Skip to main content
SpringerLink
Log in

Bilderkennung mittels künstlicher Intelligenz in der Hautkrebsdiagnostik

Artificial intelligence-based classification for the diagnostics of skin cancer

  • Leitthema
  • Published:
Die Dermatologie Aims and scope Submit manuscript

Zusammenfassung

Faltende neuronale Netzwerke (CNN) erzielen bei der Beurteilung von pigmentierten und nicht pigmentierten Hautveränderungen bereits eine vergleichbare oder sogar überlegene Leistung gegenüber Dermatolog*innen. Bei der Analyse von Bildern durch künstliche neuronale Netzwerke durchlaufen diese auf Pixelebene verschiedene Schichten mit diversen grafischen Filtern. Basierend auf hervorragenden Studienergebnissen erhielt ein erstes Deep-learning-Netzwerk (Moleanalyzer pro, Fotofinder Systems GmbH, Bad Birnbach, Deutschland) in Europa die Marktzulassung. Solche Netzwerke weisen jedoch auch wichtige Limitationen auf, seltenere Entitäten mit unzureichenden Trainingsbildern werden weniger gut klassifiziert, ebenso können Bildartefakte zu Fehldiagnosen führen. In einer Kooperation von „Mensch mit Maschine“ sind letztendlich die besten Ergebnisse zu erzielen. Im Hinblick auf die Zukunft des Hautkrebsscreenings wird das automatisierte Ganzkörpermapping evaluiert, das Ganzkörperfotografie sowie automatisierte Erfassung und Bewertung sämtlicher relevanter Hautläsionen kombiniert.

Abstract

Convolutional neural networks (CNN) achieve a level of performance comparable or even superior to dermatologists in the assessment of pigmented and nonpigmented skin lesions. In the analysis of images by artificial neural networks, images on a pixel level pass through various layers of the network with different graphic filters. Based on excellent study results, a first deep learning network (Moleanalyzer pro, Fotofinder Systems GmBH, Bad Birnbach, Germany) received market approval in Europe. However, such neural networks also reveal relevant limitations, whereby rare entities with insufficient training images are classified less adequately and image artifacts can lead to false diagnoses. Best results can ultimately be achieved in a cooperation of “man with machine”. For future skin cancer screening, automated total body mapping is evaluated, which combines total body photography, automated data extraction and assessment of all relevant skin lesions.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Log in via an institution

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Institutional subscriptions

Abb. 1
Abb. 2
Abb. 3

Literatur

  1. Argenziano G, Catricala C, Ardigo M et al (2011) Seven-point checklist of dermoscopy revisited. Br J Dermatol 164:785–790

    Article  CAS  Google Scholar 

  2. Balch CM, Buzaid AC, Soong S‑J et al (2001) Final version of the American Joint Committee on Cancer staging system for cutaneous melanoma. J Clin Oncol 19:3635–3648

    Article  CAS  Google Scholar 

  3. Blum A, Bosch S, Haenssle H et al (2020) Künstliche Intelligenz und Smartphone-Programm-Applikationen (Apps). Bedeutung für die dermatologische Praxis. Dermatologie. https://doi.org/10.1007/s00105-020-04658-4

    Article  Google Scholar 

  4. Brinker TJ, Hekler A, Utikal JS et al (2018) Skin cancer classification using convolutional neural networks: systematic review. J Med Internet Res 20:e11936

    Article  Google Scholar 

  5. Brinker TJ, Schmitt M, Krieghoff-Henning EI et al (2022) Diagnostic performance of artificial intelligence for histologic melanoma recognition compared to 18 international expert pathologists. J Am Acad Dermatol 86:640–642

    Article  Google Scholar 

  6. Dick V, Tschandl P, Sinz C et al (2018) Bildbasierte Computerdiagnose des Melanoms. Hautarzt 69:591–601

    Article  CAS  Google Scholar 

  7. Esteva A, Kuprel B, Novoa RA et al (2017) Dermatologist-level classification of skin cancer with deep neural networks. Nature 542:115

    Article  CAS  Google Scholar 

  8. Haenssle H, Fink C, Schneiderbauer R et al (2018) Man against machine: diagnostic performance of a deep learning convolutional neural network for dermoscopic melanoma recognition in comparison to 58 dermatologists. Ann Oncol 29:1836–1842

    Article  CAS  Google Scholar 

  9. Haenssle HA, Fink C, Toberer F et al (2020) Man against machine reloaded: performance of a market-approved convolutional neural network in classifying a broad spectrum of skin lesions in comparison with 96 dermatologists working under less artificial conditions. Ann Oncol 31:137–143

    Article  CAS  Google Scholar 

  10. Haenssle HA, Winkler JK, Fink C et al (2021) Skin lesions of face and scalp—classification by a market-approved convolutional neural network in comparison with 64 dermatologists. Eur J Cancer 144:192–199

    Article  Google Scholar 

  11. Lecun Y, Bengio Y, Hinton G (2015) Deep learning. Nature 521:436–444

    Article  CAS  Google Scholar 

  12. Malvehy J, Puig S (2002) Follow-up of melanocytic skin lesions with digital total-body photography and digital dermoscopy: a two-step method. Clin Dermatol 20:297–304

    Article  Google Scholar 

  13. Nelson CA, Pérez-Chada LM, Creadore A et al (2020) Patient perspectives on the use of artificial intelligence for skin cancer screening: a qualitative study. JAMA Dermatol 156:501–512

    Article  Google Scholar 

  14. Rinner C, Kittler H, Rosendahl C et al (2020) Analysis of collective human intelligence for diagnosis of pigmented skin lesions harnessed by gamification via a web-based training platform: simulation reader study. J Med Internet Res 22:e15597

    Article  Google Scholar 

  15. Salerni G, Carrera C, Lovatto L et al (2012) Benefits of total body photography and digital dermatoscopy (“two-step method of digital follow-up”) in the early diagnosis of melanoma in patients at high risk for melanoma. J Am Acad Dermatol 67:e17–e27

    Article  Google Scholar 

  16. Tschandl P, Codella N, Akay BN et al (2020) Comparison of the accuracy of human readers versus machine-learning algorithms for pigmented skin lesion classification: an open, web-based, international, diagnostic study. Lancet Oncol 20(7):938–947

    Article  Google Scholar 

  17. Tschandl P, Hofmann L, Fink C et al (2017) Melanomas vs. nevi in high-risk patients under long-term monitoring with digital dermatoscopy: Do melanomas and nevi already differ at baseline? J Eur Acad Dermatol Venerol 31:972–977

    Article  CAS  Google Scholar 

  18. Tschandl P, Rinner C, Apalla Z et al (2020) Human-computer collaboration for skin cancer recognition. Nat Med 26(8):1229–1234

    Article  CAS  Google Scholar 

  19. Tschandl P, Rosendahl C, Kittler H (2018) The HAM10000 dataset: a large collection of multi-source dermatoscopic images of common pigmented skin lesions. arXiv. https://doi.org/10.48550/arXiv.1803.10417

    Article  Google Scholar 

  20. Urban K, Mehrmal S, Uppal P et al (2021) The global burden of skin cancer: a longitudinal analysis from the Global Burden of Disease Study, 1990–2017. JAAD Int 2:98–108

    Article  Google Scholar 

  21. Vestergaard M, Macaskill P, Holt P et al (2008) Dermoscopy compared with naked eye examination for the diagnosis of primary melanoma: a meta-analysis of studies performed in a clinical setting. Br J Dermatol 159:669–676

    CAS  PubMed  Google Scholar 

  22. Winkler JK, Fink C, Toberer F et al (2019) Association between surgical skin markings in dermoscopic images and diagnostic performance of a deep learning convolutional neural network for melanoma recognition. JAMA Dermatol 155(10):1135–1141

    Article  Google Scholar 

  23. Winkler JK, Sies K, Fink C et al (2021) Collective human intelligence outperforms artificial intelligence in a skin lesion classification task. J Dtsch Dermatol Ges 19:1178–1184

    PubMed  Google Scholar 

  24. Winkler JK, Sies K, Fink C et al (2021) Association between different scale bars in dermoscopic images and diagnostic performance of a market-approved deep learning convolutional neural network for melanoma recognition. Eur J Cancer 145:146–154

    Article  Google Scholar 

  25. Winkler JK, Sies K, Fink C et al (2020) Melanoma recognition by a deep learning convolutional neural network—performance in different melanoma subtypes and localisations. Eur J Cancer 127:21–29

    Article  Google Scholar 

  26. Winkler JK, Tschandl P, Toberer F et al (2022) Monitoring patients at risk for melanoma: May convolutional neural networks replace the strategy of sequential digital dermoscopy? Eur J Cancer 160:180–188

    Article  Google Scholar 

Download references

Author information

Authors and Affiliations

  1. Universitätshautklinik Heidelberg, Im Neuenheimer Feld 440, 69120, Heidelberg, Deutschland

    Julia K. Winkler & Holger A. Haenssle

Authors
  1. Julia K. Winkler
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Holger A. Haenssle
    View author publications

    You can also search for this author in PubMed Google Scholar

Corresponding author

Correspondence to Julia K. Winkler.

Ethics declarations

Interessenkonflikt

J.K. Winkler gibt an, Vortragshonorare von BMS und Fotofinder Systems GmbH sowie Reisekostenunterstützung von Amgen, BMS, MSD, Philochem und Roche erhalten zu haben. H.A. Haenssle erhielt Honorare und/oder Reisekostenunterstützung von Scibase AB, FotoFinder Systems GmbH, Heine Optotechnik GmbH, Magnosco GmbH.

Für diesen Beitrag wurden von den Autor/-innen keine Studien an Menschen oder Tieren durchgeführt. Für die aufgeführten Studien gelten die jeweils dort angegebenen ethischen Richtlinien. Für Bildmaterial oder anderweitige Angaben innerhalb des Manuskripts, über die Patient/-innen zu identifizieren sind, liegt von ihnen und/oder ihren gesetzlichen Vertretern/Vertreterinnen eine schriftliche Einwilligung vor.

Additional information

figure qr

QR-Code scannen & Beitrag online lesen

Rights and permissions

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Winkler, J.K., Haenssle, H.A. Bilderkennung mittels künstlicher Intelligenz in der Hautkrebsdiagnostik. Dermatologie 73, 838–844 (2022). https://doi.org/10.1007/s00105-022-05058-6

Download citation

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s00105-022-05058-6

Schlüsselwörter

Keywords

Search

Navigation