Prediction of knee osteoarthritis progression using machine learning techniques
-
G. Malini Devi
Abstract
Knee osteoarthritis (OA) is a disease that increases in incidence and prevalence with advancing age, resulting in symptomatic knee OA in those over the age of 60, around 10% of men, and 13% of women. Arthritis was projected to cost the USD 336 billion in 2004, or 3% of the gross domestic product, with OA being the most prevalent form of arthritis. OA is mainly diagnosed in clinical studies using medical images. Magnetic resonance imaging (MR) is a noninvasive technique capable of producing three-dimensional images of soft-tissue intra-articular structures, including cartilage. Because of the anatomy and morphology of the knee as well as the process of MR imaging, precise and reproducible quantitative measurements from MRI scans are burdensome. Operators using cartilage segmentation software also require comprehensive training that contributes further to time and expense. For the knee, cartilage has established the segmentation of alternative MR slices or confining measurements to partial regions of cartilage approaches. Computer-aided algorithms have also been developed to assist in the segmentation of MR images by cartilage. Such techniques lack adequate precision and reliability to detect small alterations in cartilage. The proposed approach uses machine learning (ML) methods to explore the hidden biomedical information contained in the clinically used cartilage damage index (CDI), to measure the progression of knee OA disease. It uses principle component analysis as a feature selection method to find the optimum feature representation. The processed feature set is served as input to four ML methods (ANN and SVM). In the prediction of knee OA progression by CDI measures best accuracy of 96.77 is obtained by the ANN algorithm on the medial dataset.
Abstract
Knee osteoarthritis (OA) is a disease that increases in incidence and prevalence with advancing age, resulting in symptomatic knee OA in those over the age of 60, around 10% of men, and 13% of women. Arthritis was projected to cost the USD 336 billion in 2004, or 3% of the gross domestic product, with OA being the most prevalent form of arthritis. OA is mainly diagnosed in clinical studies using medical images. Magnetic resonance imaging (MR) is a noninvasive technique capable of producing three-dimensional images of soft-tissue intra-articular structures, including cartilage. Because of the anatomy and morphology of the knee as well as the process of MR imaging, precise and reproducible quantitative measurements from MRI scans are burdensome. Operators using cartilage segmentation software also require comprehensive training that contributes further to time and expense. For the knee, cartilage has established the segmentation of alternative MR slices or confining measurements to partial regions of cartilage approaches. Computer-aided algorithms have also been developed to assist in the segmentation of MR images by cartilage. Such techniques lack adequate precision and reliability to detect small alterations in cartilage. The proposed approach uses machine learning (ML) methods to explore the hidden biomedical information contained in the clinically used cartilage damage index (CDI), to measure the progression of knee OA disease. It uses principle component analysis as a feature selection method to find the optimum feature representation. The processed feature set is served as input to four ML methods (ANN and SVM). In the prediction of knee OA progression by CDI measures best accuracy of 96.77 is obtained by the ANN algorithm on the medial dataset.
Kapitel in diesem Buch
- Frontmatter I
- Contents V
- List of authors VII
- The impact of green manufacturing in Industry 4.0 for future ecosystems 1
- Custom manufacturing using Industry 4.0: cost-effective industry revolution model 17
- Cloud-based industrial IoT infrastructure to facilitate efficient data analytics 31
- An impact of robust Industry 4.0 strategy on supply chain management 53
- Machine learning based smart cloud factories 71
- Sustainable and flexible digital models for the manufacturing of ecosystem 91
- Industry 4.0: efficient supply chain management using energy-aware cloud infrastructure model 103
- Analytical models for planning and control of autonomous mobile robots for logistic management 121
- Industry 4.0: augmented reality in smart manufacturing industry environment to facilitate faster and easier work procedures 141
- Prediction of knee osteoarthritis progression using machine learning techniques 163
- Index 173
Kapitel in diesem Buch
- Frontmatter I
- Contents V
- List of authors VII
- The impact of green manufacturing in Industry 4.0 for future ecosystems 1
- Custom manufacturing using Industry 4.0: cost-effective industry revolution model 17
- Cloud-based industrial IoT infrastructure to facilitate efficient data analytics 31
- An impact of robust Industry 4.0 strategy on supply chain management 53
- Machine learning based smart cloud factories 71
- Sustainable and flexible digital models for the manufacturing of ecosystem 91
- Industry 4.0: efficient supply chain management using energy-aware cloud infrastructure model 103
- Analytical models for planning and control of autonomous mobile robots for logistic management 121
- Industry 4.0: augmented reality in smart manufacturing industry environment to facilitate faster and easier work procedures 141
- Prediction of knee osteoarthritis progression using machine learning techniques 163
- Index 173
