Evaluation of thresholding and region growing techniques in segmenting lung nodules in computed tomography images: a systematic review

Abstract

Image segmentation is an essential step in computer-aided diagnosis and treatment planning of lung nodules. Therefore, the purpose of this study was to perform a systematic review and provide an overview of the literature available on image segmentation algorithm, which is thresholding and region growing method regarding the optimization (of the different methodologies developed) of lung nodules in the lung CT scan prior for the lung nodule segmentation. This systematic review was compiled according to the Preferred Reporting Items for Systematic Reviews and MetaAnalyses (PRISMA) guidelines. A total of 687 articles were retrieved from the databases, and six articles were selected for this review. The finding showed that a 3D Automatic Lung Parenchyma
Extraction and Border Repair (ALPE&BR), which consists of an Automatic Single Seeded Region Growing (ASSRG) and a 3D Spherical region-growing method (SPRG), showed the highest sensitivity of 98.5% and 83.245%, respectively. Improvement of the existing methods or proposing a new one may be the best option. Standardization of the evaluation metrics is needed to allow a direct comparison between methods