EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Prediction of molecular subclasses of uveal melanoma by deep learning using routine haematoxylin–eosin‐stained tissue slides

Farhan Akram iD, Daniël P de Bruyn iD, Quincy C C van den Bosch, Teodora E Trandafir, Thierry P P van den Bosch, Rob M Verdijk, Annelies de Klein, Emine Kiliç, Andrew P Stubbs, Erwin Brosens iD, Jan H von der Thüsen iD

Histopathology · 2024

DOI: 10.1111/his.15271 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Histopathology, 2024. Verfasst von Farhan Akram und 10 weiteren Autor:innen. Beteiligte Einrichtung: Department of Pathology and Clinical Bioinformatics, Erasmus MC Rotterdam, Rotterdam, the Netherlands. Hauptthemen laut MEDLINE: Deep Learning, Melanoma, Uveal Neoplasms, Ubiquitin Thiolesterase. Die Arbeit wurde bisher 11-mal zitiert (Crossref) und nennt 31 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Combined Ophthalmic Research Rotterdam (CORR) Foundation, Stichting Combined Ophthalmic Research Rotterdam.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Aims

Uveal melanoma has a high propensity to metastasize. Prognosis is associated with specific driver mutations and copy number variations, and these can only be obtained after genetic testing. In this study we evaluated the efficacy of patient outcome prediction using deep learning on haematoxylin and eosin (HE)‐stained primary uveal melanoma slides in comparison to molecular testing.

Methods

In this retrospective study of patients with uveal melanoma, 113 patients from the Erasmus Medical Centre who underwent enucleation had tumour tissue analysed for molecular classification between 1993 and 2020. Routine HE‐stained slides were scanned to obtain whole‐slide images (WSI). After annotation of regions of interest, tiles of 1024 × 1024 pixels were extracted at a magnification of 40×. An ablation study to select the best‐performing deep‐learning model was carried out using three state‐of‐the‐art deep‐learning models (EfficientNet, Vision Transformer, and Swin Transformer).

Results

Deep‐learning models were subjected to a training cohort ( n = 40), followed by a validation cohort ( n = 20), and finally underwent a test cohort ( n = 48). A k‐fold cross‐validation ( k = 3) of validation and test cohorts ( n = 113 of three classes: BAP1, SF3B1, EIF1AX) demonstrated Swin Transformer as the best‐performing deep‐learning model to predict molecular subclasses based on HE stains. The model achieved an accuracy of 0.83 ± 0.09 on the validation cohort and 0.75 ± 0.04 on the test cohort. Within the subclasses, this model correctly predicted 70% BAP1 ‐mutated, 61% SF3B1 ‐mutated and 80% EIF1AX ‐mutated UM in the test set.

Conclusions

This study showcases the potential of the deep‐learning methodology for predicting molecular subclasses in a multiclass manner using HE‐stained WSI. This development holds promise for advanced prognostication of UM patients without the need of molecular or immunohistochemical testing. Additionally, this study suggests there are distinct histopathological features per subclass; mainly utilizing epithelioid cellular morphology for BAP1 ‐classification, but an unknown feature distinguishes EIF1AX and SF3B1 .

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Deep LearningMelanomaUveal NeoplasmsUbiquitin ThiolesteraseHumansRetrospective StudiesMaleFemaleMiddle AgedAgedHematoxylinEosine Yellowish-(YS)AdultTumor Suppressor ProteinsRNA Splicing FactorsEukaryotic Initiation Factor-1Aged, 80 and overPrognosisStaining and LabelingMutationPhosphoproteinsUveal Melanoma

Substanzen

BAP1 protein, humanUbiquitin ThiolesteraseHematoxylinSF3B1 protein, humanEosine Yellowish-(YS)Tumor Suppressor ProteinsRNA Splicing FactorsEukaryotic Initiation Factor-1eukaryotic peptide initiation factor-1APhosphoproteins

Autor:innen und Institutionen

  • Department of Pathology and Clinical Bioinformatics, Erasmus MC Rotterdam, Rotterdam, the Netherlands.
  • Ophthalmology, Erasmus MC Rotterdam, Rotterdam, the Netherlands.
  • Clinical Genetics, Erasmus MC Rotterdam, Rotterdam, the Netherlands.
  • Cancer Institute, Erasmus MC Rotterdam, Rotterdam, the Netherlands.

ORCID-Profile: Farhan Akram, Daniël P de Bruyn, Erwin Brosens, Jan H von der Thüsen

Förderung

  • Combined Ophthalmic Research Rotterdam (CORR) Foundation
  • Stichting Combined Ophthalmic Research Rotterdam

Zitieren

APA

Akram, F., de Bruyn, D. P., van den Bosch, Q. C. C., Trandafir, T. E., van den Bosch, T. P. P., Verdijk, R. M., de Klein, A., Kiliç, E., Stubbs, A. P., Brosens, E., & von der Thüsen, J. H. (2024). Prediction of molecular subclasses of uveal melanoma by deep learning using routine haematoxylin–eosin‐stained tissue slides. Histopathology. https://doi.org/10.1111/his.15271

Vancouver

Akram F, de Bruyn DP, van den Bosch QCC, Trandafir TE, van den Bosch TPP, Verdijk RM, et al. Prediction of molecular subclasses of uveal melanoma by deep learning using routine haematoxylin–eosin‐stained tissue slides. Histopathology. 2024. doi:10.1111/his.15271

BibTeX
@article{Akram2024,
  author = {Akram, Farhan and de Bruyn, Daniël P and van den Bosch, Quincy C C and Trandafir, Teodora E and van den Bosch, Thierry P P and Verdijk, Rob M and de Klein, Annelies and Kiliç, Emine and Stubbs, Andrew P and Brosens, Erwin and von der Thüsen, Jan H},
  title = {Prediction of molecular subclasses of uveal melanoma by deep learning using routine haematoxylin–eosin‐stained tissue slides},
  journal = {Histopathology},
  year = {2024},
  doi = {10.1111/his.15271}
}
Kostenlos starten →