EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Predicting Molecular Subtype and Survival of Rhabdomyosarcoma Patients Using Deep Learning of H&E Images: A Report from the Children's Oncology Group

David Milewski iD, Hyun Jung iD, G Thomas Brown iD, Yanling Liu iD, Ben Somerville iD, Curtis Lisle iD, Marc Ladanyi iD, Erin R Rudzinski iD, Hyoyoung Choo-Wosoba iD, Donald A Barkauskas iD, Tammy Lo iD, David Hall iD und 13 weitere

Clinical Cancer Research · 2023

DOI: 10.1158/1078-0432.ccr-22-1663 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Clinical Cancer Research, 2023. Verfasst von David Milewski und 24 weiteren Autor:innen. Beteiligte Einrichtung: Genetics Branch, NCI, NIH, Bethesda, Maryland. Hauptthemen laut MEDLINE: Deep Learning, Rhabdomyosarcoma, Rhabdomyosarcoma, Alveolar. Die Arbeit wurde bisher 23-mal zitiert (Crossref) und nennt 50 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Intramural NIH HHS, NCI NIH HHS, National Cancer Institute.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Purpose

Rhabdomyosarcoma (RMS) is an aggressive soft-tissue sarcoma, which primarily occurs in children and young adults. We previously reported specific genomic alterations in RMS, which strongly correlated with survival; however, predicting these mutations or high-risk disease at diagnosis remains a significant challenge. In this study, we utilized convolutional neural networks (CNN) to learn histologic features associated with driver mutations and outcome using hematoxylin and eosin (H&E) images of RMS. Experimental Design: Digital whole slide H&E images were collected from clinically annotated diagnostic tumor samples from 321 patients with RMS enrolled in Children's Oncology Group (COG) trials (1998–2017). Patches were extracted and fed into deep learning CNNs to learn features associated with mutations and relative event-free survival risk. The performance of the trained models was evaluated against independent test sample data (n = 136) or holdout test data.

Results

The trained CNN could accurately classify alveolar RMS, a high-risk subtype associated with PAX3/7-FOXO1 fusion genes, with an ROC of 0.85 on an independent test dataset. CNN models trained on mutationally-annotated samples identified tumors with RAS pathway with a ROC of 0.67, and high-risk mutations in MYOD1 or TP53 with a ROC of 0.97 and 0.63, respectively. Remarkably, CNN models were superior in predicting event-free and overall survival compared with current molecular-clinical risk stratification.

Conclusions

This study demonstrates that high-risk features, including those associated with certain mutations, can be readily identified at diagnosis using deep learning. CNNs are a powerful tool for diagnostic and prognostic prediction of rhabdomyosarcoma, which will be tested in prospective COG clinical trials.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Deep LearningRhabdomyosarcomaRhabdomyosarcoma, AlveolarChildHumansYoung AdultEosine Yellowish-(YS)HematoxylinPaired Box Transcription FactorsProspective StudiesPAX3 Transcription FactorPAX7 Transcription Factor

Substanzen

Eosine Yellowish-(YS)HematoxylinPaired Box Transcription FactorsPAX3 protein, humanPAX3 Transcription FactorPAX7 protein, humanPAX7 Transcription Factor

Autor:innen und Institutionen

  • Genetics Branch, NCI, NIH, Bethesda, Maryland.
  • Advanced Biomedical Computational Science, Frederick National Laboratory for Cancer Research, Frederick, Maryland.
  • Artificial Intelligence Resource, NCI, NIH, Bethesda, Maryland.
  • KnowledgeVis, LLC, Altamonte Springs, Florida.
  • Department of Pathology, Memorial Sloan-Kettering Cancer Center, New York, New York.
  • Department of Laboratories, Seattle Children's Hospital, Seattle, Washington.

ORCID-Profile: David Milewski, Hyun Jung, G Thomas Brown, Yanling Liu, Ben Somerville, Curtis Lisle, Marc Ladanyi, Erin R Rudzinski, Hyoyoung Choo-Wosoba, Donald A Barkauskas, Tammy Lo, David Hall, Corinne M Linardic, Jun S Wei, Hsien-Chao Chou, Stephen X Skapek, Rajkumar Venkatramani, Peter K Bode, Seth M Steinberg, George Zaki, Igor B Kuznetsov, Douglas S Hawkins, Jack F Shern, Jack Collins, Javed Khan

Förderung

  • Intramural NIH HHS
  • NCI NIH HHS
  • National Cancer Institute

Zitieren

APA

Milewski, D., Jung, H., Brown, G. T., Liu, Y., Somerville, B., Lisle, C., Ladanyi, M., Rudzinski, E. R., Choo-Wosoba, H., Barkauskas, D. A., Lo, T., Hall, D., Linardic, C. M., Wei, J. S., Chou, H. C., Skapek, S. X., Venkatramani, R., Bode, P. K., Steinberg, S. M., …, & Khan, J. (2023). Predicting Molecular Subtype and Survival of Rhabdomyosarcoma Patients Using Deep Learning of H&E Images: A Report from the Children's Oncology Group. Clinical Cancer Research. https://doi.org/10.1158/1078-0432.ccr-22-1663

Vancouver

Milewski D, Jung H, Brown GT, Liu Y, Somerville B, Lisle C, et al. Predicting Molecular Subtype and Survival of Rhabdomyosarcoma Patients Using Deep Learning of H&E Images: A Report from the Children's Oncology Group. Clinical Cancer Research. 2023. doi:10.1158/1078-0432.ccr-22-1663

BibTeX
@article{Milewski2023,
  author = {Milewski, David and Jung, Hyun and Brown, G Thomas and Liu, Yanling and Somerville, Ben and Lisle, Curtis and Ladanyi, Marc and Rudzinski, Erin R and Choo-Wosoba, Hyoyoung and Barkauskas, Donald A and Lo, Tammy and Hall, David and Linardic, Corinne M and Wei, Jun S and Chou, Hsien-Chao and Skapek, Stephen X and Venkatramani, Rajkumar and Bode, Peter K and Steinberg, Seth M and Zaki, George and Kuznetsov, Igor B and Hawkins, Douglas S and Shern, Jack F and Collins, Jack and Khan, Javed},
  title = {Predicting Molecular Subtype and Survival of Rhabdomyosarcoma Patients Using Deep Learning of H&E Images: A Report from the Children's Oncology Group},
  journal = {Clinical Cancer Research},
  year = {2023},
  doi = {10.1158/1078-0432.ccr-22-1663}
}
Kostenlos starten →