This is a 2-day course designed to teach pathologists the fundamentals of computational pathology, covering everything from digital imaging basics and machine learning to clinical applications and AI implementation in diagnostic workflows.

TOPICS COVERED

This comprehensive 2-day course provides pathologists with essential knowledge and practical understanding of computational pathology, from foundational concepts to advanced clinical applications. The program totals 13.75 hours of instruction delivered through lectures, demonstrations, and interactive discussions.
The course begins with core digital imaging fundamentals, covering how digital images are structured including pixels, resolution, bit depth, and color spaces. Participants learn about whole-slide imaging technology, scanner types, pyramidal image structures, and the practical implications of different file formats and compression methods for computational analysis. A dedicated session on stain physics explores the Beer-Lambert law, color deconvolution, and stain normalization approaches—critical knowledge for understanding how computational algorithms interpret histological images.
Building on these foundations, the course progresses to image processing techniques essential for tissue analysis, including intensity transformations, spatial filtering, and morphological operations. Participants explore segmentation methods for distinguishing tissue from background, nuclear and cell detection, and addressing common challenges like the reference area problem. The program covers classical feature extraction approaches, teaching attendees how to identify and quantify morphological, textural, and architectural features relevant to pathology.
Machine learning concepts are introduced systematically, beginning with classical approaches like random forests and support vector machines, then advancing to deep learning fundamentals. Participants learn about neural network architectures, training strategies, and why deep learning has transformed computational pathology. Advanced sessions cover convolutional neural networks, landmark architectures like ResNet and U-Net, Vision Transformers, and concepts like transfer learning and multiple instance learning for whole-slide analysis.
The course provides detailed coverage of emerging technologies including foundation models, self-supervised learning, and vision-language models that enable conversational AI for pathology. Practical considerations around data preparation, annotation strategies, managing class imbalance, and handling staining variability across laboratories are thoroughly addressed.
Clinical applications form a substantial component, with sessions dedicated to diagnostic support systems for cancer detection and classification across multiple organ systems, automated Gleason grading, metastasis detection, and quality control applications. Biomarker quantification sessions cover IHC analysis for Ki-67, HER2, and PD-L1, tumor microenvironment assessment, and the emerging capability to predict molecular features directly from H&E slides.
Critical topics of validation, regulation, and deployment are explored, including appropriate evaluation metrics, regulatory pathways through FDA and CE/MDR, algorithmic fairness considerations, and strategies for clinical integration. The course emphasizes model interpretability and explainable AI techniques, addressing the essential question of building pathologist trust in computational tools.
The program concludes with forward-looking discussions on multiplex imaging analysis, multimodal data integration, federated learning, and AI-augmented workflows. A dedicated session explores the evolving role of pathologists in computational pathology development, integration with omics data, career pathways, and building multidisciplinary collaborations.
Throughout both days, participants have opportunities for questions, discussion, and networking with faculty and peers. The course provides a comprehensive foundation for pathologists seeking to understand, evaluate, and potentially implement computational pathology tools in their practice.

WHO SHOULD ATTEND?

This course is designed for pathologists at any career stage seeking to understand computational pathology and artificial intelligence applications in diagnostic practice. No prior computational or programming background is required. The program is ideal for pathologists interested in evaluating AI tools for clinical adoption, participating in validation studies, collaborating in computational pathology research, or simply gaining literacy in technologies transforming the field. Whether considering implementation of digital pathology systems, wanting to understand how algorithms interpret histological images, or exploring career development in this emerging area, participants will gain comprehensive knowledge from foundational concepts through advanced clinical applications and regulatory considerations.

SPEAKERS

• RC - Runjan Chetty https://www.linkedin.com/in/runjan-chetty-6b3aa4249/ Chief Medical Officer at Deciphex/Diagnexia
• EK - Emre Köse https://www.linkedin.com/in/emrecgty/ Computational Pathologist at Deciphex
• AJ - Andrew Janowczyk  https://www.linkedin.com/in/andrewjanowczyk/ Assistant Professor at Emory University
• JW - John Weldon https://www.linkedin.com/in/john-weldon-517a52188/ Clinical AI Director at Deciphex
• JA - Jonathan Armstrong https://www.linkedin.com/in/jonathanjarmstrong/  AI Governance Lead at Deciphex
• PM - Pierre Moulin https://www.linkedin.com/in/pierre-moulin-5a95122/  MD, PhD — Chief Scientific Officer at Deciphex

For more information, please contact: [email protected].