Click an organ or subcategory dot to expand its branches; click a leaf to open its topic page.
Large subcategories branch by real pathology category (e.g. Fibroblastic, Vascular), read from the WHO classification itself — never alphabetically.
Roots with very many subcategories (Pediatric Pathology, Neuropathology, Dermatopathology, Hematolymphoid, Forensic Pathology) get one extra branch first — a disease-process bucket like Neoplastic, Infectious, or Congenital — to keep the list manageable. This grouping is a heuristic over section headers, not official WHO taxonomy.
Type in the search bar above to live-filter the tree, even on a partial word (e.g. "adeno", "sarc") — matches and their ancestors expand automatically.
Press Enter in the search bar to ask a full question instead of filtering.
Click VS next to any diagnosis to add it to the comparison tray at the bottom, then hit Compare.
Click ⌂ Home at any time to return to the Browse tree.
A few entities with no clear category are left off the tree — search still finds them.
What can I do with this JSON?
This file has everything actually used to build the page you're looking at: the synthesized answer, every textbook/WHO/Pathoutlines/video excerpt actually cited, figures, and any live literature pulled in — not just the polished prose.
Attach it to a fresh chat with any AI model and ask it to build on the material — e.g. "using this evidence, write a 10-minute lecture on this diagnosis" or "turn this into a slide deck outline."
Works well for quizzes and flashcards too: "make 15 board-style multiple-choice questions from this evidence" or "make Anki-style flashcards covering the key facts and differential diagnosis."
Because the model gets the same grounded source excerpts this page was built from (not a summary of a summary), its output stays traceable back to real textbook/WHO/literature content instead of drifting into unsourced generalities.
The debug block also shows exactly which sources were retrieved and how many — useful if you want the AI to flag where evidence was thin.