development status

Development of Medical Solutions Utilizing Artificial Intelligence (AI)

“Application of artificial intelligence (AI) to the medical field” is a medical theme with great potentials, but each stakeholder is facing challenges.

□ Doctors / Medical institutions (The entrance):

They are familiar with medical needs, abundant medical data and medical concepts but have limited AI knowledge and network with AI companies, which often prevent them from starting concrete development.

□ IT companies:

They are interested in entering the medical field, but have little network with doctors and medical institutions, and have insufficient understanding of medical needs and regulations such as Pharmaceutical and Medical Device Act. So full-scale development at the medical field is often difficult.

□ Pharmaceutical and health tech companies (The exit):

Although there is high interest in using AI, it is often difficult to handle everything from R & D to business development.

We are developing AI-based medical solutions by leveraging the perspective of focusing on understanding medical needs and development in the medical field, by networking with many physicians and clinical departments, and by utilizing the experience and know-how accumulated in investigator-initiated clinical trials. We are well-versed in the medical field and can collaborate with physicians and medical institutions that have a wealth of medical data. We can also develop medical solutions (diagnosis / treatment) that will be used in clinical practice as we can conduct clinical trials (investigator-initiated clinical trials) in accordance with the Pharmaceutical and Medical Device Act.

Currently, we are working on research and development of AI applications for medical fields including in lung disease (function test), kidney disease (dialysis medicine), and diabetes. Examples are shown below:

■ Development of diagnostic method for respiratory function test

Spirometry is a useful respiratory function test in diagnosing lung diseases, however interpreting the results can be difficult for non-specialists. We would like to provide solutions to assist non-specialists to interpret the results by making AI to learn patterns of correct answers from many test failures and many disease cases. In addition to the numerical results of the spirometer, we would also like to assist non-specialist’s interpretations by realizing the automatic analysis of the flow volume curve patterns. In July 2020, we have signed a joint development and commercialization agreement with a leading Japanese manufacturer, CHEST M.I., Inc. and distributor of medical electronic equipment for respiratory diagnosis and treatment.

■Development of dialysis management system for asymptomatic dialysis

Hemodialysis is a life-sustaining treatment for patients with chronic kidney failure. In dialysis treatment in patients with chronic renal failure, various complications such as a rapid decrease in blood pressure, poor mood, vomiting, and muscle spasm become problems. However, current technology cannot predict these events in advance, such as hypotension during dialysis, which occurs at a constant frequency. We would like to develop an AI technology to provide a safe dialysis management method that can prevent patients from those uncomfortable symptoms.
We are working with Tohoku University, dialysis clinics (St. Luke's International Hospital and several others), and a major national IT vendor to develop a system to use AI tools to predict events during dialysis. We are doing research and development by executing co-development agreements with Kanayama Clinic and Hirabari Memorial Clinic in April 2020 and with St. Luke's International University in August 2020. We added another group of hemodialysis clinics in December 2020 to start the joint research. In May 2021, we signed a collaboration agreement with Nipro Corporation for the development of the Chronic Dialysis Support System.

■Development of a Support System for Diabetes Treatment

There are more than 10 million people with diabetes in Japan, and it is already impossible for specialists alone to provide all the medical care to treat diabetes. In addition to multiple types of insulin drugs (injectable), there are numerous oral drugs for diabetes and their combination drugs. There are multiple same kinds of drugs are on the market for the treatment of diabetes. The combinations of these drugs are infinite, making drug selection a difficult task for non-specialists.
Therefore, we are collaborating with Tohoku University and major domestic IT vendors to develop an AI-based medical system that supports treatment by suggesting the most appropriately customized treatment options for each patient to assist in their treatment, taking into account their diabetes status (blood sugar control, exercise and diet), their status of visits to medical institutions, their financial situations, etc.