Evaluating the potential and problems of three-dimensional computed tomography measurements of arterial stenosis

Volume visualization is gaining widespread acceptance in medical applications. As its use increases, the issue of accuracy becomes critical. There have been very few studies examining the accuracy of volume rendering techniques. We studied the accuracy of hardware-assisted volume rendering for measurement of arterial stenosis in computed tomography (CT) data. The results of our study reveal that accurate measurements can be made from volume rendered CT data. However, error is present (absolute average error from 5.1% to 13.6%) and there is some variability, even for experts (standard deviation ranged from 4.8% to 15%). The evidence suggests that the choice of volume rendering (transfer function) parameters greatly affects the accuracy of the results. Accurate transfer function parameter selection is a difficult problem. Parameters that produce realistic images often provide inaccurate measurements. As the use of volume visualization grows and more inexperienced users begin using these tools for medical diagnosis and staging, new guidelines, aids, and techniques must be developed to ensure reliable, accurate visualization results.