Siriraj Hospital Integrates AI to Elevate Medical Research and Diagnosis, Touts YOLO Model for Centroblast Cell Counting

BANGKOK, Aug. 5, 2024 /PRNewswire/ — The Faculty of Medicine Siriraj Hospital at Mahidol University has announced the integration of artificial intelligence (AI) technology to improve medical analysis and diagnostic capabilities. This initiative aims to enhance accuracy in detecting abnormalities, diagnosing diseases, and optimizing the workflow of medical personnel.

“Detection of centroblast cells in H&E stained whole slide image based on object detection” is a key study that employs the You Only Look Once (YOLO) model, a state-of-the-art object detection algorithm, to identify and quantify centroblast cells in Hematoxylin and Eosin (H&E) stained whole slide images (WSI). This development is significant for lymphoma grading, facilitating more precise patient classification into stages 1, 2, or 3 based on the progress of the disease.

The implementation of AI has improved the centroblast cell counting, a process previously conducted manually. In this study, researchers trained the AI system using 1,669 centroblast cell images from WSIs of five patients. The YOLOv5x6 model achieved a precision of 0.808, recall of 0.776, mean average precision (mAP) at 0.5 Intersection over Union (IoU) of 0.800, and an overall mAP of 0.647.

Siriraj Hospital’s AI research extends beyond centroblast detection, encompassing various medical specialties including radiology, internal medicine, pathology, neurology, orthopedics, mental health, ophthalmology, and community healthcare. Notable applications include AI-assisted voice and face recognition for mental health assessment and the CHIVID app for COVID-19 symptom screening.

Since the initiation of its AI program in 2018-2019, Siriraj Hospital has successfully implemented AI in several areas. In radiology, the hospital has successfully deployed AI to assist in analyzing chest X-rays since 2019, enabling the detection of lung abnormalities and coronary arterial calcification. The pathology department has benefited from AI through the automated detection of centroblast cells. In ophthalmology, AI has also been utilized to screen patients for diabetic retinopathy. These advancements demonstrate Siriraj Hospital’s commitment to leveraging AI technology to enhance patient care and improve diagnostic capabilities across multiple medical specialties.

The hospital plans to expand its AI capabilities to include Mammogram and CT Brain analysis by the end of 2024.

Siriraj Hospital is committed to integrating AI technology to reflect the hospital’s dedication to advancing medical care and improving patient outcomes. This initiative not only enhances its diagnostic capabilities but also positions Siriraj Hospital at the forefront of medical innovation in Thailand and the region.

Contact information:

Faculty of Medicine Siriraj Hospital, Mahidol University

Tel.: +662-419-7000

https://www.si.mahidol.ac.th/th/ 
 https://www2.si.mahidol.ac.th/en/

About the Faculty of Medicine Siriraj Hospital

Established in 1888, the Faculty of Medicine Siriraj Hospital, Mahidol University, is Thailand’s oldest and largest healthcare and medical education institution. The faculty upholds the vision of becoming “The Medical Institute of the Nation Driving Medical Discovery and Advanced Healthcare with World Changing Impact on Health System.” Its primary mission encompasses delivering education and training to foster graduates, health professionals and specialized practitioners, providing medical and health promotion services, along with conducting research on pioneering knowledge and innovations in the health and education fields.

Source : Siriraj Hospital Integrates AI to Elevate Medical Research and Diagnosis, Touts YOLO Model for Centroblast Cell Counting

The information provided in this article was created by Cision PR Newswire, our news partner. The author's opinions and the content shared on this page are their own and may not necessarily represent the perspectives of Thailand Business News.