Computer-Aided Detection of Intracoronary Stent Location

1 Introduction

An intraluminal coronary stent is a metal mesh tube deployed in a stenotic artery during percutaneous coronary intervention (PCI), in order to prevent acute vessel occlusion, to scaffold the arterial wall after balloon angioplasty, and to restore the blood flow. After X-ray-guided stent placement, cases of underexpansion (stent correctly apposed to luminal wall but not completely expanded) or malapposition (stent not completely in contact with the luminal wall) may happen. The incorrect stent deployment due to underexpansion and malapposition may lead to restenosis.¹ These clinical cases are recognized as important risk factors and might lead to stent failure (in-stent restenosis or stent thrombosis) in the follow-up.² The identification of the strut location and the definition of the stent shape, compared with the luminal border and the vessel border, allow physicians to assess stent placement in the vessel and the need for a further balloon postdilatation.

Moreover, the assessment of the stent location and extension along the vessel axis is relevant for PCI planning, implantation, and patient follow-up. In order to have an effective deployment, a stent should be optimally placed with regard to anatomical structures such as bifurcations and stenoses. The deployment of a stent in an incorrect location may lead to restenosis¹ or bifurcation side branch occlusion as shown by Ref. 3.

Although the current reference image modality for verifying the correct positioning of a stent is intravascular optical coherence tomography (OCT), a potential alternative is intravascular ultrasound (IVUS). IVUS is a catheter-based imaging technique that provides the sequence of tomographic images (pullback) of the internal vessel morphology (see Fig. 1A). The stent placement can be deduced by the position of its struts (see Fig. 1B). The main advantage of OCT over IVUS is that the resolution is 10-fold higher, easing the visualization of the stent for the physician. However, OCT remains inferior to IVUS in matters of penetration depth (1.5 vs. 5 mm),⁴ which limits its ability to assess plaque burden, bifurcation angles,⁵ and vessel remodeling. Moreover, the guiding OCT catheter requires the intubation of the coronary ostium in order to effectively provide contrast injection, limiting the analysis of ostial and occlusive lesions. Consequently, IVUS is regarded as either complementary or preferable to OCT in complex lesions.^6–8

Sign in to download full-size image

Fig. 1. Example of IVUS pullback containing a stent (A). In (B) and (C), two examples of IVUS frames in Cartesian coordinates, and belonging to the pullback (A), are depicted. The frame (B) corresponds to the green vertical line in (A), and does not contain a stent, while (C) corresponds to the blue vertical line in (A), and contains a stent. The green circles and the green dotted line in (C) and (D) represent the estimated position of struts and stent shape obtained using the technique of Ref. 9. The classification map (D) of the IVUS frame (C) is used to guide the struts and stent estimation. The legend of figure (D) indicates the tissue type identified in the image.

In this chapter, two complementary applications of the stent assessment in IVUS pullbacks are presented. The first technique, presented in Section 2.2, focuses on strut detection and stent shape estimation in a single IVUS image. The second method, presented in Section 3.2, uses the technique described on the struts and stent estimation to assess the position of the stent in the pullback.