Professor Hoonchul Kang's Team at Severance Children's Hospital
AI-Based Drug Screening Technology Demonstrates 100-Fold Greater Efficacy Than Existing Treatments
A team of Korean researchers has successfully identified a personalized drug candidate for rare and intractable pediatric epilepsy patients by using artificial intelligence (AI) and patient-derived stem cells. This candidate demonstrated up to 100 times greater efficacy than existing treatments.
From the left, Professor Hoonchul Kang of Yonsei University Severance Children's Hospital, Researcher Jihoon Kim, Professor Dokyun Na of Chung-Ang University. Korea Health Industry Development Institute
On May 19, the Korea Health Industry Development Institute announced that Professor Hoonchul Kang’s research team at Yonsei University Severance Children’s Hospital had succeeded in discovering a patient-specific drug candidate using AI-driven drug screening technology and induced pluripotent stem cells (iPSCs) derived from patients, and had experimentally verified its efficacy.
Induced pluripotent stem cells are cells created by introducing specific genes into adult somatic cells, enabling them to differentiate into various cell types like early stem cells. These cells retain the patient’s genetic information and can be converted into brain nerve cells, making them useful for creating personalized disease models.
Pediatric epilepsy is a representative intractable neurological disorder characterized by recurrent seizures in children and adolescents under the age of 18, without clear external triggers. In Korea, approximately 250,000 people are affected, and 30 to 40 percent of patients have 'drug-resistant epilepsy,' which does not respond to conventional anticonvulsants. In particular, pediatric epilepsy patients with rare genes such as mutations in the SCN2A gene (a gene that causes seizures and developmental disabilities) experience significant differences in experimental responses between individuals, and existing treatments rarely achieve marked improvement. Moreover, there is a lack of precise disease models and appropriate drug screening technologies, making the development of patient-specific drugs urgently needed.
To address this issue, the joint research team of Professor Hoonchul Kang and Researcher Jihoon Kim from Yonsei University, and Professor Dokyun Na from Chung-Ang University, generated induced pluripotent stem cells from the patient’s blood cells to create a precision disease model that replicates the patient’s actual disease environment. To confirm that the SCN2A mutation is the direct cause of epilepsy, they used advanced gene-editing technology to correct the mutation to normal, observed the disappearance of seizure symptoms, and thus demonstrated that the mutation was indeed the cause of the seizures.
Based on this personalized neural cell model, the researchers analyzed approximately 1.6 million compounds using AI-based simulations, considering factors such as blood-brain barrier permeability, toxicity, and gene-binding affinity. They selected five optimal drug candidates, among which two exhibited about 100 times greater efficacy than the conventional treatment phenytoin, supporting the feasibility of a patient-specific precision treatment strategy.
Professor Kang stated, "This is a case that proves the effectiveness of patient-specific drug screening technology using patient-derived cells for intractable epilepsy patients," and added, "In the future, we plan to continuously expand research to develop personalized precision therapies for patients with various gene mutations beyond SCN2A."
Professor Na emphasized, "This is a case that demonstrates the practical applicability of precision medicine technology in the field of rare diseases," and said, "We expect that the results of this study will establish an innovative treatment strategy that can be applied to patients with a variety of genetic diseases in the future."
This research was supported by the 'Public Interest Medical Technology Research Project' promoted by the Ministry of Health and Welfare and the Korea Health Industry Development Institute, and was published in Volume 191 of the world-renowned medical informatics journal 'Computers in Biology & Medicine' in 2025.
© The Asia Business Daily(www.asiae.co.kr). All rights reserved.
