ASUS and Tan Tock Seng Hospital Co-Develop an AI-based Tool to Improve the Accuracy and Efficiency of Diagnosis in Blood Diseases
ASUS and Tan Tock Seng Hospital (TTSH) inked a three-year Memorandum of Understanding (MOU) earlier today to formalise their collaborative efforts on building a more robust healthcare system to meet the evolving needs of patients. Since 2021, both parties have been working on novel AI solutions to enhance the value of clinical care delivered to patients.
One notable result of this collaboration is Blade, an AI-powered software co-developed by engineers from ASUS Intelligent Cloud Services (AICS) and medical professionals from TTSH. Blade is aimed at automating peripheral blood cell identification in the laboratory setting.
The traditional method of reviewing peripheral blood film is by light microscopy, performed by a laboratory technologist. This is labour intensive and can be subject to human fatigue. Films with abnormal features or unclear diagnosis are then escalated to a haematologist for further review.
Blade can automate blood cell identification and classification with high accuracy. This means that technologists will only need to load the blood films into a digital slide scanner. The AI tool will then process and analyse the digitalized films and flag any critical findings such as leukaemia, enabling early clinical intervention.
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.
A large dataset of over 337,700 digital images of peripheral blood cells was used to develop the AI-powered software through deep learning and computer vision, with a white blood cell classification accuracy of 91.4%.
Blade is currently undergoing clinical trials and evaluation at TTSH and other collaborative sites, with plans for regulatory approval. The development team hopes to evaluate Blade in the community setting, with Hougang Polyclinic planned as the first pilot site in the second half of 2022 for telemedicine purposes.
Looking ahead, ASUS and TTSH are working on solutions for breast screening and colon cancer detection, with both parties seeing potential in jointly developing AI-based laboratory solutions for pathology, cytology and microbiology.
“A partnership between Healthcare (TTSH) and Industry (ASUS) can only be truly successful when the elements of mutual benefit and the value-adding of patient care are conspicuously present. What motivates our team of clinicians, laboratory technologists, and software engineers in this collaboration is our hope that patients will benefit from this,” said Dr Eugene Fan, Consultant, Department of Haematology, TTSH.
Formalising the strategic partnership
Singapore is facing a rapidly ageing population with an increasing burden of chronic disease. To address the constantly evolving healthcare challenges, greater emphasis has been placed on healthcare innovation. This has paved the way for the adoption of innovative technologies like AI as an enabler for healthcare transformation.
ASUS has a long-standing commitment to involving industry leaders in the development of medical solutions. “The ever-growing Singapore-based ASUS AICS team is dedicated to sharing its technical expertise with TTSH to break barriers and make a difference for patients and healthcare providers in Singapore,” said Dr Tai-Yi Huang, Corporate Vice President & Chief Technology Officer, ASUS.
“As TTSH innovates to support our nation’s transformational shift towards Healthier SG, the adoption of new technologies such as AI will enable us to reinvent our practice, to improve care and outcomes in our hospital and the community. This strategic partnership serves to connect our clinician innovators with ASUS’ engineers, so that they can translate ideas to practice, by developing and trialling new solutions which address real-world problems,” said A/Prof Tan Cher Heng, Assistant Chairman Medical Board (Clinical Research and Innovation), TTSH.