Leveraging the Power of AI in Radiology

What are your day-to-day responsibilities as the service line director of Edward-Elmhurst Health?

I am responsible for the radiology and imaging service line at two hospitals and 16 outpatient centers. We have one hospital with about 400 beds and other with roughly 300 beds that supervise about 500,000 imaging procedures a year. My financial acumen helps me oversee the day-to-day budgeting operations of two radiologist groups.

Starting as an x-ray, technologist, I became a special procedures technologist within a couple of years. Soon I was awarded a quality control (QC) coordinator position, wherein I was given more responsibilities as things changed. I was willing to take on more responsibility, and eventually became a system director, and then a service line director.

What would you say are some of the most prominent challenges in the industry today? 

The two most significant issues probably right now are the supply chain and staffing. For staffing, we partnered with four different colleges to create a relationship with students who are interested in working in healthcare as a career. We have four different school affiliations, where they work for us in the clinical setting as students. Usually, these programs last for a couple of years; if we have a position open, we hire them.

On the supply chain side, we work with different vendors to reduce some of the costs and get creative with contracts. We leverage our global purchasing organizations (GPOs) in those contracts. As the renewal comes up, we negotiate the costs down. For example, we had a computed tomography (CT) contract with a healthcare firm in Shanghai, China. Due to the pandemic, this facility was closed down. Seventy percent of our contracts were with that firm. We were fortunate and did not have to amend our protocol because we had a relationship with multiple other facilities that provided us.

What are some of the trends going on in space today?

Artificial intelligence denoising is the most recent trend today. We collaborated with a company that uses AI and deep learning to cut the noise out of the image to improve image quality. Reducing the time of study increases the noise. Applying deep learning AI algorithm, we’re able to decrease the time of the study but still maintain the same image quality. This technology reduces the MRI timing giving better results and providing better services to claustrophobic patients. We reduce the patient’s MRI time by 35 percent with this algorithm. The reduced time scan time results in reduced table time, allowing us to take up more patients.

AI software in radiology is another emerging trend in the space. Radiologists that continuously read images use AI to find potential pathology by running the image through an algorithm. If the algorithm finds any potential pathology, the image will be prioritized for the patient to get an early diagnosis.

What advice would you share with your peers and aspiring individuals in this space to navigate these complex trends and challenges?

Do your due diligence and do your homework by investigating the companies that you are working with. There are so many facilities that offer a product or service. If you engage in a contract with them and put all your eggs in one basket, then you are running a risk. In the case I mentioned about the contrast in Shanghai, China, a lot of people felt the pain from that. Having a well-versed group of contractors or vendors from whom you can purchase the same product is important.

Another piece of advice would be on AI technology. Again, do your due diligence in evaluating the product at hand, and make sure it works well for your specific organization and your specific processes. The original equipment manufacturers (OEMs) also sell AI technology. However, it would be proprietary to that specific vendor, whereas you can find an agnostic vendor that will work with any brand, and you will get more bang for your buck.