AI Revolution, Is It Possible in Healthcare Too... Market Potential Should Be Examined Based on Trust and Expectations

As 'ChatGPT' brings changes across various sectors of society, discussions about the potential of artificial intelligence (AI) are growing day by day. However, despite the medical industry being highly evaluated for its potential use of AI, it is also facing inherent challenges due to the nature of the industry, which have yet to be overcome, resulting in a lack of proper commercialization, according to ongoing analyses.

On the 26th, Bae Woong, CHO of Kakao Brain, is delivering the keynote speech at the 'Kakao Ventures Digital Healthcare Day' held at Lunit Square in Gangnam-gu, Seoul.
Photo by Lee Chunhee

원본보기 아이콘

On the 26th, the 'Kakao Ventures Digital Healthcare Day' event was held at Lunit Square in Gangnam-gu, Seoul, discussing the present and future of medical AI. The speakers presented solutions for better utilization of medical AI, and digital healthcare startups invested in by Kakao Ventures showcased business directions based on these solutions.

Wung Bae, Chief Healthcare Officer (CHO) at Kakao Brain, who delivered the keynote speech, emphasized that "for healthcare software to be well used, it must have good usability or indirect benefits such as reimbursement." To overcome this, he stressed that 'customer usability' is crucial and "human nature must be targeted." Using electronic medical records (EMR) as an example, he explained, "As handwritten records have been computerized, doctors and pharmacists are paying to use them even without reimbursement," "and if customers do not recognize intuitive utility and are reluctant to use them, considerations such as naturally integrating into existing workflows are necessary."

Continuing, CHO Bae cited 'lack of generalization (trust)' and 'lack of expectation' as reasons why AI-related healthcare software has not grown properly. When AI behaves differently from humans, trust is broken, and unlike humans who can grow through education, AI has not shown such growth. His solution to this is 'super-large AI.' He explained, "If this is utilized to create a virtuous cycle of data within existing workflows, expectations will lead to trust, which will connect to purchasing power."

Kakao Brain is currently planning an AI computer-aided diagnosis (CAD) service based on this. The strategy is for AI to provide a draft interpretation of chest X-rays (CXR), and for doctors to revise it, creating a virtuous cycle of data that enhances satisfaction with the interpretations.

On the 26th, at Lunit Square in Gangnam-gu, Seoul, Kim Chi-won, Managing Director of Kakao Ventures, is giving a lecture at the 'Kakao Ventures Digital Healthcare Day.'
[Photo by Kakao Ventures]

원본보기 아이콘

Following this, Chi-won Kim, Managing Director at Kakao Ventures and an internal medicine specialist, emphasized the importance of establishing clear criteria to evaluate the effectiveness of medical AI. He stated that it is necessary to identify which stage of the medical process?screening, risk stratification, confirmation, companion diagnosis, treatment, or monitoring?the AI product targets, and whether it produces results that change treatment outcomes.

Kim explained this point by discussing the correlation between diagnostic tests and prevalence. For example, in the case of human immunodeficiency virus (HIV), which causes acquired immunodeficiency syndrome (AIDS), despite the test's sensitivity being 99.7% and specificity 98.5%, the prevalence among those aged 15 to 49 is only 0.1%. Therefore, when testing 100,000 people, about 16,000 test positive, but only 997 are actually infected. Kim added, "During the COVID-19 pandemic, rapid antigen tests were not used before the Omicron variant spread because prevalence was low," and "the use of rapid antigen tests increased as prevalence rose."

Due to these factors, he believes that medical AI targeting patients between outpatient and intensive care unit admission stages, who can be considered actual patients rather than simple screening subjects, has higher market potential. Kim emphasized, "At the screening level, no matter how many patients are tested, prices are hard to justify, making the market size effectively 'zero.' However, at the outpatient to ICU stage, the probability of abnormalities increases, and although the total number decreases, the value rises, making it possible to generate revenue."

Noduhyun Co-nective CEO (Professor of Orthopedic Surgery at Seoul National University Hospital) is giving a presentation at the 'Kakao Ventures Digital Healthcare Day' held on the 26th at Lunit Square in Gangnam-gu, Seoul.
[Photo by Lee Chunhee]

원본보기 아이콘

Among the companies introduced as investment targets by Kakao Ventures that day, according to Kim's criteria, three are in the treatment stage (Connective, RP, DeepMetrics), and two are in the diagnostic stage (Prevenotics, NeuroXTI).

In the treatment stage, Connective's presentation, which develops AI for orthopedic knee surgery planning and artificial joint surgery robots, attracted attention. Founded in 2021 by orthopedic surgeon Doohyun Noh from Seoul National University Hospital, the company is developing ▲ 'Connective X,' an automatic X-ray interpretation software, ▲ 'Connective S,' which helps predict complications after artificial joint surgery, ▲ 'Connective P,' a surgery planning aid, ▲ 'Connective H,' an integrated rehabilitation guidance solution, and ▲ 'Connective R,' an AI-based surgical robot.

CEO Noh shared, "When patients hear about robot surgery, they often ask if it is fully automated, but currently, humans perform the surgery while the robot acts as a guide," explaining the motivation behind founding the company. Artificial joint robot surgery requires the doctor to manually set points on the bone to confirm its position, so sometimes surgeries performed by skilled doctors take less time than robot-assisted ones. Even before surgery, "patients want detailed, evidence-based explanations about their disease stage and treatment plans based on X-rays," he said. "We are building super-large AI in the musculoskeletal field to intuitively explain disease stages to patients, predict the likelihood of knee joint deterioration, enable surgery simulation, and develop devices that can be linked for use in actual surgical fields," CEO Noh added.

Managing Director Kim commented on Connective, "Initially, I was not impressed when they showed diagnostic support software, but after much consideration, they developed software connected to robots and even the robots themselves," describing it as "a 'simple' case where the value impact on treatment outcomes is easy to prove because it is directly connected to treatment."

Seong Jun-kyung, CEO of NeuroXTI, is giving a presentation at the 'Kakao Ventures Digital Healthcare Day' held on the 26th at Lunit Square in Gangnam-gu, Seoul.
[Photo by Lee Chun-hee]

원본보기 아이콘

In the diagnostic stage, NeuroXTI, which develops an MRI-based companion diagnostic platform for Alzheimer's disease treatments, recently received approval from the U.S. Food and Drug Administration (FDA) and plans to introduce it domestically soon. CEO Jun-kyung Sung explained, "Treatment prescriptions suitable for disease progression stages based on amyloid beta (Aβ)-tau (Tau) interactions are necessary," adding, "Imaging can help determine whether the treatment will be effective and if the golden time has passed."

Since Biogen and Eisai's drugs 'Aduhelm' (active ingredient aducanumab) and 'Leqembi' (active ingredient lecanemab) work by removing Aβ proteins in the brain to slow cognitive decline, patients with a certain accumulation of Aβ protein but before Tau protein propagation are the most responsive to treatment. Although such tests have been possible through positron emission tomography (PET), PET scans cost over 1 million won, whereas MRI is much cheaper, providing a competitive price advantage. CEO Sung said, "From a pharmaceutical company's perspective, this can increase the likelihood of clinical success and reduce the risk of failure at market launch," and added, "We plan to establish standard guidelines for companion diagnostics when Biogen and Eisai's drugs are introduced."

© The Asia Business Daily(www.asiae.co.kr). All rights reserved.