Uses of Intravascular Ultrasound
Characterisation of Atherosclerosis
IVUS can be used to measure plaque extent, morphology and distribution,5–7 and importantly provides information about plaque composition. This is because denser material such as calcium reflect more ultrasound waves, which results in a higher intensity image. Additionally, calcium does not allow any ultrasound waves to penetrate to deeper tissue, hence producing an acoustic shadow. On the other hand, lipid-laden lesions appear hypoechoic and fibromuscular lesions generate low-intensity or ‘soft’ echoes.3 Lipid-laden or fibromuscular lesions may exhibit a prominent echogenic fibrous cap, although most fibrous caps are too thin to be resolved by IVUS. IVUS therefore allows important decisions to be made with regard to intervention – for example, if calcium is identified, then rotational atherectomy could be considered.3
Apart from the greyscale image used for plaque interpretation, extensive research investigating ways of improving the assessment of plaque composition by IVUS has been performed. The Kawasaki group at the Gifu University Graduate School of Medicine in Japan has published studies using integrated backscatter signals from the radiofrequency signal of ultrasound, and based on the backscatter IVUS image, they have used colour to code different components of plaque.8,9 Another established technique developed by Volcano Corporation® (Rancho Cordova, CA, US) uses radiofrequency signals to determine plaque composition.10 In this technique, the distortion of radiofrequency signal by the plaque is passed through an algorithm, which is then colour-coded and superimposed on the grey image – a technique commercialised as ‘Virtual Histology intravascular ultrasound’ (VH-IVUS).10 Recent imaging technology now allows the reconstruction of VH-IVUS images in a longitudinal view, enabling a more comprehensive analysis of the total length of the plaque, its spatial orientation and its relation to the rest of the coronary artery. The potential of this imaging modality for analysing plaque vulnerability was demonstrated in a recent study where VH-IVUS backscatter data from ex vivo left anterior descending coronary arteries were recorded and compared with histological interpretation of the same sites.11 The overall predictive accuracies for VH-IVUS were 93.5 % for fibrotic tissue, 94.1 % for fibro-fatty tissue, 95.8 % for necrotic core and 96.7 % for dense calcium.11 Further data were provided by the Carotid Artery Plaque Virtual Histology Evaluation (CAPITAL) study,12 where a strong correlation between VH-IVUS plaque characterisation and characterisation following direct histological examination of the plaque (following endarterectomy) was demonstrated with a predictive accuracy of 99.4 % for thin-cap fibroatheroma (TCFA), which is thought to be the precursor lesion of plaque rupture, and 96.1 % for calcified TCFA.12
Vessel Dimensions
Although angiography allows measurement of luminal diameters in two-dimensional views, IVUS produces a tomographic view, which provides higher resolution as well as precise vessel and plaque dimensions.13 Therefore, the true minimal and maximal luminal diameter can be measured with IVUS. Furthermore, the cross-sectional area measurement of the lumen as well as the vessel can be obtained.13
In addition, IVUS has been useful in demonstrating diffuse disease in angiographically, ‘normal’ arteries, which may have as much as one-third of their cross-sectional area filled with diffuse plaque.14,15
Does Intravascular Ultrasound Use Improve Outcomes?
Identifying Vulnerable Plaque
The Providing Regional Observations to Study Predictors of Events in the Coronary Tree (PROSPECT) trial, used angiography, three-vessel greyscale and radiofrequency IVUS to evaluate the natural history of atherosclerosis in a prospective group of 697 patients with acute coronary syndromes who underwent percutaneous coronary intervention (PCI) and subsequent optimal medical therapy. During a median follow-up of 3.4 years, culprit lesions at the time of initial study were felt to be related to major adverse cardiac events (MACE) in 12.9 % of patients, with non-culprit lesions responsible in 11.6 %. Aftermultivariate analysis, non-culprit lesions associated with recurrent events were more likely to have three characteristics: a minimal luminal area of <4 mm2; a plaque burden of >70 %; or classified as TCFA. Furthermore, those lesions that were responsible for future MACE were observed to be mild when assessed by angiography (mean diameter of stenosis 32 ± 21 %), but using IVUS, these lesions had a plaque burden of 67 ± 10 %. At the time of follow-up, these lesions had progressed angiographically to a mean angiographic diameter stenosis of 65 ± 16 %.16 It is important to note, however, that while IVUS has been observed to be a validated tool to predict lesions responsible for future MACE, it is not able to image well through calcium, nor is it accurate in identifying thrombus.17
Assessment After Percutaneous Coronary Intervention
In a randomised trial studying drug-eluting stent (DES) deployments with or without IVUS guidance in 210 patients, IVUS use led to more frequent post-dilations, higher balloon inflation pressures and the use of larger balloon sizes. However, despite this there was no significant difference in MACE rates (11 versus 12 %; p = not significant) at 18-month follow-up.18 A further retrospective study found no significant differences in the rates of restenosis with and without optimal stent expansion guided by IVUS in 250 patients undergoing PCI with DES.19
Although currently insufficient evidence exists to support a reduction in the rates of restenosis with IVUS use there is some evidence supporting IVUS guidance to reduce rates of stent thrombosis. In one study of 884 patients with DES implantation, IVUS-guidance was associated with less direct stenting, more post-dilation, greater cutting balloon and rotational atherectomy use. At 30 days and 12 months, lower rates of definite stent thrombosis were seen in the IVUS group (0.5 versus 1.4 %; p=0.046 and 0.7 versus 2.0 %; p=0.014, respectively).20