Application progress of artificial intelligence in the diagnosis and treatment of scoliosis / 中华骨科杂志

ABSTRACT

Artificial intelligence (AI), a branch of computer science, is an emerging science used to develop theories, methods, technologies and application systems that can simulate human intelligence. The goal is to enable machines to solve some complex tasks that require human intelligence. With the era of big data coming, AI has been widely used in multitudinous professional fields, including machine vision, speech recognition, image understanding, genetic programming, intelligent factory and expert systems. As the most common three-dimensional deformity of the spine, scoliosis not only changes patients' appearance and body shape but also affects their mental health. Some challenges have to be solved in diagnosis and treatment of scoliosis, such as the complexity of the anatomical structure of the spine, the importance of maintaining the posture, and the long learning-curve of spine specialty. Additionally, some aspects, such as heavy workload, fatigue, high misdiagnosis rate and missed diagnosis rate, are prone to occur in large-scale scoliosis screening and diagnosis. Researchers have discovered that AI can learn inherent laws and representation of sample data in recent years. AI technology has been used in clinical practice, such as screening, diagnosis, surgical decision-making, intraoperative operation, prognosis prediction and conservative interventions. However, AI technology has serious limitations currently. There are many disadvantages in diagnosis and treatment of scoliosis, including irregular data collection, technical difficulty, inherent defects of AI, overdependency on AI, legal and ethics issues. Thus, summarizing the relevant trends and application of AI in the diagnosis and treatment of scoliosis is still required by using the literature database and data-sharing network. The aim of the present review is to provide a glimpse into the future applications of AI.