[Prognosis Prediction of Lung Cancer Patients Using CT Images: Feature Extraction by Convolutional Neural Network and...

Area of Science:

• Medical Imaging Analysis
• Computational Oncology
• Artificial Intelligence in Medicine

Background:

• Lung cancer remains a leading cause of cancer mortality worldwide.
• Accurate prognosis prediction is crucial for patient management and treatment planning.
• Current methods for predicting lung cancer prognosis from CT images have limitations.

Purpose of the Study:

• To develop and evaluate a novel method for predicting lung cancer patient prognosis using CT images.
• To leverage convolutional neural networks (CNNs) for feature extraction and machine learning for classification.
• To assess the accuracy of the developed method in predicting survival at one- to five-year intervals.

Main Methods:

• Collected CT images from 173 lung cancer patients.
• Utilized CNNs to extract features from the largest tumor slice in each CT scan.
• Applied information gain for feature selection.
• Employed artificial neural networks and Naïve Bayes classifiers to predict survival (alive/death) at yearly intervals (1-5 years).
• Validated the model using three-fold cross-validation.

Main Results:

• The developed method achieved prediction accuracies of approximately 80% for all time points from 1 to 5 years.
• The predicted survival curves closely matched the actual survival curves.
• The combination of CNN-based feature extraction and machine learning classification demonstrated predictive capability.

Conclusions:

• Feature extraction using CNNs combined with machine learning classification shows promise for predicting lung cancer prognosis from CT images.
• This AI-driven approach may enhance the ability to forecast patient outcomes, aiding clinical decision-making.
• Further research can explore refining these methods for improved accuracy and clinical integration.