Many factors influence the detection of abnormalities in medical images. Poor acquisition technique, improper processing and presentation of the image data, and sub-optimal viewing of the images are among these factors. However, even when from a technical standpoint such processes are optimal, subtle abnormalities in the image can go unrecognized. For example, subtle lesions in chest radiographs and in mammograms, and subtle fractures in muscloskeletal images are frequently missed (Muhm, 1983; Bird, 1992; Wallis, 1991; Hu, 1994). The difficulty in the detection of subtle abnormalities can be attributed to the statistical nature of decision-making in conjunction with psychophysical elements, both perceptual and cognitive, affecting the detection process. An attribute of medical images that greatly influences the psychophysics of detection is the anatomical structure within which the abnormalities are to be detected.
Anatomical structure within medical radiological images is created by the normal anatomy of the patient. In cross-sectional imaging, such as Computed Tomography (CT) and Magnetic Resonance Imaging (MRI), the structure results from the anatomy present in the cross-sectional slab of the image. In projection imaging, such as radiography and mammography, the anatomical structure is more complex, as the projection of the three-dimensional anatomy creates overlays in the two-dimensional image plane. The anatomical structure in either type of image influences the detection of abnormalities in the image, elevating the contrast threshold for detection, complicating the localization of the signals, and creating false signals.
The perceptual effects of anatomical structure on detection, though well recognized, have not been well-understood (Greening, 1954; Boynton, 1956; Smith, 1967; Neitzel, 1998). This chapter focuses on these effects from a visualâperceptual standpoint. First, the noise characteristics of anatomical structure are discussed. An attempt is then made to delineate some of the mechanisms through which anatomical structure affects detection. Finally, the influence of anatomical structure on signal detection is demonstrated in a few specific clinical examples. The chapter ends by outlining some methods for reducing the influence of anatomical structure.
Online access to SPIE eBooks is limited to subscribing institutions.