NIH Releases 100,000 Chest X-Rays

The National Institutes of Health has created a primer for artificial intelligence that will help computers learn to read chest x-rays, with the hope of providing more sophisticated and consistent diagnoses.

More than 100,000 chest x-rays from a data set of 30,000 patients have been released to the scientific community as part of an effort to teach computers how to detect and diagnose disease, the NIH Clinical Data Center announced.

The images were taken with the permission of patients at the NIH Clinical Center, and will be freely shared with researchers working to improve artificial intelligence diagnoses. The dataset was screened to remove patients’ personally identifiable information.

“By using this free dataset, the hope is that academic and research institutions across the country will be able to teach a computer to read and process extremely large amounts of scans, to confirm the results radiologists have found and potentially identify other findings that may have been overlooked,” NIH said.

Consistent automated readings of chest X-rays can be complicated by a number of factors including knowledge of anatomical principles, physiology and pathology. Ultimately, it is hoped that AI can lead to clinicians making better diagnostic decisions for patients, NIH said.   

In addition, this advanced computer technology may also be able to:

• Identify slow changes occurring over the course of multiple chest x-rays that might otherwise be overlooked
   
• Benefit patients in developing countries that do not have access to radiologists to read their chest x-rays
   
• Create a virtual radiology resident that can later be taught to read more complex images such as CT and MRI

NIH will release a large dataset of CT scans in the coming months.