AI Precisely Predicts Optimal Timing for Atrial Fibrillation-Related Stroke Treatment... 87% Agreement with Specialists

JLK's Stroke Classification Solution 'JBS-01K'
"Supporting Treatment Planning for Atrial Fibrillation-Related Stroke"

Medical artificial intelligence (AI) company JLK announced on January 16 that it has confirmed its MRI-based stroke AI solution, 'JBS-01K (JLK-DWI),' can help provide objective indicators for determining treatment criteria in patients with atrial fibrillation-related stroke.

Atrial fibrillation, which is characterized by an irregular heartbeat, is a critical variable that significantly affects the onset and prognosis of stroke, making treatment extremely challenging. In particular, determining the timing for administering anticoagulants to prevent recurrence is considered a major challenge. If administered too early, there is a high risk of bleeding; if delayed, patients are exposed to the risk of recurrent stroke. Until now, the lack of clear standard guidelines in clinical practice has led to a heavy reliance on the personal experience and judgment of medical professionals.

In response, a domestic joint research team set out to verify how clinically reliable AI-based quantitative analysis can be in establishing treatment strategies for stroke patients with atrial fibrillation. The study involved Korea University Guro Hospital, Chonnam National University Hospital, and Seoul National University Bundang Hospital.

The research team conducted a comparative analysis of the concordance between AI assessments and the diagnostic results agreed upon by a group of experienced specialists, using data from 6,734 acute stroke patients. JLK's stroke classification solution, 'JBS-01K,' served as the core analytical tool.

The analysis showed a concordance rate of 87.4% between the JLK AI and specialist assessments, with a reliability index (Cohen's Kappa) of 0.81, which is considered 'very high.' This is higher than the concordance observed when two experienced specialists interpreted the same images (Kappa 0.62). The research team concluded that, in determining the severity-a key indicator for initiating treatment in atrial fibrillation-related stroke-AI can serve as a decisive tool to reduce discrepancies among medical professionals and enhance diagnostic consistency.

Previously, medical professionals had to determine treatment strategies based on their experience, considering the size, location, and shape of the lesion. According to the results of this study, it is now expected that more systematic and standardized treatment strategies can be established based on the quantitative analysis data provided by AI.

JLK plans to further strengthen its integrated medical AI strategy encompassing both the brain and heart. Currently, JLK is collaborating with Mediana to link Holter tests and wearable devices for patients suspected of having cardioembolic stroke (CE). The company is also working to incorporate continuous monitoring for recurrence prevention, and is pursuing joint research and commercialization of AI-based diagnostic and predictive technologies that connect brain and heart data.

Kim Dongmin, CEO of JLK, stated, "JLK will continue to expand its medical AI strategy encompassing both the brain and heart through its stroke solution."