EUVIMEDEUVIMED entdecken
Open Access

Deep Learning–Based Segmentation of Lung Adenocarcinoma Whole-Slide Images for Objective Grading, Tumor Spread Through Air Spaces Identification, and Mutation Prediction

David Joon Ho, Jason C Chang, Rania G Aly, Hai Cao Truong Nguyen, Prasad S Adusumilli, Thomas J Fuchs, William D Travis, Chad M Vanderbilt

Modern Pathology · 2025-12 · 38 (12) · 100907

DOI: 10.1016/j.modpat.2025.100907 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Modern Pathology, 2025-12. Verfasst von David Joon Ho und 7 weiteren Autor:innen. Beteiligte Einrichtung: Department of Pathology and Laboratory Medicine, Memorial Sloan Kettering Cancer Center, New York, New York; Department of Public Health and AI, Grad…. Hauptthemen laut MEDLINE: Deep Learning, Lung Neoplasms, Adenocarcinoma of Lung, Image Interpretation, Computer-Assisted. Die Arbeit wurde bisher 3-mal zitiert (Crossref) und nennt 57 Literaturangaben. Der Volltext ist frei in PubMed Central verfügbar. Gefördert von: NCI NIH HHS, National Cancer Center, Warren Alpert Foundation.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Manual quantification of morphologic patterns in lung adenocarcinoma is subject to reproducibility issues due to interpathologist variability. In this study, we developed a deep learning-based multiclass segmentation model providing a modality for objective and quantitative grading of digitized lung adenocarcinoma images from resected specimens. Additionally, the model can detect tumor spread through air spaces and show enrichment of specific morphologic patterns in tumors with different genomic alterations. The study was based on 766 resected nonmucinous lung adenocarcinomas. Deep Multi-Magnification Network was trained to segment 14 tissue subtypes based on annotations of 108 internal whole-slide images at pixel level by thoracic pathologists (J.C.C. and W.D.T.). The trained model was validated on an external cohort of 130 cases for determining predominant patterns and on the remaining 528 internal cases for the 3 clinical tasks. The model graded nonmucinous lung adenocarcinomas based on the International Association for the Study of Lung Cancer Pathology Committee recommendation and successfully stratified patients into well, moderately, and poorly differentiated morphologies (P < 1 × 10-4). Pixels categorized as spread through air spaces significantly correlated with pathologists' interpretations. For molecular analysis, solid pattern was enriched with TP53 mutations and depleted of EGFR kinase domain mutations. Lepidic pattern was inversely associated with TP53 mutations. Acinar was enriched with EGFR mutations, whereas papillary was associated with RET fusions. Our study demonstrated that deep learning-based segmentation can accurately quantify histologic patterns in lung adenocarcinoma and identify additional prognostic features. By simultaneously providing an objective assessment of various tasks, our comprehensive methodology in lung adenocarcinoma paves way for deep learning-assisted pathologic diagnosis and treatment guidance.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Deep LearningLung NeoplasmsAdenocarcinoma of LungImage Interpretation, Computer-AssistedHumansMutationNeoplasm GradingFemaleMaleReproducibility of Results

Stichwörter der Autor:innen

deep learningdisease-free survivallung adenocarcinomamutationsegmentationtumor spread through air spaces

Autor:innen und Institutionen

  • Department of Pathology and Laboratory Medicine, Memorial Sloan Kettering Cancer Center, New York, New York; Department of Public Health and AI, Graduate School of Cancer Science and Policy, National Cancer Center, Goya…
  • Department of Pathology and Laboratory Medicine, Memorial Sloan Kettering Cancer Center, New York, New York.
  • Department of Public Health and AI, Graduate School of Cancer Science and Policy, National Cancer Center, Goyang, Republic of Korea.
  • Department of Surgery, Memorial Sloan Kettering Cancer Center, New York, New York.
  • Windreich Department of Artificial Intelligence and Human Health, Icahn School of Medicine at Mount Sinai, New York, New York.

Förderung

  • NCI NIH HHS
  • National Cancer Center
  • Warren Alpert Foundation
  • National Institutes of Health
  • NIH National Cancer Institute

Zitieren

APA

Ho, D. J., Chang, J. C., Aly, R. G., Nguyen, H. C. T., Adusumilli, P. S., Fuchs, T. J., Travis, W. D., & Vanderbilt, C. M. (2025). Deep Learning–Based Segmentation of Lung Adenocarcinoma Whole-Slide Images for Objective Grading, Tumor Spread Through Air Spaces Identification, and Mutation Prediction. Modern Pathology, 38(12), 100907. https://doi.org/10.1016/j.modpat.2025.100907

Vancouver

Ho DJ, Chang JC, Aly RG, Nguyen HCT, Adusumilli PS, Fuchs TJ, et al. Deep Learning–Based Segmentation of Lung Adenocarcinoma Whole-Slide Images for Objective Grading, Tumor Spread Through Air Spaces Identification, and Mutation Prediction. Modern Pathology. 2025;38(12):100907. doi:10.1016/j.modpat.2025.100907

BibTeX
@article{Ho2025,
  author = {Ho, David Joon and Chang, Jason C and Aly, Rania G and Nguyen, Hai Cao Truong and Adusumilli, Prasad S and Fuchs, Thomas J and Travis, William D and Vanderbilt, Chad M},
  title = {Deep Learning–Based Segmentation of Lung Adenocarcinoma Whole-Slide Images for Objective Grading, Tumor Spread Through Air Spaces Identification, and Mutation Prediction},
  journal = {Modern Pathology},
  year = {2025},
  volume = {38},
  number = {12},
  pages = {100907},
  doi = {10.1016/j.modpat.2025.100907}
}
Kostenlos starten →