[The evaluation of three dimensional image distortion of helical scanning CT]

We developed a computer algorithm to simulate distortion in the three-dimensional (3D) displays obtained by helical scanning. The algorithm constitutes calculation of the image profile in the longitudinal direction, which is assumed to be a convolution of the object function with the slice sensitivity profile (SSP) of helical scanning. Experiments were performed to examine the algorithm for its validity with the use of a K2HPO4 phantom. Simulated results and measurements was in a good agreement. The distortion was investigated by the computer simulation. The model simulated was a high density object (Ol) surrounded by low density tissue (Os). The helical interpolation used was 360-degree linear interpolation. Two parameters were defined: delta Lz, which is the difference in length between the 3D image and the actual object Ol in the longitudinal direction, and the cut level index (CLI), which is defined as CLI = (Cut Level CT Number-CT Number of Os)/(CT Number of OI-CT Number of Os). We found that magnitude of delta Lz increased depending on the table feed distance per 360-degree scan (Dt). When Ol was twice as long as Dt, delta Lz directly depended on CLI, but was independent of Ol length. When Ol length was longer than Dt, delta Lz was shown to be 0 at CLI not equal to 0.5 at every Dt. When Ol was shorter than Dt, delta Lz decreased remarkably depending on the length of Ol in higher CLI. The simulations, with the use of a newly developed algorithm, were demonstrated to be useful for evaluating the amount of distortion and for better understanding the characteristics of 3D displays of helical scanning.