<Korean News>Doctors vs AI in Brain Disease Prediction Showdown, Who Wins!?

A global challenge between leading stroke specialists and artificial intelligence (AI) took place in South Korea, where AI outperformed the doctors in predicting the condition of stroke patients. In just about 12 minutes, the AI analyzed the imaging data of 40 stroke patients and achieved a prediction accuracy of 0.72. In contrast, the doctors took an hour to deliver results with an accuracy of 0.50.

On November 11, nine domestic and international stroke experts and AI face off at JLK's headquarters in Gangnam-gu, Seoul, to predict the prognosis of stroke patients.

On the 11th, JLK held an "International Stroke AI Verification Seminar" at its headquarters in Gangnam, Seoul, inviting nine renowned stroke specialists from both domestic and international institutions. Stroke is a condition that occurs when the blood vessels supplying the brain either become blocked or burst, leading to a decline in brain function. A blocked vessel causes an ischemic stroke, while a ruptured vessel results in a hemorrhagic stroke.

JLK organized this event to test how accurately its newly developed AI, based on the "JLK-DWI" stroke diagnosis AI, could interpret MRI image data of the brain.

The competition format involved both the doctors and the AI diagnosing ischemic stroke and predicting the patient's condition based on MRI images. Doctors from the U.S., Japan, Taiwan, and the Philippines also participated.

JLK-DWI, an AI developed by Korean company JLK, reading images of 40 stroke patients.

The time taken to analyze all 40 sets of image data was significantly shorter for the AI. The AI completed the task in just 12 minutes and 4 seconds, taking only 15 to 20 seconds to predict the condition of each patient. In comparison, the human doctors needed an average of 45 minutes and 43 seconds, with the fastest doctor completing the task in around 21 minutes.

Between 25% and 30% of stroke patients experience worsening symptoms. When large blood vessels narrow or become blocked, a stent can be inserted to widen the vessel. However, there is no effective treatment for atherosclerosis in small blood vessels. Specialists typically predict the deterioration of a patient’s condition by reviewing symptoms and MRI images, but the accuracy of these predictions is often low and not clearly measurable.

Observers of the event remarked that this challenge "highlighted the speed and accuracy of AI technology used in stroke diagnosis," and noted that "as the global elderly population continues to grow, with increasing stroke rates, there will be a rising demand for AI-based stroke solutions that enable rapid diagnosis and treatment planning."

For Korean shopping and business agency services

www.online-korea.com