Radiologists know that precision medicine is important to moving healthcare forward, but they may not realize how important diagnostic imaging is to precision medicine.
In fact, diagnostic imaging is one of only four ways to phenotype a patient, according to a Diagnostic Imaging article on the topic. The others include clinical history with physical exam, lab testing, and histopathology and immuno pathology.
Dr. James Thrall, chief of radiology at Massachusetts General Hospital, says radiology reports are essentially a description of the imaging phenotype of disease manifestation, although they are not usually thought of in those terms.
Precision medicine relies on phenotyping systems to segment patients into different prognostic categories. For example, Thrall’s colleague developed a system to score patients with intracerebral hemorrhage according to the number of contrast spots on their CT angiogram. The five scores divide patients into risk categories based on levels of bleeding.
Molecular imaging has perhaps the greatest potential in moving precision medicine ahead, according to Thrall. For example, when Hodgkin’s lymphoma patients receive an FDG PET scan after several rounds of chemotherapy, some show no abnormal uptake of tracer, and those patients have been shown to have a normal life expectancy. Patients showing a continued presence of abnormal uptake have significantly lower life expectancy. Thrall points out that using imaging to create these two phenotypes is more accurate than the International Prognosis Scoring System. His Moreton Lecture at ACR 2015 in Washington, DC in May focused on the role of imaging in precision medicine.
Another opportunity for diagnostic imaging to participate in precision medicine lies with improved identification of follow-up treatment for smokers. Dr. Eliot Siegel, professor and vice chair of research informatics at the University of Maryland School of Medicine, tells Diagnostic Imaging this is an ideal time to make these recommendations more precise; millions of smokers aged 55 to 77 will receive low-dose lung CT screening now that Medicare covers the test.
Siegel points out the current recommendations for follow up of lung nodules are based mainly on the patient’s smoking history and size of the nodule. He suggests adding parameters such as the lobe in which the nodule is found, the nodule shape, the patient’s age, and the patient’s ethnicity using National Lung Screen Trial Data. A large database using multiple parameters would yield more precise recommendations for follow up, he says.
Both Siegel and Thrall agree that more work is needed. Siegel notes that less variability in medical imaging data will be required to create more searchable databases. Thrall has a semantic suggestion that can be put into immediate use—he urges radiologists to talk about imaging biomarkers and phenotypes rather than signs and rank.
Stay up to date on the latest news and development in diagnostic imaging by subscribing to our blog.