Improved localization of coronary stents using layer decomposition.

Abstract

Accurate placement and expansion of coronary stents is hindered by the fact that most stents are only slightly radiopaque and hence difficult to see in typical coronary X-ray images. We propose a new technique for improved image guidance of multiple coronary stent deployment using layer decomposition of coronary X-ray image sequences. We hypothesize that layer decomposition can improve the accuracy of localization of the end of a deployed stent. Layer decomposition is used to obtain good quality images of a stent in vitro. The resultant background-subtracted stent images are embedded into other cine X-ray image sequences to form a database of simulated image sequences. For each simulated sequence, the position of the stent edge is estimated from raw and layer-decomposed images using a small region of the original layer image as a template. Layer decomposition reduced median position errors in 33 of 47 image sequences (70%), including 16 of 18 sequences in which the position errors for raw and layer images differed by 5.0 pixels (0.5 mm) or more. Layer decomposition significantly reduces errors in determination of stent edge location in simulated cine X-ray image sequences.