Some radiology AI companies have started marketing their algorithms to departments outside of radiology, which illustrates the changing nature of the AI and medical imaging market. A key example of this trend is in acute care, where AI vendors like Aidoc and Viz.AI are going to neurology departments to sell stroke detection software, or pulmonology, interventional cardiology, vascular surgery or interventional radiology to sell AI for pulmonary embolism and aortic aneurism detection. These and other vendors have found new revenue streams by going directly to other medical specialty consumers of specific types of imaging.

"We are never going to exclude the radiologists, as imaging is still very central to healthcare," Parekh says. "But I think it is important to start to acknowledge other stakeholders."

He said this is why some AI vendors are looking beyond radiology. In terms of acute care teams for emergency activation of cath labs for emergency procedures, these companies are taking a multi-ology approach to reach out to the key players on these teams. While radiology pays a key role, early activation of these teams is essential for improved outcomes, so AI that can enable that is is key to several non-radiology specialties. 

"The procurement is for this type of AI is not just a radiologist-only decision or radiology department decision. It often moves up to a C-suite decision with input from these different stakeholders. The exciting thing for these AI vendors is that they are now targeting a bigger pot of money, maybe targeting an operation budget, which is far, far greater than a radiology department budget," Parekh explains. "Especially with these newer technologies, there might be more willingness to try them out."

Another aspect of this trend is AI moving beyond just radiology to help address things like population health that goes across hospital departments. AI vendors could leverage the news that the American Medical Association (AMA) released tracking CPT codes for some types of AI in 2022 to see how and where the technology is being used. Parekh said this usually points to information gathering to determine if a technology should gain reimbursement. 

Population health AI might include opportunistic screening algorithms that take a CT scan for cancer or trauma that looks all over the imaged area to assess the patient for heart disease, fatty liver disease, and an assortment of other incidental findings. If properly followed up on, this could be a boon for increasing patient encounters for preventive treatments or earlier interventions years before any of these diseases become symptomatic.   

"AI vendors can go to hospitals and say they can impact patient care from a broader population health angle rather than directly to diagnostic radiology only. And that is really exciting, because you are looking at improving patient care and intervening earlier in a patient's pathway and improving patient outcomes," Parekh says. 

If these types of algorithms gain reimbursement, it would be a double win for both patients and the healthcare systems using them.