Medical AI pioneer VUNO sets sights on global frontier

Industry

Published: 21 Apr. 2024, 16:06

Medical AI pioneer VUNO sets sights on global frontier

VUNO CEO Lee Yeha poses for a photo during an interview with the Korea JoongAng Daily at the company's headquarters in southern Seoul. [PAKR SANG-MOON]

Game Changer

Hospitals are busy — a given regardless of season or current events — with doctors and nurses tirelessly attending to patients around the clock.

With a chronic medical staff shortage persisting against the backdrop of the world's lowest fertility rate, things are likely to get even more hectic for Koreans with a healthcare crisis on the horizon.

This is where artificial intelligence (AI) comes in as the ultimate sidekick, says Lee Yeha, CEO and co-founder of VUNO.

From predicting potential heart attacks to helping diagnose Alzheimer's disease, VUNO's medical AI software has already been deployed in over 80 hospitals across Korea.

"It is nearly impossible for healthcare workers to keep an eye on every patient at every moment, and therefore, the statistics show that a cardiac arrest occurring in a regular ward leads to death in most cases," said Lee.

"With VUNO's deep learning-based software detecting patients with high risks in advance, [hospitals] will be able to use their resources more efficiently and appropriately to provide better care for patients."

VUNO, a Seoul-based medical AI startup, is making quick strides in the market as one of the pioneering players. The company achieved many "firsts" since its foundation in 2014: the first to receive approval from Korea's Ministry of Food and Drug Safety in 2018, the first to get a green light from the European authorities with Conformité Européenne (CE) certification in 2019 and the first to be certified as an innovative medical device company in Europe in 2020.

VUNO’s products “can discern which patients are at greater risk and help medical workers care for patients even amid a staff shortage in the essential healthcare sector,” Lee said.

Its VUNO Med-DeepCARS, for example, predicts a patient's risk of cardiac arrest within 24 hours based on vital signs. It was named the first “breakthrough device” among medical AI devices by the U.S. Food and Drug Administration (FDA) in June of last year.

The Kosdaq-listed company reported a 60 percent on-year jump in revenue to 13.3 billion won ($9.6 million) in 2023. VUNO is expecting annual turnaround in 2025 and aims to achieve a quarterly operating surplus in the latter half of this year.

The goal this year is to double its revenue, mainly on the back of overseas expansion. The U.S. market is its prime destination.

In October of last year, its VUNO Med-DeepBrain earned FDA approval. It parcels the brain into more than 100 regions to provide quantified information on the degree of atrophy within a minute, which helps doctors diagnose patients with dementia stemming from major degenerative brain diseases, such as Alzheimer's disease.

Medical imaging of VUNO Med-DeepBrain [VUNO]

VUNO recently clinched a supply deal for its Med-Chest X-ray with a major Spanish X-ray device manufacturer, further accelerating its penetration into the European market.

The Korea JoongAng Daily sat down with Lee at VUNO’s headquarters in southern Seoul to discuss the company’s journey so far and its plans for the global expansion.

This interview has been edited for length and clarity.

Q. VUNO has been a pioneering figure in Korea's medical AI sector. Can you walk us through the journey of developing the company's deep learning model?
A. Before we founded VUNO, I worked as a researcher at Samsung Electronics’ Samsung Advanced Institute of Technology for around three years. Based on my experience there, I decided to start a business with the belief that deep learning technology will be able to be of significant value in various sectors.

By the time we had to come up with a deep-learning engine framework, there was no framework such as TensorFlow or PyTorch widely available like there are now, so we had to develop our own — VunoNet.

Of course, these days we are using other development tools as well, but because we have previous experience and technological capabilities from establishing our own deep-learning engine framework, we are more capable of bringing products to a commercial level or optimizing devices.

Can you give an overview of VUNO’s products?
We are focusing on medical imaging tools and biosignals.

For the former, we have solutions that can analyze medical images such as X-rays, CTs and MRIs to discern any abnormality to help improve diagnostic accuracy. There are six products, including VUNO Med-BoneAge, which helps assess a child's bone age.

Regarding biosignals, our product analyzes data from vital signs including breathing, blood pressure, pulse and body temperature to assess the possibility of a sudden deterioration in condition or cardiac arrest.

VUNO Med-DeepCARS, which predicts possibility of an acute deterioration in patient conditions based on biosignals [VUNO]

What is important in medical practice is a tool's clinical effectiveness, such as how well medical professionals are able to improve their diagnostic accuracy using our products and whether that leads to better patient outcomes.

For example, our VUNO Med-DeepCARS analyzes the biosignals of a patient in a regular ward. It predicted an acute deterioration in condition with an 87 percent success rate according to our studies conducted at four hospitals. The studies also proved that the accuracy of the predictions remains similar regardless of the patient's age or sex.

How do VUNO’s products contribute to mitigating the healthcare worker shortage?
At intensive care units, medical professionals are able to keep track of each patient up close. But in regular wards, doctors and nurses check on their patients once every eight hours, so about three times a day.

If there were a lot of medical staff that could keep an eye on every patient, there would be no problem at all. But with the healthcare worker shortage worsening and the demand for healthcare services growing due to aging demographics, it is nearly impossible for healthcare workers to keep an eye on every patient at every moment.

With VUNO's deep learning-based software detecting high risk patients in advance, hospitals will be able to use their resources more efficiently and appropriately to provide better care for patients.

Moreover, DeepCARS can also be used for pediatric and adolescent patients, meaning that it can help medical professionals provide better care despite the ongoing shortage of pediatricians, which has been particularly severe.

VUNO has attended various global conferences, including the annual meeting of the Radiological Society of North America (RSNA) held in Chicago last year, where you delivered an on-site presentation. Has there been any change in the industry that has stood out to you over the years?
We have attended the RSNA conference every year since 2015, and the response has steadily been getting better.

I think it is not solely about VUNO's performance but rather the endorsement and perception of AI technology advancement in radiology, clinical studies and such. With the progress of research and the adoption of technology in actual clinical trials, the industry is accumulating empirical evidence that the technology helps medical staff in practice and improves patient outcomes.

As for VUNO, we got our first FDA approval in 2023 — for VUNO Med-DeepBrain — which led to bigger interest in our other products. Our overseas revenue in 2023 doubled from a year earlier, hitting new milestones one by one. It was great to discuss more solid business opportunities with potential partners and also talk about the actual deployment of the FDA-approved product at local hospitals.

VUNO was granted approval for DeepBrain by the FDA last October. What’s your plan for global expansion, especially for U.S. market penetration?
The plan is to gain FDA approval for all of our other products as well, perhaps one or two of them every year. We hope to launch the product as early as the first half of this year, or maybe by the third quarter.

The ultimate goal, obviously, is to expand overseas revenue to an extent that far surpasses that of domestic revenue, considering that the U.S. market is about 10 times the size of Korea's. It is the same for European nations, Japan and the Southeast Asian region as well.

We doubled our overseas revenue in 2023, and hope to keep the pace in 2024 as well. Given that growth is much steeper in our global sales, overseas revenue will likely catch up with domestic revenue in about two to three years and eventually outpace our business at home.

What is VUNO’s strength over global competitors?
There are a lot of medical AI companies in the United States and also in the European region. Our competitive edge against those companies, in my opinion, comes from Korea's research environment — favorable for developing AI-based medical software. Medical professionals in Korea are very competent, and the medical services are of high quality as well. This means that the data sets available for deep learning are of high quality.

Moreover, it was very convenient to collaborate with medical professionals because major hospitals in Korea are located close to each other and the level of digitization was very high.

Thanks to such factors, Korean companies including ourselves could get an early lead and quickly advance in the global market.

Furthermore, nobody will ever believe in a medical AI device's performance if there is no evidence besides lab results, no matter how good they seem, so it is important to accumulate as much real-world data as possible, which Korean players did well.

What is VUNO’s ultimate vision?
Our ultimate goal is to help people live healthier lives using technology. Doctors and hospitals are there to save patients — so are we. We hope to aid medical professionals in improving diagnostic accuracy and efficiency amid the healthcare worker shortage, but we also aim to take a step further to directly help patients manage their health with our service.

BY SHIN HA-NEE [shin.hanee@joongang.co.kr]