The evaluation of breast cancer characteristics, long done by hand under a microscope by pathologists, could soon by aided by a computer program as researchers have created a model that can accurately and automatically analyze breast cancer microscopic images, according to a paper published Nov. 9 in Science Translational Medicine.
Computer scientists at the Stanford School of Engineering and pathologists at the Stanford School of Medicine in Palo Alto, Calif., created a model called Computational Pathologist, or C-Path, a machine-learning-based method for automatically analyzing images of cancerous tissues and predicting patient survival.
To train C-Path, the researchers used existing tissue samples taken from patients whose prognosis was known. The computers measured various tumor structures to use that information to predict patient survival. By comparing results against the known data, the computers adapted their models to better predict survival and gradually figured out what features of the cancers matter most and which matter less in predicting survival.
“In essence, the computer learns,” Daphne Koller, PhD, professor of computer science at Stanford, said in a statement.
Three specific features have traditionally been used for evaluating breast cancer cells: what percentage of the tumor is comprised of tube-like cells, the diversity of the nuclei in the outermost (epithelial) cells of the tumor and the frequency with which those cells divide. These three factors are judged by sight with a microscope and scored qualitatively to stratify breast cancer patients into three groups that predict survival rates.
While pathologists have been trained to look for those factors of known clinical importance, there are many other features of which the clinical significance is not known, according to the researchers.
“The computer strips away that bias and looks at thousands of factors to determine which matter most in predicting survival,” said Koller.
C-Path, in fact, assesses 6,642 cellular factors. Once trained using one group of patients, C-Path was asked to evaluate tissues of cancer patients it had not checked before and the result was compared against known data. The results from C-Path were a statistically significant improvement over human-based evaluation.
The computers were also able to identify characteristics of the stroma that mattered as much or more than the factors traditionally analyzed by pathologists with regard to predicting patient outcomes. Certain stromal features had a stronger association with patient survival than epithelial characteristics, according to the researchers.
“Through machine learning, we are coming to think of cancer more holistically, as a complex system rather than as a bunch of bad cells in a tumor,” said Matt van de Rijn, MD, PhD, a professor of pathology at Stanford. “The computers are pointing us to what is significant, not the other way around.”
Van de Rijn does not see computers replacing pathologists. “We’re looking at a future where computers and humans collaborate to improve results for patients across the world,” he said.
The researchers said that machine learning processes like C-Path could eventually lead to computers that can evaluate cancers in underserved areas of the world where trained professionals are scarce. In other areas, machine learning may reduce the variability in results, improving the accuracy of prognoses for all breast cancer victims, and might even be applied to predict the effectiveness of various forms of treatment and drug therapies.
“If we can teach computers to look at a tumor tissue sample and predict survival, why not train them to predict from the same sample which courses of treatment or drugs a given patient might respond to best? Or even to look at samples of non-malignant cells to predict whether these benign tissues will turn cancerous,” said Koller. “This is personalized medicine.”