Abstract:

Osteoporosis is a disease of bones that leads to an increased risk of fracture and it is characterized by low bone mineral density and micro-architectural deterioration of bone tissue. In this article, the dataset consists of 3426 subjects (1083 pathological and 2343 healthy cases) whose diagnosis was based on laboratory and osteal bone densitometry examination. In all cases, four diagnostic factors for osteoporosis risk prediction, namely age, sex, height and weight were stored for later evaluation with the selected classifiers. In order to categorize subjects into two classes (osteoporosis, non-osteoporosis), twenty machine learning techniques were assessed, based on their popularity and frequency in biomedical engineering problems. All classifiers have been evaluated using the well-known 10-fold cross validation method and the results are reported analytically. In addition, a feature selection method identified that with the use of only two diagnostic factors (age and weight), similar performance could be achieved. The scope of the proposed exhaustive methodology is to assist therapists in osteoporosis prediction, avoiding unnecessary further testing with bone densitometry.
Keywords: dataset, deep learning algorithm

Introduction:

Osteoporosis is the prevailing bone’s disease, and its features are low bone density mass and the modification of their micro-architecture structure, so that bones’ tolerance is reduced and the risk of fracture is increased. In osteoporosis, the Bone Mineral Density (BMD) is reduced; the bone micro-architecture is disrupted whereas the concentration and the variety of proteins in bones are altered. The classic osteoporotic fractures are hip, vertebral and wrist fractures. Osteoporotic fractures are defined as occurring at a site associated with low BMD and which at the same time increased in incidence after the age of 50 years .
Apart from the direct physical implications of a fracture, such as pain and inconvenience, osteoporotic fractures are a major cause of morbidity and mortality. The lifetime risk in the United States for a hip, spine, or forearm fracture at the age of 50 years has been estimated to be 40% in women and 13% in men. In Sweden, the corresponding percentages are 46% for women and 22% for men. Caucasians and Asians are at increased risk, since African and Americans have 6% higher BMD. In the European Union one person breaks a bone because of osteoporosis every fifteen seconds. It is a fact that a percentage as high as 75% of the women with osteoporosis disregards this disorder.
There are two types of osteoporosis, the primary (idiopathic) osteoporosis, which is a most frequent disease for women after menopause and is called postmenopausal osteoporosis. This type also includes the senile osteoporosis that may also be developed in men. The secondary osteoporosis, which may occur on anyone in the presence of particular hormonal disorders and other chronic diseases, as a result of medications, specifically glucocorticoids or other conditions causing increased bone loss by various mechanisms. In this case the disease is called steroid or glucocorticoid induced osteoporosis . Often the first apparent symptom of osteoporosis is a broken bone, which is why the condition is also known as “the silent crippler”, as people do not realize they have osteoporosis until it’s too late. However early detection and treatment of osteoporosis can decrease the fracture risk of a person to a minimum. For these reasons, there are studies using Artificial Intelligence techniques that are used for predicting whether a person has osteoporosis or not.

Osteoporosis is a prevalent and debilitating bone disease characterized by reduced bone density and structural deterioration, leading to an increased risk of fractures. It primarily affects older adults, especially postmenopausal women, but can also impact men and younger individuals with specific risk factors. Early detection is crucial for effective management and prevention of severe complications. Traditional Diagnostic Methods Traditional diagnostic methods for osteoporosis include: The Role of Deep Learning in Osteoporosis Detection Deep learning, a subset of machine learning and artificial intelligence (AI), has shown tremendous potential in medical imaging and diagnostics. By leveraging large datasets and advanced neural networks, deep learning algorithms can automatically learn and identify complex patterns in medical images, potentially outperforming traditional methods in accuracy and efficiency.