JoVE - Journal of Visualized Experiments
JoVE Visualize
Contact Us

Predicting Gene Comutation of EGFR and TP53 by Radiomics and Deep Learning in Patients With Lung Adenocarcinomas

  • 0Department of Radiology, the Affiliated Huaian No. 1 People's Hospital of Nanjing Medical University, Huaian.
Journal of Thoracic Imaging +

|

September 25, 2024

|

September 25, 2024

View abstract on PubMed

Summary

This summary is machine-generated.

This study developed radiomics and deep learning models to predict EGFR and TP53 mutations in lung adenocarcinoma, aiding in identifying patients for targeted therapy and predicting prognoses.

Area Of Science

  • Oncology
  • Radiology
  • Bioinformatics

Background

  • Lung adenocarcinoma (LUAD) treatment is increasingly personalized.
  • Identifying patients with specific genetic mutations like EGFR and TP53 is crucial for effective targeted therapy.
  • Accurate prediction of these mutations can guide treatment decisions and improve patient outcomes.

Purpose Of The Study

  • To construct and evaluate progressive binary classification models using radiomics and deep learning.
  • To predict the presence of epidermal growth factor receptor (EGFR) and TP53 mutations in LUAD patients.
  • To assess model performance in identifying patients suitable for TKI-targeted therapy and those with poor prognoses.

Main Methods

  • Retrospective analysis of 267 LUAD patients with genetic testing and noncontrast chest CT.
  • Feature extraction from regions of interest (ROIs) and construction of clinical, radiomics, deep learning (DL), and ensemble models.
  • Model performance evaluated using Area Under the Curve (AUC), sensitivity, specificity, accuracy, precision, and F1 score.

Main Results

  • The DL-rad-clin model achieved the highest AUC (0.783) for predicting EGFR status.
  • For TP53 status in EGFR-mutated patients, the rad-clin model showed the highest AUC (0.811).
  • Ensemble models combining clinical, radiomics, and DL features demonstrated superior predictive capabilities.

Conclusions

  • Radiomics and deep learning-based classification models offer a valuable tool for predicting EGFR and TP53 mutations in LUAD.
  • These models can serve as a reference to complement clinical identification of TKI responders and patients with poor prognoses.
  • The findings support the integration of advanced imaging analysis in precision oncology for lung cancer.

Keywords: