Related Experiment Video
Updated: Nov 15, 2025
13:01
Industrialized, Artificial Intelligence-guided Laser Microdissection for Microscaled Proteomic Analysis of the Tumor Microenvironment
Published on: June 3, 2022
4.2K
Whole-slide margin control through deep learning in Mohs micrographic surgery for basal cell carcinoma
Mike C M van Zon1, José D van der Waa2, Mitko Veta1
1Medical Image Analysis Group, Eindhoven University of Technology, Eindhoven, The Netherlands.
Experimental Dermatology
|March 3, 2021
Summary
Deep learning models can automate basal cell carcinoma (BCC) detection in histopathology slides from Mohs micrographic surgery (MMS), potentially improving workflow and reducing costs.
Area of Science:
- Pathology
- Artificial Intelligence
- Dermatology
Background:
- Basal cell carcinoma (BCC) is the most common skin cancer, with rising incidence.
- Mohs micrographic surgery (MMS) requires extensive histological analysis of frozen sections for BCC treatment, increasing healthcare costs.
Purpose of the Study:
- To develop and evaluate a deep learning model for automated detection and classification of BCC in histopathology slides.
- The model aims to identify BCC-negative slides and classify BCC presence using whole-slide images (WSIs).
Main Methods:
- Two deep learning models were created using 171 digitized H&E frozen slides from 70 patients.
- A U-Net model performed BCC segmentation, followed by a convolutional neural network for slide-level BCC or BCC-negative classification.
Main Results:
- The BCC segmentation model achieved a Dice score of 0.66.
- The slide-level classification model demonstrated an area under the receiver operating characteristic curve (AUC) of 0.90.
Conclusions:
- Deep learning models applied to WSIs show promise for enhancing clinical workflows in BCC histological analysis.
- This approach may reduce the economic burden associated with histological analysis in MMS for BCC.
