Medical professionals have identified the scope of improving success rates and quickness in treatments while reducing mortality with the help of AI tools. This opens up opportunities for AI solution providers to find a high demand market.
In August 2018, Microsoft and Apollo Hospitals announced a collaboration in healthcare. To predict the risks and patterns of cardiovascular diseases (CVDs) in Indian citizens, they leveraged on artificial intelligence (AI) to create a CVD Risk Score API platform. The CVD vulnerability rate, causes and patterns in the Indian population are quite different from the rest of the globe. The AI-powered API helps predict these risk scores and patterns based on parameters like diet, tobacco consuming and smoking habits, body fitness, psychological stress and others, precisely as per Indian conditions. This allows hospitals to plan diagnoses and treatments in time for incoming patients.
This is a small example of how AI and cognitive technologies are creating a shift in healthcare today.
Manipal Hospitals are using IBM Watson cognitive platform in their oncology department for faster and better treatments. Doctors are looking at smart technologies for various reasons—be it for more accurate diagnoses, quick reference to remedies or for consultation on-the-go. We delve deeper into this and look at solutions currently being used.
Real instances of AI in healthcare
The main expectation of healthcare specialists from AI-based medical approaches is to promote precision treatment and, in turn, improve their success rate with regards to treating patients. To provide treatment suggestions and success predictions, cognitive platforms feed on a variable array of data across many medical cases, including medical histories, genomics of patients, treatment processes, medicines and cures used, patient’s behavioural patterns, financial estimates and more.
Keeping this in mind, AI is also being looked at as a tool to find, or at least support in finding, cures for difficult diseases like different forms of cancer, Fragile X syndrome, AIDS and so on. For example, England-based startup Healx is working on drug delivery for rare diseases through their proprietary AI-based platform called HealNet. The five-year-old company—with Dr David Brown, co-inventor of Viagra, as one of its chiefs—has seen success in Fragile X treatment during an ongoing project where the aim was to identify, validate and repurpose approved drugs for the disease, using AI, while reducing production costs.
Fragile X is a genetic disability that leads to various behavioural, developmental and intellectual limitations in a person. So far, no proper medication or cure has been found.
With the help of HealNet platform, eight drugs were identified as candidates, and in a span of six months of experimentation on Fragile X-affected mice, three of the drugs showed substantial improvement. Studies were carried out based on four behavioural parameters in mice, namely, open field venturing, nesting, fear conditioning and sociability. One of the three drugs has been selected for further in-depth studies.
In another development last year, Department of Dermatology at Dr D.Y. Patil Medical College, Pune, in collaboration with US-based AI solution provider Polyfins Technology, developed an AI-based platform for treating skin and hair disorders. The outcome of the project is an application called Tibot, which can diagnose 12 high-level dermatological disorders, including bacterial or fungal infections, tumours, eczema, alopecia and psoriasis, among others, in addition to 90 secondary conditions.
The machine learning platform uses data based on images of the affected skin area and questions answered by the user on the app, to predict top-three skin conditions the user may be suffering from. It provides confidence score and urgency index for every analysis, enabling the user to prioritise the ailment that requires to be diagnosed and treated quickly. Dr D.Y. Patil Medical College is using the platform in its dermatology section, headed by Dr Sharmila Patil, for trials, and is receiving good results.
More research is on the way that will soon have real-life use cases. Researchers at IBM have come up with a small sensor with a microprocessor that can be attached to the fingertip to predict patients’ health deterioration. It has received success in Parkinson’s detection so far. The sensor can also be used to measure grip strength or grasping forces, to detect users’ emotional and behavioural patterns.
The Industrial Technology Research Institute of Taiwan has developed an AI platform that, based on image analysis and data mining, can detect the stage and complexity of diabetic retinopathy—a disease where long-term diabetic patients are at risk of vision loss. The platform is set to be utilised in hospitals across Taiwan to evaluate the effectiveness of the solution. The team behind the platform plans to include cloud technology in the platform, which will also allow doctors to provide remote consultation.
Diabetic retinopathy can be analysed using AI
Opportunities in hand
Medical professionals have identified the scope of improving success rates and quickness in treatments while reducing the mortality rate with the help of AI tools. This opens up opportunities for AI solution providers to find a high demand market. A 2018 report by Frost and Sullivan suggests that AI for healthcare will globally become a US$ 6.16 billion market by 2022, with a CAGR of 68.5 per cent. It will bring about savings of US$ 150 billion for the healthcare industry. However, only 15 per cent of healthcare customers have adopted AI properly, which leaves a massive opportunity of improvement to be filled by healthcare businesses, and a huge demand for AI developers.
India has a strong software base and is growing its embedded hardware capabilities, too, promising a good scope for creating the desired AI-based healthcare ecosystem. How the players seize the day will be an important factor.
