedly by 2020.

Performance of phalangeal quantitative ultrasound parameters in the evaluation of reduced bone mineral density assessed by DX in patients with 21 hydroxylase deficiency

The purpose of this study was to verify the performance of quantitative ultrasound (QUS) parameters of proximal phalanges in the evaluation of reduced bone mineral density (BMD) in patients with congenital adrenal hyperplasia due to 21-hydroxylase deficiency (21 OHD). Seventy patients with 21 OHD (41 females and 29 males), aged between 6-27 y were assessed. The QUS measurements, amplitude-dependent speed of sound (AD-SoS), bone transmission time (BTT), and ultrasound bone profile index (UBPI) were obtained using the BMD Sonic device (IGEA, Carpi, Italy) on the last four proximal phalanges in the non-dominant hand. BMD was determined by dual energy X-ray (DXA) across the total body and lumbar spine (LS). Total body and LS BMD were positively correlated to UBPI, BTT and AD-SoS (correlation coefficients ranged from 0.59-0.72, p < 0.001). In contrast, when comparing patients with normal and low (Z-score < -2) BMD, no differences were found in the QUS parameters. Furthermore, UBPI, BTT and AD-SoS measurements were not effective for diagnosing patients with reduced BMD by receiver operator characteristic curve parameters. Although the AD-SoS, BTT and UBPI showed significant correlations with the data obtained by DXA, they were not effective for diagnosing reduced bone mass in patients with 21 OHD.

Bone ultrasound and physical activity in postmenopausal Spanish women

We studied the effects of moderate physical activity on bone mass in healthy postmenopausal women as assessed by quantitative bone ultrasound. We enrolled a total of 326 postmenopausal women (60.95 ± 8.51 years old and 13.02 ± 9.52 years since menopause) and categorized them by weekly physical activity in their leisure time (sedentary, moderate, and active). We administered a questionnaire on general health and diet and measured bone density by ultrasound on phalanges II-V in the nondominant hand. We found no significant difference in the amplitude-dependent speed of sound (Ad-SoS; p > .05) between sedentary and active women. We concluded that the changes induced by moderate physical activity on bone mass are minimal in healthy postmenopausal women.

Quantitative Ultrasound in Spanish Children and Young Adults With Cystic Fibrosis

We studied the relationship between bone density and cystic fibrosis in Spanish children and young adults. We measured the phalangeal bone amplitude-dependent speed of sound (Ad-SoS) in 35 patients with cystic fibrosis and in 30 healthy controls matched for gender, age, and body mass index (BMI). Participants were subjects with normal levels of 25(OH) Vitamin D. We found no difference in Ad-SoS between patients and controls. The only difference between the groups was that the patients had a significantly higher daily caloric intake than the controls ( p < .05) as a result of the patient group's greater intake of fats ( p < .05). There was a positive correlation between Ad-SoS and weight ( p < .0001), but after adjusting for potential confounding factors such as age, the correlation was lost. The percentage of ideal weight did not differ between the two groups. We conclude that well-nourished CF patients had similar Ad-SoS to controls.