This story is over 5 years old.

Digital Imaging and AI Mean Fewer Surgeries for Breast Cancer Patients

Advances in medical imaging, coupled with artificial intelligence, may obviate the need for additional surgeries.
Image: LarsZ/Shutterstock

A project led by Pennsylvania's Lehigh University might help cancer patients avoid undergoing multiple surgeries. The results can be crucial since according to recent studies, 25 percent of women who undergo breast-saving lumpectomy surgery will require a second operation, which incurs a median cost of $16,000 and causes further health complications.

After removing the tumor from the breast, doctors have to check the operated area for residual cancerous cells. Currently, the process of examination involves taking tissue samples from the margin of the operated area, freezing them with liquid nitrogen, sectioning them to thin slices, and sending them to a lab for examination. The process, known as histopathology, can take as long as a week. If the histopathological analysis finds cancerous cells in the tissues, the patient must undergo another surgery.


But the diagnostic method devised by the researchers at Lehigh University uses advanced imaging techniques and artificial intelligence algorithms to speed up the process and enable real-time scanning and evaluation of the operated margin without the need for extracting tissues.

"If used during surgery, this technique has the potential to significantly reduce the need for a second breast cancer surgery," said Chao Zhou, assistant professor of electrical engineering at Lehigh who is leading the effort, in an email to Motherboard.

"If used during surgery, this technique has the potential to significantly reduce the need for a second breast cancer surgery."

The computer-aided diagnostic uses Optical Coherence Microscopy (OCM), an advanced imaging technique that can create high-resolution visualizations of biological tissues at the cellular level, without the need for invasive techniques such as biopsies. The obtained images are then fed to AI algorithms that scan them for presence of cancerous cells.

"OCM images reveal distinctive texture features of benign and malignant tissues, which can be subtle to distinguish," Zhou said. "Computers are more efficient in spotting these differences than humans."

AI algorithms have already proven their efficiency at reducing costs and error margins in diagnosing and treating various forms of cancer.

The new imaging apparatus. Image: Chao Zhou/Lehigh University

"The process takes a large number of images, and labels the types of tissue in the sample," said Sharon Huang, associate professor of computer science and engineering at Lehigh. "For every pixel in the image, we know whether it is fat, carcinoma or another cell type. In addition, we extract thousands of different features that can be present in the image, such as texture, color or local contrast, and we use a machine learning algorithm to select which features are the most discriminating."


The researchers trained the algorithms with a large number of OCM images obtained from the patients, and tested the results against those obtained from the histopathological analysis of the same patients.

The results of their experiments, published in a paper in the Medical Image Analysis Journal, show that the technique yields 93 percent accuracy in classifying benign and cancerous cells.

However, the researchers don't think their method can yet be considered a replacement for histopathology yet. "Histopathology has been used for a hundred years and is the current gold standard for clinical diagnosis," Zhou said. "For our OCM results, we were using histopathology as the ground truth to get the sensitivity and specificity results. Our results are very good (close to 94 percent accuracy), but that is still compared to ground truth set by histopathology. We can not say OCM is more accurate compared to histopathology."

The true value of OCM diagnosis, Huang pointed out in a phone call with Motherboard, is its superior speed, which makes it a highly efficient complement to histopathology by detecting residual cancerous tissues in the surgery room before biopsies are sent for examination.

OCM diagnosis is still in its development and testing phase and hasn't been deployed in full yet. The Lehigh University team plans to deploy the technology at a local hospital, where its effectiveness in treating and diagnosing breast cancer patients will be evaluated.

Get six of our favorite Motherboard stories every day by signing up for our newsletter .