Researchers at the Dartmouth-Hitchcock Norris Cotton Cancer Center have announced the development of a machine learning method that promises to expedite the process used to identify cancerous esophagus tissue on microscopy images.

To date, histopathology image analysis has required a laborious, high-cost, manual annotation process by pathologists called "region-of-interest annotations." 

According to a statement, the research team, led by Saeed Hassanpour, PhD, has developed “a novel attention-based deep learning method that automatically learns clinically important regions on whole-slide images to classify them.”

“Data annotation is the most time-consuming and laborious bottleneck in developing modern deep learning methods,” Hassanpour said in the statement. “Our study shows that deep learning models for histopathology slides analysis can be trained with labels only at the tissue level, thus removing the need for high-cost data annotation and creating new opportunities for expanding the application of deep learning in digital pathology.”

The team tested their method for identifying cancerous and precancerous esophagus tissue on high-resolution microscopy images without training on region-of-interest annotations, then proposed the network for Barrett esophagus and esophageal adenocarcinoma detection and found that their method achieved better results than the traditional method.

“The result is significant because our method is based solely on tissue-level annotations, unlike existing methods that are based on manually annotated regions.”” said Hassanpour.

According to the report, which was published in JAMA Network Open, the study has some limitations. First, “all experiments were conducted on slides collected from a single medical center and scanned with the same equipment. Second, the data set was relatively small compared with conventional data sets in deep learning.”

With those limitations in mind, Hassanpour's team is planning to validate their model further by testing it on data from other institutions and running prospective clinical trials. They also plan to apply the proposed model to histological images of other types of tumors and lesions for which training data are scarce or bounding box annotations are not available.

“Our method would facilitate a more extensive range of research on analyzing histopathology images that were previously not possible due to the lack of detailed annotations,” Hassanpour concluded. “Clinical deployment of such systems could assist pathologists in reading histopathology slides more accurately and efficiently, which is a critical task for the cancer diagnosis, predicting prognosis, and treatment of cancer patients.”