Deep Learning for Skin Melanocytic Tumors in Whole-Slide Images: A Systematic Review
Problem
Histological interpretation of melanocytic tumors is challenging and subject to interobserver variability between dermatopathologists.
Why it matters
Diagnostic variability in melanoma interpretation affects consistency of patient pathology assessments.
What the AI did
Systematic review (PRISMA) of 28 studies (2015–July 2022) applying deep learning to whole-slide images of melanocytic tumors, grouped by clinical objective including pathologists vs deep learning comparisons.
Results
28 studies included and grouped into four clinical objective groups; pathologists vs deep learning comparisons n=10.
Impact
Summarizes evidence that deep learning shows promising performance as a support tool in melanocytic tumor pathology.
Record
| Venue | Cancers |
|---|---|
| Published | 2022 |
| Work type | review |
| Authors | Andrés Mosquera‐Zamudio, Laëtitia Launet, Zahra Tabatabaei, Rafael Santiago Parra-Medina, Adrián Colomer, Javier Oliver Moll, Carlos Couceiro Monteagudo, Emiel A. M. Janssen et al. |
| Citations | 33 (OpenAlex, at import time) |
| References | 77 |
| Model or method | Deep learning; specific architectures not stated in source |
| Organization | not stated in source |
| Primary source | https://doi.org/10.3390/cancers15010042 |
| Open access | gold |
| Keywords | melanoma, deep learning, whole-slide images, systematic review, dermatopathology |
Importance rating
real but narrower advance
Interobserver variability in melanocytic tumor interpretation is a real but narrow pathology advance; review only, no scale stated.