AI boosts healthcare access in Thailand
Hundreds of hospitals in Thailand are using AI Chest 4 All, a disease screening technology developed by Thammasat University and the Department of Medical Services, to improve patient outcomes
Sponsored by Thammasat University
AI is improving access to healthcare and helping diagnose disease in Thailand, says Charturong Tantibundhit, a professor of electrical and computer engineering at Thammasat University in Thailand.
“In Thailand, we have only about 1,400 radiologists and most of them live in big cities such as Bangkok,” he says. Facilities and experts are rare in the country’s rural areas. “That’s why we wanted to build AI Chest 4 All so that it can be used for free across Thailand.”
Thammasat University researchers and collaborators analysed more than 200,000 chest X-rays, which had been labelled by radiologists, to train their deep-learning model to detect abnormalities in chest scans.
Doctors can then upload chest X-rays to the AI Chest 4 All system, which puts the X-ray into one of six categories: normal; suspected active tuberculosis; suspected lung malignancy or cancer; heart abnormality; intrathoracic abnormalities, such as Covid-19 or pneumonia; and abnormalities outside of the thorax, such as a bone or spinal cord abnormality.
The researchers found that their model achieved over 92 per cent accuracy in detecting heart disease, lung cancer and tuberculosis. “Since its launch in December 2021, the web-based system has been made available free of charge and adopted by more than 300 hospitals across Thailand,” says Tantibundhit.
One of the major advantages of AI Chest for All is how quickly it can analyse a scan. “A radiologist can take up to 10 minutes to analyse one image,” explains Tantibundhit. “That takes a lot of time. But AI Chest for All takes 50 milliseconds per image, which means we can service hundreds of hospitals in a second.”
Tantibundit and colleagues developed the model further, employing it in a screening protocol that focuses on disease diagnosis from chest X-rays, rather than simply identifying an abnormality. The researchers collaborated with radiologists to reclassify the labels of public datasets into eight classes of disease that are common in Thailand, such as pneumonia.
The research was made possible by interdisciplinary collaboration between institutions, something that Thammasat University actively promotes, Tantibundit says. “Thammasat University’s culture had a significant role in supporting the development of not just AI Chest 4 All but also other projects,” he says. “They support collaboration with universities around the world, and interdisciplinary collaboration,” such as in the AI Chest 4 All project, which involves collaborating with Udonthani Cancer Hospital, the National Chest Institute of Thailand and Rajavithi Hospital.
As AI continues to develop, so will its applications, Tantibundit says: “In Thailand, AI is enhancing healthcare outcomes by not only improving diagnostics but also personalising treatments, increasing operational efficiency and expanding access to care.”