South Korean startups have developed AI (Artificial Intelligence) solutions that read tuberculosis through X-ray images. It even developed X-rays for chest that are easy to carry and inexpensive, and released them all together. It is supplying lung tuberculosis deaths to developing countries such as Southeast Asia and Latin America to create a “K-Medical” Korean Wave.
Radisen (CEO Na Yoon-joo) has developed a small low-dose chest X-ray and a pulmonary tuberculosis reader. Chest X-ray devices are for medical checkups only. It has been reduced in size to suit outdoor or simple medical conditions. It is a low-dose equipment that consumes less electricity in preparation for poor electricity conditions.
The key point is software (SW). At the same time as chest X-rays, AI solutions have been introduced to determine whether tuberculosis exists or not. By using its own deep learning technology, the probability of filtering out patients suspected of tuberculosis has been raised to 98 percent. In Vietnam, where there are many tuberculosis patients as well as in Korea, thousands of video information were secured and learned. It is mounted on chest X-rays and operated in one unit.
“The existing AI reading solution learns the final results of X-ray imaging, but Radisen learns the final derived power data to increase sensitivity and accuracy,” said CEO Na Yoon-joo of Radisen. “We work with places with confirmed tuberculosis patients such as Vietnam National Pulmonary Hospital to increase product reliability.”
It was designed to target developing countries and underdeveloped countries from product development stage. Considering poor electrical facilities and inconvenient traffic, it was produced in low doses and small sizes. AI reading solution utilizes embedded database (DB) without Internet. Existing AI reading solutions are processed by accessing clouds or internal servers. It supports all readings with internal DBs targeting countries with poor Internet infrastructure.
Radisen was the only Korean company to participate in the 49th International Conference on Lung Health held in Hague, Netherlands from the 24th to the 27th of last month and introduced the solution to delegations from each country.
Starting with chest X-rays, the Ministry of Food and Drug Safety is applying for permission for medical devices. Although AI solutions will be used as diagnostic support tools for the time being, it is planning to obtain permission as SW for medical devices in the first half of next year. Targets are Southeast Asian and Latin American underdeveloped countries such as Vietnam and Indonesia.
According to the World Health Organization (WHO), 10 million new tuberculosis cases were reported in the world last year. Among them, 1.6 million people died. India has the largest number of new patients with 2.74 million, with 890,000 in China, 840,000 in Indonesia, 580,000 in the Philippines and 530,000 in Pakistan. As of 2016, the incidence of tuberculosis was 77 per 100,000, the highest among OECD members.
Tuberculosis is a fast infection but difficult to diagnose. Usually, after a chest X-ray test, if there is a suspected element, a precise test such as DNA extraction is performed. X-rays are not common in Southeast Asia, Central Asia, and Africa, and the cost is also very high. Even if it is difficult to film, there is absolutely no intention to read it.
“We have to pay up to $30 per chest X-ray and reading in underdeveloped countries and developing countries, and it’s a huge amount for them,” CEO Na said. “We will save time and money by providing cheap X-ray devices and AI reading solutions for up to 100 people.”