A scoping review of deep learning approaches for lung cancer detection using chest radiographs and computed tomography scans

Abstract

Lung cancer remains the most lethal cancer, primarily due to late diagnoses. Thus, early detection of lung cancer is critical to improving patient outcomes. While conventional methods like Chest X-rays (CXRs) and computed tomography (CT) scans are widely used, their effectiveness can be limited by subjective interpretation and variability in the detection of subtle lesions. Recent advancements in deep learning (DL) have shown the potential to enhance the accuracy and reliability of lung cancer diagnosis through medical image analysis. This review provides a comprehensive overview of current DL approaches applied to CXRs and CT scans for lung cancer detection. Various DL techniques and their ability are explored to address challenges such as data scarcity, imbalanced datasets, and overfitting. The current state of research, including the most utilized datasets and popular DL training methods, is also examined. Future directions for integrating DL into clinical practice are discussed. The findings are based on a review of peer-reviewed literature published between January 2023 and July 2024, aiming to offer insights into the evolving landscape of DL in lung cancer detection and to outline potential pathways for future research and clinical implementation.