The rise of artificial intelligence (AI) has thrown pharmaceutical companies and medical device into a race to develop new products and services to open the age of AI healthcare.
Korea also has been approving, slowly but steadily, new AI medical devices for use in the clinical field while setting up new regulations and application methods to speed up the new sector’s development.
Not all AI-based medical devices are showing optimal results as promised, however.
Most recently, IBM's Watson for Oncology, an AI-led diagnosis program, is quickly losing popularity, as hospital staffs say that the device has met a brick wall not least because of the Korean healthcare system's limitations bound by the inflexible rules of the Health Insurance Review and Assessment Service (HIRA).
Such problems have raised the need for an AI-based device that can fit in well with Korea's system.
One of the companies aiming to break such limitations and offer optimal treatment guides is Vuno, a startup that provides AI-based medical imaging services.
"Vuno is Korea's first AI-based medical diagnosis service company," Vuno’s chief strategy officer, Kim Hyun-jun, said in an interview with Korea Biomedical Review. "While many AI-based medical device companies focus on imaging alone, we collect data from all sources, including radiology, pathology, patient's bio-signal, and electronic medical records, in a comprehensive way."
As doctors treat patients based on their test results, Vuno's goal is to help and speed up the process with its technology, Kim added.
Vuno is already well-known in Korea's AI healthcare field as it obtained the nation’s first AI-based medical equipment approval from the Ministry of Food and Drug Safety with its VUNOmed-BoneAge in May last year.
The device determines a patient's bone age by recognizing patterns on an X-ray image and marking its similarities with 58 standard reference images -- 31 males and 27 females. Combined with hormone levels and other information, doctors can use the machine's information to diagnose precocious puberty and slow growth.
The company used 20,000 images from Asan Medical Center to train its device.
"Previously, doctors had to flip through a reference book manually to find an age-labeled bone image to compare with the standard image, a process taking five minutes or more," Kim said. "However, most general hospitals have 500 cases a month, while insurance data shows that Korea annually produces more than 500,000 cases."
Therefore, the task in diagnosing symptoms by siftings through reference books was cumbersome for doctors, Kim added.
With Vuno's device, however, doctors only have to insert the X-ray image and can receive the top three suggestions.
"The correct diagnosis being in one of the top three suggestions is over 98 percent," Kim said. "The device has also shortened diagnosis time by 40 percent and increased accuracy by 10 percent."
The device has received positive opinions from local hospitals, and the company is working to install the system at more than 50 hospitals, Kim added.
More recently, the company has also received approval for other AI-based medical devices, including VUNO Med-Chest X-Ray, VUNO Med – DeepBrain.
"VUNO Med-Chest X-Ray, a chest-disease screening system that assists diagnosis by providing screening results and suspected radiological findings from chest radiographs, received a nod in August," Kim said. "VUNO Med – DeepBrain, a device uses the brain magnetic resonance images for diagnosing Alzheimer's disease by using a pre-learned model, won the go-ahead in June.
Based on such a wide range of AI-medical devices, the company has also set its eye toward global markets, Kim stressed.
Even before marketing the product in Korea, the company has already succeeded in exporting VUNO Med-TriVu, an AI-based treatment solution for sexually transmitted infections, to Mongolia and the Philippines.
"VUNO has developed the microscope-based AI solution Med-TriVu for the diagnosis of sexually transmitted infections since 2017 with the support of the Korea International Cooperation Agency (KOICA)," Kim said. "Last year, the company successfully conducted demonstration projects at 10 national medical institutions in Mongolia."
Currently, Mongolia Ministry of Health is adopting VUNO's AI solution as the national standard inspection method, he added.
Based on its success in Mongolia, the company was able to export the system to the Philippines and plans to facilitate export deals for other AI-based medical devices, including VUNO Med-TriVu.
Asked about what makes Vuno's AI device different from other AI-based medical devices, Kim said that in healthy market competition, the purpose of using AI medical devices is similar.
"However, what differentiates Vuno from other companies is that a lot of our products are market-driven," Kim said. "In terms of the products that are released, they are more focused on solving the problems that doctors in the field face than marketability."
Another difference is that while most AI products provide a diagnosis, Vuno's product offers the finding process that led to the diagnosis, Kim added.
"One of the most important issues with AI-device is the so-called "black box" issue, where the device only provides the final diagnosis without the process of reaching such conclusions," Kim said. "Such a process might be okay for experienced physicians, but we believe that the role of AI-based devices is to help less-experienced doctors."
If the device only provides the solution, it may be of little use to less-experienced doctors as they may not understand how the machine landed on the conclusion, he added.
AI not out to replace doctors
Not all people welcome the rapid rise of artificial intelligence. Among the skeptics are people who are directly involved in the medical field.
Kim stressed, however, that Vuno and other AI-based medical device companies are not out to replace physicians.
"AI-based medical device companies are not aiming to replace doctors or for our devices to make a final decision on a patient's diagnosis," Kim said. "In the AI industry, legal responsibility ultimately falls to humans, so manufacturers do not want to enter that territory."
The company's goal is to remain as just a supplemental tool to help physicians improve the quality of medical diagnoses, Kim added.
Kim stressed that AI is not an almighty platform that can solve every problem.
"It's just a machine that can heal doctor's work more efficiently in the field," Kim said. "If the tool is efficient and safe, there is no reason not to use AI, and Vuno is certain that its technology is safe and effective."
If a customer has any questions or doubts about Vuno’s products and technologies, the company invites them to feel free to use the devices to evaluate or test any of the company's products, he added.