Future Tense: Radiology’s Clinical Pathway

Twitter icon
Facebook icon
LinkedIn icon
e-mail icon
Google icon

Roderic Pettigrew, MD, PhD, is director of the National Institute of Biomedical Imaging and Bioengineering (NIBIB) of the National Institutes of Health (NIH). He states his organization’s goal simply: developing technology that can detect disease early, even at the molecular level, long before physical symptoms begin to appear.

“We want to achieve the earliest possible detection so the disease can be treated at its earliest stage,” he says. “In addition, in the past, those processes have been separate: You detect; later, you decide how to treat. We are focused on merging diagnosis and treatment in a single setting, with the ultimate goal of preventing disease.”

Pettigrew defines imaging broadly as the science of observation, and he is not alone in his perception of imaging as a clinical tool that far exceeds the confines of what is known today as radiology. Hedvig Hricak, MD, PhD, chair of the department of radiology at Memorial Sloan-Kettering Cancer Center (MSKCC), New York, New York, and a past president of the RSNA, takes a similarly open view. “Rather than being owned by one specialty, integrated diagnostics will be a collaborative effort. Large amounts of data (generated by each of the involved specialists) will be integrated, with the help of analytical computers, in order to provide the most comprehensive and detailed overview of a certain pathology or disease,” she predicts.

The dovetailing of clinical specialties will be made possible by advanced IT—technology so powerful that it will make today’s informatics systems look like filing cabinets by comparison, Hricak says. “In health-care centers today, we use computers largely for storing patient data,” she says. “Tomorrow, we will use them as intelligent systems that think with us to help us integrate large quantities of information productively.”


Integrated Diagnostics

Models for this productive integration of clinical data are already being developed in leading institutions around the country. Hricak and her colleagues at MSKCC have a collaborative agreement with IBM to work with the Watson artificial-intelligence system in a clinical setting; on the opposite coast, the University of California–Los Angeles (UCLA) recently formed a Radiology Pathology Center headed by Dieter Enzmann, MD, chair of radiology for Ronald Reagan UCLA Medical Center in Los Angeles.

“A lot of imaging is ordered for either the detection of cancer or the monitoring of cancer treatment,” Enzmann says. “To round out the data, we need something in addition to the imaging phenotype. As molecular therapies become more sophisticated, the clinician choosing between these therapies needs molecular information. Just being aware of a tumor’s size and where it is may not be enough to determine the proper therapy or to follow that therapy.”

While physically integrating the two specialties was comparatively simple—pathologists were located with radiologists in an imaging center—the informatics side of the equation presents a bigger challenge, Enzmann says. “What we are trying to do is create an integrated report from different databases in different departments, and those databases aren’t currently designed to communicate with one another,” he notes. “Then, there is the question of how referring physicians should interact with the integrated report. The goal is to make it easy to use, understandable, and interactive. It’s a work in progress.”

At the University of Florida College of Medicine, Anthony Mancuso, MD, chair of the department of radiology, and Christopher Sistrom, MD, MPH, PhD, CIO in the department of radiology, are collaborating on a conceptual model of diagnostic imaging that maps clinical scenarios and imaging procedures in a 2D matrix. “It’s a way of picturing the domain of imaging. It’s a big space, with millions of plausible cells; any one radiologist might be able to know only a small cluster of cells in the matrix. This is a huge domain that requires a team to figure out,” Sistrom explains.

Daniel Sullivan, MD, professor of radiology at Duke University, heads another such initiative, the Quantitative Imaging Biomarker Alliance (started by the RSNA in 2007). It brings together academic researchers, device manufacturers, and information from clinical trials and clinical practice to develop quantitative measures for imaging. “We can’t get good studies if we just use subjective interpretations of imaging tests,” Sullivan notes. “We need objective measures. For radiology to be relevant