Predicting Breast Cancer Recurrence After 10-Years with MRI Scans
Researchers at Penn Medicine have developed an imaging technique that can provide non-invasive characterisation of tumour heterogeneity. They used Magnetic Resonance Imaging (MRI) and radiomics – an emerging field of medicine that uses algorithms to extract large amounts of features from medical images. Published in Clinical Cancer Research, the technique can be used to identify the heterogeneity of cancer cells to better understand the causes and progression of the disease and possibly predict its recurrence.
Current detection methods to diagnose breast cancer include blood tests, mammograms, MRI, and breast biopsy. Whilst these can detect the presence of cancer overall, understanding the cellular makeup of the tumour is important to understand but difficult to do. Tumours can vary between people, but also within the same tumour itself. In addition, biopsies only capture a small number of cells, which may not be representative of the whole tumour environment.
There are some cases where patients are over-treated and getting therapy that may not be beneficial, or some who need more aggressive therapy who may not end up receiving it. It’s important for patients to get the right treatment for them, and since current detection methods are not perfect, the more we can develop these approaches the better it will be for patients.
The researchers extracted 60 radiomic features from the images of 95 women with primary invasive breast cancer using MRI and from each scan, a “signal enhancement ratio” (SER) map was generated. The patients were followed for 10 years and it was found that pre-treatment scans showing the highest tumour heterogeneity were more likely to have a greater risk of tumour recurrence. This algorithm was then compared to an independent sample of 163 patients with breast cancer to validate their findings, which was supported and showed that the imaging technique could be used to successfully predict cancer recurrence.
Imaging has the potential to capture the whole tumour’s environment without carrying out a procedure that is invasive or limited by sampling error. What is currently the “gold standard” for diagnosing breast cancer is debated, and even though imaging may not completely replace the need for tumour biopsies, radiologic methods could allow patients to get a more personalised treatment by giving a more detailed profile of the tumour.