John S. Harris, MD received his medical degree from Temple University School of Medicine Philadelphia, PA. He is currently a fellow in the Department of Cardiology, Temple University Hospital.

Brian O'Murchu, MD is an Assistant Professor of Medicine and the Associate Director of the Cardiac Catheterization Laboratory, Department of Cardiology, Temple University Hospital, Philadelphia, PA.

Although it has become the central modality for the diagnosis and management of coronary disease, coronary angiography has its limitations. Intracoronary injections provide information about the coronary artery lumen without assessing the entire artery or the physiological importance of a stenosis. Three more recently developed techniques address these limitations: the measurement of the coronary flow reserve, the measurement of the fractional flow reserve, and intravascular ultrasound.

Since its origins in the early 1960s, ^1,2 selective coronary angiography has remained the central imaging modality used by cardiologists in the diagnosis and management of coronary artery disease (CAD). Early randomized studies demonstrated the importance of angiographic data in the management of CAD. ³ The initial technology of percutaneous coronary catheterization developed into balloon angioplasty in the 1970s ⁴ and later to stenting ⁵ and atherectomy. ⁶ Current and emerging therapeutic technologies include brachytherapy, gene therapy, and the direct administration of antimitotic therapy by drug-eluting stents. The available technologies continue to have significant limitations, however, and management of CAD will improve as new technologies emerge.

One well-recognized limitation of contrast angiography is that it provides a two-dimensional visual representation of the three-dimensional vessel lumen. ⁷ Angiography frequently fails to provide information about either the vessel wall or the physiologic importance of a lesion. As revealed by a better understanding of coronary artery disease progression, angiography may underestimate the true degree of stenosis. In current diagnostic and interventional cardiology, newer techniques are being used to overcome these limitations, including intravascular flow measurements, pressure measurements, and ultrasound. This article will review these technologies and their current applications to the diagnosis and management of CAD.

Anatomic aspects

As an introduction to a discussion of the techniques used to improve the diagnostic accuracy of coronary angiography, a brief review of certain features of the pathophysiology of CAD progression is appropriate. First shown in the simian model, ⁸ and later in humans, ⁹ an artery that has developed plaque will demonstrate a change in dimension prior to any encroachment upon its lumen, a process called remodeling. Glagov et al ⁹ showed that before a stenosis reaches 40% of the "potential lumen area," the lumen remains essentially unchanged. A compensatory enlargement of the artery attenuates the degree of stenosis, and only after the lesion grows past 40% of the potential lumen area will the lumen begin to narrow (Figure 1).

With this concept of remodeling in mind, the main pitfall associated with conventional coronary angiography becomes apparent: the intracoronary injection of contrast into an artery provides information chiefly about the lumen, not about the entire vessel. Thus, an angiographically normal or slightly irregular lumen may conceal significant CAD. Also, because a "normal" appearing vessel segment adjacent to a stenosis is typically used as a reference to grade lesion severity and to guide interventional therapy, the operator will compare two diseased segments to each other.

Physiologic aspects

An understanding of the response of the microvasculature is needed to interpret data derived from the measurement of coronary flow and pressure in the setting of CAD. When an epicardial coronary lesion compromises oxygen delivery to the myocardium, the microvasculature responds by dilating. This dilation maintains basal flow at a level appropriate to meet the needs of myocardial oxygen demands during resting states. However, in the setting of increased oxygen demand or hyperemia, whether pharmacologically or exercise-induced, the ability of resistance vessels to dilate further is limited as compared with a healthy vascular bed. This translates into a decreased coronary flow velocity reserve ¹⁰ and fractional flow reserve, which are discussed in the following sections.

Coronary flow velocity reserve

The concept of directly measuring coronary flow with Doppler ultrasound is not a new one. The initial use of Doppler coronary flow assessment was during direct external examination of the epicardial arteries, an application limited to animal models ^11,12 and open-chest patients. Next, transesophageal echocardiography (TEE) Doppler technology was used to evaluate coronary artery flow and, while minimally invasive, it became apparent that this application was limited to the proximal portion of the artery. ^13,14 Later, catheter-based, closed-chest intravascular coronary Doppler measurement was developed. Although initial catheters were small enough to enter a coronary artery, they were large enough to impede flow and, thereby, distort the data obtained. Finally, a Doppler probe was attached to a standard angioplasty wire, and this is the method used today.

Assessment of flow in a coronary artery begins with the delivery of a Doppler device to the site of a stenosis. Typically, the patient is prepared in a fashion similar to that of angioplasty preparation. Anticoagulants are administered and a Doppler-tipped 0.014-in guidewire is advanced though a 6F to 8F coronary catheter and placed distal to the stenosis of interest. Next, the time velocity integral (TVI) of coronary flow is recorded both at rest and at maximal hyperemia, achieved with an intracoronary injection of a potent vasodilator, such as adenosine. The ratio of the measured maximal flow under hyperemic conditions to baseline flow defines the coronary flow velocity reserve (CFVR).

Studies in angiographically normal coronary arteries have determined that the normal range of CFVR for both men and women is >= 2.7 ± 0.6. ¹⁵ In patients with angiographically evident coronary artery disease, a poststenotic CFVR of >2.0 correlated well with negative myocardial perfusion studies. Miller et al ¹⁶ compared CFVR measurements with ^99m Tc-sestamibi imaging in symptomatic patients undergoing diagnostic cardiac catheterization. A CFVR ¾ 2.0 was found in association with at least one reversible defect in all 14 patients and results of CFVR and ^99m Tc-sestamibi imaging agreed in 24 of 27 patients (89%). Using a cutoff of CFVR ¾ 2.0, Joye et al ¹⁷ compared coronary flow measurements with single-photon emission computed tomographic (SPECT) thallium-201 imaging and found a sensitivity of 94%, a specificity of 95%, and overall predictive accuracy of 94%. Deychak et al ¹⁸ found a strong concordance (96%) between a reversible thallium defect and a CFVR <1.8.

The clinical application of CFVR measurement to the management of CAD has been investigated. Blinded to the results of noninvasive testing, Kern et al ¹⁹ studied 88 patients with anginal symptoms in whom angioplasty was deferred on the basis of a normal translesional hemodynamic evaluation. Compared with a cohort of patients who underwent angioplasty after the finding of an abnormal CFVR, a decision to defer intervention was found to be both safe and clinically effective. Similarly, the DEBATE study evaluated the use of CFVR after angioplasty as a predictor of late outcomes. ²⁰ A CFVR >2.5, in combination with a percent diameter stenosis (PDS) ¾ 35%, was predictive of a low recurrence of symptoms or ischemia at 1 and 6 months and a lower rate of restenosis and need for revascularization. DEBATE II, however, a study that investigated the cost and clinical role of using CFVR to guide the need for stenting after balloon angioplasty, did not support these clinical findings. ²¹ In this trial, patients were randomized to receive either primary stenting or balloon angioplasty. The latter group underwent stent placement only when the postprocedure CFVR was <2.5 and the PDS was >36%. After 1 year, there was no clinical difference between the two groups, but "provisional angioplasty" carried a higher economic cost.

Although easily measured, there are limitations to the clinical application of CFVR measurement. The finding of a truly normal CFVR requires both nonsignificant epicardial artery disease and a normal microvascular circulation. Microvascular disease--as found in left ventricular hypertrophy, chronic or acute ischemia, and diabetes--can confound the findings. ¹⁰ An abnormal CFVR cannot differentiate between epicardial and microvascular components. This limitation can be overcome by comparing CFVR in the artery of interest with that measured in a normal vessel; this is called relative coronary flow velocity reserve (rCFVR). In this way, the confounding effect of a diseased microcirculation is negated. Baumgart et al ²² reported that, compared with the CFVR, rCFVR has a strong relationship to lesion percent area stenosis. Additionally, the clinical usefulness of CFVR is limited when the target artery supplies infarcted myocardium, since the assumption of a normal microvasculature is no longer valid. In patients with multivessel disease, a normal reference artery may not be present, which precludes the assessment of rCFVR.

Fractional flow reserve

As an alternative to CFVR, a technique for direct catheter-based measurement of intracoronary pressure has been developed. With this technique, the mean intracoronary pressure distal to a stenosis is compared with the mean aortic pressure during conditions of maximal hyperemia. Fractional flow reserve (FFR) is then derived from these values and is defined as the ratio of maximal blood flow to the myocardium in the presence of a coronary artery stenosis to the theoretical flow that could be achieved without that stenosis. ²³

First described by Pijls et al, ²⁴ FFR differs from CFVR in that it is independent of changes in blood pressure, heart rate, and other conditions known to affect baseline myocardial flow. Normal FFR for every coronary artery is 1.0. This value, unlike CFVR, can be measured easily in multivessel disease without the need for a reference vessel. Several investigators have evaluated the diagnostic accuracy and clinical relevance of FFR and have compared it with noninvasive imaging including positron emission tomography (PET), ²⁵ exercise electrocardiography, ²⁶ nuclear imaging, and stress echocardiography. ²³ An FFR <0.75 correlates well with findings of myocardial ischemia.

In a study by Pijls et al, ²³ patients with an intermediate stenosis were assessed with both noninvasive imaging and FFR. Every patient with an FFR <0.75 had findings of ischemia on at least one noninvasive modality, and FFR reverted to normal after revascularization in all patients. Conversely, 21 of 24 patients with an FFR >0.75 had negative noninvasive stress testing. Sensitivity was 88%, specificity was 100%, positive predictive value was 88%, and negative predictive value was 93%.

Beyond assessment of an individual lesion, FFR has value in defining a culprit lesion in the setting of multivessel disease, as well as in pinpointing the point of severity in a diffusely diseased segment. By slowly withdrawing the pressure wire across a diffusely diseased segment, one can identify the point at which the FFR falls to <0.75; it may then be possible to direct an intervention accordingly. Likewise, by interrogating multiple diseased vessels, a culprit lesion may be identified in the setting of multivessel disease.

Several studies have looked at the safety of basing clinical decisions on the results of FFR calculations. Bech et al ²⁷ found that coronary artery bypass graft surgery could be deferred safely in patients with angiographic left main disease of 40% to 60% when FFR was >= 0.75. Survival and freedom from events during a 3-year follow-up were excellent. Likewise, basing therapy decisions on FFR data has been justified in single-vessel disease, ²⁸ while disregarding a positive FFR has proven to be unwise. Chamuleau et al ²⁹ showed that patients with an FFR <0.75--but with a negative SPECT--in whom angioplasty was deferred had a higher incidence of adverse outcomes as compared with those patients with FFR >= 0.75.

Fractional flow reserve also may be used to assess adequacy of stent deployment. Fearon et al, ³⁰ comparing intravascular ultrasound (IVUS) data to FFR measurements, found that an FFR <0.96 measured after stent deployment predicted a sub-optimal result by IVUS. In contrast, however, they found that an FFR >= 0.96 did not regularly predict a good outcome by IVUS. Bech et al ³¹ examined the clinical outcome of patients with a postangioplasty FFR >= 0.90 and reported an excellent prognosis after 2 years. Likewise, a small study from The Netherlands examining coil stents, found a correlation with an FFR of >0.94 and optimal stent deployment by IVUS. ³² Though the data conflict somewhat, it is clear that the closer the postrevascularization FFR is to 1.0, the better the stent deployment and clinical outcome. Thus, FFR can be a useful adjunctive tool in assessing the success and clinical efficacy of coronary intervention.

Collateral flow has long been known to contribute to myocardial blood flow, but the exact extent of this contribution has been difficult to quantify. Recently, FFR has been used to address this question. Defining recruitable collateral blood flow (Q _c ) as a fraction of normal maximal myocardial perfusion (Q ^N ), Pijls et al ³³ proposed the pressure-derived fractional collateral blood flow (Q _c /Q ^N ). This is defined as (P _w ­ P _v )/(P _a ­ P _v ) where coronary wedge pressure (P _w ), mean central venous (P _v ), and arterial (P _a ) pressures are measured simultaneously. Among 119 patients undergoing elective coronary angioplasty, a Q _c /Q ^N of ¾ 23% (inferring a lack of collateral flow) occurred in 82 of 90 patients experiencing ischemia during intracoronary balloon inflation, while a Q _c /Q ^N >24% was present in all of the 29 patients who did not experience ischemia. During follow-up, 15 of 16 patients with ischemic events fell into the group with poorer collateral circulation. Matsuo et al ³⁴ duplicated these findings in a recent publication.

As with CFVR, however, measurement of FFR is limited in the setting of microvascular disease. Small vessel disease may restrict the increase in blood flow that occurs during maximum hyperemia induced by adenosine. This translates into an attenuated decline of distal pressures and may cause an underestimation of lesion severity. ^24,35 Recent studies, however, suggest that this limitation may not bear greatly on the effect of FFR calculations, thereby allowing its application to a more general population. ³⁶

Intravascular ultrasound

Intravascular ultrasound, in use for well over two decades, is performed using a catheter that incorporates a miniaturized ultrasound transducer to directly image the vessel wall. Given the close proximity of the catheter to the vessel wall, current IVUS catheters emit high frequency ultrasound (30 to 40 mHz) to obtain an axial resolution of up to 100 µm and a lateral resolution of up to 250 µm. ³⁷

Ultrasonography of a normal artery may identify three layers: the intima, media, and adventitia. In diseased vessels the intima/media border may be difficult to separate due to the relatively high reflectivity of the intima, thus giving a two-layered appearance. ³⁸ The inner layer is defined by the vessel lumen internally and by the external elastic membrane externally (which separates the media from the adventitia); the outer layer represents the adventitia. Intravascular ultrasound in diseased vessels produces images of three general types. Low-density echo images characterize lipid-infiltrated lesions. Bright echo density lesions indicate dense fibrous disease, and bright lesions with acoustic shadowing are found with calcification. ³⁷ By outlining the borders of the endovascular lumen and the external elastic membrane, a ratio of the areas as measured by planimetry produces a percent area stenosis. Additionally, by measuring the maximal and minimal plaque thickness, the eccentricity of a plaque can be further described.

Intravascular ultrasound is exquisitely sensitive for the detection of coronary disease, ³⁹ and its measurements of disease tend to be more severe than estimates from angiography. Alfonso et al ⁴⁰ performed IVUS evaluations of 25 patients with angiographically normal coronary arteries and found that 80% had at least some degree of plaque burden. Likewise, Mintz et al ⁴¹ used IVUS in angiographically normal reference segments in 884 patients undergoing percutaneous transluminal coronary angioplasty (PTCA) for symptomatic coronary disease; on IVUS, only 6.8% of these vessel segments were found to be disease free. These studies underscore the failings of coronary angiography in the estimation of CAD severity and have revealed that segments that appear angiographically normal may be significantly diseased (Figure 2). Also, studies from Mintz et al ⁴² and Hoffman et al ⁴³ provided data that were pivotal in advancing the understanding of the mechanism of restenosis following either angioplasty or stenting.

As with FFR and CFVR, data derived from IVUS may contribute to the planning of optimal coronary intervention. Intravascular ultrasound is more sensitive in detecting calcium and postangioplasty dissections than standard angiography. ⁴⁴ Additionally, IVUS can be used to assess the degree and quality of stent expansion and deployment. As previously noted, IVUS performed after stent deployment has revealed that between 40% and 70% of stents that appear well-deployed angiographically are actually underexpanded in at least one segment as compared with the rest of the stent or as compared with the reference vessel. ³⁰ In as many as 90% of patients with subacute stent thrombosis, IVUS revealed an abnormality that could explain the event, compared with abnormalities defined angiographically, which were seen in only 25%. ⁴⁵ Thus, IVUS can help direct larger and higher pressure balloon inflations of a stent that would otherwise have remained underdeployed. In a randomized trial, Fitzgerald et al ⁴⁷ showed that IVUS-directed angioplasty leads to less restenosis and less repeat revascularization. Some other studies that have investigated this finding failed to reach a primary end point. Also, in the case of failed stent deployment, stent dislodgement, or stent embolism, IVUS can be an invaluable tool used to locate the stent and confirm its relationship to the guidewire, thereby facilitating decision making about salvage.

Recent uses of IVUS have also involved the differentiation of vulnerable versus stable plaque. Angioscopy studies have revealed that yellow plaques are associated more frequently with acute coronary events than white plaques. ⁴⁸ By quantifying the echodensity of the plaque, several investigators have attempted to predict the degree to which a plaque is vulnerable to rupture. However, some studies produced conflicting results. ^49,50 One suggested that plaques with a high distensibility and a compensatory enlargement, as diagnosed by IVUS, may be more vulnerable to rupture. ⁴⁹

Despite these advances, IVUS technology has its limitations. ⁵¹ It can only be used to assess one vessel at a time and can be used on only those vessels that are large enough to accommodate the IVUS catheter. Also, IVUS is limited by artifacts and shadowing.

Conclusion

Conventional angiography can both underestimate and overestimate the extent of CAD. Furthermore, it does not reveal the physiologic significance of a stenosis. Pressure and flow wires and IVUS are additional tools that contribute to the diagnosis, management, and our understanding of coronary atherosclerotic disease.