Developed by qure.ai, qXR is a point-of-care automated chest X-ray interpretation aid —it tells a health worker whether the chest X-ray is normal or abnormal, lists any abnormalities that have been found and highlights them on the chest X-ray for review. 


All that are expensive and sometimes difficult to find. Built with deep learning technology, qXR was trained with 2.5 million X-rays. This means that it generalizes well, providing accurate results in any population setting. With high sensitivity of 95% on GeneXpert, no abnormal cases are missed and yield for active case finding is improved. A recent study by the Stop TB partnership was published in Nature Scientific Reports concluded that qXR had a 95% sensitivity and 84% specificity on GeneXpert outperforming radiologists in the evaluation in Nepal and Cameroon. The study measured cost savings and demonstrated that the qXR algorithm could reduce GeneXpert cartridge consumption by 66% while maintaining a sensitivity of 95%.

It has been tested and trained on various types of X-ray device types including cartridge-based systems and fully digital X-ray systems, and works out of the box with new types of X-ray hardware.

With qXR you can not only detect a host of pulmonary, pleural, mediastinal and cardiac abnormalities, but it also comes with a dedicated tuberculosis screening module. Like a radiologist, qXR’s TB screening module combines the size, location and type of chest abnormality to decide whether an X-ray has signs that suggest TB.

Whether you are using X-rays for triage or systematic screening, you can use qXR to automate decision making about the next steps (like molecular confirmation). This allows your program to run faster and more efficiently, all while optimizing your costs.  


qXR is currently deployed across 15 countries, with fixed and mobile X-ray systems at public and private hospital settings as well as on mobile vans.


qXR is CE marked and HIPAA compliant