The answer is yes, of course. Through artificial intelligence can realize bone age fast, accurate detection. Not long ago, the radiology department of the Children's Hospital of Zhejiang University School of Medicine had more AI bone age reading machines. After installing the relevant software, it can automatically recognize the X-ray film of children's hands and read out the bone age, and the whole process only takes a few seconds. According to professional doctors' assessment, the accuracy rate is very high.
Bone age detection blue ocean just waiting for AI to dig gold
Bone age because it can reflect the degree of growth and development of the human body, is the most basic pediatric diagnosis of a tool, in medicine through the detection of bone age can be diagnosed and monitor the children's endocrine diseases and growth disorders disease. Moreover, many parents believe that their children are short because of delayed development, and will naturally grow taller during the developmental period, but when they find out that their children have dwarfism, they have already missed the treatment period. In fact, the childhood stage is the best time to treat dwarfism, bone age test has a great guidance significance to the treatment.
With the increase in health awareness among parents, the number of children undergoing bone age tests is increasing day by day. However, at the same time, due to the shortage of specialized medical institutions and doctors, bone age testing has been in short supply and the market space is huge. Nowadays, more and more hospitals and Internet medical institutions are focusing their research on the field of artificial intelligence.
The software installed in the Children's Hospital of Zhejiang University School of Medicine is developed and completed by the hospital and EtuTech***. According to Fu Junfen, vice president of the Children's Hospital of Zhejiang University School of Medicine, with the help of the Children's Hospital of Zhejiang Province, more than 10,000 healthy children's physical examination of the bone age piece of data, ettu science and technology to use deep learning technology, trained this set of bone age reading machine. Weining Health also has a similar exploration, they use deep neural networks in line with the TW3 method of bone age characteristics of the region of deep learning, and then the bone age image features and clinical data fusion training bone age assessment model.
AI can make up for the shortcomings of traditional testing
Yang Xiujun, director of the imaging department at Shanghai Children's Hospital, said the application of AI in bone age testing will solve the actual pain points.
Clinically, bone age is usually determined by observing the development of each bone in the hand through X-rays of the left wrist. The GP method is the earliest and most complete method for identifying bone age, and it is mainly used by doctors to make visual comparisons based on a paper booklet of images against X-rays of the hand bones," said Mr. Fu, who is also a member of the team. This method is highly subjective, with poor clinical accuracy, and the results of the test are subject to considerable error. the TW3 method has significantly improved in accuracy, and it takes at least 15 minutes to detect a child's bone age, which may not seem like a long time for a single test, but combined with the fact that doctors are strapped for resources and the testing population is huge, the method is not able to satisfy the current market demand in China.
With the intervention of artificial intelligence, a bone-age test that is both accurate and fast has been introduced.
Automatically calculating children's bone age through intelligent film reading, this artificial intelligence bone age reading machine in Children's Hospital of Zhejiang University School of Medicine has significantly reduced the doctor's workload; in accordance with the algorithmic model integrating mainstream bone age standards such as GP/TW3, it can read each hand bone characteristics, avoiding errors in bone age diagnosis, with an accuracy of up to 0.1 years of age, and the doctor's calculation of the value of the bone age compared to the error less than half a year accounted for 98%. Ni Hao, president of Etu Medical, told reporters that the bone age reading machine shortens the original 15-minute chart recognition time to seconds.
Massive data may "feed" a new standard for bone age reading
Artificial intelligence bone age detection, big data is the core.
Lin Qiang, head of the bone age program at Etu Technology, believes that the core of medical artificial intelligence is high-quality labeled data. Take the bone age reading machine as an example, the basis of its detection is massive data, and high-quality labeled data is the material for algorithms to learn, summarize, and extract experience. "After the intelligent system has established the model, it has to be continuously improved and cultivated by manual means, so as to establish a complete medical knowledge architecture and diagnostic standards." Lin Qiang said that only through continuous learning can a set of highly accurate detection system be gradually formed.
From the current point of view, this black technology can already replace the past bone age detection method. Fu Junfen said that next, they will also verify and calibrate it among normal people in schools and other scenarios, and hope to expand it to various hospitals across the country.
The R&D team of the AI bone age reader also called for the development of a new standard for bone age interpretation in China in the future, which will be more far-reaching in value and significance.