A Curated Resource for Modern Pathology

The Pathology Notebook is an open-access educational resource that provides well-organized information for every subspecialty. It is a foundational tool for trainees and faculty preparing for board review or daily sign-out.

Explore the Notebook
Table of Contents for the Pathology Notebook

Introducing the Knowledge Hub Initiative

Building on the open-access resources of the Pathology Notebook, the Knowledge Hub is our next initiative: a project to create and perfect a scalable platform for academic knowledge management. The core of this work is our documented pipeline, which transforms a department's disconnected internal archives into a structured, RAG-ready knowledge base. Our vision is to provide a low-cost, efficient model that other academic departments can adopt to build their own internal "Googles," thereby creating the foundational data needed for the next generation of precision education tools.

A clean search bar for the Pathology Knowledge Hub.

A Working Prototype in Action

The Hub is not just a concept—it's a functional prototype that unifies disparate content types and provides deep contextual linking directly to the authoritative source.

A unified grid of search results.

1. Unified Results: The Hub returns an expert-vetted grid of results from both textbooks and lectures, replacing hours of manual searching.

A modal showing a textbook figure with a link to the PDF page.

2. Deep Linking to Textbooks: Clicking a figure provides a high-resolution view and a direct link to the exact page in the source PDF.

A modal showing a lecture slide with a link to the video timestamp.

3. Deep Linking to Videos: Clicking a lecture slide opens its transcript and a link to play the source video starting at the precise timestamp.

Use Case: Generating Intelligent Study Materials

The true power of the Hub's structured data is its ability to fuel next-generation learning tools. We can programmatically generate sophisticated, media-rich **Anki flashcards** that are impossible to create manually.

Card Example 1: Annotated Figure

Front of an Anki flashcard. Back of the flashcard with annotations. Back of the flashcard with linked text.

Card Example 2: Linked Lecture

Front of a second Anki flashcard. Back of the flashcard with a lecture slide. Back of the flashcard with a YouTube link.

Our pipeline can auto-generate study cards that combine active recall with rich, contextual answers, including **annotated images, linked video segments, and relevant text** from the Pathology Notebook.

The Technology: Building a Reusable Platform

The Hub is powered by the Knowledge Pipeline (v3.2), a sophisticated data processing pipeline developed in Python. This is an infrastructure project that uses established analytical tools to create a structured, permanent "Golden Asset" of institutional knowledge.

  • Automated Transcription using OpenAI's Whisper.
  • Computer Vision Segmentation using OpenCV and the SSIM algorithm.
  • Multimodal OCR & Enhancement using Google's Gemini models.
  • RAG Linking to the Pathology Notebook using Sentence-Transformers.

The Team

Ronald Balassanian, MD - Principal Investigator (Professor & Director, Pathology Residency Program, UCSF)

Charlie Herndon, DO - Co-Investigator & Lead Developer (Resident Physician, UCSF)

© 2025 The Pathology Knowledge Hub Initiative