MDSpire - Takeaway
From the Journals

AI Model Differentiates BCC vs cSCC Subtypes

External validation identifies calibration shift in COBRA cohort.

Share

  • 1

    Weakly supervised model achieves 100% accuracy in cancer classification.

  • 2

    Utilizes multiple-instance learning and vision transformer techniques.

  • 3

    Internal test set comprised 335 slides with dual dermatopathologist confirmation.

  • 4

    High sensitivity and specificity noted in external cohort validation.

  • 5

    Performance adjustment addressed calibration shifts in external data.

  • 6

    Study highlights importance of reliable AI deployment in dermatopathology.

Original Source(s)

AI Model Differentiates BCC vs cSCC Subtypes

  • by Kathryn Wighton

  • March 23, 2026

Related Content

Postmortem Blood Cultures May Aid Autopsy Analysis

Study examines bacterial findings in 100 medico-legal autopsies

Value-based Healthcare: How Do We Get There?

Experts discuss ways to achieve a new model of proactive, population-level healthcare

What Are Foundation Models Doing for Pathology?

Andy Beck, CEO of PathAI, explains how foundation models are accelerating computational pathology and enabling a new generation of AI-powered products