Abstract
Thailand is entering an aging society completely. Osteoporosis is another disease that may become more frequent as it is more common among the elderly around the world. Osteoporosis increases the risk of bone fractures and, when fractures occur, is a contributing factor in increased mortality for the elderly. If fractures can be predicted for older adults with osteoporosis and solutions to prevent fractures can be found before they occur to the elderly, these will help to reduce the danger, loss of property, wasted time, and death of the elderly. The medical images obtained either as a direct diagnosis of fractures or in the diagnosis of other diseases can be further enhanced by combining with computer modeling techniques for biomechanically analysis based on finite element method, which are used in the diagnosis, treatment, and may include prognosis of various medical conditions widely, to analyze and predict fractures for elderly people with osteoporosis. However, this analysis technique is used mainly in research laboratory. This is due to a lot of resources required. If a sufficiently large number of analytical results can be obtained and used as a sample set input for the machine learning system to learn, then the current operating procedures may be reduced. It may result in faster diagnosis or prediction with adequate accuracy compared to traditional techniques. Therefore, the objective of this study was to predict fractures in the elderly due to various activities and conducting a study on the feasibility of collecting the biomechanical analysis results to be used as a sample input for future machine learning systems to use in learning with the aim of replacing the current procedure for performing part of the finite element analysis. From the study, it could be confirmed that the condition of osteoporosis greatly affected the force transmissibility. The external load was mainly experienced by the bone tissue which has a high bone mineral density, but this type of bone tissue is less common in patients suffering from osteoporosis. Therefore, occurring of osteoporosis affect bone strength directly. Moreover, the study was likely to find a relationship between the bone mineral density values obtained from medical imaging and stress values obtained from finite element analysis under various stress conditions. This may be the possibility of using machine learning to replace engineering analysis using the finite element method leading to the ability to reduce the process of finite elements in order to get a faster but still accurate diagnosis or prediction result.
