It reads routine chest X-rays to highlight up to 29 abnormalities. Including the patient’s risk of lung cancer so a radiologist can take a look. Usually, human radiologists scour X-rays by eye. But 20% of lung cancer risks go undetected this way, causing a potential delay in diagnosis and poorer outcomes. The new tool was developed by Qure.ai and is being scaled up in partnership with AstraZeneca. The tool was trained using over 4 million X-rays. In trials, it was 17% more sensitive than a human. Early data indicates impressive accuracy for identifying risk of malignancy