Deep Learning-Assisted Computer-Aided Diagnosis System for Early Detection of Lung Cancer

Area of Science:

• Medical Imaging
• Artificial Intelligence in Oncology
• Computational Pathology

Background:

• Lung cancer is a leading global cause of cancer-related mortality.
• Accurate diagnosis depends on tumor characteristics, lesion presence, cancer type (malignant/benign), and patient mental health.
• Computer-assisted diagnosis (CAD) systems are crucial for improving lung cancer detection.

Purpose of the Study:

• To develop and evaluate a novel computer-assisted diagnosis (CAD) model for enhanced lung cancer detection.
• To improve diagnostic accuracy and sensitivity compared to existing methods.

Main Methods:

• The proposed model utilizes a modified neural network architecture with 10 levels, adapted from ImageNet.
• Preprocessing involves a Fast Nonlocal Means Filter (FNLM).
• Feature extraction is performed using the Binary Grasshopper Optimization Algorithm (BGOA).

Main Results:

• The proposed model achieved a high accuracy of 99.53% and sensitivity of 98.95%.
• Performance was evaluated against deep learning techniques using the Modèle dataset.
• The model demonstrated superior effectiveness, indicated by its proximity to true positive values on the ROC curve.

Conclusions:

• The developed CAD model significantly outperforms existing techniques in accuracy and sensitivity for lung cancer diagnosis.
• The proposed deep learning strategy offers a promising advancement in early and accurate lung cancer detection.
• The model's high performance suggests its potential for clinical application in improving patient outcomes.