Infervision: Using Medical AI and Deep Learning to Battle COVID-19

With the recent COVID-19 outbreak, the huge surge in demand for medical services has presented the Chinese healthcare system management with unprecedented challenges. Infervision‘s newly launched InferRead™ CT Pneumonia uses Artificial Intelligence (AI) to support doctors on the front lines of the virus outbreak to diagnose infected people with urgency. By calculating the volume, density and distribution of suspected infections in less than 10 seconds as opposed to the 15 minutes required by a manual read of a CT scan, this AI solution has detected over 32,000 novel coronavirus cases from more than 34 hospitals so far. Infervision is fully committed to supporting doctors and patients in managing the pandemic and it is bringing us one step closer to control the COVID-2019 outbreak.

 

Infervision and HIEx

The collaboration between the Health Innovation Exchange (HIEx) and Infervision started over eighteen months ago and has generated several partnership opportunities for health impact. Participation at the HIEx Innovations Marketplace, on the margins of the World Health Assembly in May 2019, helped Infervision meet several Ministers of Health and to better understand the gaps and challenges in health care, particularly in Africa.  The potential role that AI could play in addressing the urgent challenges became clearer when the HIEx facilitated a visit of innovators to Ethiopia where the team met with the Minister of Health and several health care providers. This resulted in an official partnership between the Ethiopian government and Infervision in building a national AI infrastructure to support healthcare, which was then extended to include other areas such as education, trade and agriculture. 

Other exciting opportunities emerged with the new year when HIEx invited Infervision to participate at an event in Davos in January 2020 at the World Economic Forum. Infervision shared its initial efforts to repurpose its AI for TB work to address the rapidly emerging COVID-19 infections in China.  HIEx helped disseminate the information about the role of AI in combatting COVID-19 and helped Infervision in articulating the need for using new technologies in developing a broader toolbox in responding to COVID-19. HIEx has been working closely with Infervision to share the lessons from the use of AI and has helped reach key partners who are developing strategies in responding to COVID-19. 

 

The Health Innovation Exchange has provided a very good platform for us to communicate and coordinate

wisdom so that we can actually provide better technologies for the society"

 

 

Kuan Chen, known to his friends as CK, is the founder and dynamic CEO of Infervision. He is committed to SDG3 and wants to use his expertise in AI to impact lives of people around the world. If you want to know more about Infervision, read Kuan’s story below!


What are the main challenges in the Chinese healthcare system?

Since a very young age, I have been exposed to the challenges of the Chinese healthcare system as my mother’s family has a lineage of being barefoot doctors in rural China, while my father almost died of infection as an infant due to lack of medical resources in the great Chinese famine.  I could see that if in big cities hospitals are often confronted with huge demand from patients, working long hours under very stressful environment, in rural areas, trained physicians are in short supply and price of medical services can be very high. 

So, after studying for some years in the US on how to use Machine Learning and AI modelling in Economics and Finance, I dropped my PhD to apply the nascent but powerful AI technology to areas closer to my heart. This is when I went back to China and started Infervision. Although the development of the initial prototype was quite challenging as no hospitals in China had heard of AI and Deep Learning, the potential of AI in healthcare was soon recognized. Infervision, is an AI software for image analysis for automated identification and diagnosis based on chest CT scans. This technology has resulted to be particularly effective in the early detection of infectious diseases like TB as well as it has been used for cancer prevention due to its ability to detect early lung nodules. Today, after 5 years, Infervision is being adopted by 300 hospitals, conducting AI-enhanced diagnoses of more than 50,000 cases per day in 9 different countries.


What do you see as the role of Artificial Intellingence (AI) in Healthcare?

In order to understand how AI works, it is useful to think of AlphaGo. It is a computer program that learns from millions of chess games, especially by the top chess players, so that it can learn to play professional chess and ultimately becoming better than any chess players in the world.

Similarly, but with the objective to improve medical imaging analysis, Infervision uses machine learning to learn from millions and millions of past cases from the top physicians and top hospitals, so that it can learn to detect, measure, evaluate, diagnose, treat just like top physicians. After the initial learning process, AI can be used in any setting, where soft medical capabilities are most needed tackling the problems of excessive workload for healthcare workers and doctors in some areas and the lack of adequate services in others.

A great strength of AI is its ability to continuously adapt and learn. In fact, as long as AI has learned in 1 hospital, it can bring improvement and enhancement to all hospitals with AI deployment, breaking down the geographic boundary of capability improvement. AI adapts as the virus or infections adapt and helps us keep one step ahead.
 

What has been Infervision doing in responding to the COVID-2019 outbreak?

Having tested Infervision with different diseases, since the early stage of the coronavirus outbreak, we started working with Wuhan Tongji Hospital, Shenzhen Third Hospital and the Wuhan Zhongnan Hospital to implement an AI solution for the coronavirus pneumonia. During the entire Chinese New Year vacation, our team worked around the clock to develop the technology with the hospitals. We finally made it work in the hospital and are deployed our AI across several hospitals in China and now in other countries.
 

 

Infervision and broader AI has multiple benefits in the fight against this current outbreak:

1. Cross-infection is a huge problem in this current outbreak, a recent article from JAMA states that 41% of infection within Wuhan Zhongnan Hospital are hospital-related transmission, of which 29% is hospital staff. From an article a few days ago in China, more than 1000 hospital staff in the city of Wuhan has been confirmed infected. From a lung CT scan, AI quickly detects lesions in the image of possible coronavirus pneumonia, measures its volume, shape, density. While a manual read of a CT scan can take up to 15 minutes, AI can finish reading the image in 10 seconds. It speeds up the diagnosis and gives a fast warning to the doctors if AI finds a probable case, where immediate quarantine action can be taken to prevent further cross-infection. 
 

2. Diagnosing and monitoring patients’ progress accurately can be challenging and time consuming for doctors, in very intense outbreak environments. AI can automatically measure and compare changes in the lung lesion from the image and give out a quantitative report to assist doctors making judgements.
 

3. The testing process in the outbreak epicentres, like Wuhan for instance, where there is a high prevalence of infection, can be difficult as the PCR testing can take up to a week to receive the confirmation. In this context, CT+AI serves as a surrogate for doctors when fast judgement is needed. CT examination with the help of AI can identify high-probable cases quickly and shall be the surrogate gold standard for decisive action, while waiting for PCR confirmation.

The implementation of new healthcare technologies is showing great potential to improve pandemic management and maximize the use of human and financial resources. The speed, sensitivity and adaptability of AI are essential features in improving the screening process of patients in a context of an epidemic outbreak. Infervision and its team is on the front line to support medical professionals and patients in controlling the COVID-2019 outbreak.