RSNA 2017: Economics, quality, tech regulations determine value of imaging

What determines value in medicine will differ based on specialty, but what should radiologists consider when trying to maximize value in imaging? "The Value of Imaging," a Nov. 26 session at RSNA 2017 in Chicago, aimed to provide answers.

Three experts explored what constitutes value in medical imaging, specifically the economics of information, measuring imaging quality, appreciation of quality progress and the regulation of new imaging technology by the U.S. Food and Drug Administration (FDA).  

Saurabh Jha, MD, an associate professor of radiology at the Hospital of the University of Pennsylvania, opened the session by introducing the economics of imaging by analyzing the basics of information theory, how to appreciate the purpose of medical information and the appreciation of threshold basis decision-making.  

He explained that currently the value of healthcare can be aligned with uncertainty. The value of imaging is based on what radiologists have achieved rather than what they are actually doing and saying, according to Jha. He also argued the underlying problem with value in imaging itself is its subjective nature. However, the value also lies in discernment.

"There is value in discernment, the value of actually getting to the right diagnosis," Jha said. "It is paradoxically the harm in treatment which drives the benefit of the value of imaging. The harm of imaging actually yields the value of radiologists."

Diagnostic accuracy, he asserted, is necessary in imaging as it has the power to alter clinical decision making.   

Continuing the conversation, Jeffrey Robinson, MD, MBA, assistant professor of radiology at the University of Washington, discussed imaging quality. Measuring quality is possible, but it is dependent on the correct interpretation of findings and accurate diagnosis.  

"Imaging opens doors and revels diagnosis or absence, information reduces possibilities, and information (positive or negative) must change management," Robinson said.  

Currently, many assessments of accuracy suffer bias, Robinson explained. The two different unbiased approaches radiologists should use to measure accuracy is blind panel prospective review and blind expert retrospective review.  

However, Robinson warned that ensuring accuracy and measuring error in a diagnosis are key components to the quality of imaging, describing examples of images that led to misdiagnoses and detrimental patient outcomes due to inaccuracy to attendees.   

"Quality is more than what can easily be measured," Robinson concluded. "I submit that accuracy really is the key to quality for radiologists. What's important to us is diagnostic accuracy and is the key to objectivity in peer review when you're trying to determine accuracy or what should or shouldn't have been identified."  

Lastly, H. Benjamin Harvey, MD, JD, assistant professor of radiology at Harvard Medical School, ended the session by discussing FDA regulatory structure for approving medical devices into radiology.  

The charge of the FDA, Harvey explained, is to protect and promote public health and reduce regulatory burdens while simultaneously trying to reduce the time to get devices on the market.  

According to Harvey, medical devices under inspection are classified by the FDA into three classes loosely based on their risk, with most radiology devices going through 510 (k) clearance.  

"There's significant differences in premarket performance and policies for different types of radiology devices, so it's important to know what your device is and try to manipulate or adapt your intended use to get through a process that will be most efficient and most profitable and effective," Harvey said.

One way to do this is to utilize the pre-submission process to seek feedback prior to market submission. Harvey recommended this process for new artificial intelligence-related radiology devices trying to pave way into the market. AI radiology devices and technology are more likely to go through a completely different FDA clearance process due to the level of innovation, such as CADe (computer aided detection) and CADx (computed aided diagnosis) software. However, also known as the FDA's De Novo Classification Process, Harvey explained that one can claim that the newness and novelty of AI imaging technology needs to be down classified to have a chance be cleared. 

"Recent decisions of FDA movement demonstrate that they really proactive in trying to foster rather than stifle the development in this field [radiology]," Harvey concluded.