On May 10, an AI medical software aimed for X-ray image-based Bone Fracture AI-assisted diagnosis by a Chinese company—Huiying Medical Technology Co., Ltd. (HY Medical) was approved with Class III medical device certificate issued by the National Medical Products Administration of China (NMPA), which is the first approval of similar products in China and the only X-ray fracture AI product dually certified by CE and NMPA.
Misdiagnosed fractures are the most common Emergency Department incidents. Missed fractures often occur due to interpretational errors, resulting in treatment delays leading to poor outcomes. This has been listed as one of the leading causes of malpractice claims against physicians.
Auxiliary Interventional Fracture Diagnosis
How does AI empower?
Clinically, the gold standard for fracture diagnosis has been X-ray imaging equipment. Digital Radiography (DR) can show the change of the whole external structure around the lesion, with high spatial resolution and low price, significantly superior to other imaging examinations such as CT in the identification of fractures.
Under normal circumstances, well-trained doctors can basically determine fracture condition through DR images, still with a risk of missed diagnosis. According to statistics, the missed diagnosis rate of X-ray fracture is 15%, especially in the limb considering the joints.
Dr. Li Aiyin, chief physician of the Radiology Department of Qianfoshan Hospital in Shandong, China, commented that DR film, as an imaging of the overall structure, occasionally has an overlap of the structure of the fracture and normal site, affecting the judgment.
As a radiologist, in his 21 years of practice, he has seen many cases of missed diagnoses which nearly led to accidents and doctor-patient disputes. "There was a patient who came to the hospital at the night clinic complaining of ankle pain, and the doctor examined the patient based on the patient's description of pain, while the radiologist only examined the ankle joint for observation after taking the film. As a result, the patient's fracture site was in the tarsal bone while there was no obvious structural overlap, so it was a missed diagnosis." Aiyin said, “We were able to communicate with the family in a timely manner, but fortunately there were no serious consequences and we finally got the understanding of the other party.”
"This is a very typical case which reflects the current problems in the imaging diagnoses of emergency visits in radiology departments consisting of vague patient complaints, inexperience of junior doctors, and physical fatigue of night shifts can also affect the concentration of doctors."
So, is this “New Thing” helpful? Reliable?
In order to alleviate the work pressure of doctors, make up for the defects caused by inexperience, and improve the accuracy of fracture imaging examination, Qianfoshan Hospital has introduced the fracture software of HY Medical in 2020.
However, many orthopedic doctors and radiologists, including Aiyin, still have doubts about the reliability of this "new thing" after the product has been introduced.
Regarding concerns, this hospital conducted a trial specifically to re-perform AI-assisted screening of past fracture DR films through monthly case tracing, later comparing them with previously issued reports.
As a result, Aiyin and colleagues were surprised to find that in more than 30 DR films that could not be accurately judged by young doctors, this software accurately labeled the lesion site, including the cases that nearly caused disputes. This led to the realization that the application of AI technology in the medical field will bring great help to the clinical work of doctors.
Upon the completion of DR films, the AI software will directly issue a report in the background, marking all suspected fractures in red, only requiring seconds. Once completed, doctors will make a second judgment on the marked area, and finally issue a report.
"Another highlight of this product is that it basically covers over 80% of the body (hand, foot, wrist, shoulder, hip, ankle, femur, hip, pelvis, radius and ulna, tibiofibula, etc.) in terms of bone examination, which will be especially useful for patients with suspected multi-site trauma.” Aiyin stated that, “If in the case of a single fracture, AI is merely a ‘patch’ for a junior doctor. In the case of a multi-site injury, it directly identifies the suspected fracture all over the body, relieving the time and effort that would otherwise be required to screen each site in detail.”
Aiyin believes that based on the clinical advantages of this software, it can become the most powerful assistant for doctors. Especially in primary hospitals, AI software can be seen as the "semi-instructor" for junior doctors. Presently, the automatic identification rate and sensitivity of this product have exceeded 95%.
Today, HY Medical’s fracture software has been deployed in more than a hundred domestic and overseas hospitals.
References:
(1) https://www.nature.com/articles/s41746-020-00352-w
(2) https://www.researchgate.net/publication/349469430_Artificial_intelligence_application_in_bone_fracture_detection
(3) https://www.ncbi.nlm.nih.gov/pmc/articles/PMC1386703/
Company Name: Huiying Medical Technology Co., Ltd.
Contact Person: Claire Zhang
Email: zhangfengyan@huiyihuiying.com
Country: China
Website: http://en.huiyihuiying.com/