Radiologists Already Practicing Precision

The medicine of today is evidence-based, James Thrall, MD, said during the Moreton Lecture at ACR 2015. Its practice is based on proven relationships between diagnosis, treatment, and outcomes.

The medicine of tomorrow, Thrall said, is precision medicine.

A 2011 National Research Council

defined precision medicine (or personalized medicine) as “the tailoring of medical treatment to the individual characteristics of each patient…the ability to classify individuals into subpopulations that differ in their subspecialty to a particular disease or their response to a specific treatment.”

Thrall recalled his own epiphany of the need for precision medicine when he realized there were now 60 different variants of lymphoma. When Thrall started practicing medicine in the 60s, he said there were five variants.

There are two overarching concepts from the NRC white paper that radiologists need to be aware of, Thrall said: the basic point of the need to operate in more precise patient subpopulations, and the way we do this is based on understanding both the phenotype of disease manifestation and genotype and gene expression.

Imaging Phenotype
Thrall described phenotype as observable characteristics of an individual in relation to gene expression and exposure to other environmental factors. The term includes phenome made observable by technologies like imaging, laboratory testing, and pathology testing.[[{"type":"media","view_mode":"media_crop","fid":"38707","attributes":{"alt":"","class":"media-image media-image-right","id":"media_crop_6540026961632","media_crop_h":"0","media_crop_image_style":"-1","media_crop_instance":"3876","media_crop_rotate":"0","media_crop_scale_h":"0","media_crop_scale_w":"0","media_crop_w":"0","media_crop_x":"0","media_crop_y":"0","style":"float: right;","title":"James Thrall, MD","typeof":"foaf:Image"}}]]

“So the idea is that we want to combine both molecular and genomic information with new and better phenotyping methods, that’s us in imaging, to more precisely define disease taxonomy,” Thrall said.

Radiologists are accustomed to using classification grading and scoring systems, which are really ways of defining subpopulations, Thrall said. “Every classification grading or scoring system actually defines subphenotypes of disease manifestation, that’s the reason they are developed in the first place, to try to distinguish patients with different prognoses that might benefit from different therapies.”

“Image findings or imaging biomarkers taken together constitute the imaging phenotype of disease manifestation,” he said. “If we follow that one step further, the radiology reports that we’ve all generated over the years are really no more or less of a written description of the imaging phenotype of a disease or a condition.”

The point Thrall wanted to drive home was that radiologists have been practicing precision medicine all along.  He acknowledges that the practice of radiology hasn’t been thought of in these terms, but he urged the audience to keep in mind that they are “fundamentally in the business of creating imaging phenotypes.”

Thrall commented on the comprehensiveness of today’s image biomarker toolkit with functional, microenvironmental, metabolic, and molecular biomarkers.

“We can literally directly measure dozens of parameters,” he said. “It is through the systematic interrogation of these parameters that we’ve put together the imaging phenotypes of disease manifestation that can be used to help classify and define subpopulations of patients to guide therapy and research.”

Through imaging, phenotypes are identified which then guides the selection of treatments, which can then have its therapeutic response assessed through the application of imaging biomarkers. Thrall also mentioned the importance of imaging as a prognostic indicator. He compared the International Prognostic Scoring System, a commonly used tool in predicting prognosis, to using an FDG PET scan as a prognostic indicator. His example showed that no matter what the International Prognostic Scoring System score was, if the follow up metabolic imaging study is negative, the likelihood of progression-free survival was extremely high. A positive follow-up metabolic imaging study correspondingly indicated a low likelihood of progression-free survival.

“This is precision imaging,” Thrall said. “There is no other method of medicine to provide this kind of information; the pathologists cannot make this prediction, the geneticists cannot make this prediction, we make this prediction by imaging.”

Imaging Genotype/Gene Expression
Two years ago, an oncologist said to Thrall’s team at MGH, “You radiologists are out of date, you’re still talking about lung cancer, colon cancer, and kidney cancer; you should be talking about EGFR, KRAS, BRAF, and L3 arrangement and MEC.”

In heritable mutations, the role of imaging is not to make a genetic diagnosis, Thrall said, but to provide surveillance of the occurrence, location, extent, and severity of disease. In monogenic disorders, FDG PET is the only medical method available to demonstrate the total body burden of disease, he said.

In somatic mutations, Thrall gave examples of specific gene mutations that responded to particular therapies. It was through imaging that researchers were able to identify the response to therapy and extract that all of the patients who responded to a therapy had the same mutation. The standard of care in cancer centers is now to biopsy the cancer tissue to determine if the patient has a mutation that will respond to a therapy.

“We are our own little island in radiology,” Thrall said. “We are still talking about roentgen signs and image findings. We need to talk about imaging biomarkers and we need to learn the language that other physicians are using in the era of precision medicine. It will be a challenge for us all.”

There are opportunities for imaging in precision medicine, though, Thrall said. Imaging can be the earliest and most efficient way to assess response to therapy; radiology has information that’s unique that no other laboratory, pathological, or genetic testing method can duplicate.

“We are already doing more than we think,” he said. “We are already one of the supreme methods for phenotyping, and we are learning more and more about the relationship between image phenotype and genotype.”  

“How will imaging play the game? I like our chances,” he said.