BioMind (COVID-19)

1. RT-PCR was initially considered the standard technique for diagnosing COVID-19. However, the long processing time has made evident the need for alternatives. Reagents are often depleted, resulting in a long lead time which has fueled the spread. Increasing evidence has proven the higher sensitivity of CT scans than PCR, this is a good complement. CT is also used to confirm dubious cases.

2. Our application uses deep learning trained on low dose NCCT scans. It is able to analyse CT scans in seconds, detect COVID-19 pneumonia, differentiate from non-COVID-19 pneumonia, perform triage and quantify disease progression. The intelligent report is then generated automatically for physicians to confirm.

3. As the virus is new, the ability to differentiate it from other types of pneumonia is difficult. Our application empowers physicians to make accurate decisions, facilitate early intervention and expedite patient management. This also helps to overcome manpower shortage.

1. Sudden Surge in Number of Cases with COVID-19

Time taken for nucleic acid-based laboratory testing to produce a diagnosis is long and unable to meet the needs of the increased number of cases with COVID-19.

2. Critical Need for Early Diagnosis
Given the highly contagious nature of the virus, there is a critical need for early diagnosis to prevent infecting a larger population.

3. High Number of False Negative Results
Clinical publications reported false negative nucleic acid test results, resulting in leakage to the community.

Radiological features corresponding to the COVID-19 pneumonia can be detected on CT scan accurately and rapidly with our AI solution, leading to early detection and intervention measures, directly combating the viral spread.

Our solution analyses the CT scans for COVID-19 pneumonia and generates a report for physicians. We use deep learning technology and trained on a huge collection of non-contrast computed tomography scans which were confirmed by RT-PCT tests. It is able to detect COVID-19 pneumonia, performs triage, quantitative assessment of the lungs and disease progression. The diagnostic report is then generated automatically for physicians to review and confirm. This application can then be expanded to other infectious diseases that we are developing.

Our application is targeted at the radiology department in hospitals. Radiologists can use our application to assist them to analyse medical scans and generate accurate reports. Our system can also be integrated into the existing PACS of the hospital while ensuring a seamless flow of information from the existing scanners to our prediction server. Report generated by our AI then goes back to the PACS, forming a closed loop.

Being a fast and accurate solution, BioMind (COVID-19) is also able to aid in the prevention of further spread of the virus as it has a much higher sensitivity and turnaround time as compared to the RT-PCT test, thus it can serve as a good complement.  

This application will help to speed up the diagnosis of COVID-19. This growth in demand for diagnosis is faster than the growth of newly trained radiology professionals to interpret the scans. The long turnaround time that PCR test kits need, i.e. 1-3 days,  is also unable to address the high demand for early and fast diagnosis. 

It is also reported that there are a high number of false negative results from using the PCR test due to low viral load in early stages and other issues faced with PCR processing. With the adoption of our application as a complement, accuracy of diagnosis of medical scans will be enhanced, resulting in fewer misdiagnosis and missed diagnosis, allowing early interventions and right treatment approaches for COVID-19. 

Lastly, urgent cases are also prioritized as it offers triaging of severe conditions.

Our application uses deep learning technology trained on a huge collection of non-contrast computed tomography scans which were confirmed by RT-PCT tests. 

This application has a high sensitivity and high specificity rate for COVID-19 Pneumonia detection.

Our plan is to scale our solution across countries that are affected by the COVID-19 virus. Subsequently a lot of learning, sharing and workshops will be conducted. Improved patient outcomes are shared, and best practises can be translated to new pilot sites. 

NA

Raymond is the CEO of BioMind® . He graduated from the prestigious National University of Singapore (NUS)’s Engineering Double Degree Program and has served as a scholar and technology advisor in big data engineering since 2008. He has advised over 50 organisations to achieve high growth rate and productivity driven by technology adoptions.

CTO Joe graduated from NUS with a Double Major in Physics and Mathematics. He obtained a Masters degree in Computer Science from McGill University. In 2013, he collaborated with Deep Learning Pioneers at University of Montreal, MILA lab. Previously from Microsoft Research and A*STAR, he has since worked with many scientists in the study of deep learning techniques. 

MIT CSAIL (Alliance Partner)
We are honoured to be an AI alliance partner of MIT CSAIL, whereby CSAIL is the most prestigious computer vision laboratory in the world. MIT also has a long heritage in AI since the 1950s, and is one of the world’s most important centers of AI research.

National University of Singapore (AI Research Partner)
A leading global university and ranked first in Asia Pacific, the National University of Singapore (NUS) is Singapore’s flagship university. It offers a global approach to education and research, and has a deep Engineering and Computing Science core since inception. We launched the world’s first Healthcare Machine Learning PhD Program with NUS, sponsoring talented individuals for PhD studies through our scholarships. Our scholars are sent for overseas for machine learning training once every year.

Hôpitaux Robert Schuman (European Flagship Partner)
In 2019, a MOU was signed between BioMind and the renowned Hôpitaux Robert Schumann in piloting BioMind’s AI solution in Europe. The launch was conducted in the presence of Luxembourg’s Deputy Prime Minister and Minister of Health, Etienne Schneider. 

Our company develops AI applications to address the needs of healthcare. These include AI applications in diagnosis, prevention, prognosis as well as patient rehabilitation. Our product includes BioMind, an AI diagnostic support software which uses deep learning technology to automatically analyse medical images for the brain, and BioMind (COVID-19) which offers AI detection of coronavirus on CT scans of the lungs. 

Our COVID application is able to assist Radiologists in the analysis of medical scans to pick up COVID-19 virus and aid in the report generation. It helps Radiologists to meet reporting time, reduce diagnosis errors and lessens their heavy workload to prevent doctor's burnt out. 

Finalists in our Challenges will pitch their solutions to a live audience of 400+ leaders and expert judges at Solve Challenge Finals in September during UN General Assembly Week in New York City. Those that are ultimately selected as a Solver will:

• Join a supportive community of peers, funders, and experts to help advance their innovative work through Solve's nine-month program;
• Receive mentorship and strategic advice from Solve and MIT networks;
• Attend Solve at MIT, our annual flagship event in May; and
• Receive access to more than $1 million in prize funding for the 2020 Challenges.

Are these reasons for our participation?

We hope to look for test bed partners (hospitals) and distributors with medical device experience to extend the reach of our AI solution.