The promises of artificial intelligence (AI) have captured the interest and imagination of just about every industry, and healthcare is no exception. Radiology is an area that has already seen an impact from AI innovation. In fact, roughly a third of radiologists are currently utilizing AI in their practices according to a recent ACR Data Science Institute survey. However, those practices are only employing an average of slightly over a single algorithm, suggesting a limited number of available tools and significant room for growth once users find the right AI applications for their needs.
Although AI is still in relative infancy, it is already proving useful in a variety of tasks and in service of some important goals in the healthcare field. Here are four ways AI is reshaping radiology as we know it.
1. Flagging irregularities and prioritizing workflows
A radiologist brings the combination of knowledge, experience and a good eye to the process of reading medical images, often taking in a number of different details in an image to draw conclusions. Artificial intelligence is still a way off from being able to replicate this ability, but AI has proven able to detect some potential issues in medical images. For example, AI tools like MammoScreen and CMTriage from CureMetrix can assess the likelihood of malignancy in mammograms.
While this capability can’t necessarily be used for a diagnosis, it can help to organize and prioritize images in order to optimize a radiologist’s workflow. By using AI as a helper to categorize images or draw attention to especially worrisome ones, radiologists can use their time more efficiently.
This functionality was put to the test by CureMetrix, which conducted an “AI Test Drive Challenge” at the recent Radiological Society of North America meeting, demonstrating that radiologists with AI assistance performed 40 percent faster on reads and had 25 percent improved accuracy.
2. Triaging emergencies
Over the course of the Covid crisis, different regions have become inundated with patients and have needed help doing triage — this was especially the case in the early months of the pandemic as healthcare workers were still learning how to diagnose and manage the disease.
Hospitals and clinics overwhelmed by patients presenting with Covid symptoms needed to be able to quickly determine which patients required immediate medical attention, and imaging — particularly chest x-rays — proved to be one good method for recognizing troubling symptoms.
This need inspired healthcare start-up Qure.ai to re-purpose its AI-powered chest x-ray tool so that it could look for signs of Covid — enabling overwhelmed frontline clinicians around the globe to better manage cases. This is a fantastic example of how AI can be deployed to assist healthcare workers facing emergency situations, and how a smart company can evolve its own technology quickly to meet a new demand.
3. Providing access to care in underserved communities
The ongoing problem of hospital closures in rural areas in the U.S., and a lack of enough specialists both in the States and in remote areas abroad, means that there can be a backlog in reviewing medical imaging. So, even if it’s possible to provide equipment like CT scanners and ultrasound machines to areas in need, there is no guarantee that an experienced person will be readily available to interpret the results.
AI is providing needed support in these resource-strapped areas. AI programs that are able to go through images and flag ones that appear to show something amiss can help prioritize images that may more urgently require the attention of a radiologist. The program can even be designed to automatically send flagged images electronically to a radiologist in another location who can examine them.
4. Enabling easy image sharing among professionals and patients
Have you ever been asked by a new provider to bring your previous medical scans with you? Some patients are shocked to discover they have to travel to a radiology lab and sit there while an administrator burns their images onto a CD (a near defunct medium) and then physically bring the CD along to the new appointment. Fortunately, AI is making it possible to forgo this costly, time-consuming and very out-of-date method of image sharing. More and more radiology departments are putting images online to take advantage of new tech innovation, making it easy for radiologists to share imaging with other healthcare providers and patients alike.
To do this requires the adoption of a cloud computing platform, a sector that is exploding as businesses across industries seek to become more interoperable.
Some radiology practices, especially smaller ones, may be reluctant to adopt a new platform because of concerns about cost or the effort needed to get everyone adjusted to a new system. However, cloud platforms are essentially the critical infrastructure required to make good use of AI, so there is a powerful business case to be made for investing in such a transition — especially since AI will only continue to become more prevalent.
by Morris Panner
When reviewing an idea, my advice to leaders is to focus on strategic alignment. Leaders need to provide clear strategic guidance on the arenas they want their innovation teams to explore. It is on the basis of this guidance that they can then evaluate whether an idea is worth investing in.
Last month, the FDA announced the 12 winners of it’s Low or No-Cost Tech-Enabled Traceability Challenge, which was part of the New Era for Smarter Food Safety initiative. The initiative aims at achieving end-to-end traceability throughout the supply chain.
Digital tools enabled by advanced analytics, artificial intelligence, and machine learning can help companies uncover the fastest and most effective path to abating the O&G industry’s greenhouse gas emissions.