Optimizing analysis, visualization, and navigation of large image data sets: one 5000-section CT scan can ruin your whole day.

UNLABELLED: The technology revolution in image acquisition, instrumentation, and methods has resulted in vast data sets that far outstrip the human observers' ability to view, digest, and interpret modern medical images by using traditional methods. This may require a paradigm shift in the radiologic interpretation process. As human observers, radiologists must search for, detect, and interpret targets. Potential interventions should be based on an understanding of human perceptual and attentional abilities and limitations. New technologies and tools already in use in other fields can be adapted to the health care environment to improve medical image analysis, visualization, and navigation through large data sets. This historical psychophysical and technical review touches on a broad range of disciplines but focuses mainly on the analysis, visualization, and navigation of image data performed during the interpretive process. Advanced postprocessing, including three-dimensional image display, multimodality image fusion, quantitative measures, and incorporation of innovative human-machine interfaces, will likely be the future. Successful new paradigms will integrate image and nonimage data, incorporate workflow considerations, and be informed by evidence-based practices. This overview is meant to heighten the awareness of the complexities and limitations of how radiologists interact with images, particularly the large image sets generated today. Also addressed is how human-machine interface and informatics technologies could combine to transform the interpretation process in the future to achieve safer and better quality care for patients and a more efficient and effective work environment for radiologists. SUPPLEMENTAL MATERIAL: http://radiology.rsna.org/lookup/suppl/doi:10.1148/radiol.11091276/-/DC1. RSNA, 2011