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 hp
Sign In
Advertise with Us

Download Mobile App

Generative AI Tool Enables Timely Interpretation of Chest X-Rays by ED Physicians

By MedImaging International staff writers
Posted on 06 Oct 2023
Print article
Image: Generative AI can assist in chest radiograph interpretation in the emergency department (Photo courtesy of 123RF)
Image: Generative AI can assist in chest radiograph interpretation in the emergency department (Photo courtesy of 123RF)

Quick and accurate interpretation of diagnostic X-rays is essential in emergency departments (ED), but not all facilities have round-the-clock radiology services. Generative AI technologies can potentially bridge this gap by offering nearly immediate interpretations of medical images, and handling a high number of cases without getting tired or needing additional staff. Now, a new AI model has been developed to help emergency doctors identify life-threatening conditions in chest X-rays.

A team of researchers at Northwestern University (Chicago, IL, USA) set out to create and assess a generative AI tool specifically designed for interpreting chest X-rays in an emergency setting. This AI model belongs to a newer category of generative AI known as transformer models. These models combine large language models, similar to ChatGPT, with deep learning techniques to analyze images. In simple terms, the AI tool works as an encoder-decoder model that takes chest X-ray images and generates a corresponding radiology report. The team trained the model using 900,000 chest X-ray reports that contained textual findings from radiologists.

To evaluate the model's effectiveness, the researchers took a set of 500 X-rays from their own ED, which had been previously reviewed by both a remote teleradiology service and an in-house radiologist between January 2022 and January 2023. These reports were individually compared to the AI-generated reports by six emergency department physicians, using a 5-point Likert scale for assessment. The sample consisted of 336 normal X-rays (67.2%) and 164 abnormal ones (32.8%), with frequent findings including issues like infiltrates, pulmonary edema, and pleural effusions, among others. Upon evaluation, the team found that the AI model's accuracy and quality of text were almost on par with the traditional methods.

“The generative AI model produced reports of similar clinical accuracy and textual quality to radiologist reports while providing higher textual quality than teleradiologist reports,” the group wrote. “Implementation of the model in the clinical workflow could enable timely alerts to life-threatening pathology while aiding imaging interpretation and documentation.”

Related Links:
Northwestern University

Gold Member
Solid State Kv/Dose Multi-Sensor
Ultrasound Color LCD
Remote Controlled Digital Radiography and Fluoroscopy System
Eco Track-DRF - MARS 50/MARS50+/MARS 65/MARS 80
Ultrasound System

Print article



view channel
Image: The pathways in the brain highlighted are those most affected by concussion (Photo courtesy of Benjamin Hacker et al)

AI Model Diagnoses Traumatic Brain Injury from MRI Scans With 99% Accuracy

A concussion is a type of traumatic brain injury that may lead to temporary disruptions in brain function. Occurring from incidents such as sports injuries, whiplash, or a simple bump to the head, many... Read more


view channel
Image: The new FDA-cleared AI-enabled applications have been integrated into the EPIQ CVx and Affiniti CVx ultrasound systems (Photo courtesy of Royal Philips)

Next-Gen AI-Enabled Cardiovascular Ultrasound Platform Speeds Up Analysis

Heart failure is a significant global health challenge, affecting approximately 64 million individuals worldwide. It is associated with high mortality rates and poor quality of life, placing a considerable... Read more

Imaging IT

view channel
Image: The new Medical Imaging Suite makes healthcare imaging data more accessible, interoperable and useful (Photo courtesy of Google Cloud)

New Google Cloud Medical Imaging Suite Makes Imaging Healthcare Data More Accessible

Medical imaging is a critical tool used to diagnose patients, and there are billions of medical images scanned globally each year. Imaging data accounts for about 90% of all healthcare data1 and, until... Read more

Industry News

view channel
Image: Calantic Digital Solutions is an orchestrated suite of AI radiology solutions that aims to transform radiology (Photo courtesy of Bayer)

Bayer and Rad AI Collaborate on Expanding Use of Cutting Edge AI Radiology Operational Solutions

Imaging data constitutes approximately 90% of all medical data, with the volume of such data continuously expanding, thereby significantly increasing the workload for radiologists amid existing resource limitations.... Read more
Copyright © 2000-2024 Globetech Media. All rights reserved.