The Tan Tock Seng Hospital in Singapore and Taiwan-based ASUS Intelligent Cloud Services have co-developed an AI-powered software that automates the identification of peripheral blood cells in laboratories.

Recently, TTSH and ASUS signed a memorandum of understanding that formalises their research collaborations which started last year.

In one of their projects, they have built Blade, an AI tool for automating blood cell identification and classification. It processes and analyses blood films loaded into a digital slide scanner and then flags any critical findings like leukaemia.

WHY IT MATTERS

According to TTSH, peripheral blood films are traditionally reviewed by a laboratory technologist using light microscopy. It is said that the method is labour intensive and can be subjected to human fatigue. 

"By assisting laboratory technologists and haematologists with the reporting of peripheral blood films, Blade aims to accelerate the overall review duration by 50%, thus translating into better productivity and faster diagnosis," it said in a statement.

Blade was developed using a dataset of nearly 400,000 digital images of peripheral blood cells. It has achieved a 91.4% accuracy for classifying white blood cells. 

The technology is still being tried out and evaluated at TTSH and other collaborative sites. The research team is also seeking to evaluate Blade in a community setting with the Hougang Polyclinic planned as its first pilot site in the second half of the year.

TTSH is now working to secure regulatory approval for its software.

In the future, TTSH and ASUS plan to also develop solutions for breast screening and colon cancer detection, as well as AI-based laboratory solutions for pathology, cytology and microbiology.

THE LARGER TREND

TTSH has been involved in research projects that are employing AI to enhance disease diagnosis. Last year, its clinicians developed an automated glaucoma screening method with scientists at the Nanyang Technological University. In a study, the AI method was found to have achieved 97% accuracy and 95% sensitivity in diagnosing glaucoma.

In other news, late last year AICS introduced some of its latest AI medical solutions, including its cloud-based hospital information system called xHIS, which integrates AI services and optimises clinical processes. It also unveiled an EMR search platform that analyses medical records and examination reports and an AI-assisted drug safety system which can flag abnormal dosages in prescriptions.