Dosy Health uses AI to analyze brain CT scans for four stroke-related aspects in one system, detecting brain hemorrhage, blood clots, and ischemic brain tissue from non-contrast CT scans. The system is already in trial use at hospitals.
Dosy Health, a Thai medical AI startup, presented an AI system for analyzing brain CT scans of stroke patients during the Thai Medical AI Showcase at the Global Medical AI Summit 2026 on 30 Oct 2026 at True Digital Park. This system builds on research from Mahidol and Thammasat Universities to analyze non-contrast brain CT scans, a cost-effective test widely available in hospitals. It covers all four diagnostic aspects in one system. The company noted that its service fees are a fraction of foreign AI systems, making it accessible to hospitals previously unable to afford such AI.
Dosy Health's system detects brain hemorrhage and classifies it into five types, identifies the location of blood clots blocking vessels with confidence levels, and assesses early-stage ischemic brain tissue damage. This information helps physicians distinguish between hemorrhagic and ischemic stroke cases, as treatments differ. The system also screens for risk factors.Hydrocephalus (fluid accumulation in the brain's ventricles).Specifically, Normal Pressure Hydrocephalus (NPH), a treatable cause of dementia symptoms. The NPH capability is currently in the research phase.
Sathat Asawadilokchai, CEO and co-founder of Dosy Health, stated that stroke is a race against time. Every minute of delayed diagnosis results in loss of about two million brain cells, potentially preventing patients from returning to normal life. The Department of Disease Control's epidemiological reports from 2020–2024 show stroke has the highest death rate among the top five non-communicable diseases in Thailand, with a continuing upward trend.
Cost is another barrier. Sathat said foreign AI systems often charge $10,000–50,000 USD or more per capability, mostly cloud-based, with high maintenance fees in Thailand. Large private hospitals may afford this, but many others cannot. In contrast, Dosy Health charges about 400,000 baht annually for all four diagnostic aspects, installs on a small computer costing around $900 USD (about 31,000 baht) within hospitals, operates offline without internet, and completes analysis within 7–10 minutes per case.
The system connects with hospital information and medical image storage and transfer systems. After scanning, images are sent to the AI for analysis, which then automatically returns results to the existing image viewing software used by radiologists. It provides heatmaps and confidence scores for each image slice, showing physicians the AI's decision points, while final decisions remain with the doctors.
The CEO added that tests show resident doctors using AI-assisted reading achieve accuracy close to that of radiologists. The company believes this system will help non-specialist doctors interpret results nearly at radiologist-level standards.
This technology derives from four software products registered by Mahidol and Thammasat Universities, for which Dosy Health holds exclusive rights for 10 years. The system is currently under trial at Siriraj and Rajavithi hospitals, integrated with their PACS systems. The work has also received a certificate at the International Exhibition of Inventions in Geneva, Switzerland.
Besides the brain CT reading system, which Dosy Health said is ready for commercial service, the company plans to expand AI CT analysis to heart, lung, liver, and bone scans in partnership with foreign medical technology firms. The heart CT reading system has FDA approval; the lung CT system is under approval process; and liver and bone CT systems are preparing submissions for approval.
The Global Medical AI Summit 2026 was organized by RISE together with the Artificial Intelligence Entrepreneurs Association of Thailand. Dosy Health was among the Thai startups presenting their work on the main stage during the Thai Medical AI Showcase.