Three AI tools revolutionising healthcare in Japan

The market for diagnostic and therapeutic AI healthcare tools in Japan is expected to be worth around $114 million by 2027. Startups founded by medical professionals are driving much of this momentum.

These entrepreneurial physicians know a lot about healthcare challenges. They are transforming the industry by using extensive disease data and their clinical expertise to teach new AI healthcare tools. And AI makes an excellent assistant for these physicians.

Here are three examples of how AI healthcare tools developed in Japan are revolutionising diagnostics for the country's doctors.

Endoscopes are used for early cancer detection and Japanese companies hold a 98% share of the global market for these tools. But diagnosing cancer from images captured by endoscopy is highly challenging – even specialised physicians typically require a decade of experience. But Japanese startup AI Medical Services (AIM) is harnessing the skills of these doctors through AI.

It has developed an AI tool that analyses images of patients’ stomachs and colons captured by endoscopy to indicate their likelihood of developing cancer in the future. The endoscopic AI has been trained with more than 200,000 high-resolution videos of the stomach and other organs collected from more than 100 medical institutions across Japan. As a result, it can analyse still images and videos in real-time, helping physicians to confirm suspicious areas while they conduct endoscopic examinations.

Doctors perform thousands of image diagnoses daily to identify cancer cases, but around 20% of early-stage stomach cancers are still overlooked. This tool aims to reduce the workload associated with diagnosing using images and data, and improve accuracy. It can analyse a single image in just 0.02 seconds – much faster than the four seconds specialist physicians typically take. It can also determine the presence or absence of cancer with approximately 94% accuracy.

The final diagnosis remains the responsibility of the physician, however. Dr. Tomohiro Tada, the founder of the company and a gastroenterologist, emphasises the collaborative aspect: "the combination of human and AI inspections can enhance the accuracy of cancer detection". Recognising its innovative contributions to addressing crucial global challenges, the World Economic Forum selected the company as a Technology Pioneer in 2021.

When looking for signs of flu, a doctor usually examines distinctive swellings in the throat, known as follicles, which develop when infected. Iris, a Japanese startup aims to replicate the expertise of skilled physicians by training AI with over 500,000 throat images, supplementing the visuals with information from a patient questionnaire to provide a diagnosis.

Its AI medical device, nodoca, analyses throat images, body temperature and self-reported symptoms to diagnose influenza within seconds. Unlike traditional nasopharyngeal tests, nodoca doesn’t require a sample collection from patients, ensuring a pain-free experience.

The company was founded by Dr. Sho Okiyama, an emergency medicine specialist who wants to bridge the technological gap among physicians using the power of AI.

A “super stethoscope” developed by another startup, Acute Medical Innovation (AMI), aims to help physicians diagnose both the presence and type of heart disease in patients.

Using a sensor applied to the chest for just 10 seconds, the super stethoscope digitises heart sounds and electrocardiograms, transmitting the data. This opens up the possibility of using AI healthcare tools to perform remote diagnosis for online medical care.

Even as the development of AI medical devices accelerates, lengthy approval processes for practical use are a significant hurdle in Japan. As of February 2023, approximately 30 AI medical devices have obtained pharmaceutical approval and certification in the country. But getting approval to use new AI healthcare tools can take over a year. Each upgrade of a device requires a new round of scrutiny, even after the approval.

It's also difficult for healthcare providers to get insurance when using AI medical devices. If public health insurance doesn't apply to AI devices used in medical practices, healthcare organisations cannot receive reimbursement from the public insurance system for the services they provide.

A consortium of 16 companies has been formed to urge the Japanese government to establish a review and approval scheme for AI medical devices, as well as expedited approval processes. The government has also launched the DASH for SaMD strategy to expedite approval and strengthen support for AI healthcare tools.

Public-private collaborations are also encouraging the use of more AI healthcare tools by Japan's medical industry. In 2017, the Ministry of Health, Labour and Welfare identified six focal points – including diagnostic imaging and medical product development – for the introduction of AI tools that could optimise healthcare.

The next year, the government launched an AI Hospital scheme, a public-private project involving five companies, including Hitachi, IBM Japan and SoftBank. They're working on initiatives such as establishing a highly secure medical information database and developing systems for AI-assisted informed consent and voice recognition for medical record updates.

The progress made by such collaborations and by Japan’s medical industry aligns with wider moves to make AI healthcare tools more accessible globally. In October, for example, the World Health Organization introduced new guidelines to ensure ethics and human rights are key considerations in the design, deployment, and use of AI.

Such moves could help boost the momentum of AI use in healthcare. For Japan, this could help to address the country’s chronic workforce shortages, aging population and the regional disparities in healthcare. This would certainly revolutionise the future of healthcare.