A Systematic Review of Recent Chest Radiograph Bone Suppression Techniques

Abstract

Chest radiograph (CXR) are essential diagnostic tools to visualize the thoracic cavity's anatomical structures, particularly the lungs. However, the interpretability of these radiographs can be compromised by the presence of overlying bones, such as the ribs and clavicles, which may obstruct the view of the lung regions. Recently, the bone suppression technique applied to CXR has shown promise in aiding radiologists and computer-aided diagnosis systems in detecting lung diseases. Numerous studies have indicated that employing bone-suppressed images (BSIs) provides clinical evidence of enhancing diagnostic accuracy and confidence. This systematic review paper provides a recent of CXR bone suppression techniques and highlights their respective results. The preference for systematic analysis over traditional literature review stems from its capability to mitigate research bias. Recently, researchers increasingly favour using deep learning methodology to suppress bone structures. Implementing these techniques opens a pathway for various applications, particularly in lung nodule detection or pathology assessment through radiological analysis of CXR.