An algorithm for automatic measurement of KI-67 proliferation index in digital images of breast tissue

This paper proposes an algorithm aimed at quantifying the expression of KI-67 protein in digital images of breast biopsy tissue samples obtained through an optical microscope. The algorithm allows to obtain a report on the quantity of non-proliferating and proliferating cells through the detection and quantification of KI-67. The sample analysis via software aims to reduce the level of subjectivity in the diagnosis of diseases such as breast cancer. The algorithm proposed involves the application of statistical image processing techniques, adaptive thresholds, object segmentation and color filtering. A method of analysis and quantification of overlapping cells is also proposed to improve the efficiency of the algorithm. The results of the method proposed were satisfactory, as they were highly correlated with those obtained by visual inspection by pathologists.