Algorithms, MRI scans help predict 10-year breast cancer recurrence
Researchers at Penn Medicine have used radiomics, a method that leverages algorithms to extract large amounts of features from medical images, to provide more personalized tumor characterization for breast cancer patients.
In the process, the algorithms were able to successfully predict recurrence-free survival after 10 years, according to a study published in the journal Clinical Cancer Research.
Researchers contend that the use of imaging to characterize the genetic makeup of tumors can pave the way for individualized, non-invasive treatment.
“Our study shows that imaging has the potential to capture the whole tumor’s behavior without doing a procedure that is invasive or limited by sampling error,” says Rhea Chitalia, lead author of the study and a doctoral candidate in the School of Engineering and Applied Science at the University of Pennsylvania. “Women who had more heterogeneous tumors tended to have a greater risk of tumor recurrence.”
Researchers extracted 60 radiomic features—or biomarkers—from 95 women with primary invasive breast cancer and, after following up with the patients 10 years later, they found that an MRI scan that showed at the time of diagnosis a high diversity of cells (high tumor heterogeneity) could successfully predict a cancer recurrence.
“We’ve just touched the tip of the iceberg,” says principal investigator Despina Kontos, an associate professor of radiology in the Perelman School of Medicine at the University of Pennsylvania. “Our results and the validation study give us confidence that there are many opportunities for these markers to be used in a prognostic and potentially a predictive setting.”
Although medical imaging may not completely replace the need for tumor biopsies, radiomics could enhance the current “gold standard” of care by providing a more detailed profile of a patient’s disease and guiding personalized treatment, contends Kontos.
“If we’re only taking out a little piece of a tissue from one part of a tumor, that does not give the full picture of a person’s disease and of his or her response to specific therapies,” adds Kontos. “We know that in a lot of instances, patients are over-treated, getting therapy that may not be beneficial. Or, conversely, patients who need more aggressive therapy may not end up receiving it. The method we currently have for choosing the appropriate treatment for patients with breast cancer is not perfect, so the more steps we can take toward more personalized treatment approaches, the better.”
Going forward, researchers plan to expand their analysis to a larger patient cohort and delve into which specific markers are more predictive of particular outcomes.