Detection and Evaluation of Ki-67 Proliferation Index of Breast Cancer Cells Using Deep Learning Technique

The Ki-67 proliferation index (PI) is one of the important evaluation indices in histopathological diagnosis, especially for tumors. It is calculated based on the presence of the Ki-67 protein using immunohistochemical methods. PI is regularly assessed through visual evaluation of samples by pathologists. Ki-67 expression is closely related to cancer proliferation during the development of breast cancer tissue. The level of Ki-67 expression is crucial in determining the treatment direction for breast cancer patients, such as whether or not to proceed with therapeutic regimens. Therefore, routine Ki-67 measurement is widely performed in the pathological tumor evaluation process. The Ki-67 index is a predictive factor commonly used to make treatment decisions in breast cancer patients. Ki-67 plays an important role in predicting whether therapy is necessary and in planning treatment during the formation and monitoring of breast cancer. The role of the nuclear protein Ki-67 as a marker of cell proliferation activity, determined by the proportion of Ki-67 positive cells (expressed as a percentage), is often referred to as the proliferation index (PI). The expression of the nuclear protein Ki-67 is visualized by immunohistochemistry (IHC), which is performed for each newly diagnosed breast cancer during routine histopathological examination. In this paper, we propose a method for evaluating the Ki-67 proliferation index of breast cancer cells using deep learning techniques and cell detection. Breast cancer tissue is identified and then used to segment relevant cells through image processing methods, followed by evaluating the Ki-67 proliferation index.