AI Healthcare Technology Begins to Innovate in Analysis and Medical Assistance for Human Health (op-ed)

AI Healthcare Technology Begins to Innovate in Analysis and Medical Assistance for Human Health

The use of AI in the medical field has had limitations from the very beginning, unlike other fields. This is primarily due to the fact that medicine deals with life, so accountability must be clear. This is why autonomous AI technology cannot be used in the same way. Additionally, AI technology relies heavily on data collection, and each country has different medical laws, which affects how AI can be used. Despite these limitations, the healthcare AI market continues to experience explosive growth, as the unique potential of AI technology remains the same across different sectors.

Limitations in Data Collection

Unlike shopping or search, the medical field cannot use a common global dataset. This is because each country has different genetic traits and health information about its citizens. Therefore, it is difficult to apply AI trained on foreign health data directly in a domestic setting. Unlike other fields, medical AI needs to focus on developing AI that utilizes health data from its own citizens. While international trends in AI can provide valuable insights, it is crucial to train AI on data from the local population. For this reason, many countries are highly sensitive to foreign AI systems collecting and utilizing their citizens' health data. As a result, the use of AI for health data in the medical field is often highly restricted.

A Rapidly Growing Healthcare AI Market Despite Limitations

For these two reasons, healthcare AI is still mainly used in limited areas. Most of the healthcare AI developed so far has been used as a support tool for physicians or for analysis purposes. Unlike OpenAI's conversational AI "ChatGPT," AI cannot take a leading role in medical practices. In other words, due to regulatory frameworks in various countries, AI cannot perform medical procedures independently of physicians.

Nevertheless, healthcare institutions worldwide are showing strong interest in adopting AI, and the global market is expected to grow rapidly. According to a market research firm, Precedence Research, the global healthcare AI market, which was valued at $1.07 billion (approximately 1.47 trillion KRW) in 2022, is expected to grow nearly 20 times to $21.7 billion (approximately 29.78 trillion KRW) by 2032.

Tech Giants Entering the Medical AI Field

As a result, major global IT companies have entered the medical AI development space. Google developed a conversational AI for the healthcare field, "MedPaLM2," through Google Cloud. MedPaLM2 uses data from generative AI to answer healthcare-related questions. The Mayo Clinic in the U.S. has implemented MedPaLM2 for training medical staff. Naver has also developed "Smart Survey," which uses AI to automatically record patients' consultation details in medical terms.

Meta has developed an AI called "ESM Fold" to assist in the development of disease treatments by analyzing and predicting protein structures. Microsoft (MS), in partnership with the global pharmaceutical company Novartis, established the "AI Innovation Lab" in 2019 to conduct AI-based drug research. Apple is also developing an AI called "Quartz" to provide health advice based on health data collected from users' smartwatches, such as electrocardiograms.

Reducing Drug Development Time with AI

Global pharmaceutical companies are also actively utilizing AI to reduce the time and cost involved in drug development. Janssen has partnered with the UK-based AI startup Benevolent to develop AI-driven treatments for difficult-to-treat diseases. Merck has also invested in Benevolent and Exscientia and, in collaboration with Atomwise, has discovered potential compounds for Ebola treatment using AI.

Among domestic pharmaceutical companies, Daewoong Pharmaceutical has set up an AI-driven drug research team, while Yuhan Corporation has partnered with the AI startup Eisen Science for cancer drug research. SK Chemical has collaborated with an AI startup to develop candidates for treating hepatitis and pulmonary fibrosis.

AI Technologies Enhancing the Efficacy of Treatments

A startup called Portree, founded by four doctors from Seoul National University College of Medicine, uses AI to analyze biological information and assist in drug development. They analyze "spatial transcriptomics," a process that provides information on the location of tissues in the human body, much like a postal address. By understanding the spatial transcriptome, AI can pinpoint the exact location of cancer cells, making it possible to deliver drugs more accurately, thereby enhancing treatment effectiveness. The AI developed by this company helps track cancer cells and also assists in developing new drug substances to deliver the medication precisely to cancer cells.

Medical AI Still Has a Long Way to Go

Despite the growing use of medical AI in various fields, many challenges remain. The biggest issue is accountability when AI is involved in medical procedures. If a problem arises after AI-assisted treatment, questions of responsibility between the physician, the hospital, and the AI developer inevitably arise. Since healthcare directly concerns a patient’s health and life, expanding AI to medical practices beyond limited areas such as image interpretation, data analysis, and tracking is challenging. To address this, it is crucial to establish clear validation and national certification systems for AI diagnostics.

Diseases Identified by AI

1. Lung Diseases

In South Korea, AI startups are actively involved in medical image analysis for cancer diagnosis. A representative company is Lunit. The AI "Lunit Insight" developed by the company analyzes radiography and CT scans to detect signs of lung cancer. It is particularly effective in identifying small abnormalities that are difficult to spot with the naked eye, thus assisting physicians in diagnosis. The company plans to develop AI capable of diagnosing all types of cancer.

Deepnoid’s "DeepAI" also assists in analyzing chest radiography and brain MRA images to support medical professionals’ interpretations. Coreline Soft offers various AI solutions for analyzing CT scans of the chest and head and neck. Notably, their "AIVUE LCS Plus" solution detects lung nodules, emphysema, and coronary artery calcification. Coreline Soft is also participating in a European lung cancer screening project led by five European countries, including Germany and Italy.

2. Brain Diseases

JLK specializes in analyzing brain imaging data. The company uses AI to analyze CT and MRA data to diagnose various brain-related diseases such as brain hemorrhages, strokes, aneurysms, and dementia. To develop their AI, JLK signed an exclusive contract with the Korean Brain MR Imaging Data Center, utilizing over 1.4 million brain images collected over 10 years.

AI is also used in dementia diagnosis. The Korea Electric Institute's Medical Equipment Research Division has developed AI that analyzes language and brainwave data from elderly patients to detect signs of Alzheimer's disease. The research team conducted a validation experiment on about 100 participants and identified six cases of mild cognitive impairment and seven suspected cases.

3. Heart Diseases

Vuno has developed a medical device called "Heartib P30," which uses AI to analyze users' electrocardiograms and detect abnormalities such as atrial fibrillation and tachycardia. The company has also established a healthcare AI startup in Japan, M3 AI, in collaboration with Japan’s largest medical information company, M3, to expand its presence in the global market.

*This guest column represents the views of the author, not the editorial board.