Precision Medicine Paves the Way for Patient-Specific Care
Monday, Nov. 30, 2015
In the not-too-distant future, imaging will no longer be applied generically for specific clinical indications, but instead will be tailored to meet the individual needs of the patient, according to the presenter of a Sunday session.
Presenter Ella Kazerooni, M.D., said radiologists will be able to leverage informatics to extract information from imaging alone or together with the huge of amount of information available through the electronic health record (EHR) and other social data to deliver more patient-specific care.
"Think about Google Maps and think about the many layers of data that exist in Google Maps," Dr. Kazerooni said during her presentation. "A vast amount of data goes into what we see as a very superficial display and take for granted. But can we do that with healthcare—can we integrate vast arrays of data to bring those to patient and provider?"
Instead of making a generic recommendation—that, for instance, a patient with a cancer needs to have an annual PET or CT scan—providers can use more personalized data to say this particular patient is a bit more at risk for a recurrence and should be imaged more frequently, she said. "Or the patient could have a lower risk and need imaging less frequently."
This approach can also make a difference in the type of imaging utilized, said Dr. Kazerooni, a professor of radiology, associate chair for clinical affairs, director of cardiothoracic radiology and service chief of diagnostic radiology at the University of Michigan.
In terms of thoracic CT scans, for example, protocols can be written in many different ways—depending on the individual questions being asked, she said.
Something as simple as shortness of breath can be linked to any number of conditions, "but, by knowing more about the patient, we can do a more disease-specific protocol rather than using a generic one-size-fits-all approach," Dr. Kazerooni said. "We can provide more precise—often quantitative—information to help follow a patient's disease over time."
While the idea of using "Big Data" to provide more precise medical care in imaging makes sense, the approach is still in the concept stage, she said.
"The proof isn't out there yet—how it works in practice to lead to better outcomes is something that we need to see on a broader scale," Dr. Kazerooni said.
Gradually, the approach is taking radiologists from the era of description and largely qualitative reporting into a quantitative mindset—an approach that could require changes in day-to-day practice, Dr. Kazerooni said. "It could mean that radiologists will have to change from being more descriptive in the way they report their exams to being more quantitative," she said.
She pointed out that RSNA developed the Quantitative Imaging Biomarkers Alliance (QIBA), with the idea of transforming patient care by making radiology a more quantitative science.
"When you can start reporting metrics quantitatively, you have the data points you can extract from imaging that are much more precise," she said.
Structured Reporting Key to Data Analysis
Structured reporting is critical to extracting and analyzing this kind of data. As an example, Dr. Kazerooni pointed to a software tool she and her colleagues use at the University of Michigan to view patients' lung tissue.
Instead of reporting that a patient has mild emphysema that is centrilobular or paraseptal in its disease type, Dr. Kazerooni and her colleagues have quantitative metrics of chronic obstructive pulmonary disease (COPD) they can put into their structured report, which can include total lung volume, the percentage that is normal lung tissue, the percentage that is emphysema, and—most importantly—the percentage of the tissue that is functional small airway disease or what is essentially pre-emphysema.
"I can take individuals with the same clinical stage of disease and I can show that they have very different lung imaging signatures," she said. "And that while one already has emphysema, the other patient has no emphysema. And if we can find all those population risks [for that patient without emphysema] that exist in the EHR, and all the social data and exposures—identify them and potentially treat them—we can prevent what we know as emphysema or late-stage tissue destruction.
"We're extracting information that has been in CT scans for years, but we haven't been able to measure, describe and report it in a very precise manner," Dr. Kazerooni said. "These are the kinds of software tools that are being developed and commercialized that radiologists will be able to use—and the one for COPD is just one of many in development."