Idea
Clinical-grade AI model improving intraoperative pathology accuracy and efficiency for surgical decision-making.
Research Paper
Core Innovation
This paper introduces CRISP, a foundation model trained on a large, diverse frozen-section dataset from multiple centers, enabling robust, generalizable intraoperative pathology diagnosis. It uniquely supports a wide range of diagnostic tasks and rare cancers with clinical-grade accuracy validated prospectively.
Why It Matters
Intraoperative pathology is critical for precision surgery but limited by diagnostic complexity and scarce high-quality data. CRISP enhances diagnostic accuracy and reduces workload in real-world surgical settings, accelerating AI adoption in clinical workflows and improving patient outcomes at scale.
Market Size (TAM)
$10–20B TAM for AI-driven pathology diagnostics; $2–5B SAM from hospitals and surgical centers. Driven by rising demand for precision surgery and AI integration in clinical workflows.
Potential Customers & Pain Points
- Hospitals – Need faster accurate intraoperative pathology
- Surgical centers – Require reliable decision support
- Pathology labs – Face workload and resource constraints
- Medical device companies – Seek integrated AI solutions for surgery.
Business Model
Subscription-based SaaS platform licensing CRISP AI diagnostics to hospitals and surgical centers, with options for integration into pathology lab software and surgical decision support systems.
Competitive Landscape
- PathAI
- Paige.AI
- Proscia
- Ibex Medical Analytics
Implementation Challenges
- Regulatory approval for clinical AI tools
- Integration into existing surgical workflows
- Data privacy and interoperability challenges
- Clinician trust and adoption resistance
Validation Strategy
- Conduct multi-center prospective clinical trials to confirm diagnostic accuracy and impact on surgical outcomes
- Partner with leading hospitals for pilot deployments and workflow integration
- Obtain regulatory clearances (FDA
- CE) for clinical use
- Collect real-world usage data to refine model and demonstrate cost savings
Research Paper Overview
A Clinical-grade Universal Foundation Model for Intraoperative Pathology
Summary
CRISP is a clinical-grade AI model trained on over 100,000 frozen pathology sections from multiple centers, designed to support intraoperative pathology diagnosis. It was validated on 15,000+ slides across diverse tasks and institutions, showing robust accuracy and generalization. In a prospective study with 2,000+ patients, CRISP informed surgical decisions in 92.6% of cases, reduced diagnostic workload by 35%, and improved micrometastases detection, enabling practical AI integration into surgical workflows.