Abstract:Deep Learning-based computational pathology algorithms have demonstrated profound ability to excel in a wide array of tasks that range from characterization of well known morphological phenotypes to predicting non human-identifiable features from histology such as molecular alterations. However, the development of robust, adaptable and accurate deep learning-based models often rely on the collection and time-costly curation large high-quality annotated training data that should ideally come from diverse sources and patient populations to cater for the heterogeneity that exists in such datasets. Multi-centric and collaborative integration of medical data across multiple institutions can naturally help overcome this challenge and boost the model performance but is limited by privacy concerns among other difficulties that may arise in the complex data sharing process as models scale towards using hundreds of thousands of gigapixel whole slide images. In this paper, we introduce privacy-preserving federated learning for gigapixel whole slide images in computational pathology using weakly-supervised attention multiple instance learning and differential privacy. We evaluated our approach on two different diagnostic problems using thousands of histology whole slide images with only slide-level labels. Additionally, we present a weakly-supervised learning framework for survival prediction and patient stratification from whole slide images and demonstrate its effectiveness in a federated setting. Our results show that using federated learning, we can effectively develop accurate weakly-supervised deep learning models from distributed data silos without direct data sharing and its associated complexities, while also preserving differential privacy using randomized noise generation. We also make available an easy-to-use federated learning for computational pathology software package: http://github.com/mahmoodlab/HistoFL.

Personalized Stain Style Transfer Layers for Distributed Histology Classification

A Federated Learning System for Histopathology Image Analysis With an Orchestral Stain-Normalization GAN

Style Normalization in Histology with Federated Learning

FedDBL: Communication and Data Efficient Federated Deep-Broad Learning for Histopathological Tissue Classification

Histopathological Stain Transfer using Style Transfer Network with Adversarial Loss

Virtual stain transfer in histology via cascaded deep neural networks

Federated learning for computational pathology on gigapixel whole slide images

Unpaired Multi-Domain Stain Transfer for Kidney Histopathological Images

RandStainNA: Learning Stain-Agnostic Features from Histology Slides by Bridging Stain Augmentation and Normalization

Structure-Preserving Multi-Domain Stain Color Augmentation using Style-Transfer with Disentangled Representations

Unpaired Stain Style Transfer Using Invertible Neural Networks Based on Channel Attention and Long-Range Residual

Stain Style Transfer of Histopathology Images Via Structure-Preserved Generative Learning

Federated Impression for Learning with Distributed Heterogeneous Data

Pseudo-Data based Self-Supervised Federated Learning for Classification of Histopathological Images

Self-Supervised Digital Histopathology Image Disentanglement for Arbitrary Domain Stain Transfer.

Federated learning with hyper-network—a case study on whole slide image analysis

Multi-domain stain normalization for digital pathology: A cycle-consistent adversarial network for whole slide images

General stain deconvolution of histopathology images with physics-guided deep learning

StainDiff: Transfer Stain Styles of Histology Images with Denoising Diffusion Probabilistic Models and Self-ensemble

StainFuser: Controlling Diffusion for Faster Neural Style Transfer in Multi-Gigapixel Histology Images

Learning to Predict the Optimal Template in Stain Normalization for Histology Image Analysis