Electronic Thesis/Dissertation
 

CNN-Based Transfer Learning for EGFR Mutation Prediction in Lung Cancer Histopathology Whole Slide Images

Open Access Deposited

Evaluation of histopathology is routine in clinical practice for diagnosing lung cancer. However, the association between genetic mutations that serve as essential biomarkers, such as Epidermal Growth Factor Receptor (EGFR), and image features remains largely unknown. This study aims to investigate whether EGFR mutation status can be predicted directly from histopathology image data, bypassing the need for genetic testing. This study proposes a framework for utilizing transfer learning on convolutional neural network (CNN) models to classify EGFR-mutated tissue patches from lung cancer patients. The study was performed on a total of 20 whole-slide images (WSIs) that were transformed into image patches, each labeled with EGFR mutation status. The model was evaluated using 5-fold cross-validation and demonstrated the ability to identify EGFR mutations accurately. The findings demonstrate that CNNs could be utilized to analyze WSIs to assist in detecting EGFR gene mutation in lung cancer, suggesting that somatic mutations could present subtle changes in the tumor microenvironment whose features can be identified. Ultimately, this approach has the potential to contribute to precision medicine in lung cancer treatment by integrating imaging data into the clinical decision-making process for rapid detection of mutations, facilitating the choice of targeted therapies.

Author Language Date created Type of Work License
  • All rights reserved
Rights statement GW Unit Degree Advisor Committee Member(s) Persistent URL

Notice to Authors

If you are the author of this work and you have any questions about the information on this page, please use the Contact form to get in touch with us.

Thumbnail Title Date Uploaded Visibility Actions
Preview of Bogle_gwu_0075M_17037.pdf Bogle_gwu_0075M_17037.pdf 2025-04-11 Open Access