PROJECT SYNOPSIS

We are creating software for automatically segmenting breast cancer biopsy and pathology slides. Segmentation is the process by which regions of suspicious tissue are identified and labeled. It is a critical and time-consuming process which is typically done manually by trained pathologists. Our goal is to create decision support software that can perform the labor-intensive segmentation task while leaving the analysis and diagnosis to a trained pathologist. A high performance segmentation algorithm is being developed that uses a combination of artificial intelligence and a large database of breast cancer biopsy images.

IMPACT

Our goal is to create software to assist pathologists in the painstaking and highly sophisticated task of assessing biopsy slide images. Our technology identifies areas containing abnormal tissue and provides estimates of cancer grade and stage. By reducing time spent on repetitive tasks, AI tools can enable pathologists to serve more patients, especially in underserved communities where generalist pathologists may lack domain-specific expertise. It is also likely that the diagnostic requirements of precision medicine may require diagnostics that exceed human visual acuity.

The data and technology being developed in this project is made available as open source resources. Please go here to learn more about the data and corpora available.