A multi-model deep learning approach to detect tumors on Mohs micrographic surgery slides
Signal, Image and Video Processing, cilt.20, sa.14, 2026 (SCI-Expanded, Scopus)
- Yayın Türü: Makale / Tam Makale
- Cilt numarası: 20 Sayı: 14
- Basım Tarihi: 2026
- Doi Numarası: 10.1007/s11760-026-05559-0
- Dergi Adı: Signal, Image and Video Processing
- Derginin Tarandığı İndeksler: Science Citation Index Expanded (SCI-EXPANDED), Scopus, Aerospace Database, Compendex, INSPEC, zbMATH, Technology Collection (ProQuest)
- Anahtar Kelimeler: Basal cell carcinoma, Deep learning, Histopathology, Mohs micrographic surgery
- Açık Arşiv Koleksiyonu: AVESİS Açık Erişim Koleksiyonu
- Hacettepe Üniversitesi Adresli: Evet
Özet
Mohs micrographic surgery (MMS) is the goldstandard technique for removing high-risk nonmelanoma skin cancers (NMSC); however, intraoperative histopathological examination requires significant time, effort, and specialist expertise. The objective of this study was to develop a deep learning model to detect basal cell carcinoma (BCC) and artifacts on Mohs slides. The objective of this study was to develop a deep learning model to detect basal cell carcinoma (BCC) and artifacts on Mohs slides. A total of 731 Mohs slides from 51 patients with BCC were used in this study, of which 91 contained tumor and 640 were classified as artifact. The dataset was used to train U-Net-based models to segment tumor and non-tumor regions on the slides. The segmented patches were classified as tumor or non-tumor to generate predictions for wholeslide images (WSIs). During segmentation, the tumor and artifact models achieved Dice scores of 0.70 and 0.67 and area under the curve (AUC) values of 0.98 and 0.96, respectively. For tumor classification, the patch-level and slide-level AUCs on the test set were 0.98 and 0.91, respectively. We present an AI system that can detect tumor and artifact regions in Mohs slides with promising performance. Deep learning can aid Mohs surgeons and dermatopathologists in making more accurate decisions.