Clinical Report: Evaluating EGFR Gene Mutations in Lung Adenocarcinoma
Overview
This study demonstrates the effectiveness of a combined model utilizing 18F-FDG PET/CT radiomics and tumor habitat analysis to predict EGFR mutation status in lung adenocarcinoma. The combined model achieved an AUC of 0.862, outperforming other predictive models.
Background
EGFR mutations are critical in the management of lung adenocarcinoma, influencing treatment decisions and patient outcomes. Accurate prediction of these mutations can guide personalized therapy, particularly in advanced stages of the disease. The integration of advanced imaging techniques with radiomics offers a promising approach to enhance predictive accuracy.
Data Highlights
Model
AUC
95% CI
Combined Model
0.862
0.80–0.93
Habitat Model
0.831
0.76–0.90
Key Findings
The combined model outperformed all other models in predicting EGFR mutations (AUC = 0.862).
The habitat model also showed strong predictive performance (AUC = 0.831).
Peritumoral models with a 6 mm expansion had the highest AUC among peritumoral analyses.
SHAP analysis revealed that 16 of 17 key features in the habitat model were derived from specific tumor habitat subregions.
Approximately two-thirds of the top predictive features were based on CT imaging.
Clinical Implications
The findings suggest that integrating 18F-FDG PET/CT radiomics with tumor habitat analysis can significantly enhance the prediction of EGFR mutation status. This approach may facilitate more personalized treatment strategies for patients with lung adenocarcinoma.
Conclusion
The study underscores the potential of advanced imaging techniques in predicting EGFR mutations, which is crucial for guiding targeted therapies in lung adenocarcinoma. Further validation in clinical settings is warranted.
