ENG | |
Our Stories
Innovating for Social Good

Novel system beats the clock to save lives of stroke patients

“The clock is ticking for stroke patients. Chances of recovery decrease with every minute passing by.”

In Hong Kong, about 3,500 people die from stroke each year, making it the fourth leading cause of death in our city. Worldwide, stroke affects 15 million people a year and is the second leading cause of death among people above the age of 60.

With the rapid ageing of Hong Kong’s population, stroke detection has become increasingly important for saving lives. Thanks to Dr Tang Fuk-hay and his research team at the Department of Health Technology and Informatics of The Hong Kong Polytechnic University (PolyU), a valuable new diagnostic tool has been developed — a computer-aided system that can diagnose acute strokes in three minutes within the three “golden hours” of the onset of stroke.

The system developed at PolyU combines artificial intelligence, pathology and a sophisticated algorithm. Through analyzing 80-100 X-ray images taken by a CT scan and from the information received, it can detect whether the patient has suffered an ischemic stroke, the most common form of stroke, or subtle increases in the brain’s tissue density indicating a haemorrhagic stroke. For both types of stroke, the system is able to make diagnoses with 90% accuracy.

The computer-aided detection stroke technology developed by Dr Tang’s team also provides a second opinion for frontline doctors allowing them to make timely treatment decisions for stroke patients.

“It is important to identify stroke patients and help them get the urgent treatment they need,” said Dr Tang. As millions of brains cells are destroyed each minute after a stroke, life or death decisions must be made quickly and accurately. “The clock is ticking for stroke patients. Chances of recovery decrease with every minute passing by.”

With the stroke detection system, experienced specialists and even non-specialist physicians can quickly detect subtle, minute changes in the brain, with less chance of missed diagnoses. The system also helps physicians eliminate false-positive and false-negative cases, as well as other conditions that mimic a stroke. This is particularly critical in hospitals where overworked medical staff might misdiagnose and prescribe unnecessary treatment.

The system is also capable of improving its diagnoses over time. With its built-in artificial intelligence, it can learn from experience using every scan together with feedback from specialists to enhance accuracy.