PlanPath: Exploration-Guided Orchestration for Patient-Level Pathology Reasoning
What it is
PlanPath is an agentic framework for patient-level pathology visual question answering that converts questions into exploration plans to route queries to relevant slides. It uses coarse-to-fine acquisition tools to measure tissue and cell-level features, aggregating evidence across multiple slides to form traceable patient-level conclusions.
The problem it solves
Existing pathology agents analyze slides in isolation and describe morphology rather than measuring it, failing to integrate complementary findings across the multiple slides that constitute a single patient case.
What you could use it for
- Answering complex clinical questions that require integrating evidence from multiple pathology slides
- Performing patient-level diagnosis for renal and gastric cancer cases
- Quantifying morphological features across a patient's entire set of whole slide images
Read the source ↗Back to the library →
Summary written by an AI model from the source; check the source before you rely on it.