One Vietnam-based AI startup is helping doctors around the world to speed up diagnoses by automating clinical interpretations of medical data

More than 100 hospitals in Vietnam, Myanmar, New Zealand and the USA have been using AI to facilitate clinicians to make better diagnoses.

Using what is said to be the only AI software for automated X-ray diagnostics in South-east Asia, and what is among the first AI platforms to be cleared by the US Food and Drug Administration to detect features suggestive of collapsed lungs from chest X-rays, these hospitals are also contributing data to train the AI for continual improvement.

The software applies AI analysis to medical images for more than 120,000 patients each month. Since doctors cannot make diagnoses based on any single factor, they also evaluate other data such as patient histories, symptoms, and laboratory and radiology reports. AI technology responsible for automated analyses of X-rays will soon be applied to help clinicians speed up evaluation of vital signs, blood tests, medical images such as MRI scans, and more.

The firm behind such medical AI tech is VinBrain, a Vietnam-based health-tech startup funded by one of Vietnam’s largest public firms. The startup trains its AI models—including those for medical imaging, intelligent video analytics, automatic speech recognition, natural language processing and text-to-speech—using NVIDIA AI solutions that can achieve near-linear-level speedups for model training, achieving 100x faster training compared with CPU-only training and significantly shortening the turnaround time for model development.

According to VinBrain CEO Steven Truong: “Our medical imaging models can analyze chest X-rays and make automated observations about abnormal findings in a patient’s heart, lungs and bones. Doctors could then combine the X-ray analysis with a large language model that reads health records to learn the patient also has a fever: this can help clinicians quickly determine a more specific diagnosis of pneumonia.”