Article Text

Download PDFPDF

Peripheral arterial disease: the evolving role of non-invasive imaging
  1. A R Owen1,
  2. G H Roditi2
  1. 1Barwon Medical Imaging, Geelong, Victoria, Australia
  2. 2Glasgow Royal Infirmary, Glasgow, UK
  1. Correspondence to Dr Andrew Owen, Barwon Medical Imaging, Geelong Hospital, Ryrie Street, Geelong, Victoria 3226, Australia; andrewowen{at}


Peripheral arterial disease is usually secondary to stenotic or occlusive atherosclerosis and is both common and increasing in western society. The majority of symptomatic patients have intermittent claudication and only a minority (<2% and typically those with diabetes mellitus or renal failure) progress to critical limb ischaemia, heralded by the onset of rest pain and/or tissue loss. Imaging is largely reserved for patients with disabling symptoms in whom revascularisation is planned. In these patients, accurate depiction of the vascular anatomy is critical for clinical decision making as the distribution and severity of disease are key factors determining whether revascularisation should be by endovascular techniques or open surgery. Driven by advances in technology, non-invasive vascular imaging has recently undergone significant refinement and has replaced conventional digital subtraction angiography for many clinical indications. In this review, the relative merits and limitations of duplex ultrasound, CT angiography, and magnetic resonance angiography are discussed, emerging imaging techniques are described, and complications relating to the use of intravascular contrast agents are highlighted.

  • Cardiovascular imaging
  • interventional radiology
  • vascular medicine
  • vascular surgery
View Full Text

Statistics from


Peripheral arterial disease is usually secondary to stenotic or occlusive atherosclerosis and is both common and increasing in western society. The prevalence of symptomatic disease is approximately 5% at the age of 60 years and increases with age.1 The majority of symptomatic patients have intermittent claudication and only a minority (<2% and typically those with diabetes mellitus or renal failure) progress to critical limb ischaemia, heralded by the onset of rest pain and/or tissue loss. The diagnosis of peripheral arterial disease has important prognostic implications; among these patients there is a 5–7% annual rate of major cardiovascular events.2

If a patient is suspected of having peripheral arterial disease, measurement of the ankle: brachial pressure index (ABPI) is a mandatory part of the clinical examination.3 An ABPI <0.9 demonstrates 95% sensitivity for the presence of angiographically proven peripheral arterial disease and a normal result is 100% specific for the absence of disease. Imaging is reserved for those patients in whom revascularisation is planned or when the diagnosis is in doubt.

In such patients, accurate depiction of the vascular anatomy is critical for clinical decision making as the distribution and severity of disease are key factors in determining the optimal revascularisation strategy. In the UK, invasive treatment is reserved for those with critical limb ischaemia or disabling claudication who have failed conservative management. The choice of treatment is guided by the Transatlantic Intersociety Consensus (TASC) recommendations. The TASC classification places peripheral arterial lesions into one of four categories (A–D) based on their length and anatomical location. The reader is referred to the TASC document for a thorough review of this topic.4 In brief, aortoiliac and femoropopliteal disease that is short segment and does not involve the internal iliac, common femoral or infrageniculate popliteal artery is considered best treated endovascularly, whereas long segment disease is best treated with surgical bypass. Imaging findings therefore dictate intervention.

Digital subtraction angiography (DSA) is the historical reference imaging standard on the basis of its superior spatial and temporal resolution, and is used as the arbiter when non-invasive imaging is inconclusive and in clinical trials. This position is, however, increasingly contentious. Arterial stenosis on DSA, particularly when eccentric, shows poor correlation with the reduction in luminal cross sectional area that is best demonstrated using three dimensional techniques and which is the determinant of haemodynamic effect.5 6 DSA is also invasive, with attendant risks of haemorrhage, vessel injury, cholesterol embolism syndrome, and contrast induced nephropathy.7 It is expensive and both labour and time intensive. Non-invasive imaging of the peripheral arteries is evolving rapidly and DSA, as a routine diagnostic technique, looks set to become obsolete.8 Duplex ultrasound (DU), CT angiography (CTA), and magnetic resonance angiography (MRA) have demonstrated significant improvements in diagnostic accuracy that have paralleled technological advances. In this review, these modalities are discussed, including their relative merits and limitations. Evolving techniques are described and there is focus on contrast-induced nephropathy and nephrogenic systemic fibrosis (table 1).

Table 1

Key differences between the non-invasive vascular imaging modalities

Duplex ultrasound

DU (the term duplex refers to the combination of B mode and colour Doppler ultrasound) is typically the initial imaging modality of choice.8 DU is widely available and requires no exposure to either ionising radiation or potentially toxic intravascular contrast agent. It is highly portable and in the emergency setting permits a rapid targeted examination, potentially even on the ward or in the operating theatre. It can also provide detailed haemodynamic information; an increase in the peak systolic velocity ratio >2 across a stenosis indicates a reduction in cross-sectional calibre of >50% (figure 1). One specific advantage of DU is that once a bypass target site has been identified, it provides the opportunity to mark this position on the skin, thereby limiting the size of the surgical incision.9

Figure 1

A duplex ultrasound (A and B) in a patient with short distance claudication demonstrates a greater than 10-fold increase in peak systolic velocity (from 39.1 cm/s in panel A to 445.6 cm/s in panel B) in the left popliteal artery. The subsequent digital subtraction angiography (C) confirms the presence of a focal tight stenosis in the popliteal artery (arrow) just below the level of the knee joint.

DU is, however, both operator and patient dependent and diagnostic accuracy may be limited in certain vascular territories (eg, assessment of the iliac arteries may be compromised by tortuous arterial anatomy or overlying bowel gas). The sensitivity (88%) and specificity (95%) of DU are documented to be lower than both CTA and MRA for the detection of haemodynamically significant lesions (ie, >50% stenosis or occlusion).10

DU is considered relatively inexpensive, but this has been recently challenged by the Netherlands based Diagnostic Imaging of Peripheral Arterial Disease (DIPAD) trial. In this study, 514 patients were randomised to DU or one of either MRA (1.5 T) or CTA (16 detectors). Referring clinicians were asked to score their clinical confidence in making a treatment plan on the images provided, and were free to request additional imaging as required. A significantly higher clinical confidence and less additional imaging were reported for MRA and CTA compared with DU, and total costs (which comprised both initial and supplementary imaging costs) were significantly higher for MRA and DU than for CTA. Interestingly, there was no significant difference in functional outcome or quality of life between the three groups. High DU costs were explained by the need for additional imaging in 43% of patients (in comparison, additional imaging was required in 23% of patients who had MRA and 15% of patients who had CTA). This result may simply reflect the fact that vascular surgeons have increased confidence in images that permit an angiographic display analogous to DSA, but it suggests that the ‘true’ cost of DU may be higher than expected.11 Nevertheless, DU remains the mainstay of peripheral arterial imaging in most centres and, when combined with accurate clinical findings and non-invasive assessment, allows a treatment plan to be formulated in the majority of patients.8 12

CT angiography

The rapid evolution of multidetector CT (MDCT) technology has driven advances in CTA. CT scanners with 64 detector rows (64 MDCT) are now in widespread clinical use and enable the entire peripheral vasculature to be rapidly examined with high diagnostic accuracy.13 This technique not only displays the vessel lumen but provides important information on the vessel wall (figure 2) and the nature of atheromatous plaque.

Figure 2

CT angiography of the lower leg arteries in a male patient with a prior right below knee amputation. The maximal intensity projection image (A) produces a angiogram analogous to digital subtraction angiography, but the left popliteal artery aneurysm (arrows) is much better appreciated on the axial (B) and curved planar reformatted images (C).

High spatial resolution and optimal contrast enhancement are prerequisites for diagnostic angiography. The velocity of a contrast bolus varies significantly between patients with peripheral arterial disease. In a study by Fleischmann et al, the range of contrast velocities was 29–177 mm/s and did not clearly relate to disease severity.14 Contrast travel times are largely determined by cardiac output with the effect of peripheral vascular disease secondary. If the speed of the patient through the scanner (ie, the speed of image acquisition) does not match the speed of the contrast bolus through the peripheral vasculature, diagnostic inaccuracy will ensue. If the imaging acquisition is significantly faster than the speed of the contrast bolus, patent vessels that are not opacified with contrast will be imaged and considered occluded (this is termed ‘outrunning the bolus’). In the converse situation, venous enhancement (‘contamination’) will result and may impair arterial assessment.

In most clinical settings the CTA protocol is standardised such that image acquisition is triggered a set time after the bolus reaches the abdominal aorta. A further delayed image acquisition can be built into the protocol and initiated by the imaging technologist if there is asymmetrical disease or if there is concern that the imaging acquisition has ‘outrun’ the contrast bolus. An alternative technique for obtaining optimal contrast enhancement is the adaptive acquisition method. Using this technique, the transit time between the suprarenal abdominal aorta and the midpoint of the popliteal arteries is calculated using small serial bolus injections. The table speed and gantry rotation time can then be altered to match the speed of the bolus, thereby tailoring the examination to the patient's haemodynamics. This technique is more cumbersome, but is reported to provide high quality imaging (particularly of the pedal vessels) that is free of venous contamination.15 The contrast injection should be immediately followed by a saline bolus which not only flushes contrast out of the arm veins, thereby maintaining a tight contrast bolus, but also contributes to patient hydration to ameliorate the risk of contrast induced nephropathy.16

CTA generates large amounts of data and post-processing the volumetric information into a clinically accessible form is challenging and time consuming (typically 30–45 min).17 Volume rendered, maximal intensity projection and curved planar reformatted images allow rapid and comprehensive review. Despite the delay inherent in post-processing, the ready availability of CTA permits its use in the acute setting (figure 3). Volume rendered images allow clear appreciation of the spatial relationships between artery, soft tissue and adjacent bony landmarks which aids surgical and endovascular planning (figure 4). It should be remembered that post-processed images are susceptible to artefact, particularly during attempts to remove calcification and bone. In the latter case, an adjacent opacified vessel may be automatically subtracted with the bone giving the impression of vessel occlusion.18 Experimental studies stress the importance of careful review of the transverse images to maximise diagnostic accuracy (figure 5).5 19

Figure 3

CT angiography (A) and corresponding digital subtraction angiography (B) demonstrate acute embolic occlusion of the left common femoral artery (arrow).

Figure 4

Volume rendered (A) and curved planar reformatted (B) images of bilateral internal iliac artery aneurysms (arrows). The curved planar reformatted images not only demonstrate the size and morphology of the left iliac artery aneurysm but also its precise relationship to the common iliac artery bifurcation and its inflow and outflow vessels. This information is critical for endovascular treatment planning.

Figure 5

A volume rendered image of the left tibial vessels (A) demonstrates an apparent focal occlusion of the posterior tibial artery (arrow) that resulted from an error in post-processing. Review of the axial images and the subsequent digital subtraction angiography (B) confirmed vessel patency.

A recent meta-analysis of the diagnostic performance of CTA versus DSA concluded that CTA is highly accurate (overall sensitivity 95–97%, specificity 91–98%) in the assessment of significant stenosis or occlusion.20 These data included trials of CTA versus DSA across three different generations of CT technology (ie, scanners with 4, 16, and 64 detector row) and with technological advancement there has been a corresponding improvement in diagnostic accuracy (sensitivity and specificity has increased from 75–99% and 83–99% with 4 detector rows to 98–99% and 96–99% for 64 detector rows). Diagnostic accuracy is lower for the smaller distal arteries compared with the larger proximal vessels, but the diagnostic performance below the knee remains good (sensitivity 85–99%, specificity 79–97%). Interobserver agreement is good to excellent (κ values >0.8) in most studies.

The great majority of patients in trials of peripheral arterial imaging are claudicants and there are limited data in patients with critical limb ischaemia. In a recent study of 28 patients with critical limb ischaemia who were evaluated with 16 detector row CTA, 23 had treatment plans confidently formulated on the basis of the CTA alone, and in the five additional DSAs that were requested there were no supplementary findings that altered management.21 This research group reported similar clinical utility of 16 detector row CTA among patients with intermittent caludication.22

New developments in CTA

Despite the impressive diagnostic performance with current multidetector CTA, vascular calcification remains a significant problem. High density calcium attenuates the x ray beam and results in ‘blooming’ or loss of image contrast in the lumen which may lead to overestimation of stenosis.17 Calcium can be post-processed from the image by automated or semi-automated techniques, but these are currently imperfect. Dual energy CT (DECT) offers a potential solution. In this technique, two x ray sources that emit photons of different energy (typically 80 and 140 kV) are mounted on the same scanner gantry and separated by 90°. The technique may also be performed using a single source and dual layer detectors with different sensitivities. Interpolation of the generated data produces greater soft tissue differentiation which is most pronounced between materials of high atomic number (ie, between iodinated contrast, calcified atheromatous plaque, and bone) Therefore, those voxels containing calcium (ie, calcified plaque and bone) can be rapidly subtracted from the image, thereby shortening post-processing time and creating a ‘luminogram’ analogous to DSA.

Early data on peripheral DECT supports claims of faster bone removal (although total post-processing time is unaltered) and suggest that plaque subtraction is not yet entirely reliable in the below knee arteries.23 The technique will undoubtedly evolve and further studies are awaited with interest.

In-stent restenosis

In-stent restenosis is the major limitation of peripheral vascular stenting and increases over time.24 In the presence of stents, CTA suffers from beam hardening artefact and contrast enhanced MRA suffers from susceptibility artefact. In vitro this has been shown to manifest itself as both under- and overestimation of stenosis on CTA whereas it will result in overestimation of stenosis on MRA (figure 6).25 There are limited data on the diagnostic accuracy of CTA in quantifying in-stent restenosis in the lower limbs. In a recent report of 81 stents evaluated with 64 detector row CTA and using DSA as the reference standard, Li et al reported an overall sensitivity of 85% for the detection of in-stent restenosis. However, in this study 23.5% of stents were deemed to be ‘unassessable’, principally due to metallic artefact.26 Extrapolating from the coronary literature, CTA should have good performance if performed at sub-millimetre resolution at least for larger (iliac and femoral) stents, although images generated with a high spatial frequency reconstruction from the raw data should be used.

Figure 6

Contrast enhanced magnetic resonance angiography of the aortoiliac arteries demonstrates artefactual bilateral external iliac artery occlusions resulting from the presence of stainless steel stents. Note the complete signal voids corresponding to the stents, the sharp demarcation, and the ‘non-pathophysiological’ appearances.

Radiation dose

The average radiation dose reported in the CTA literature is 7.47 mSv,13 although average doses as high as 13.7 mSv have been reported in some series.27 In a trial of 16 detector row CTA versus DSA, Willmann et al reported a fourfold higher radiation dose for DSA compared with CTA.28 To place these doses in context, the average annual background radiation exposure is 3 mSv.20 It has been suggested that patient radiation dose issues are of limited concern in patients with advanced peripheral vascular disease, as their life expectancy is significantly less than the latent period of a radiation induced malignancy.29

Contrast induced nephropathy

Contrast induced nephropathy (CIN) is defined as an increase in baseline creatinine of >25% or 44 μmol/l that occurs within 3 days of the intravascular administration of contrast in the absence of an alternative aetiology.30 The most important risk factor for the development of CIN is the presence and severity of pre-existing renal impairment, particularly in the presence of diabetes.30 31 The risk of CIN is dose dependent32 and is higher when contrast is administered intra-arterially than when given intravenously.30 In a recent study of patients with renal impairment (glomerular filtration rate (GFR) <60 ml/min) undergoing MDCT, the incidence of CIN was 0.6% in those with a GFR >40 ml/min and 7.8% in those with a GFR<30 ml/min. The clinical implications for the development of CIN are not fully understood. Only a minority go on to require renal replacement therapy (<1%), but in a retrospective review of over 16 000 inpatients exposed to contrast media, in-hospital mortality rates were fivefold higher (34% vs 7%) among patients who developed CIN, even after adjustment for comorbidity.33 A full discussion of CIN is beyond the scope of this article and interested readers are directed to the European Society of Urogenital Radiology for more information (

Contrast enhanced MRA

In contrast to CTA, in which the peripheral vasculature is imaged in one continuous uninterrupted passage of the patient through the scanner, contrast enhanced MRA (CEMRA) is currently produced from a series of overlapping body segments or ‘stations’. The peripheral arteries are typically divided into aortoiliac, femoropopliteal, and calf (or below knee) stations and each is imaged separately during first passage of an intravenously administered gadolinium based contrast agent (GBCA).

The typical delay between contrast arrival in the common femoral artery and the ankle is reported to be approximately 12 s.34 Even with optimised hardware, software and technique, the latest CEMRA acquisition times are typically 10–15 s per station and are separated by table movement times of 5 s per station.35 If the distal vessels are the final station in a bolus-chase protocol the result is a progressive time lag of imaging acquisition behind the contrast bolus which may result in venous contamination and reduced diagnostic accuracy.36–38 This is usually not a problem in claudicants where the arteriovenous transit times are long and there is significant volume of dilution in the venous compartment. However, it may well be an issue in patients with critical ischaemia where arteriovenous shunting is pronounced. Using a single bolus-chase technique, Wang et al reported diagnostic accuracy of 96–100% in the pelvis and thigh, but only 43% in the calf.39 To overcome this problem, ‘hybrid’ MRA techniques have evolved, whereby the calf (± foot) station is imaged first using maximal spatial resolution. Real-time tracking of the contrast bolus allows accurate image acquisition in the arterial phase, thereby minimising venous contamination and increasing diagnostic accuracy.35 36 38 The aortoiliac and femoropopliteal stations are subsequently imaged consecutively following a second contrast injection.

Recent developments in the way that the MR signal is detected and stored for display (eg, parallel imaging techniques such as sensitivity encoding and alterations in k-space ordering) have allowed faster imaging without loss of spatial resolution or coverage. Sensitivity encoding uses the spatial information in multiple receiver coils to allow a reduction in the number of phase encoding steps, hence shortening imaging time for a given spatial resolution and therefore reducing movement artefact. Alternatively, for a given imaging time, sensitivity encoding can be used to improve spatial resolution. Although parallel imaging tends to reduce signal, this is seldom an issue in CEMRA where contrast is generated from exogenously administered agents that produce profound T1 shortening. When incorporated into a bolus-chase protocol, sensitivity encoded CEMRA has demonstrated increased diagnostic accuracy compared with standard CEMRA and depicts more patent below knee segments.40

Further refinement is the use of time resolved imaging for acquisition whereby with the use of view sharing/keyhole techniques and in combination with parallel acceleration, very rapid dynamic three dimensional CEMRA can be acquired. This not only gives dynamic flow information akin to DSA (although with three dimensional datasets) but also obviates timing issues since data are acquired throughout the bolus transit in the region of interest. This is typically applied to the below knee arteries where it is combined with a standard stepping table acquisition for the aortoiliac and femoropopliteal stations, but it can be applied at all stations (figures 7 and 8).

Figure 7

Contrast enhanced magnetic resonance angiography (CEMRA) in a patient with bilateral thigh and calf claudication, worse on the left. A montage of selected images (A) from time resolved three dimensional CEMRA of the below knee arteries (providing dynamic information analogous to digital subtraction angiography) shows normal arteries bilaterally. (B) Overview three station CEMRA showing severe stenotic right iliac disease and complete left common and external iliac occlusions but preserved infra-inguinal run-off.

Figure 8

Contrast enhanced magnetic resonance angiography (CEMRA) in a patient with bilateral critical lower limb ischaemia. A montage of selected images from time resolved three dimensional CEMRA of the below knee arteries (A) shows diseased right popliteal artery with single vessel peroneal run-off and critically stenosed left popliteal artery with essentially single vessel peroneal run-off. (B) Overview three station CEMRA shows preserved aortoiliac arteries but severe stenotic femoropopliteal disease and essentially single vessel peroneal run-off bilaterally.

Contrast agent choice for CEMRA has been expanded with the advent of the specific blood-pool contrast agent gadofosveset trisodium. This high relaxivity agent allows CEMRA at low dose and with its extended blood pool residence time permits the acquisition of ultra-high resolution (eg, 125 μm voxel volume) steady state imaging.

The latest trials of CEMRA report sensitivity and specificity of >90% for the detection of significant stenosis or occlusion (figure 1).35 41 Clinical confidence in the technique is high and increases with experience.42 CEMRA is proven to identify vessels that are not demonstrated by DSA, resulting in alterations in clinical decision making.43–47 As has been the experience with CTA, the majority of CEMRA data have been obtained in patients with claudication so its accuracy in patients with critical lower limb ischaemia is less clearly established. However, in a study of 30 patients with critical limb ischaemia who underwent CEMRA with high resolution imaging of the calf vessels, diagnostic accuracy was similar to DSA.48

Nephrogenic systemic fibrosis

In 2006 reports were published linking the use of GBCA with the hitherto little known condition, nephrogenic systemic fibrosis (NSF), a scleromyxoedema-like skin thickening in patients on dialysis.49 In a minority of patients the disease is relentlessly progressive and can be fatal. As yet there are no effective treatments, although in some cases regression has followed improvement in renal function, particularly after transplantation.50

The link to GBCA was first noted in a cohort of patients with NSF that had all recently undergone CEMRA studies with gadolinium DTPA-BMA (gadodiamide).51 Given that most patients in renal failure do not develop NSF, it is suspected that the aetiology may be multifactorial. However, as yet no other potential causative factor (eg, erythropoietin administration, proinflammatory events, etc) has been conclusively linked to the development of NSF.52 53

All of the GBCAs are chelates of ionic gadolinium (free Gd3+ is highly toxic) and their stability appears to be important in the pathogenesis of NSF. Chelates of low stability release free Gd3+ that bind with endogenous anions to form insoluble salts which deposit in tissues and may initiate NSF.54 The great majority of reported cases have been with gadodiamide (Omniscan, GE Healthcare, USA), with significantly fewer cases reported with gadopentate dimeglumine (Magnevist, Bayer Schering Pharma AG, Germany) and gadoversetamide (Optimark, Covidien, USA). These agents are all linear chelates which bind gadolinium less strongly than the cyclic chelates. There have been no confirmed reports of NSF following the sole administration of gadobenate dimeglumine (Multihance, Bracco, Italy—one of the linear chelates with higher stability indices and some hepatobiliary excretion) or following administration of any of the cyclic chelates: gadoteridol (Prohance, Bracco), gadoterate meglumine (Dotarem, Guerbet, France) or gadobutrol (Gadovist, Guerbet). The GBCA doses associated with NSF have mainly been in the 0.2–0.3 mmol/kg range (ie, double and triple 'standard' dose) as had often been used for CEMRA. There are few reports of NSF following single/standard dose administration.52 In practice the implications of the association between GBCA and NSF are that the patients estimated GFR (eGFR) must be known before CEMRA. Patients with an eGFR <30 ml/min should not receive the linear chelates Omniscan, Optimark or Magnevist, and alternative imaging techniques including non-contrast enhanced MRA should be considered. If a GBCA is to be used then the lowest dose feasible is advocated (eg, half usual dose) and therefore GBCAs with increased specific relaxivity are advantageous. When the scan is planned then close liaison with the patient's renal physician is important, ideally with dialysis pre- and immediately post-procedure where practicable. Elimination of GBCA is slower with peritoneal compared with haemodialysis and these patients are therefore at increased risk.52

Unenhanced MRA techniques such as ‘time-of-flight’ (TOF) were introduced in the 1980s, but were largely superceded by CEMRA in the peripheral vasculature due to faster image acquisition and improved diagnostic accuracy.55 The causal link between GBCAs and NSF combined with hardware and software developments has led to the re-emergence of unenhanced MRA techniques. Even in the era of CEMRA, two dimensional TOF has continued to be used and has accuracy comparable to DSA below the knee.44 56 Alternative sequences such as ECG gated, partial Fourier, fast spin echo acquisition57 are emerging but are yet to be fully validated. The longer acquisition times of non-contrast techniques render them susceptible to movement artefact, which is a particular problem in patients with rest pain who find it difficult to both lie flat and remain still.


Non-invasive peripheral arterial imaging has rapidly evolved and has been widely disseminated, largely replacing DSA as a diagnostic test. In skilled hands, DU is robust and, given its availability and excellent safety profile, should remain the initial imaging modality for most clinical applications. If DU fails to answer the clinical question, CTA and MRA both provide comprehensive and accurate assessment. It is currently unclear if the superior diagnostic accuracy of CTA and CEMRA over DU results in meaningful differences in patient outcomes and this should be a focus for future research.

The faster acquisition of CTA is a potential advantage over MRA, but CTA requires the use of potentially nephrotoxic contrast agents and ionising radiation. CTA may also suffer from post-processing difficulties if the arteries are heavily calcified.

MRA may be precluded by the presence of metallic implants, pacemakers, etc, and analysis of in-stent stenosis is impaired by susceptibility artefact. In general, the risks of GBCAs appear to be much smaller than those of iodinated contrast. Unenhanced MRA remains an alternative in patients with severe renal failure (ie, eGFR <30 ml/min/1.73 m2). In contrast to current CTA technology, CEMRA is the only modality that allows assessment of flow dynamics, analogous to DSA. Advances in coil technology, higher field strengths, faster gradient systems, and improved contrast agents will lead to faster image acquisition, improved temporal resolution and will allow lower contrast doses.

While the authors believe that the choice between CTA and MRA should ideally be governed by the clinical scenario, it is likely that it will be strongly influenced by local factors (eg, equipment availability, local expertise and cost) and by national guidance.

Multiple choice questions (true (T)/false (F); answers after the references)

1. Regarding peripheral arterial disease (PAD):

  1. The prevalence of symptomatic disease is approximately 15% at the age of 60 years

  2. The typical symptomatic patient progresses from intermittent claudication to critical limb ischaemia in a stepwise fashion

  3. Among patients with PAD there is an 1–2% annual rate of major cardiovascular events

  4. Digital subtraction angiography (DSA) remains a central part of the work up of the patient with PAD

  5. An ankle: brachial pressure index (ABPI) below 0.9 demonstrates 70% sensitivity for the presence of angiographically proven PAD

2. Regarding duplex ultrasound (DU):

  1. Duplex refers to the combination of M mode and colour Doppler ultrasound

  2. For the detection of significant stenosis or occlusion, DU has lower sensitivity than both CT angiography (CTA) and magnetic resonance angiography (MRA)

  3. In the work-up of the patient with PAD a ‘DU first’ approach is the most cost effective

  4. Clinicians have more confidence in CTA than DU

  5. A doubling of the peak systolic velocity at the level of an arterial stenosis equates to a 50% narrowing

3. Regarding CTA:

  1. The move from 16 to 64 slice scanners has seen corresponding increases in diagnostic accuracy

  2. B Following intravenous contrast injection, the velocity of the contrast bolus is inversely proportional to disease severity

  3. If the imaging acquisition is significantly faster than the contrast bolus, venous contamination will result

  4. Vessel wall calcification leads to overestimation of stenosis

  5. The average radiation dose of CTA is approximately 20-fold higher than the average annual background radiation exposure

4. Regarding contrast enhanced MRA (CEMRA):

  1. CEMRA is proven to identify vessels that are not demonstrated by DSA, resulting in alterations in clinical decision making

  2. CEMRA tends to underestimate stenosis

  3. In patients with critical limb ischaemia, significant arteriovenous shunting may lead to venous contamination and impair diagnostic accuracy

  4. Parallel imaging techniques tends to reduce signal but this is seldom an issue in CEMRA where contrast is generated from exogenously administered agents that produce profound T1 shortening

  5. The latest trials of CEMRA report sensitivity and specificity of >90% for the detection of significant stenosis or occlusion

5. Regarding contrast agents:

  1. Contrast induced nephropathy (CIN) is defined as an increase in baseline creatinine of >25% or 44 μmol/l that occurs within 3 days of the intravascular administration of contrast in the absence of an alternative aetiology

  2. The most important risk factor for the development of CIN is diabetes mellitus

  3. Gadolinium based contrast agents administered to patients with severe renal impairment may produce nephrogenic systemic fibrosis. This is a scleromyxoedema-like skin thickening that may be associated reduced joint mobility and pulmonary, hepatic and cardiac fibrosis

  4. Nephrogenic systemic fibrosis (NSF) characteristically spares the head and neck

  5. NSF is irreversible and progressive

Main messages

  • Non-invasive peripheral vascular imaging has undergone significant refinement such that conventional digital subtraction angiography (DSA) is now largely obsolete as a diagnostic tool.

  • A normal ankle: brachial pressure index (ABPI) virtually excludes the presence of significant peripheral vascular disease and should be a mandatory part of vascular assessment.

  • In patients in whom revascularisation is planned, the choice between endovascular techniques and open surgery is strongly influenced by the severity and distribution of arterial disease and accurate imaging is therefore critical to clinical decision making.

  • Duplex ultrasound (DU) is robust and is the preferred initial imaging modality, although it is both operator and patient dependent and has documented lower accuracy than both CT angiography (CTA) and magnetic resonance angiography (MRA).

  • CTA is rapidly acquired, but patients are exposed to the risks of ionising radiation and potentially nephrotoxic contrast agents. Heavy vascular calcification impairs diagnostic accuracy.

  • MRA is slower to acquire than CTA and luminal assessment within stents may be severely limited. MRA does not result in exposure to ionising radiation. The risks of gadolinium based contrast agents appear to be significantly smaller than those of iodinated contrast agents.

  • Diagnostic accuracy continues to improve in parallel with technological advances. Dual energy CTA is a potential solution to the problem of vascular calcification. Improved hardware, software, and contrast agents hold promise for MRA.

Key references

▶ Norgren L, Hiatt WR, Dormandy JA, et al. Inter-Society Consensus for the Management of Peripheral Arterial Disease (TASC II). Eur J Vasc Endovasc Surg 2007;33 (Suppl 1):S1–75.

▶ Met R, Bipat S, Legemate DA, et al. Diagnostic performance of computed tomography angiography in peripheral arterial disease: a systematic review and meta-analysis. JAMA 2009;301:415–24.

▶ Pereles FS, Collins JD, Carr JC, et al. Accuracy of stepping-table lower extremity MR angiography with dual-level bolus timing and separate calf acquisition: hybrid peripheral MR angiography. Radiology 2006;240:283–90.

▶ Collins R, Burch J, Cranny G, et al. Duplex ultrasonography, magnetic resonance angiography, and computed tomography angiography for diagnosis and assessment of symptomatic, lower limb peripheral arterial disease: systematic review. BMJ 2007;334:1257.

▶ Ouwendijk R, de Vries M, Stijnen T, et al. Multicenter randomised controlled trial of the costs and effects of noninvasive diagnostic imaging in patients with peripheral arterial disease: the DIPAD trial. AJR Am J Roentgenol 2008;190:1349–57.


  1. A (F, 5%); B (F, only 2%); C (F, 5–7%); D (F); E (F, 95%)

  2. A (F, B mode); B (T); C (F, see the DIPAD trial); D (T); E (T)

  3. A (T); B (F, see Fleischman reference14); C (F, outrunning the bolus); D (T); E (F, twice)

  4. A (T); B (F, overestimate); C (T); D (T); E (T)

  5. A (T); B (F, pre-existing renal impairment); C (T); D (T); E (F)


View Abstract


  • Competing interests None declared

  • Provenance and peer review Commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.