Abstract
Previous studies applied deep learning to Mohs micrographic surgery (MMS) by developing algorithms able to classify digitized histopathology slides as BCC positive or negative. 1-3 Because the histological location of residual tumour is necessary to guide subsequent excision, we aimed to develop and evaluate an algorithm using convolutional neural networks (CNNs) that would automatically localize and point out BCC tumour islands in digitized MMS histology slides.