Not every doctor has enough years of practice and experience to diagnose a disease. A neural network with access to scientific literature and millions of case studies can quickly classify a case, relate it to a similar one, and generate proposals for a treatment plan. Currently, AI algorithms cannot solve complex tasks that are beyond the capacity of a doctor: for example, diagnosing a person and prescribing an effective treatment. However, relatively simple tasks can be solved: for example, defining whether there is a pathology in the X-ray image or whether there are cancer cells in the cytological material. At the same time, the accuracy of medical data assessments for MRI, ultrasound images, and mammograms  already exceeds 90%.

IBM Project: Watson

The most famous example of introducing AI into medical diagnostics is the IBM Watson system. In 2015, IBM even created a separate division – IBM Watson Health, which is engaged in the application of AI technology in health care. The Watson computer has access to various data sources: encyclopedias, databases of scientific articles, and knowledge anthologies. Thanks to its enormous computing power, IBM Watson was able to analyze 30 billion medical images and 50 million anonymous electronic medical records.

 Initially, IBM applied AI for oncology. However, recently, Watson developers began to cooperate with the American Heart Association. Now, according to the ultrasound imaging data, the cognitive platform can look for signs of stenosis of the aortic heart valve (a very common type of heart disease).

Watson Health technologies are available through the Watson Health Cloud cloud platform. It is intended for doctors, researchers, insurance agents, and medical companies.

Google Projects: DM Health and Verily

Naturally, IBM is not the only technological giant that takes advantage of AI solutions. A division of Google – DeepMind Health – uses technologies of the IT giant in medicine. Today, DM Health is working with the Moorfields Eye Hospital, an eye clinic in London. With the help of AI, doctors want to analyze thousands of anonymous eye shots to find the primary symptoms of blindness. In addition, DM Health is developing a product that can automatically distinguish between healthy and cancerous tissues in the head and neck area. 

Verily, formerly a division of Google X and now an independent subsidiary of Alphabet Inc, is engaged in a similar project. The company uses AI and Google search engine algorithms in order to analyze what makes a person healthy.

 

Israeli startup: MedyMatch Technology

According to statistics, the number of errors in making a diagnosis from computed tomography data exceeds 30%. It means that the wrong treatment plan is prescribed for  every third patient.

Based on AI and big data, the Israeli company has developed a solution by which doctors can more accurately diagnose a stroke. The MedyMatch system compares a snapshot of a patient’s brain with hundreds of thousands of other snapshots in its cloud. MedyMatch’s AI recognizes the smallest deviations from the norm, which the specialist is not always able to notice, thus minimizing the probability of error in diagnosis. MedyMatch employs only 20 people.

Los Angeles based startup HealPoint.io also utilizes AI to minimize diagnostic errors by generating doctor consensus based on conclusions submitted by multiple doctors. Ethereum blockchain is used along with AI to do that.

 

Market Assessment and Perspectives of AI in Healthcare

 

According to the Frost & Sullivan consultants, revenues of companies in the medical AI market will reach $ 6.1 billion by 2021. IBM’s market share will be about 45%. Frost & Sullivan drew up a roadmap for the development of the industry for several years ahead. By the early 2020, AI systems will begin to offer medical care to patients based on data that both patients and doctors will have access to. Chronic diseases such as cancer and diabetes will be diagnosed within a few minutes using cognitive systems that visualize the physiological characteristics of the scan of the human body. Optimists from F & S suggest that by 2025, AI systems will be involved in 90% of US clinics and approximately 60% of the world’s largest hospitals. Experts hope that 70% of the world’s population will have access to AI powered healthcare solutions.