By Taylor Kubota
Treatments for common but devastating diseases that occur in the chest, such as pneumonia, often rely heavily on how doctors interpret radiological imaging. But even the best radiologists are prone to misdiagnoses due to challenges in distinguishing between diseases based on X-rays.
Researchers at Stanford University in California have developed CheXNet, an algorithm that offers diagnoses based off chest X-ray images. CheXNet can diagnose up to 14 types of medical conditions and is reportedly able to diagnose pneumonia better than expert radiologists working alone. A paper about the algorithm was recently published on the open-access, scientific preprint website arXiv.
“Interpreting X-ray images to diagnose pathologies such as pneumonia is very challenging, and we know that there’s a lot of variability in the diagnoses radiologists arrive at,” says Pranav Rajpurkar, a PhD student in the Stanford Machine Learning Group and colead author of the paper. “We became interested in developing machine learning algorithms that could learn from hundreds of thousands of chest X-ray diagnoses and make accurate diagnoses.”
The work uses a public data set recently released by the National Institutes of Health Clinical Center. That data set contains 112,120 frontal-view chest X-ray images labeled with up to 14 possible pathologies. It was released in tandem with an algorithm that could diagnose many of those 14 pathologies with some success, designed to encourage others to advance that work.
As soon as they saw these materials, the Machine Learning Group knew it had found its next research direction. The researchers, working with Matthew Lungren, MD, MPH, an assistant professor of radiology, had four Stanford radiologists independently annotate 420 of the images for possible indications of pneumonia. The researchers have chosen to focus on this disease, which brings 1 million Americans to the hospital each year, according to the Centers for Disease Control and Prevention, and is especially difficult to spot on X-rays, the researchers indicate. In the meantime, the Machine Learning Group got to work developing an algorithm that could automatically diagnose the pathologies.
Within one week, the researchers had an algorithm that diagnosed 10 of the pathologies labeled in the X-rays more accurately than previous state-of-the-art results. In just more than one month, their algorithm could exceed these standards in all 14 identification tasks. In that short time span, CheXNet also outperformed the four Stanford radiologists in diagnosing pneumonia accurately.
“The motivation behind this work is to have a deep learning model to aid in the interpretation task that could overcome the intrinsic limitations of human perception and bias and reduce errors,” says Lungren, who is coauthor of the paper. “More broadly, we believe that a deep learning model for this purpose could improve health care delivery across a wide range of settings.”
After about one month of continuous iteration, the algorithm outperformed the four individual Stanford radiologists in pneumonia diagnoses. This means that the diagnoses provided by CheXNet agreed with a majority vote of radiologists more often than those of the individual radiologists. The algorithm now has the highest performance of any existing work related to the National Institutes of Health chest X-ray data set.
Also detailed in their arXiv paper, the researchers have developed a computer-based tool that produces what looks like a heat map of the chest X-rays but, instead of representing temperature, the colors of these maps represent areas that the algorithm determines are most likely to represent pneumonia. This tool could help reduce the amount of missed cases of pneumonia and significantly accelerate radiologist workflow by showing them where to look first, leading to faster diagnoses for the sickest patients.
In parallel to other work the group is doing with irregular heartbeat diagnosis and EMR data, the researchers hope CheXNet can help people in areas lacking easy access to a radiologist.
“We plan to continue building and improving upon medical algorithms that can automatically detect abnormalities, and we hope to make high-quality, anonymized medical data sets publicly available for others to work on similar problems,” says Jeremy Irvin, a graduate student in the Machine Learning Group and colead author of the paper. “There is massive potential for machine learning to improve the current health care system, and we want to continue to be at the forefront of innovation in the field.”
— Taylor Kubota is a science public information officer with Stanford News Service.