人工智能在诊断骨质疏松症中的研究进展

Advances in Artificial Intelligence in the Diagnosis of Osteoporosis

摘要: 骨质疏松(osteoporosis, OP)是一种以骨转换增加和骨量减少为特征的疾病，伴有骨脆性，导致骨折风险增加的疾病。骨质疏松症影响10.2%的50岁以上成年人，预计到2030年将增加到13.6%。随着全球人口老龄化趋势进展飞速，骨质疏松症和骨质疏松症相关骨折是老年人发病和死亡的常见原因，然而在日常诊疗中骨质疏松并没有明确诊断。作为诊断骨质疏松金标准的双能X线吸收测定法，在中国，年龄 ≥ 20岁的人群中，只有2.8%的人进行了检测，而年龄 ≥ 50岁的人群中，这一比例为3.7%。因此为更好地诊断骨质疏松和骨密度减少，有必要寻求有效、安全、成本低廉的替代品。近年来，机器学习(Machine learning, ML)已成为一种很有前途的预测风险技术。据报道，目前人工智能已积极应用于医学诊断和筛查领域，以提高诊断准确性，如肺结节的诊断、骨折的检测等。本文就目前人工智能应用于骨质疏松的现状进行阐述，以期能为未来临床骨质疏松的诊断提供新方向。

Abstract: Osteoporosis (OP) is a disease characterised by increased bone turnover and decreased bone mass, accompanied by bone fragility, leading to an increased risk of fracture. Osteoporosis affects 10.2% of adults over the age of 50 and is expected to increase to 13.6% by 2030. With the global trend of population ageing progressing at a rapid pace, osteoporosis and osteoporosis-related fractures are a common cause of morbidity and mortality in the elderly, yet osteoporosis is not clearly diagnosed in routine practice. Dual-energy X-ray absorptiometry, the gold standard for the diagnosis of osteoporosis, is performed in only 2.8% of people aged ≥ 20 years in China, compared with 3.7% of people aged ≥ 50 years. Therefore for better diagnosis of osteoporosis and reduced bone density, it is necessary to seek effective, safe and cost-effective alternatives. In recent years, machine learning (ML) has emerged as a promising technique for predicting risk. Currently, AI has been reported to be actively applied in the field of medical diagnosis and screening to improve diagnostic accuracy, such as the diagnosis of lung nodules and the detection of bone fractures. This paper describes the current status of AI application to osteoporosis, with the aim of providing a new direction for future clinical osteoporosis diagnosis.