Using Convoluted Neural Networks in Diagnosing Lung Cancer on Computed Tomography Scans.

Area of Science:

• Medical Imaging
• Artificial Intelligence
• Oncology

Background:

• Lung cancer is a leading cause of mortality worldwide, necessitating advanced diagnostic tools.
• Early detection and accurate classification of lung nodules are critical for effective patient management.
• Computer-assisted diagnostic systems can enhance the analysis of complex medical imaging data.

Purpose of the Study:

• To develop and evaluate a novel convoluted neural network (CNN) for classifying lung nodules in computed tomography (CT) images.
• To assess the performance of the CNN model in differentiating between malignant and benign lung lesions.
• To explore the feasibility of implementing AI-driven tools for lung cancer diagnosis.

Main Methods:

• A deep learning model (CNN) was trained using augmented CT images from 176 patients.
• The dataset included cases with confirmed lung masses and positive pathology or follow-up.
• Model performance was evaluated using accuracy, recall, and precision metrics.

Main Results:

• The CNN model achieved a validation accuracy of 77.01% on an imbalanced dataset.
• Recall was 79.31% and precision reached 62.16% in classifying malignant versus benign lung lesions.
• The study successfully enrolled 176 patients, with most tumors located in the right lung.

Conclusions:

• A CNN model can be effectively implemented on standard hardware for classifying malignant and benign lung lesions from CT scans.
• AI-driven tools show promise for improving the accuracy and efficiency of lung lesion imaging diagnosis.
• Physician oversight remains essential for medical management decisions in lung cancer cases.