We use cookies to understand how you use our site and to improve your experience. This includes personalizing content and advertising. To learn more, click here. By continuing to use our site, you accept our use of cookies. Cookie Policy.

Features Partner Sites Information LinkXpress
Sign In
Advertise with Us

Download Mobile App




Events

ATTENTION: Due to the COVID-19 PANDEMIC, many events are being rescheduled for a later date, converted into virtual venues, or altogether cancelled. Please check with the event organizer or website prior to planning for any forthcoming event.

AI Diagnostic Tool Performs On Par with Radiologists in Detecting Diseases on Chest X-Rays

By MedImaging International staff writers
Posted on 19 Sep 2022
Print article
Image: New tool overcomes major hurdle in clinical AI design (Photo courtesy of Unsplash)
Image: New tool overcomes major hurdle in clinical AI design (Photo courtesy of Unsplash)

Most artificial intelligence (AI) models require labeled datasets during their “training” so they can learn to correctly identify pathologies. This process is especially burdensome for medical image-interpretation tasks since it involves large-scale annotation by human clinicians, which is often expensive and time-consuming. For instance, to label a chest X-ray dataset, expert radiologists would have to look at hundreds of thousands of X-ray images one by one and explicitly annotate each one with the conditions detected. While more recent AI models have tried to address this labeling bottleneck by learning from unlabeled data in a “pre-training” stage, they eventually require fine-tuning on labeled data to achieve high performance. Now, scientists have developed an AI diagnostic tool that can detect diseases on chest X-rays directly from natural-language descriptions contained in accompanying clinical reports.

The new model named CheXzero that was developed by scientists at Harvard Medical School (Boston, MA, USA) and colleagues at Stanford University (Stanford, CA, USA) is self-supervised, in the sense that it learns more independently, without the need for hand-labeled data before or after training. The step is deemed a major advance in clinical AI design because most current AI models require laborious human annotation of vast reams of data before the labeled data are fed into the model to train it. The model relies solely on chest X-rays and the English-language notes found in accompanying X-ray reports. The model was “trained” on a publicly available dataset containing more than 377,000 chest X-rays and more than 227,000 corresponding clinical notes.

Its performance was then tested on two separate datasets of chest X-rays and corresponding notes collected from two different institutions, one of which was in a different country. This diversity of datasets was meant to ensure that the model performed equally well when exposed to clinical notes that may use different terminology to describe the same finding. Upon testing, the researchers successfully identified pathologies that were not explicitly annotated by human clinicians. It outperformed other self-supervised AI tools and performed with accuracy similar to that of human radiologists. The approach, the researchers said, could eventually be applied to imaging modalities well beyond X-rays, including CT scans, MRIs, and echocardiograms.

“We’re living the early days of the next-generation medical AI models that are able to perform flexible tasks by directly learning from text,” said study lead investigator Pranav Rajpurkar, assistant professor of biomedical informatics in the Blavatnik Institute at HMS. “Up until now, most AI models have relied on manual annotation of huge amounts of data - to the tune of 100,000 images - to achieve a high performance. Our method needs no such disease-specific annotations.”

“With CheXzero, one can simply feed the model a chest X-ray and corresponding radiology report, and it will learn that the image and the text in the report should be considered as similar—in other words, it learns to match chest X-rays with their accompanying report,” Rajpurkar added. “The model is able to eventually learn how concepts in the unstructured text correspond to visual patterns in the image.”

Related Links:
Harvard Medical School 
Stanford University 


Print article
CIRS -  MIRION
Sun Nuclear -    Mirion

Channels

Ultrasound

view channel
Image: EG-740UT ultrasound endoscope combined with ARIETTA 850 provides outstanding ultrasound image quality (Photo courtesy of FUJIFILM)

Next-Gen Ultrasound Endoscope Enables Complex Diagnostic and Therapeutic Procedures

Endoscopic ultrasound is a specialist procedure performed utilizing an endoscope equipped with an ultrasonic transducer which emits and receives ultrasonic waves within the gastrointestinal tract, such... Read more

General/Advanced Imaging

view channel
Image: New guidance standardizes care for patients presenting with acute chest pain in the ED (Photo courtesy of Pexels)

New Guidance Recommends Coronary CTA as First-Line Test when Treating Acute Chest Pain in ED

Diagnosis and triage of emergency department (ED) patients with suspected acute coronary syndrome (ACS) consume a large and increasing amount of healthcare resources. ED overcrowding is associated with... Read more

Imaging IT

view channel
Illustration

Global AI in Medical Diagnostics Market to Be Driven by Demand for Image Recognition in Radiology

The global artificial intelligence (AI) in medical diagnostics market is expanding with early disease detection being one of its key applications and image recognition becoming a compelling consumer proposition... Read more
Copyright © 2000-2022 Globetech Media. All rights reserved.