Abstract:
The purpose of this study was to generate contrast detail (CD) curves for low contrast mass lesions embedded in images obtained in head and neck CT examinations. Axial head and neck CT slice images were randomly chosen from patients at five different levels. All images were acquired at 120 kV, and reconstructed using a standard soft tissue reconstruction filter. For each head CT image, we measured detection of low contrast mass lesions using a 2 Alternate Forced Choice (2-AFC) experimental paradigm. In an AFC experiment, an observer identifies the lesion location in one of two regions of interest. After performing 128 sequential observations, it is possible to compute the lesion contrast corresponding to a 92% accuracy of lesion detection (i.e., I92%). Five lesion sizes were investigated ranging from 4 mm to 12.5 mm, with the experimental order randomized to eliminate learning curve as well as observer fatigue. Contrast detail curves were generated by plotting log[I92%] versus log[lesion size]. Experimental slopes ranged from ~ -0.1 to ~ -0.4. The slope of the CD curve was directly related to the complexity of the anatomical structure in the head CT image. As the apparent anatomical complexity increased, the slope of the corresponding CD curve was reduced. Results from our pilot study suggest that anatomical structure is of greater importance than quantum mottle, and that the type of anatomical background structure is an important determinant of lesion detection in CT imaging.
Duke University * Arts & Sciences * Physics * Faculty * Staff * Grad * Researchers * Reload * Login
Copyright (c) 2001-2002 by Duke University Physics.