Automatic Classifying of Patients With Non-Traumatic Fractures Based on Ultrasonic Guided Wave Spectrum Image Using a Dynamic Support Vector Machine

Abstract

Bone fractures are caused by diseases or accidents and are a widespread problem throughout the globe. Worldwide, 1.6 millions of hip fractures occur every year and are expected to rise to 6.3 millions in 2050. The current gold standard to assess fracture risk is the Dual-energy X-ray Absorptiometry (DXA), which provides a projected image of the bone from which areal bone mineral density is extracted. Ultrasound techniques have been proposed as non invasive alternatives. Recently, estimates of cortical thickness and porosity, obtained by Bi-Directional Axial Transmission (BDAT) in a pilot clinical study, have been shown to be associated with non-traumatic fractures in post menopausal women. Cortical parameters were derived from the comparison between experimental and theoretical guided modes. This model-based inverse approach failed for the patients associated with poor guided mode information. Moreover, even if the fracture discrimination ability was found similar to DXA, it remained moderate. The goal of this paper is to explore if these two limitations could be overcome by using automatic classification tools, independent of any waveguide model. To this end, a dynamic machine learning approach based on a Support Vector Machine (SVM) has been used to classify ultrasonic guided wave spectrum images measured by BDAT on post menopausal women with or without non-traumatic fractures. This approach has then been improved using parameters tuned by Bat Algorithm Optimization (BOA). The applied methodology focused on the extraction of texture features through a gray level co-occurrence matrix, structural comparison and histograms. The results accuracy was assessed using a confusion matrix. The effectiveness of the learning approach reached an accuracy of 92.31%. INDEX TERMS cortical bone, osteoporosis, fracture discrimination, quantitative ultrasound, guided waves, support vector machine, bat algorithm, guided wave spectrum image.

Machine Learning Solutions for Osteoporosis—A Review

“…Among the 32 studies that investigated fracture detection (Table 3), (66–97 ) 11 were on vertebral fractures, ( 66–76 ) 17 hip fractures, ( 74–90 ) and 10 other fracture sites such as humerus or wrist. ( 75,76,90–97 ) Nineteen studies developed CNN models for image analysis. ( 66,67,71,72,77–86,90,91,93–95 ) Others used features extracted from images or collected from non‐imaging data.…”

“…( 66,67,78,79,84,90,94 ) However, several studies did not use techniques that allowed for adequate understanding of the models' decision‐making (eg, data visualization). ( 68,70–76,80,82,83,85–89,92,93,95,97 ) In some studies, class unbalancing was handled using sampling methods, ( 74,97 ) and overfitting was controlled with data augmentation. ( 80,83,84,95 ) Mutasa and colleagues demonstrated how data augmentation using digitally reconstructed and generated images was able to improve the test AUC from 0.80 (95% CI, not reported) to 0.92 (95% CI, not reported).…”

“…DL models were not used in any of the studies comprising less than 300 samples given that large sample sizes are a prerequisite for the DL model implementation. ( 69,70,73,75,86–89,92,96 ) Three studies shared their code for model development replication. ( 67,75,77 )…”

“…( 66,67,71,72,77–86,90,91,93–95 ) Others used features extracted from images or collected from non‐imaging data. ( 68–70,73–77,87–89,92,96,97 ) Studies reported average best AUC of 0.92 (range 0.63 to 1.00) and average best accuracy of 89.8% (range 78.4% to 99.1%]. Surprising findings were that hospital‐related variables, such as the scanner device model, were better predictors of fractures than patients' characteristics or their images (AUC = 0.89 [95% confidence interval (CI) 0.87–0.91] versus 0.79 [95% CI 0.75–0.82] and 0.78 [95% CI 0.74–0.81], respectively), ( 77 ) which the authors acknowledged as a potential bias induced by the triage process.…”

A new approach for the detection of pneumonia in children using CXR images based on an real-time IoT system

“…CXR is the most common method to detect pneumonia worldwide due to its low cost, speed, and availability. In addition to pneumonia, radiography can be used to diagnose other pathologies, such as the detection of fractures [ 23 , 31 , 33 ], cavities [ 16 ], and tumors [ 32 ].…”