CheXpert (AI system of recognition of pneumonia)

2019: Announcement

In the middle of October, 2019 researchers from group of Intermountain Healthcare hospitals and Stanford University provided a new AI system which is capable to detect symptoms of pneumonia in 10 seconds. The model easily adapts to georgaficheky features of the population of a certain region and can be used in any hospital of the world.

The CheXpert model as the instrument of support of adoption of clinical solutions is intended for emergency doctors and also for general practitioners. Use of such AI system considerably reduces time of diagnostics and confirmation of the diagnosis and allows to begin treatment of heavy patients as soon as possible.

Researchers from group of Intermountain Healthcare hospitals and Stanford University provided a new AI system which is capable to detect symptoms of pneumonia in 10 seconds

The CheXpert system developed by Stanford Machine Learning Group represents the automated model of interpretation of x-rays of a thorax. It studied for 188,000 images received by the Stanford medical center. Researchers checked system accuracy, using it for assessment of x-rays in several departments of emergency aid. The CheXpert model identified all key symptoms of pneumonia on x-rays in 10 seconds, at the same time its results were confirmed by radiologists. Besides, an AI system is capable to generate independently reports, excepting errors in interpretation of its data.

Now researchers are going to test an AI system in department of emergency aid of hospital of Salt Lake City. In case of success the program will be deployed in several more hospitals of the Intermountain Healthcare group. The dl of the beginning of CheXpert will reveal only pneumonia, but further researchers hope to use it and for detection of other diseases, such as heart failure and HOBL.^[1]

Notes