Comparing Image Analysis Platforms Using Estrogen and Progesterone Receptors in Breast Cancer

Throughout their lifetimes, women have a 1 in 8 chance of developing breast cancer, which is the most common cancer among women. Using immunochemistry and tissue microarrays (TMAs), the level of core staining for certain antibodies can be studied and quantified using image analysis platforms. Some of the most common studied for clinical research are estrogen receptors (ER) and progesterone receptors (PR). As there are many different image analysis platforms, this study analyzed three image analysis platforms and compared the data to determine if different platforms produced the same results. The breast cancer TMAs, with about 80 patient samples, were produced and evaluated using Aperio positive pixel count, QuPath imaging software, and Halo imaging software. These image analysis platforms quantify positivity of stained cells using a brown versus blue algorithm, where brown is positive and blue is negative. When comparing the data for three platforms for ER and PR, the correlation coefficients were similar indicating a respectable consistency among them.

Supervisor: Dr. George Sandusky
Department: Pathology and Laboratory Medicine