While digital breast imaging through mammography is effective in detecting breast cancer even in its early stages, it can lead to abnormal findings that necessitate an additional visit to obtain additional views, what is called “recalls”. This undoubtedly leads to anxiety and additional costs, even when the additional confirmatory imaging shows that there was no suspicious breast lesion that needs to be biopsied, i.e. “false recalls”.
A study from the University of Pittsburgh that involved radiologists, bioinformaticians, bioengineers and artificial intelligence experts shows that convolutional neural network models (a type of artificial intelligence algorithms) can identify nuanced findings on mammograms and distinguish truly benign findings from malignant findings, which may eventually lead to a computerized clinical toolkit that works side by side with the radiologist to help reduce the necessity of false recalls.
Dr. Shandong Wu, an assistant professor of Radiology, Biomedical Informatics and Bioengineering at the University of Pittsburgh, and principal investigator of the study, said “We showed that there are imaging features unique to recalled-benign images that deep learning can identify and potentially help radiologists in making better decisions on whether a patient should be recalled or is more likely a false recall” in a statement. “Based on the consistent ability of our algorithm to discriminate all categories of mammography images, our findings indicate that there are indeed some distinguishing features/characteristics unique to images that are unnecessarily recalled. Our [artificial intelligence] models can augment radiologists in reading these images and ultimately benefit patients by helping reduce unnecessary recalls.”
The study is published in Clinical Cancer Research journal in October 2018. The entire article link can be found here.