In recent years, through its ability to collect and swiftly analyze huge volumes of data generated by imaging studies, artificial intelligence (AI) has been revolutionizing the practice of radiology.
Throughout the field, applications leveraging AI are being used to improve diagnostic accuracy, imaging consistency, workflow efficiency, and patient care by automating many formerly tedious, time-consuming, and manually performed tasks.
Traditional methods for implementing AI have relied primarily on machine-learning algorithms based on expert programming of predefined rules. More recent advances, however, have given rise to superior algorithms that learn through direct navigation of data to “recognize” potentially suspicious findings. These “deep learning” algorithms are advantageous because they operate with minimal human expert intervention, and instead collect and process data in raw form through the use of artificial neural networks.
Current and future clinical applications of AI in radiology range from improving image recognition and suspicious lesion identification to streamlining reporting through predictive analytics.
Ryan Lee, MD, Section Chief of Neuroradiology and Vice Chair of Safety and Quality at Einstein Healthcare Network, described how Einstein has been using AI for the past two years in lung nodule detection.
“In this specific instance, what we have found is that it doesn't actually save time, but it has allowed our chest radiologists to have increased confidence, knowing that this will pick up the lung nodules and they can then decide what it is. So, there are lots of things besides time-savings that AI and machine learning can do,” Dr. Lee said.
In addition to optimizing images, AI-based tools are being used to manage radiation and contrast doses. Click here to see Dr. Lee discussing the utility of these applications with Lawrence Tanenbaum, MD, Vice President and Chief Technology Officer, Medical Director Eastern Division, Director of CT, MRI and Advanced Imaging, at RadNet, Inc.
The holistic integration of AI tools can further enhance clinical workflow consistency, safety, and efficiency, ultimately providing patients with better care and treatment experiences. For example, software that controls and monitors contrast and radiation dose, and seamlessly connects to RIS, PACS, and EMR systems at the enterprise level, gives radiology practices easy access to data, standardized reporting formats, and improved performance capabilities.
“I think you can generate some very powerful data for protocol optimization and quality control, by being able to actually integrate the information about the radiation dose, contrast injection, and the image quality,” said Daniele Marin, MD, Medical Director of Multi-Dimensional Image Processing Laboratory at the Duke University School of Medicine.
NEXO [DOSE]®\Multi-Modality Radiation Informatics captures and records, enterprise-wide, real-time information needed to monitor and track trends, improve operating efficiency through “rules-based alerts” notification and implement of ALARA protocols, keeping a facility at the forefront of radiation safety and compliance.
NEXO® Contrast Management Solution is designed to improve compliance and track performance; provide real-time KPI capture to improve safety; streamline clinical workflow; and increase productivity in CT by monitoring patient safety, keeping up-to-date patient records, and seamlessly connecting and synchronizing data with RIS/PACS, EMRs, and voice dictation.
While challenges to implementation exist, progress in the use of AI in imaging is likely in the coming years, positioning radiologists to improve upon best practices and patient care.
“Some people refer to artificial intelligence as the new electricity,” Dr. Tanenbaum noted. “This is a bit of hyperbole, but in our daily activities, it appears all around us and these machine learning capabilities, this artificial intelligence augmentation of human capabilities, is going to help us be better radiologists.”
“AI is about concepts and constructs. It all goes back to the patient. We can become more patient-centric than ever. It’s not a tool. It’s how we approach AI. It’s personalized medicine, precision medicine, and it’s what’s best for the patient. That’s really why it’s inevitable.”
- Matthew J. Kuhn, MD, FACR
Back To Top
Artificial Intelligence and Imaging Optimization . Appl Radiol.