EUVIMEDEUVIMED entdecken
Crossref

Weakly supervised deep learning image analysis can differentiate melanoma from naevi on haematoxylin and eosin‐stained histopathology slides

Nigel G Maher, Homay Danaei Mehr, Cong Cong, Nurudeen A Adegoke, Ismael A Vergara, Sidong Liu, Richard A Scolyer iD

Journal of the European Academy of Dermatology and Venereology · 2024

DOI: 10.1111/jdv.20307 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Journal of the European Academy of Dermatology and Venereology, 2024. Verfasst von Nigel G Maher und 6 weiteren Autor:innen. Beteiligte Einrichtung: Melanoma Institute Australia, The University of Sydney, Sydney, New South Wales, Australia. Hauptthemen laut MEDLINE: Melanoma, Skin Neoplasms, Eosine Yellowish-(YS), Deep Learning. Die Arbeit wurde bisher 12-mal zitiert (Crossref) und nennt 30 Literaturangaben. Gefördert von: Royal College of Pathologists of Australasia, University of Sydney, National Health and Medical Research Council.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Background

The broad histomorphological spectrum of melanocytic pathologies requires large data sets to develop accurate and generalisable deep learning (DL)‐based diagnostic pathology classifiers. Weakly supervised DL promotes utilisation of larger training data sets compared to fully supervised (patch annotation) approaches.

Objectives

To evaluate weakly supervised DL image classifiers for discriminating melanomas from naevi on haematoxylin and eosin (H&E)‐stained pathology slides.

Methods

A representative H&E slide for 260 naevi and 260 melanomas from mucocutaneous sites at one tertiary institution was digitized. Clinicopathological features were recorded for each case including thickness and histological subtype. Whole‐slide or whole‐tissue section labels were applied. The ground truth was established by consensus diagnosis from two pathologists. Multiple‐instance learning models, Trans‐MIL, CLAM and DTFD‐MIL were evaluated at 10×, 20× and 40× magnifications using stratified fivefold Monte Carlo cross‐validation, with 80/10/10 splits for training/validation/test groups, to predict melanoma from naevus. Heatmaps were generated to understand model performance.

Results

Naevi cases were younger (median age: 51 years; melanoma median age: 71.5 years), with more balanced sex distribution (males: 48.8%, melanoma male subgroup: 64.2%). The most frequent histological subtypes of naevi and melanomas were dysplastic compound ( n = 99, 38.1%) and superficial spreading ( n = 124, 47.7%), respectively. Average AUC (±1 SD) for Trans‐MIL, CLAM and DTFD‐MIL across test groups were 0.9952 ± 0.006, 0.9925 ± 0.0052 and 0.9708 ± 0.0328, at 20× magnification, respectively. Performance of the models varied according to the magnification used. Heatmaps from the two best performing models, Trans‐MIL and CLAM, generally indicated attention on appropriate tissue regions for interpretation.

Conclusions

Weakly supervised DL on pathological slides of common mucocutaneous melanocytic tumours provides highly accurate diagnostic value for discrimination of melanomas and naevi. External validation and further assessment on less frequently occurring histologic subtypes and borderline cases using this method is required.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

MelanomaSkin NeoplasmsEosine Yellowish-(YS)Deep LearningHematoxylinHumansMiddle AgedDiagnosis, DifferentialMaleFemaleAgedAdultNevusAged, 80 and overStaining and Labeling

Substanzen

Eosine Yellowish-(YS)Hematoxylin

Autor:innen und Institutionen

  • Melanoma Institute Australia, The University of Sydney, Sydney, New South Wales, Australia.
  • Faculty of Medicine and Health, The University of Sydney, Sydney, New South Wales, Australia.
  • Tissue Pathology and Diagnostic Oncology, Royal Prince Alfred Hospital and NSW Health Pathology, Sydney, New South Wales, Australia.
  • Centre for Health Informatics, Australian Institute of Health Innovation, Macquarie University, Sydney, New South Wales, Australia.
  • Charles Perkins Centre, The University of Sydney, Sydney, New South Wales, Australia.

ORCID-Profile: Richard A Scolyer

Förderung

  • Royal College of Pathologists of Australasia
  • University of Sydney
  • National Health and Medical Research Council
  • Melanoma Institute Australia
  • Tour de Cure Ltd

Zitieren

APA

Maher, N. G., Danaei Mehr, H., Cong, C., Adegoke, N. A., Vergara, I. A., Liu, S., & Scolyer, R. A. (2024). Weakly supervised deep learning image analysis can differentiate melanoma from naevi on haematoxylin and eosin‐stained histopathology slides. Journal of the European Academy of Dermatology and Venereology. https://doi.org/10.1111/jdv.20307

Vancouver

Maher NG, Danaei Mehr H, Cong C, Adegoke NA, Vergara IA, Liu S, et al. Weakly supervised deep learning image analysis can differentiate melanoma from naevi on haematoxylin and eosin‐stained histopathology slides. Journal of the European Academy of Dermatology and Venereology. 2024. doi:10.1111/jdv.20307

BibTeX
@article{Maher2024,
  author = {Maher, Nigel G and Danaei Mehr, Homay and Cong, Cong and Adegoke, Nurudeen A and Vergara, Ismael A and Liu, Sidong and Scolyer, Richard A},
  title = {Weakly supervised deep learning image analysis can differentiate melanoma from naevi on haematoxylin and eosin‐stained histopathology slides},
  journal = {Journal of the European Academy of Dermatology and Venereology},
  year = {2024},
  doi = {10.1111/jdv.20307}
}
Kostenlos starten →