Tutorial for Code-free development and deployment of deep segmentation models for digital pathology

Tutorial for "Code-free development and deployment of deep segmentation models for digital pathology" (http://arxiv.org/abs/2111.08430). All scripts, trained models, and the full dataset of 251 WSIs with ~31k epithelium annotations at https://github.com/andreped/NoCodeSeg We present a code-free pipeline utilizing free-to-use, open-source software (QuPath, DeepMIB, and FastPathology) for creating and deploying deep learning-based segmentation models for computational pathology. We demonstrate the pipeline on a use case of segmenting epithelium from stroma in colon biopsies. A dataset of 251 annotated WSIs (140 hematoxylin-eosin (HE)-stained and 111 CD3 immunostained colon biopsy WSIs), were developed through active learning using the pipeline. A mean intersection over union score of 96.6% and 95.3% was achieved for the hold-out test set of 36 HE and 21 CD3-stained WSIs. We demonstrate that pathologists without programming experience can create near state-of-the-art deep learning based segmentation training and direct WSI prediction visualization for histopathological WSIs using only free-to-use software. All scripts, trained models, a video tutorial, and the full dataset of 251 WSIs with ~31k epithelium annotations are made openly available at https://github.com/andreped/NoCodeSeg. Authors: Henrik Sahlin Pettersen*,1,2,3, Ilya Belevich4, Elin Synnøve Røyset1,2,3, Erik Smistad5,6, Eija Jokitalo4, Ingerid Reinertsen5,6, Ingunn Bakke2,3, and André Pedersen2,5 1. Department of Pathology, St. Olavs hospital, Trondheim University Hospital, Trondheim, Norway. 2. Department of Clinical and Molecular Medicine, Faculty of Medicine and Health Sciences, NTNU - Norwegian University of Science and Technology, Trondheim, Norway. 3. Clinic of Laboratory Medicine, St. Olavs hospital, Trondheim University Hospital, Trondheim, Norway. 4. Electron Microscopy Unit, Institute of Biotechnology, Helsinki Institute of Life Science, University of Helsinki, Helsinki, Finland. 5. Department of Health Research, SINTEF Digital, Trondheim, Norway. 6. Department of Circulation and Medical Imaging, Faculty of Medicine and Health Sciences, NTNU - Norwegian University of Science and Technology, Trondheim, Norway. * Correspondence: Henrik Sahlin Pettersen [email protected] Keywords: digital pathology, artificial intelligence, code-free, histopathology, segmentation, U-Net, open datasets, deep learning, semantic segmentation.