Detection of lung cancer mutation based on clinical and morphological features using adaptive boosting method

Lung cancer is a leading cause of cancer-related mortality worldwide, and accurate detection of epidermal growth factor receptor mutations is essential for personalized treatment. However, non-invasive identification of these mutations remains challenging due to the complexity of clinical and morphological patterns. This study develops an adaptive boosting (AdaBoost)-based machine learning model for detecting lung cancer mutations using clinical and morphological data. The dataset consists of clinical and morphological attributes from 80 patients, which processed through comprehensive preprocessing steps, including imputation, outlier removal, and feature selection. One-hot encoding increased the feature count beyond the original 28, and analysis of variance was employed to retain the most relevant 33 features. AdaBoost was trained with optimized hyperparameters, including learning rate and the number of estimators, which were tuned using grid search to ensure robustness. The model’s performance was evaluated using an 80/20 train-test split and k-fold cross-validation to assess generalization capability. Experimental results demonstrated that AdaBoost outperformed other models, achieving an accuracy of 83% and an area under the curve of 0.90 after feature selection. The model maintained superior cross-validation scores compared to Naive Bayes, decision tree, K-nearest neighbors, and support vector machine, reinforcing its reliability in mutation detection. The study highlights the significance of preprocessing steps in improving classification performance and suggests that AdaBoost can serve as an effective, non-invasive tool for assisting clinical decision-making in lung cancer mutation detection.