AI Model Differentiates BCC vs cSCC Subtypes - Takeaways - 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

  • 1

    A weakly supervised deep learning model achieved perfect classification of infiltrative basal cell carcinoma and poorly differentiated cutaneous squamous cell carcinoma.

  • 2

    The model demonstrated 100% accuracy, sensitivity, and specificity on an internal test set, with an area under the curve of 1.0.

  • 3

    In external cohorts, the model maintained high performance, achieving an AUC of 1.0 in the Queensland cohort and 0.92 in the COBRA cohort.

  • 4

    Attention heatmaps indicated that the model focused on tumor regions, differentiating between infiltrative atypical squamoid cells and basaloid cells.

  • 5

    The study highlights the need for careful calibration and domain adaptation for reliable deployment of AI models across different institutions.

Original Source(s)

AI Model Differentiates BCC vs cSCC Subtypes

  • by Kathryn Wighton

  • March 23, 2026

Related Content

FDA Approves Expanded Liquid Biopsy Panel

Updated liquid biopsy combines genomic and epigenomic profiling for companion diagnostic use across multiple tumor types.

Urine Markers May Aid Psych Diagnosis

Genetically predicted urinary metabolite levels were associated with schizophrenia, bipolar disorder, attention-deficit/hyperactivity disorder, and anorexia nervosa in a Mendelian randomization analysis.

FDA Approves New First-Line TNBC Regimen

Phase 3 results showed longer progression-free survival with pembrolizumab plus sacituzumab govitecan-hziy than with pembrolizumab plus chemotherapy in previously untreated PD-L1-positive advanced triple-negative breast cancer.