基于深度学习的骨质疏松影像学辅助诊断研究进展

Advances in imaging-assisted diagnosis of osteoporosis based on deep learning

骨质疏松症是由于多种原因导致的骨密度和骨质量下降,骨微结构破坏,造成骨脆性增加,从而容易发生骨折的常见全身性骨病。以双能X线吸收测定法为代表的影像学诊断方法无法确切定量评估骨质量,同时受限于成像特点容易出现误差。深度学习技术在计算机辅助诊断中的应用,为精准骨质疏松状态评估提供了可能。本研究对深度学习技术在骨质疏松影像学辅助诊断的研究进展进行综述并展望。

Osteoporosis is a common systemic bone disease caused by multiple causes of decreased bone density and bone mass and destruction of bone microarchitecture, resulting in increased bone fragility and thus susceptibility to fracture.Diagnostic imaging methods represented by dual-energy X-ray absorptiometry are unable to quantitatively assess bone quality with certainty, while being prone to errors due to imaging characteristics. The application of deep learning technology in computer-aided diagnosis offers the possibility of accurate osteoporosis status assessment. This study reviews and outlooks the research progress of deep learning technology in imaging-assisted diagnosis of osteoporosis.