Told by 12Sigma (图玛深维) CEO ZHONG Xin (钟昕), 12Sigma would explore and develop cloud in 2019. Introduced in previous articles, 12Sigma will extend its business to North America and Europe this year while it already has a strong R&D relationship with research institutes and universities in these regions.
Healthcare resources in China are distributed heavily in top-tier cities and the unbalanced distribution causes that other cities’ healthcare services are far from expectation. China’s hospitals and clinics are categorized into three major grades and in a total of 10 levels (三级十等). Third-grade Class-A (三甲) is the highest level in China’s hierarchical medical system that has the best healthcare resources among all hospitals.
Third-grade Class-A hospitals have the most experienced doctors and the best facilities. Reported by National Health Commission of the PRC, the total hospital number was 32,476 by the end of November 2018, in which 12,072 hospitals were state-owned and 20,404 were privately owned. Among these hospitals, only 2,498 were valued as Third-grade Class-A hospitals.
From the Third-grade Class-A hospital number, it took only 7.7% of the total hospital number. Nevertheless, the patients that Third-grade Class-A hospitals accepted in 2018 Q3 was approximately 51.7% of the total visits – 463.4 million visits out of a total 897.0 million visits were accepted Third-grade Class-A hospitals. Mentioned by Zhong Xin, 12Sigma’s potential hospital clients are majorly the Third-grade Class-A hospitals. Third-grade Class-A hospitals took the main responsibility in healthcare services even though the size of them do not match with flows of accepted patients. For AI medical image reading service, it aims to assist radiologists to simplify the image reading process and relieve the work stress.
In non-Third-grade-Class-A hospitals, the healthcare condition varies. Medical imaging units include X-ray machines, computed tomography (CT) units, magnetic resonance imaging (MRI) units and so on. The x-ray machine is the mostly-equipped machines in China’s hospitals, but only top-class hospitals can afford for CT and MRI units, especially MRI units due to the high costs. AI medical image reading technology lies in the data generated from these medical image generators. The scarcity of medical image generator in lower-class hospitals makes the AI technology difficult to show its powerful function.
The retained quantity of MRI machines in China was only 8,289 by 2017. Statista’s 2017 data of MRI units indicates that the U.S’ number of MRI units was as high as 37.56 per million population. China’s number of MRI units per million population was only 5.96, which is only 15.9% of the number of the U.S.
MRI unit is one type of the medical image generators and cannot represent the overall situation of medical image generator in different countries, but it can provide some information about the healthcare service level in the country. China’s healthcare service level is yet to be developed. If China could reach the level of U.S. MRI units per million population in 2017, the MRI units would increase to 52,208, which is 629.9% of 2017’s number and the medical images would accordingly increase based on the increase of MRI units. The potential growth of medical image generators (not only MRI units) signals the potential future of AI medical image reading’s market.
Shortage of radiologists and machines in towns and small cities might result in a blank in the medical image diagnosis market, but the need is still there, especially in an aging society. With optimistic thinking, the medical image cloud service can relieve the restriction caused by radiologist shortage – towns and small cities can purchase medical image generators while no experienced radiologists are in place if the hospital has technicians that know how to take medical pictures.
Medical image cloud service allows facilities to upload medical images to the storage cloud, and the AI medical image reading service providers can use their technologies to assist radiologists to diagnose based on the medical image stored in cloud, which might be sent from a hospital in rural area that no one knows how to interpret the image. With the medical image cloud, radiologists can offer telediagnosis service. The combination of 5G network and cloud computation might even be able to actualize the streaming of telediagnosis process between the radiologists and hospital ignoring the geographical restrictions.
To develop cloud service is a significant task for 12Sigma in 2019. The digitalization in the healthcare area has been going on through years in China, but a systematic database that can integrate all hospitals’ digital data is yet to be built. Currently, hospitals and even within hospitals, different databases are used and the communication among these databases breaks down – databases are independent and hence the restructuring process of the database becomes an impossible task if there is no intermediator in the process.
The government is pushing and encouraging the standardized digitalization process of healthcare data. 12Sigma’s plan on medical image cloud might be benefited by the preferred policy while the cloud service allows it to expand its business into a broader scale.
Scoping on the MRI units per million, China has a tremendous space to grow and this will apply to other medical image generators like X-ray machines and CT scanners, though these two might not have similar growth potential as MRI machines do. The growth of medical image generators will contribute greatly to the total number of medical images generated and the booming of the quantity will inversely push the development of medical image cloud since the population growth of radiologists cannot catch up with the growth of image produced.
Adding the aging issue of China, the healthcare service will enter a fast-growing status and the demand for medical image diagnosis will correspondingly be stimulated. Cloud service and AI-assisted diagnosis are critical for medical image vertical. With other technologies’ support like the 5G telecommunication technologies, the medical image diagnosis process can be improved in speed, accuracy, and even affordability.