A method to compare supra-pubic ultrasound and CT images of the prostate: technique and early clinical results.

Abstract

We describe a unique method that allows the comparison of spatially registered ultrasound (SRUS) images and computed tomography-derived contours (CTDCs) that were acquired with a minimal time lapse. As such, we have a tool that will provide validation of the spatial accuracy of the US system and that will allow comparison of anatomical boundaries derived via the two different imaging modalities. We describe the method by which the commercial US system is mechanically registered to a CT simulator and a unique data processing procedure. This data processing procedure circumvents the standard data acquisition and manual contouring sequence, thus reducing the time lapse from CT to US image acquisition to 10 minutes on average. Verification using a phantom demonstrated the method to be spatially accurate to within +/- 1 mm in the anterior-posterior (AP) and lateral directions and +/- 3 mm in the inferior-superior (IS) direction. Early clinical results gathered on 8 patients demonstrated alignment between the US and the CTDCs to be 0 mm in the AP and lateral directions and 2 mm in the IS direction, on average. The technique was used to compare the appearance of the prostate using US and CT imaging. The lateral dimension of the prostate indicated by the CTDCs was larger than that indicated by US imaging in all cases and on average by 0.9 cm. The height of the prostate in the AP direction was larger on average by 0.3 cm using CTDCs than US, and was larger by 5 mm or more in 3 out of 7 cases. The role of uncertainties in the determination of the CTDCs is examined as a possible cause and implications for treatment planning are described.