The Diagnexia Computational Pathology Course

Diagnexia Computational Pathology Symposium

📍 Reuben College, Oxford
📅 Date: 2nd September

Morning Session

• 09:00 AM - 09:05 AM | Welcome
• 09:05 AM - 09:15 AM | Opening Remarks
• 09:15 AM - 09:45 AM | Introduction to Computational Pathology: Concepts and Importance
  Overview of computational pathology, its key concepts, and its relevance to modern pathology practices.
• 09:45 AM - 10:15 AM | Understanding Digital Images: The Building Blocks of Computational Pathology
  Explanation of images, pixels, resolution, color spaces, stain vector estimation, image representation, compression, whole slide imaging (WSI), and pyramidal images.
• 10:15 AM - 10:45 AM | Image Processing in Pathology: Techniques and Whole Slide Imaging
  Fundamental image processing techniques such as thresholding, filtering, morphological operations, segmentation, object identification, and feature extraction.
• 10:45 AM - 11:05 AM | ☕ Coffee Break (20 mins)
• 11:05 AM - 11:35 AM | Feature Extraction in Computational Pathology: Feature Engineering
  Introduction to heuristic feature extraction techniques, including texture, shape, intensity, and morphology, and their applications in computational pathology.
• 11:35 AM - 12:05 PM | Introduction to Deep Learning
  Overview of neural networks, forward and backward propagation, loss function, gradient descent, training, inference, validation, test datasets, and performance measurements in clinical contexts.
• 12:05 PM - 12:35 PM | Neural Networks in Computational Pathology
  Introduction to convolutional networks (ConvNets), Transformers, Vision Transformers (ViTs), and their relevance in pathology.
• 12:35 PM - 01:35 PM | 🍽️ Lunch Break (1 hour)

Afternoon Session

• 01:35 PM - 02:05 PM | Data Preparation and Challenges in Computational Pathology
  Discussion on data preparation steps (digitization, annotation, cleaning) and challenges such as staining variability, dataset biases, and annotation requirements. Explain augmentation vs. real-world variation.
• 02:05 PM - 02:35 PM | Building and Training Models in Computational Pathology
  Overview of training approaches, including fully-supervised, multiple instance learning (MIL), and unsupervised learning.
• 02:35 PM - 03:05 PM | Validation and Regulation of Computational Pathology Tools
  Exploring validation requirements, regulatory approval processes (FDA, CE marking), and quality assurance.
• 03:05 PM - 03:25 PM | ☕ Coffee Break (20 mins)
• 03:25 PM - 03:55 PM | Pathology as a Data-Heavy Science
  Discussion on the vast data volumes in pathology compared to other medical disciplines and why pathologists must master data analysis techniques.
• 03:55 PM - 04:25 PM | Emerging Trends and Future Technical Directions in Computational Pathology
  Highlighting cutting-edge innovations such as multimodal AI, chatbots, and the evolving role of pathologists in advancing the field.
• 04:25 PM - 04:55 PM | Case Studies: Computational Pathology in Action
  Real-world applications of computational pathology in diagnostics, research, and prognostic analysis.
• 04:55 PM - 05:25 PM | Ethics and the Future of Computational Pathology
  Discussion on ethical considerations, patient data privacy, medical responsibility, and training for a future with AI-assisted diagnostics.
• 05:25 PM - 05:55 PM | Open Discussion & Q&A
  An interactive forum for participants to discuss key takeaways, ask questions, and share insights.
• 05:55 PM - 06:00 PM | Closing Remarks