Radiomics-Clinical AI Model with Probability Weighted Strategy for Prognosis Prediction in Non-Small Cell Lung Cancer
Fuk-Hay Tang1, Yee-Wai Fong1, Shing-Hei Yung1
1School of Medical and Health Sciences, Tung Wah College, Hong Kong, China.
Biomedicines
|August 26, 2023
Summary
A new radiomics clinical probability-weighted model accurately predicts non-small cell lung cancer (NSCLC) prognosis. Combining imaging features and clinical data, this approach enhances survival prediction and aids personalized treatment for NSCLC patients.
Area of Science:
- Oncology
- Radiology
- Medical Imaging
- Machine Learning
Background:
- Non-small cell lung cancer (NSCLC) prognosis prediction remains challenging.
- Accurate prognostic models are crucial for guiding personalized treatment strategies.
Purpose of the Study:
- To develop and evaluate a radiomics clinical probability-weighted model for NSCLC prognosis.
- To assess the model's performance in predicting patient survival.
Main Methods:
- Radiomic features were extracted from CT images of 422 NSCLC patients.
- A voted ensemble machine learning (VEML) model optimized five algorithms.
- A probabilistic weighted approach integrated radiomic and clinical data for risk scoring.
Main Results:
- The combined model achieved an AUC of 0.949, outperforming radiomic (0.941) and clinical (0.856) models.
- The enhanced model significantly improved 1-year survival prediction (p < 0.05).
Conclusions:
- The proposed radiomics clinical probability-weighted model demonstrates high accuracy in NSCLC prognosis.
- This model has the potential to improve patient outcomes and personalize treatment decisions.
