Cancer Diagnosis from Histopathology Images Using Deep Learning: A Review
Vijaya Gajanan Buddhavarapu, J. Angel Arul Jothi · 2022
Histopathology image analysis is often called the gold standard for cancer diagnosis. Pathologists are tasked with analyzing tissues under microscope in order to diagnose the presence or absence of cancer and various other diseases. Healthcare providers such as hospitals process large number of slides a year. Therefore, pathologists, who analyze slides, may become inundated with data that needs to be processed. Moreover, the prognosis of cancer treatment is favourable with early and accurate diagnosis. Consequently, computer aided diagnosis (CADx) and computer aided detection (CADe) systems for the interpretation of medical images have been the focus of development in the past few decades. These systems assist medical professionals by performing tasks that accelerate the diagnostic process. Recently, CADe/CADx systems of histopathology images using Deep Learning (DL), a newer paradigm of machine learning that uses neural networks, have become popular. Therefore, CADe/CADx systems are being researched and developed for various histopathology tasks such as classification of image data and segmentation of region of interest (ROI). This article aims to give an overview of DL, detail the applications of deep learning on histopathology images for cancer diagnosis, and discusses recent trends and future scope of DL in histopathology image analysis.