Pipelines in Image Analysis

The advent of quantitative image analysis has spawned the collection and organization of many cohorts of combined imaging and clinical data. Researchers want to be able to access these cohorts and use them to discover combinations of features that can be used to predict clinical variables, such as response to specific treatments and/or overall survival. There are several software platforms that can derive image features from regions within multi-dimensional image volumes, but the challenge is to be able to perform these analyses efficiently for large cohorts of patients. Quantitative imaging pipelines result from the integration of tools into a package with a convenient interface that can be run to build predictive models incorporating imaging data. This chapter discusses the building blocks employed by these pipelines and gives two examples of pipelines that are currently available and still evolving.