Method Could Potentially Automate Breast Cancer Detection

An international, multidisciplinary team of researchers have potentially created an automated way to detect breast cancer in histolopathology images of cells.

The team of engineers, mathematicians, and physicians used a technique for finding underwater marine structure damage to detect cancer cells in biopsy tissue images. This technique could potentially automate screening and improve detection, according to a study published in PLOS ONE.

Currently, the prognosis of breast cancer is determined using the Bloom-Richardson grading system. This process relies on visual examination of tissue biopsy under a microscope by a pathologist, which leaves room for potential variation in grades and mistakes.

The creation of digital pathology and digital slide scanners has also created a way to automate the process through image-processing methods. However, the image-processing methods may not accurately diagnose high-grade breast cancer cells because they are typically clustered, and have vague boundaries that make it difficult to detect, according to the study.

“Detection of cancerous nuclei in high-grade breast cancer images is quite challenging and this work may be considered as a first step towards automating the prognosis,” said Joy John Mammen, MD, head of the Department of Transfusion Medicine & Immunohaematology at Christian Medical College, Vellore, India.

The novel imaging method seeks to overcome that difficulty and provide an accurate and automated way to diagnose high-grade breast cancers.

“This unique research group could draw on a broad and deep knowledge base. Experts in numerical methods and image-processing liaised with medical pathologists, who were able to offer expert insight and could tell us precisely what information was of value to them,” said researcher Bidisha Ghosh, PhD, civil engineering professor at Trinity College Dublin. “It is an excellent example of how multidisciplinary research collaborations can address important societal issues.”

The method has been traditionally used to detect damage on underwater structures, such as bridge piers, off-shore wind turbine platforms, and pipelines, according to the study. Researchers applied this method to histopathology images of cells from breast biopsy.

“The potential for this technology is very exciting and we are delighted that this international and inter-disciplinary team has worked so well at tackling a real bottle-neck in automating the diagnosis of breast cancer using histopathology images,” said lead author Maqlin Paramanandam, MCA, mathematician at Madras Christian College, India.

They first considered the likelihood of every point in the image being near a cell center or cell boundary, the researchers wrote. Then, they used a belief propagation algorithm and traced out the cell boundaries.

These findings could be used to create faster, and more uniform prognoses for patients with breast cancer. This could also speed up time to treatment, which could save lives.

“Coming from a civil engineering background where most of our image-processing tools were designed to assess structural damage, it was nice to discover some cross-over applications and find areas where we could lend our expertise,” said researcher Dr Michael O’Byrne. “We all found it particularly rewarding to contribute towards breast cancer research.”