AI Features in Radiology to Adopt Today
Reading Time: 9 minutes read
AI delivers enhanced imaging and improved workflow for medical imaging.
Discussions about potential applications for artificial intelligence in radiology have dominated medical imaging conferences and news headlines for years. All the noise and hype can make it hard for imaging providers to know which AI-based applications will deliver value. In this blog, I’ll give you an overview of AI features for radiology that you can adopt today to help improve image quality, optimize dose and increase workflow efficiency.
Carestream is applying AI – and in other cases sophisticated proprietary algorithms – in our imaging software to deliver clinical value in Imaging Intelligence and Workflow Efficiency. Let’s look at the specific features in each of these two fundamental areas of medical imaging.
Imaging Intelligence: AI in image capture for radiology
First, let me introduce Eclipse, the engine that powers our imaging software. Eclipse uses AI technology and proprietary algorithms that help improve image capture. Beneficial outcomes of this are improved visualization and consistent presentation. It also provides features that enable our customers to better optimize radiation dose, and to create custom preference looks. We call this suite of imaging solutions “Imaging Intelligence”. Let’s look at some of these specific AI-driven features for image capture.
Carestream’s newest noise reduction method – Smart Noise Cancellation (Pending 510(k) clearance) – is ground-breaking technology that leverages AI to remove noise while simultaneously preserving image structures. The amount of noise subtracted can be adjusted by your imaging facility to meet your desired image clarity. Carestream is among the first medical imaging companies to use AI to separate and remove noise from an image. Smart Noise Cancellation will be available starting with our DRX-Evolution and the DRX-Evolution Plus Systems.
Another AI-powered application for radiology that is available today is our Bone Suppression software. Based on a powerful form of AI called Random Forests, this tool creates a companion image that suppresses the appearance of bone and enhances the visualization of soft tissue to increase clinical confidence of pathology assessments. This feature requires no additional exposure to the patient.
In addition, we offer other Imaging Intelligence features that apply algorithms to improve image quality and optimize dose, bringing added value to clinical decision support.
- Tube and Line provides a companion image with optimized processing for clearer, easier visualization of PICC lines and tubes. It increases confidence that tubes and lines are placed correctly and remain in position.
- With EVP Plus, radiologists have the freedom to adjust, manipulate and view images in harmony with their preferences. EVP Plus supports multiband frequency processing to provide better noise control, sharpness, contrast and density while minimizing artifacts.
- Black Surround masks the white collimated area to eliminate foreground flare and automatically define your area of interest. It allows manual selection of complex anatomical structures for rapid, accurate collimation verification.
- Pediatric Image Optimization and Enhancement provides default acquisition techniques and image-processing parameters optimized specifically for each patient body size, from the smallest neonatal patient to the largest adolescent.
- Smart Grid software provides image quality comparable to images acquired with anti-scatter grid in bedside chest imaging.
- Auto Long Length Imaging captures a wide range of vertebral or long-bone images with the patient in the upright or supine position. The captured component images are automatically aligned and stitched to create a composite image.
- Dual Energy allows for improved detection sensitivity for abnormalities in the lung, and improved assessment sensitivity of bone abnormalities.
- Digital Tomosynthesis uses lower dose than standard CT exams and provides better depth localization compared to conventional projection radiography.
- Pneumothorax Visualization creates a companion image from original exposure that is optimized to accentuate the appearance of free air in the chest cavity.
All these advanced processing features help deliver the best possible image quality at the dose chosen by the operator. Improved image quality not only helps with diagnostics, it can also ease the burden on overworked radiologists. Some studies find that the average radiologist must interpret one image every three to four seconds to keep up with their increasing workloads. AI technologies in radiology that subtract noise within an image or suppress the appearance of bone can help make images easier to read – which in turn can help reduce physician fatigue.
Workflow Efficiency: AI in radiology workflow
AI applications available for radiology today also can help automate workflow steps. Carestream’s solution set, that we call Workflow Efficiency, can help improve workflow and productivity for the radiographer.
Automation is especially beneficial in busy departments with limited resources. A few seconds of saved time can help patients be triaged more quickly. I believe the need for automation is more important than ever today. Radiographers, along with other hospital staff, are experiencing burnout from the stress of imaging COVID-infected patients. A few additional automation features can speed up exam time, and give your radiographers a bit of relief, too.
Here are our AI-driven radiology solutions for Workflow Efficiency – none of which require additional training:
AI for chest imaging
We know that chest imaging is one of the most common exams performed so we’ve applied AI to create improvements specific to this medical imaging exam. Some of the features help the radiographer acquire images in a consistent manner from patient to patient, enabling more consistent presentation and quality. Other features help speed up workflow; and some of them do both. Let’s take a closer look at few of the AI features in our Workflow Efficiency set.
- Anatomy Clipping Software uses Artificial Intelligence to outline areas of the chest anatomy that may have been collimated off the image. This anatomy clipping feature will be especially beneficial in a teaching institution with students or inexperienced radiographers.
- Chest Image Auto-Correct helps saves time by applying AI to automatically display chest images in the proper orientation. This can accelerate workflow for radiographers, and images can be sent to the PACS more quickly.
Advancing automation with smart workflow
Recently, Carestream introduced new AI-based Smart features that take workflow efficiency for imaging exams to the next level. Our “smart room” has features that help optimize equipment positioning, patient positioning and radiographer workflow. In addition to increasing patient throughput, these features can help improve image consistency and reduce retakes.
These smart features can benefit both highly skilled and less-experienced radiographers. Advances in automation also should decrease the time that a radiographer needs to be in close contact with patients for positioning — a huge benefit in the presence of infectious diseases. Here are AI-driven automated features within our Workflow Efficiency solution set that you can purchase and put to use today:
- Smart Auto Position automatically moves the wall stand and tube head according to the active view and patient height. This reduces the need for physical contact between the radiographer and patient.
- Smart Patient Positioning augments a live patient view that indicates the detector boundary, along with patient position and pose verification. It also provides suggested correction to the radiographer on the console. The radiographer can then provide guidance to the patient on needed adjustments.
- Smart Collimation adjusts the collimator blades in accordance with the detector height and detected width of the patient in front of the wall stand bucky. There are 2 preset widths and 4 widths collimation sizes available.
- Smart Technique automatically detects the patient size to apply an appropriate APR setting that has been previously configured by the facility.
AI features in radiology to adopt today
I hope that this blog has demonstrated that there are very real and beneficial AI applications that can be applied today in medical imaging. If you’re still not sure which ones are right for you, I recommend that you consider your biggest pain points. For example, do your radiologists have a high volume of studies to review that is causing a backlog and fatigue? In that case, our Intelligent Imaging features can help make images easier to read.
Is your facility facing an increased number of exams? In this case, AI-driven Workflow Efficiency can help your radiographers complete exams more efficiently and more consistently.
Is the transmission of infectious diseases a concern? Then our suite of Smart Workflow features can help shorten the time that radiographers spend in close contact with patients.
I encourage you to look beyond the hype and take a closer look at AI technology that is available to you today that can help increase the value of the entire imaging chain, from capture to diagnosis.
*Software featured in this blog is available only on limited Carestream products.
Download the Eclipse Brochure.
Learn about our investigation of NVIDIA Clara AGX — an embedded AI platform for medical devices.
Read our blogs on:
Ron Muscosky is a Product Line Manager at Carestream Health. He has more than 30 years of experience in healthcare/medical imaging.