EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

EGFR Mutation Detection in Whole Slide Images of Non‐Small Cell Lung Cancers Using a Two‐Stage Deep Transfer Learning Approach

Michele Zanoletti iD, Filippo Ugolini, Laila El Bachiri iD, Valeria Pasini, Marco Laurino iD, Francesco De Logu, Eleonora Melissa, Carolina Marchi, Maria Colombino, Daniela Massi, Guido Rindi iD, Camilla Eva Comin und 2 weitere

Cancer Medicine · 2025-09 · 14 (18)

DOI: 10.1002/cam4.71249 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Cancer Medicine, online erschienen am 18.09.2025. Verfasst von Michele Zanoletti und 13 weiteren Autor:innen. Beteiligte Einrichtung: Institute of Clinical Physiology, National Research Council, Pisa, Italy. Hauptthemen laut MEDLINE: Carcinoma, Non-Small-Cell Lung, Lung Neoplasms, Deep Learning, Mutation. Die Arbeit wurde bisher 3-mal zitiert (Crossref) und nennt 55 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: PNRR M4/C2 Research Program of the Innovation Ecosystem, THE-Tuscany Health Ecosystem, Regione Autonoma della Sardegna.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Background

Lung cancer (LC) is the leading cause of cancer death worldwide. Non‐small cell lung cancer is the most frequent and includes adenocarcinoma and squamous cell carcinoma. Currently, LC treatment is based on tumor molecular profiling. LC may display Epidermal Growth Factor Receptor (EGFR) gene mutation. Detecting mutations in the EGFR gene is crucial for the tyrosine kinase inhibitory therapy.

Methods

This study used a computer‐based methodology with two Convolutional Neural Networks (CNNs) based on InceptionResNet‐V2, applied to Whole Slide Images, to distinguish healthy from cancerous tissue and then EGFR mutated tumor tissue samples. We also integrated an Explainable AI technique (Grad‐CAM) to clearly visualize insights into the model's decision‐making process. The analysis was conducted on 259 LC cases collected from three different centers (Florence, Rome, and Sassari).

Results

This methodology achieved an accuracy of 96.17% in distinguishing healthy from cancerous tissue, with specificity of 87.89%, sensitivity of 98.43%, an F1 score of 97.59% and an AUC of 0.99. Additionally, Cohen's Kappa indicated a consistency of 0.7982, and the confusion matrix showed a correct classification rate of 96.2%. For EGFR mutation detection in cancer tissue, slide‐level performance after aggregation reached an accuracy of 76.67% with specificity of 80.77%, sensitivity of 73.53%, an F1 score of 78.12%, a consistency of 0.5583 of Cohen's Kappa and an AUC of 0.77. The confusion matrix showed 76.7% as a correct classification rate.

Conclusion

The two tested CNNs showed potential for assisting LC diagnosis, especially in distinguishing healthy from tumor tissue. While the direct detection of EGFR mutational status remains challenging, the results suggest that relevant predictive signals can still be extracted from routine H&E slides.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Carcinoma, Non-Small-Cell LungLung NeoplasmsDeep LearningMutationHumansErbB ReceptorsFemaleDNA Mutational AnalysisMaleNeural Networks, Computer

Stichwörter der Autor:innen

CNNEGFR mutationWSIadenocarcinomaartificial intelligencedeep learningdigital pathologyexplainabilitynon‐small cell lung cancer

Substanzen

ErbB ReceptorsEGFR protein, human

Autor:innen und Institutionen

  • Institute of Clinical Physiology, National Research Council, Pisa, Italy.
  • Section of Pathological Anatomy, Careggi University Hospital, Florence, Italy.
  • Unit of Pathological Anatomy and Histology, Azienda Ospedaliero Universitaria, Department of Biomedical Sciences, University of Sassari, Sassari, Italy.
  • Section of Pathological Anatomy, Department of Health Sciences, University of Florence, Florence, Italy.
  • Section of Clinical Pharmacology and Oncology, Department of Health Sciences, University of Florence, Florence, Italy.
  • Department of Experimental and Clinical Biomedical Sciences "Mario Serio", University of Florence, Florence, Italy.

ORCID-Profile: Michele Zanoletti, Laila El Bachiri, Marco Laurino, Guido Rindi, Giuseppe Palmieri, Antonio Cossu

Förderung

  • PNRR M4/C2 Research Program of the Innovation Ecosystem, THE-Tuscany Health Ecosystem
  • Regione Autonoma della Sardegna

Zitieren

APA

Zanoletti, M., Ugolini, F., El Bachiri, L., Pasini, V., Laurino, M., Logu, F., Melissa, E., Marchi, C., Colombino, M., Massi, D., Rindi, G., Comin, C. E., Palmieri, G., & Cossu, A. (2025). EGFR Mutation Detection in Whole Slide Images of Non‐Small Cell Lung Cancers Using a Two‐Stage Deep Transfer Learning Approach. Cancer Medicine, 14(18). https://doi.org/10.1002/cam4.71249

Vancouver

Zanoletti M, Ugolini F, El Bachiri L, Pasini V, Laurino M, Logu F, et al. EGFR Mutation Detection in Whole Slide Images of Non‐Small Cell Lung Cancers Using a Two‐Stage Deep Transfer Learning Approach. Cancer Medicine. 2025;14(18). doi:10.1002/cam4.71249

BibTeX
@article{Zanoletti2025,
  author = {Zanoletti, Michele and Ugolini, Filippo and El Bachiri, Laila and Pasini, Valeria and Laurino, Marco and Logu, Francesco De and Melissa, Eleonora and Marchi, Carolina and Colombino, Maria and Massi, Daniela and Rindi, Guido and Comin, Camilla Eva and Palmieri, Giuseppe and Cossu, Antonio},
  title = {EGFR Mutation Detection in Whole Slide Images of Non‐Small Cell Lung Cancers Using a Two‐Stage Deep Transfer Learning Approach},
  journal = {Cancer Medicine},
  year = {2025},
  volume = {14},
  number = {18},
  doi = {10.1002/cam4.71249}
}
Kostenlos starten →