AI Model Differentiates BCC vs cSCC Subtypes - Scorecard - MDSpire

AI Model Differentiates BCC vs cSCC Subtypes

External validation identifies calibration shift in COBRA cohort.

  • By

  • Kathryn Wighton

  • March 23, 2026

  • 3 min

Share

Clinical Scorecard: AI Model Differentiates BCC vs cSCC Subtypes

At a Glance

CategoryDetail
ConditionBasal Cell Carcinoma (BCC) and Cutaneous Squamous Cell Carcinoma (cSCC)
Key MechanismsWeakly supervised deep learning model using clustering-constrained attention and feature extraction from a vision transformer.
Target PopulationPatients with infiltrative basal cell carcinoma and poorly differentiated cutaneous squamous cell carcinoma.
Care SettingDermatopathology

Key Highlights

  • Model achieved 100% accuracy, sensitivity, and specificity on internal test set.
  • External validation showed AUC of 1.0 in Queensland cohort and 0.92 in COBRA cohort.
  • Attention heatmaps indicated tumor localization in histopathology images.
  • Fine-tuning of HistoGPT model improved accuracy to 98% with high sensitivity and specificity.
  • Calibration and domain adaptation are crucial for reliable deployment across institutions.

Guideline-Based Recommendations

Diagnosis

  • Utilize weakly supervised deep learning models for accurate classification of BCC and cSCC.

Management

  • Implement careful calibration and domain adaptation for model deployment.

Monitoring & Follow-up

  • Regularly assess model performance across different cohorts and settings.

Risks

  • Potential for calibration shifts and performance variation due to diagnostic subtype differences and image quality.

Patient & Prescribing Data

Patients with diagnosed BCC and cSCC requiring histopathological evaluation.

Deep learning models can enhance diagnostic accuracy in challenging cases.

Clinical Best Practices

  • Ensure thorough validation of AI models in diverse clinical settings.
  • Monitor model performance and adjust thresholds based on specific cohort characteristics.

Related Resources & Content

Original Source(s)

AI Model Differentiates BCC vs cSCC Subtypes

  • by Kathryn Wighton

  • March 23, 2026

Related Content

Narrow DCIS Margins May Not Always Warrant Reexcision

An ancillary analysis found less than 2-percentage-point differences in 10-year recurrence across margin thresholds, with no statistically significant adjusted association.

High-Dose Vitamin C Could Pose Risk in Severe Burns

“High-dose intravenous vitamin C administered early after injury did not reduce organ dysfunction or [mortality] and should not be a treatment in severe burns.”

Does Structured Reporting Improve Lung Cancer Reports?

A structured reporting tool improved report completeness, reduced classification errors, and achieved high physician adoption during routine lung cancer pathology reporting.