31 October 2026 Published by Silicon Valley Innovation ไทย

Thai Health Business

Hospitals, clinics and wellness in Thailand

AI in hospitals

Thai hospitals scale up AI chest X-ray reading

Thailand has about 2,000 radiologists for more than 30 million X-rays a year, and the national health fund is paying for the rollout through the 30 baht scheme.

อ่านภาษาไทย

Black and white frontal chest X-ray showing both lungs with hazy shadows in the lower fields
A chest X-ray, the type of image the tools read. Illustrative; not a Thai patient image.Photo by CDC on Unsplash

Siriraj Hospital says its artificial intelligence chest X-ray reader has now been used on more than 500,000 patient cases and handed over to more than 145 Ministry of Public Health hospitals, with a target of 450 in 2026, The Nation reported in February 2026. It is the largest of several Thai-built tools that are moving from pilots into routine use.

The rollout has since run well past that target. In August 2026 Siriraj said the reader was in use at more than 610 public health hospitals nationwide, after the National Health Security Office board approved it as a covered benefit in April 2026, Thai Post reported.

Why chest X-rays

There are too few radiologists. Microsoft, writing about Ramathibodi Hospital's tool in April 2026, says about 2,000 radiologists read more than 30 million X-rays a year, roughly 15,000 each. A Ramathibodi thoracic radiologist, Dr Narissara Chobarun, told Microsoft a radiologist often reviews 60 to 200 chest X-rays a day. Microsoft adds that caseloads can pass 300 a day in busier hospitals.

The stakes are highest for lung cancer. It is Thailand's second biggest cause of cancer deaths, with 17,222 new cases a year, and more than 60% are found at stage 4, The Nation reported from the HealthTech Summit at Techsauce Global Summit 2025. Survival is 62% when found early and 4% at stage 4.

Two Thai tools

Siriraj's reader was developed by its radiology department, starting during the Covid-19 pandemic, and The Nation's lung screening report names Perceptra as its development partner. The Royal College of Radiologists of Thailand certified accuracy above 95%, and the hospital says it returns a result in under 10 seconds. Next, Siriraj plans AI for brain CT scans and for mammography.

Ramathibodi's RAMAAI runs on Microsoft Azure, screens for up to 16 conditions including lung nodules, pneumonia and COPD, and has a tuberculosis module. It was in 8 public hospitals in Bangkok, Samut Prakan, Chonburi, Lampang and Chiang Rai, serving about 2,000 medical staff, and had processed more than 500,000 images. Ten more hospitals were due by June 2026. The Department of Medical Services plans to pilot it across its own hospitals.

The accuracy claims come from the developers and the college's certification, not from an independent trial.

Who pays

Payment is what makes a national rollout possible. The National Health Security Office, which runs the 30 baht scheme, plans AI-assisted chest X-ray screening at 167 hospitals in fiscal 2025 for 55 million baht a year, 445 hospitals in fiscal 2026 for 135 million baht a year, and 887 hospitals in fiscal 2027 for 225 million baht a year, according to the board decision as reported by Matichon. A peer-reviewed review of radiology AI notes that before this, hospitals received no specific reimbursement for AI-assisted imaging.

The hospital's own validation still matters. Passakorn Wanchaijiraboon, deputy director of Phrapokklao Hospital, told the summit that hospitals should test AI on their own data because performance varies by setting.

Private hospitals

Private groups started earlier. Bumrungrad announced in July 2022 that it uses Lunit INSIGHT CXR for chest X-rays and INSIGHT MMG for mammograms, delivered by Fujifilm Thailand on Microsoft Azure, covering more than 100,000 images a year. BDMS lists "Smart Healthcare", which combines AI with its data systems, as one of four priorities for 2026 in its annual filing, though it names no specific tool there.

Next to watch: whether the National Health Security Office funds the third phase, 887 hospitals, in fiscal 2027, and whether Ramathibodi's RAMAAI reaches the 18 hospitals implied by its eight plus ten more.

FAQ

How accurate is Siriraj's AI chest X-ray reader?

Siriraj says it is above 95%, certified by the Royal College of Radiologists of Thailand. The figure is from the hospital and the college, not an independent study.

How many Thai hospitals use AI to read chest X-rays?

Siriraj said in August 2026 that its tool is in more than 610 public health hospitals, up from more than 145 in February. The National Health Security Office had planned 167 hospitals in fiscal 2025 and 445 in fiscal 2026.

Does AI replace radiologists in Thailand?

No. The tools prioritise cases and flag findings for doctors, and Ramathibodi describes RAMAAI as an assistant.

Sources

AIradiologySirirajRamathibodiNHSOchest X-ray

More from Thai Health Business