vRad AI models are at work today prioritizing critical cases like strokes, improving reporting speed and accuracy, and assisting billing compliance. Below, neuroradiologist Josh Morais, MD recalls how AI helped promptly identify a stroke victim in need of immediate treatment, and offers his perspective on the daily impact he sees from AI.
Invisible assistance: Improving care
A non-contrast head CT that I read became the subject of a recent artificial intelligence case study, complete with my smiling face on the cover. The case study represents how vRad is using AI to help accelerate care delivery in critical cases, such as detecting the potential presence of intracranial hemorrhage.
At the time, I had no idea that AI was involved with this case. I was simply at my workstation, as usual, and selected the next emergent case to be read, which was highlighted in purple at the top of my worklist. The woman’s images revealed acute intracranial hemorrhage. I was able to alert the ordering physician just 2.9 minutes after the CT images had been uploaded, and the patient was quickly prepared for emergency surgery. The speed of care delivery helped significantly improve her results.
Later, I learned that AI had been at work in the background. The AI model had detected potentially critical findings, and automatically escalated the case to the top of my worklist. As a result, I was able to deliver a diagnosis to the physician about 10 minutes sooner than if the case had processed through my normal queue.
Everyone involved with stroke victims knows: Time is brain. Enabling a care team to act decisively 10 minutes sooner has the potential to hugely improve outcomes for victims of stroke and other cerebrovascular incidents. So, even if it’s invisible to me as I manage my caseload, it’s great to know AI is working to ensure the appropriate case is read next.
Tangible support: Making life easier as a radiologist
Typically, attention has been focused on radiology AI applications designed to process images and flag potential diagnoses, as in the case above. However, in my experience, there are a lot of practical applications beyond diagnostics.
Every day, AI helps me be a more effective, more efficient radiologist. Actually, vRad workflows blend AI, machine learning and just plain smart technology into an integrated system that streamlines tasks and supports everything I do.
For example, our dictation system and our report facilitator. I dictate brief positive findings into the system for each case, then I click “process.” A whole report is generated that includes my findings, complete with everything else relevant to the case filled in. It speeds things up, meaning the diagnosis is delivered faster to the clinician, and I can move on to the next case more quickly. What’s more, if the integrated natural language processing recognizes critical findings in the report, it will automatically alert me and autodial a call to the ordering physician, and connect me when she’s on the line.
Meanwhile, our program filters review each order, to ensure emergent cases receive priority attention at the front end, and to check for proper terminology to ensure billing compliance on the back end. Additionally, system workflows direct specialized cases to available radiologists with appropriate expertise or credentials. So, if a case is at the top of my list, I trust it deserves my immediate attention. And, as a neuro-specialist, I am confident that I will see cases when they require that expertise.
I’m glad to know vRad is investing in AI. Frankly, I think it’s essential. Any general medicine or radiology practice that is not exploring the potential of AI is going to get left behind.