Evaluation of section sensitivity profiles and image noise in spiral CT.

Abstract

Spiral computed tomography (CT) offers continuous volume scanning of complete organs or body sections within a single breath hold. Almost all image quality characteristics of spiral CT are identical to those of conventional section-by-section CT; however, there is a change in pixel noise values and degradation in the shape of the section sensitivity profiles (SSPs). Computer simulations, phantom measurements, and clinical studies were used in evaluating the SSP and noise characteristics of two new section-interpolation algorithms. The results were compared with standard CT and spiral CT data processed with the commonly employed linear section-interpolation algorithm. Degradation of SSP quality was insignificant for a table feed distance per 360 degrees revolution equal to the section thickness when the new algorithms were applied; noise values, however, increased. SSP width increased for table feed distances greater than the section width, the effect being less pronounced with the new algorithms. The value of these algorithms is primarily seen in the improved quality of multiplanar reformations and cine and three-dimensional displays.