ABSTRACT
Computed tomography (CT) is the imaging modality of choice for screening and staging of lung cancers. This chapter explores the existing review information in order to provide an overview of the analysis methods implemented to process CT data to automatically extract useful information. It provides a description of the basic algorithmic components of a computer-aided detection (CAD) system, of the methodological issues to consider in CAD performance evaluation, and of the potential impact of a CAD on the clinical radiology workflow. CAD systems integrated in routine clinical practice of a radiology department could assist physicians in the diagnosis of a large variety of pathologies. A CAD can process the complex information encoded in biomedical images and highlight abnormalities to a human reader who is in charge of annotating the exams. CAD developers have to take into account the finite size of available cases to train, test, and validate the CAD system, and that the sample size affects the classifier performance.
