Comparing Machine Learning Algorithms to Digitally Stain Breast Tissue Biopsies

Histopathology, the traditional method for cancer diagnosis, has many limitations in distinguishing between malignant and benign tissue. Specifically, staining and pathological diagnosis are time consuming, and diagnosis are often observer dependent. In light of modern analysis techniques, there exist novel, more effective means to recognize and diagnose cancerous tissue. Through the implementation of various machine learning algorithms in concurrence with infrared imaging techniques, breast cancer tissue biopsies can be digitally stained to resemble H&E stained biopsies and used in clinical application to diagnose cancer autonomously.