Computational Imaging and Analysis of Ki-67 for Breast Cancer in Digital Pathology
Justin Lee, Massachusetts Institute of Technology
Ki-67 is a monoclonal antibody used to immunohistochemically label the nuclei of proliferating cells in breast cancer tissue. The proportion of Ki-67-positive cells in a tumor is also referred to as its proliferation index (or ratio). Higher proliferation ratios confer a worse outcome for patients. This prognostic marker is currently assessed manually by counting positive cells under the microscope, which is inaccurate and time-consuming. Here, we develop an open-source, web/browser-based software solution to quantify cell proliferation ratios and facilitate physician and computational analysis of Ki-67. We have validated our software in a clinical trial and are investigating extensions of our application for analysis of multi-focal plane whole-slide images (and to specify metrics from these quantitative analyses to improve the prognostic and predictive potential of Ki-67 for breast cancer patients).
Abstract Author(s): Justin Lee, George Barbastathis, Elena Brachtel, Yukako Yagi