Imaging for peripheral artery disease

Summary

The goal of imaging in peripheral artery disease (PAD) is to acquire angiographic images that allow for accurate treatment planning for atherosclerotic disease of the lower limbs. In patients with intermittent claudication, a full diagnostic study will demonstrate the arteries bilaterally from the distal aorta to the ankles in both legs. In critical limb ischaemia patients, one is interested in the affected side only and imaging should be done unilaterally from the distal aorta up to and including the level of the pedal arteries.

Digital subtraction angiography (DSA) is the time-honoured reference stand- ard for angiography. However, as an invasive procedure, DSA is costly and carries a small risk of complications.

Magnetic resonance (MR) angiography is a readily available and accurate non-invasive imaging technique. New time-resolved sequences have improved the accuracy for imaging the arteries in critical limb ischaemia. MR imaging of metallic vascular stents is not possible. As a major advantage, vessel calcifications do not limit MR angiography.

Contrast-enhanced computed tomography (CT) angiography is another widely available technique. Images tend to be sharper than those of MR angiography. Imaging of in-stent restenosis is possible. This technique is limited in calcified vessels and carries the risks associated with iodinated contrast agents. It is unfortunate that vascular calcifications and decreased renal function are especially present in PAD patients.

Duplex Doppler ultrasound (DUS), also widely available is limited to evaluation of the aorto-iliac and femoral arteries. Although the leading initial imaging modality, this may not be a cost-effective use of the technique.

The choice of imaging modality in PAD depends on whether endovascular therapy seems likely. If more extensive work-up is needed for treatment planning or if surgery seems likely, then either CT or MR angiography may be the best choice.

Digital subtraction angiography

VALUE OF DIGITAL SUBTRACTION ANGIOGRAPHY FOR IMAGING PAD

Catheter-based digital subtraction angiography (DSA) is the time-honoured method to image the peripheral artery run-off ( Figure 1 ).

A complete imaging study in patients with intermittent claudication will show the peripheral arteries from the distal aorta (i.e., the abdominal aorta from several centimetres above its bifurcation) up to the level of the ankles ( Figure 2). The hallmark complaint of pain in the calf during walking is related mainly to obstructive peripheral artery disease (PAD) lesions in the aorto-iliac, femoral arteries or popliteal arteries and it is generally taught that revascularisation treatment, aimed at improving quality of life, should not be done in arteries beyond the distal popliteal artery. It is therefore not necessary to image the distal arteries at the ankle or in the foot. Following treatment of the symptomatic leg, patients with ntermittent claudication will often resume normal walking exercise and then experience claudication complaints of the contralateral leg. Therefore since PAD generally affects both legs to a similar extent, imaging is preferably done of both legs.

In critical limb ischaemia (CLI) the culprit PAD lesions are frequently at multiple levels and often include lesions in the below-the-knee arteries. The hallmark complaints are rest pain in the foot and non-healing ulcers of the foot, and treatment is aimed at limb salvage by restoring straight arterial inflow from the aorta to the ischaemic part of the foot. It is therefore necessary to image the peripheral run-off from the distal aorta up to and including the arteries of the foot (i.e., the dorsal foot artery and the distal posterior tibial artery, the plantar arteries and the pedal arch). The foot arteries are best imaged in lateral projection( Figure 3 ). CLI patients generally have more extensive atherosclerosis than patients with intermittent claudication and have a higher incidence of diabetes and kidney failure. For this reason it is prudent to avoid inducing kidney failure with too large a dose of iodinated contrast agent and it is therefore wiser to limit imaging to the affected side.

DSA provides high spatial resolution images of the arterial lumen and in addition gives semi-quantitative information on flow direction and flow velocity. Most angiography suites used for endovascular interventions are also well equipped for performing high-quality diagnostic DSA. Minimal requirements for diagnostic DSA are:

1) A dedicated angiography system with a C-arm allowing a wide range of angulations and allowing DSA at variable frame rates of 0.5 – 2 frames per second. Variable size input screens to allow both to overview images of large areas (of the order of 40 cm) as well as zoomed-in spot views (of approximately 15 cm in size).

2) An angiographic power injector.

3) Physiological monitors for on-line invasive pressure measurements, preferably with two simultaneous channels.

Other requirements relate to the hygiene and safety measures of an invasive procedure ( View chapter ):

1) Separate adjoining rooms to scrub before the procedure, separate rooms to stock sterile angiography materials and to store used materials after the procedure.

2) An area close to the angiography suite appropriate for patient observation following the procedure with facilities for emergency resuscitation. At this location there should be skilled personnel to provide care and immediate resuscitation.

One can witness a trend towards stricter demands on sterility issues, including requirements for laminar flow units over the operating area, increased ventilation rates, sluices to create three-compartment workflow areas, etc. These demands are driven by the fact that ever larger devices can be implanted via the percutaneous route such as large-size stent (-grafts) for aneurysm repair and for heart valve replacement. This, and the emergence of hybrid procedures combining percutaneous techniques with open surgery, has initiated enthusiasm for the hybrid angiography suite fully integrated into the conventional surgical operating theatre( Figure 4 ).

In most patients with PAD multiple midstream aortic flush injections in antero-posterior directions with additional oblique views of the area of interest provide good quality diagnostic images. To resolve remaining diagnostic questions, selective catheterisation with selective contrast injections and zoomed-in views may be useful. When in doubt about the haemodynamic significance of an aortic or iliac artery stenosis, one should liberally add invasive pressure measurements, ideally by using a two-channel system, to measure simultaneously pressures proximal and distal to the stenosis at the catheter tip (proximal to the lesion) and the tip of the introducer sheath (distal), respectively. A pressure difference of 10 mmHg in mean arterial pressure (or alternatively, a difference of 10% in systolic peak pressure) at rest or after arterial vasodilatation with papaverine or tolazoline is considered proof of haemodynamically significant obstruction.

ROTATIONAL ANGIOGRAPHY – CONE BEAM CT

It is possible to obtain a 3D data set of the arteries by using rotational angiography. This is obtained by performing a motorised movement of the C-arm at a constant speed around the patient during a continuous contrast injection. The 3D volume obtained in this way can be rotated and viewed in any direction, and optimal tube positioning (angulation, skew) can be chosen ( Figure 5 ) [11. van den Berg JC, Moll FL. Three-dimensional rotational angiography in peripheral endovascular interventions. J Endovasc Ther. 2003 10:595-600. ]. In fact, the rotating x-ray tube and flat panel detector can be considered to be a CT scanner. This technique of so-called cone-beam CT (XperCT [Philips Healthcare, Best, The Netherlands], DynaCT [Siemens AG, Munich, Germany], Innova CT [GE Healthcare, Waukesha, WI, USA]) uses a wider beam of x-rays as compared to the classical (multi-detector) CT scanners, in combination with all the individual detectors of the flat panel detector. Rotational angiography / cone beam CT has in this author's opinion limited added value for the imaging of peripheral artery disease as discussed here. Yet, this technique has proved indispensible for oncological interventional radiology; in particular, for the pre-treatment identification of the tumour feeding vessels during transcatheter intra-arterial chemo-embolisation in liver tumors.

DIGITAL VARIANCE ANGIOGRAPHY

A recent technical variation on DSA is digital variance angiography [22. Gyánó M, Góg I, Óriás VI, Ruzsa Z, Nemes B, Csobay-Novák C, Oláh Z, Nagy Z, Merkely B, Szigeti K, Osváth S, Sótonyi P. Kinetic imaging in lower extremity arteriography: Comparison to digital subtraction angiography. Radiology. 2019 290:246-53. ]. Image formation in digital variance angiography is based on a different phenomenon than in standard DSA. Digital variance angiography was developed to acquire information about functional enhancement. The imaging sequence consists of a run of multiple underexposed images rather than fully exposed images as in standard DSA. Statistical analyses of the full collected image series and their noise allow for the calculation of the standard deviation of the X-ray attenuation for every pixel. Digital variance angiography thus involves a global analysis of the entire imaging series. Because multiple frames contribute to image formation rather than one as in standard DSA, this makes more efficient use of the radiation dose. The improved signal-to-noise may be used to obtain better angiographic image quality with the ability to view smaller vessels, or alternatively, to obtain the same quality angiographic images as in standard DSA but by using lower radiation dose and/or lower contrast-agent dose. However, this comes at a price: Digital variance angiograph is more sensitive to motion artifacts than standard DSA. The method relies on capturing the rise, plateau and fall of enhancement at the pixel level caused by the passing contrast bolus. Absence of vessel motion and other motion artifacts during this acquisition period is assumed. Digital variance angiography shows promise but at present is an experimental technique and not in common use.

DSA IMAGING PROBLEMS

Metallic hip or knee prostheses may obscure the view of the adjacent arteries in DSA and it can be cumbersome to find a projection angle to image the artery appropriately. In the case of extensive arterial occlusion, the flow in the distal vessels may be slow and diminished to the extent that patent vessels distal to the occlusion fail to opacify when using midstream aortic flush injections. Selective antegrade injections in the ipsilateral iliac or common femoral arteries will generally show the majority of such vessel segments to better advantage, but even then some patent vessels segments may not be visualised ( Figure 6 ). In general, any distal vessel with a potential clinical value as a landing zone for a surgical bypass graft (thus a vessel segment with run-off to the foot) can be imaged with selective antegrade DSA. It is not unusual

during endovascular revascularisation procedures however, to engage isolated patent vessel segments not identified earlier with DSA - but which are of no further clinical relevance. In cases of bilateral femoral occlusion or bilateral iliac occlusion, performing diagnostic DSA via the groin may be difficult, in which case a brachial approach can be used though this increases procedure time and carries the risk of slightly more complications.

A major drawback of DSA is that it is an invasive technique which, especially in older, fragile patients, is done as an in-patient procedure thereby adding to the cost. DSA carries a small but real risk of complications at the puncture site and also in the vessels engaged during catheter manipulation (e.g., dissection, vessel thrombosis, embolisation of thrombotic material, haematoma and false aneurysm formation). The rate of serious complications at the puncture site has been reported as 0.04-1.5% and catheter-induced complications as 0.4-1.0% [33. Singh H, Cardella JF, Cole PE, Grassi CJ, McCowan TC, Swan TL, Sacks D, Lewis CA Society of Interventional Radiology Standards of Practice Committee. Quality improvement guidelines for diagnostic arteriography. J Vasc Interv Radiol. 2003 14:S283-8.
The rate of serious complications at the puncture site has been reported as 0. 04-1. 5% and catheter-induced complications as 0 4-1.].

IODINATED CONTRAST AGENTS AND ADVERSE REACTIONS

A specific risk of DSA and also computed tomography (CT) angiography is in the use of iodinated contrast agents. A contrast agent is made up of molecules that contain iodine atoms, the higher the iodine concentration, the denser the contrast. Commercially available contrast agents vary with respect to the particular compound that is used to bind the iodine atoms and can be either ionic or non-ionic (organic). As a consequence, they also vary in viscosity and osmolality, both properties influencing the rate of adverse reactions. The high-osmolality agents have a 5-8 fold, and low-osmolality, a 1-3 fold higher osmolality than the 290 mOsm/kg osmolality of blood plasma.

High-osmolality contrast agents, which include most ionic contrast agents, are associated with more adverse reactions than low-osmolality (mostly non-ionic) contrast agents. In a large case series (337, 647 cases), the risk of a severe adverse drug reaction was 0.2% for ionic and 0.04% for non-ionic contrast agents; and the risk of a very severe adverse drug reaction was 0.04% for ionic contrast agents and 0.004% for non-ionic contrast agents [44. Katayama H, Yamaguchi K, Kozuka T, Takashima T, Seez P, Matsuura K. Adverse reactions to ionic and nonionic contrast media. A report from the Japanese Committee on the Safety of Contrast Media. Radiology. 1990 175:621-8.
In a large case series (337,647 cases), the risk of a severe adverse drug reaction was 0. 2% for ionic and 0. 04% for non-ionic contrast agents, and the risk of a very severe adverse drug reaction was 0.04% for ionic contrast agents and 0.004% for non-ionic contrast agents.]. It should be mentioned that these data may underestimate the risk of using iodinated contrast agents in PAD because these were obtained in all patients undergoing a contrast–enhanced (CT) examination. Patients with PAD generally have impaired renal function which adds to the risks, especially of nephrotoxicity.

Adverse contrast reactions are classified as either anaphylactic (allergic, idiosyncratic) or physiochemotoxic (non-idiosyncratic). Anaphylactic reactions may vary in intensity from mild urticarial itching, to severe and potentially life-threatening arrhythmias, hypotensive shock, pulmonary oedema and bronchospasm. The proposed mechanism of anaphylactic reactions includes enzyme induction, causing the release of vasoactive substances such as histamine and serotonin and the activation of a physiological cascade and eventually the complement system.

Physiochemotoxic reactions include a sense of warmth, metallic taste, nausea, vomiting, bradycardia, hypotension, vasovagal reactions, neuropathy, and the false feeling of spontaneous bladder voiding. These physiochemotoxic reactions are believed to result from the ability of the contrast media to upset the homeostasis of the blood circulation.

Nephrotoxicity is one particular manifestation of physiochemotoxic reactions and is thought to be mediated by a combination of pre-existing haemodynamic alterations: renal vasoconstriction, possibly through mediators such endothelin and adenosine, and direct cellular toxicity of the contrast agent itself. Contrast-induced nephropathy (CIN) is estimated to occur in 2%-7% of patients when defined as an elevation of the serum creatinine level that is more than 0.5 mg/dL (44 mmol/L) or more than 0% of the baseline level at 1-3 days following the contrast load [55. Sing J, Daftary A. Iodinated contrast media and their adverse reactions. J Nucl Med Technol. 2008 36:69-74. ]. The creatinine level usually returns to baseline in 10-14 days but as many as 25% of patients with this nephropathy have a sustained reduction in renal function. Patients with pre-existing renal insufficiency have 5-10 times the risk of CIN than those without insufficiency. Patients with renal insufficiency based on diabetic nephropathy are at the greatest risk. In general, the higher the pre-existing serum creatinine level, the greater the likelihood of CIN.

Diabetic patients who use the oral antidiabetic drug metformin (proprietary brands, Glucophage, Riomet, Metaglip, Avandamet): have an additional risk of developing lactic acidosis. Metformin blocks the uptake of lactate by the liver, which under normal circumstances is balanced because lactate is then preferentially cleared by the kidneys. However, whenever the renal function becomes impaired in reaction to a contrast agent, the clearance of both lactate and metformin is impaired potentially causing lactic acidosis because of a build-up of lactate. It is generally advised that metformin medication is stopped at the time of or prior to the administration of contrast agent and should be withheld for 48 hours after the procedure.

Contrast nephrotoxicity can be reduced (though not completely avoided) by limiting the amount of contrast agent used, by using a non-ionic contrast agent and by discontinuing other nephrotoxic medication. Patients at risk should be well-hydrated before a contrast-enhanced study and hydration should be continued for at least two hours after the procedure. Reducing the overall amount of contrast agent used can also be achieved by decreasing, the contrast injection rate or by using diluted contrast agent during DSA injections. In particular the flat panel detectors used in modern angiography equipment are quite sensitive to contrast and allow good quality images to be obtained with less contrast agent than was possible with earlier angiography systems.

An alternative contrast agent in DSA is carbon dioxide. As a gas, it is a negative contrast agent by means of which the vessel lumen is visible because the gas bubbles replace the blood or float on top of the blood (buoyancy). Carbon dioxide is low-cost and completely non-toxic since it is completely eliminated from the body during the first pass through the lungs. However, the handling of this invisible gas is cumbersome and gas emboli to organs may result in blood flow stasis, i.e., vapour lock. The gas bubbles may then cause ischaemic pain and even necrosis when trapped in the highest parts of the toe arteries, the small bowel and also the brain in unrecognised atrial septal defects. Carbon dioxide also provides less reliable images. Carbon dioxide DSA of the lower limb was not diagnostic to the level of the popliteal artery in 26% of non-selective studies (injection in the aorta) and 16% of selective studies (injection in the lower limb) [66. Bees NR, Beese RC, Belli AM, Buckenham TM. Carbon dioxide angiography of the lower limbs: initial experience with an automated carbon dioxide injector. Clin Rad. 1999 54:833-38. ]. This author is not aware of any clinical unit performing carbon dioxide angiography on a routine clinical basis.

Magnetic resonance angiography

VALUE OF MR ANGIOGRAPHY FOR IMAGING PAD

Contrast enhanced first-pass magnetic resonance (MR) angiography provides accurate, high resolution images of the peripheral arteries from the aorta until the pedal arch in the foot ( Figure 7 ) [77. Kramer H, Nikolaou N, Sommer W, Reiser MF, Herrmann KA. Peripheral MR angiography. Magn Reson Imaging Clin N Am. 2009 17:91-100.
Review of state-of-the-art MR angiography of PAD]. The required 1.0 tesla (T) or 1.5 T MR scanners are the standard in modern radiology departments. Dedicated hardware for MR angiography, including MR surface coils, help to increase the accuracy and robustness of the method. MR angiography of the peripheral vessels makes use of static, high-resolution contrast-enhanced MR imaging sequences. Various MR techniques are available for this purpose, however, most commonly employ a three-dimensional (3D) acquisition scheme. First, an unenhanced mask image is made of the area of interest. Then, a 10 ml bolus of MR contrast agent is injected in the elbow vein and MR images are acquired during the first pass of the contrast agent when there is a peak concentration in the arterial lumen. Digital subtraction of the unenhanced mask from the contrast-enhanced MR image yields a 3D map of the arterial tree. Usually, the entire peripheral run-off is imaged in three subsequent imaging steps requiring separate contrast injections. Alternatively, one makes use of a stepping or continuous moving table technique whereby a single contrast bolus can be physically followed through the various imaging stations.

The 3D MR imaging map is commonly viewed just like a conventional DSA image, that is, by using a maximum intensity projection MIP viewing method whereby all arteries are projected onto a single imaging plane irrespective of whether they are near or far from this plane. The resulting image therefore does not contain information of depth. Multiple projectional views are made, each at 5 or 10 degree incremental steps of rotation, which can be displayed one after another as a movie and this rotational display allows one to assess the arteries from all angles.

First-pass MR angiography thus provides clean images of the arterial lumen only. The vessel wall is subtracted and is not seen. Furthermore, calcifications in the vessel wall are never visible because calcium does not contain protons that form the basis of the MR signal.

MR IMAGING ARTEFACTS AND PROBLEM AREAS

Metallic objects, including vascular stents, pose a problem in MR angiography. MR imaging requires a perfectly homogeneous magnetic field. Para- or ferromagnetic metal objects such as hip and knee prostheses may cause local magnetic field disturbances resulting in signal loss in the vicinity of these objects. Stainless steel vascular stents, e.g., most balloon-expandable stents, are also paramagnetic and will cause loss of MR signal. Self-expandable stents are generally made of Nitinol (a metal alloy of nickel and titanium), are not paramagnetic, and therefore cause less signal loss. However, the inside of all metal stents, including nitinol stents, are generally obstructed from view because of the Faraday cage effect which also causes MR signal loss.

Even when pushing the envelope of MR imaging-technique, using state-of-the-art acquisition techniques (parallel imaging, k-space view sharing, etc.) to accelerate imaging, the imaging process in first-pass MR imaging remains inherently time-inefficient. There is therefore a trade-off between imaging volume, spatial resolution and signal-to-noise ratio. The imaging volume required for peripheral angiography is large and this is at the expense of both spatial resolution (the detail in the image) and long acquisition times. Acquisition times also need to stay within the arterial window when the vessels are enhanced by the bolus of contrast agent. Spatial resolution is in the order of 1.3-1.5 mm, which is not as good as in DSA or CT angiography though clinically adequate. Image acquisition time tends to be long, in the order of 1 minute or more, which leads to two specific problems :

1) Venous contamination: during the long imaging period the contrast agent may have entered the venous circulation. An abundance of contrast enhanced veins may hamper the evaluation of the arteries in the final MR angiogram.

2) Motion artefacts: during the imaging period required to obtain the mask image and the contrast-enhanced image, the patient is required to lie perfectly still. Any motion of the pelvis or the legs will result in motion artefacts within each image and imperfect subtraction of the two image sets. This may be even more pronounced in stepping table techniques, in which the entire peripheral run-off is imaged at one go and there is a longer period between mask and contrast-enhanced image.

Dedicated hardware addresses these two problems. First, the legs may be immobilised by the use of bandages, bean bags, etc.. Second, venous contamination may be limited by using an inflated venous compression cuff around the thigh [88. Koenigskam-Santos M, Sharma P, Kalb B, Carew J, Oshinski JN, Martin D. Lower extremities magnetic resonance angiography with blood pressure cuff compression: quantitative dynamic analysis. J Magn Reson Imaging. 2009 29:1450-56. ].

As a more radical solution to avoid the time-inefficiency of first-pass MR angiography, it has been proposed that MR acquisition schemes be used which obtain imaging data during the equilibrium or steady-state phase after contrast injection. Steady-state MR imaging is done over the course of minutes and provides much better spatial resolution and signal-to-noise than first-pass MR imaging where acquisition is done in a matter of seconds [99. Hadizadeh DR, Gieseke J, Lohmaier SH, Wilhelm K, Boschewitz J, Verrel F, Schild HH, Willinek WA. Periperal MR angiography with blood pool contrast agent: prospective intraindividual comparative study of high-spatial-resolution steady-state MR angiography versus standard-resolution first-pass MR angiography and DSA. Radiology. 2008 249:701-11. ]. However, there are two caveats. First, one cannot use the conventional gadolinium-DTPA contrast agents for steady-state MR angiography: blood-pool contrast agents should be used instead. This is because gadolinium-DTPA will, after minutes, rapidly diffuse from the vessel lumen into the tissue interstitium. Blood pool agents, by contrast, only slowly diffuse out of the vascular lumen compartment and will therefore make the blood vessels stand out from the surrounding tissues for a longer period after injection. The first of the novel class of blood-pool contrast agents has been approved for clinical use in the USA only recently, in 2008: Gadofosveset Trisodium (Vasovist®). Another problem is that steady-state MR angiograms will show the contrast-enhanced veins brightly (venous contamination) because minutes after contrast injection all vascular compartments will contain the blood-pool contrast agents at similarly high concentrations. Because of these issues, in clinical practice MR angiography is still almost exclusively done with first-pass imaging techniques.

In comparison with DSA and CT angiography, MR angiography is not as robust and relatively prone to imaging artefacts because of small, unpredictable disturbances of the magnetic field. Often, fat suppression is suboptimal and inhomogeneous. Such technical imperfections require the attention and active intervention of the MR technicians to adjust MR acquisition parameters. More than in DSA and CT angiography, the quality of MR angiography depends on the quality and dedication of the radiology technicians operating the MR scanner.

Several contraindications should be considered. Patients with pacemakers are generally not imaged with MR angiography because the strong magnetic field of the MR scanner may alter the functioning and settings of the pacemaker. Claustrophobic patients should probably not undergo MR angiography in a small bore tunnel as this may provoke anxiety and panic attacks. Currently, manufacturers are marketing special large bore MR imaging scanners intended for claustrophobic and obese patients.

TIME-RESOLVED MR ANGIOGRAPHY

MR angiography, other than DSA, does not show the dynamic inflow of a contrast agent, but shows the arterial tree at the equilibrium phase at a moment when it is completely filled with contrast agent ( Figure 7 ). On the one hand, this is an advantage as MR angiography will also show the patent vessel segments distal to long occlusions that may not be opacified at DSA, thus allowing better appraisal of the true length of occluded vessel segments. On the other hand, this is at the loss of the potentially dynamic useful information about flow velocities and the preferential flow channels of the blood.

This has prompted the development of time-resolved MR methods which carry the same dynamic information as DSA( Figure 8 ). As in DSA, the leg is imaged during inflow of the contrast agent. Time-resolved MR sequences are based on repetitively acquiring the same imaging volume at a rate of one image every two or three per second. Whereas it is technically not possible at this speed to acquire the complete data set required to reconstruct the full imaging volume, it is possible to obtain the relevant information on (changes in) image contrast provided by the contrast agent. Technically this is done by refreshing only the central lines of k-space during subsequent imaging runs (key-hole imaging) and copying in the remaining peripheral k-lines from the baseline acquisition. Time-resolved MR angiography of the distal calf and pedal vessels has been shown to be superior to standard MR angiography regarding the number of diagnostic segments and assessment of the degree of luminal narrowing [1010. Andreisek G, Pfammatter T, Goepfert K, Nanz D, Hervo P, Koppensteiner R, Weishaupt D. Peripheral arteries in diabetic patients: standard bolus-chase and time-resolved MR angiography. Radiology. 2007 242:610-20.
Time-resolved MRA has added value in diabetic patients over static MR angiography]. Also, in patients with CLI, time-resolved MR angiography proved superior to DSA for imaging of the pedal arteries [1111. Langer S, Krämer N, Mommertz G, Koeppel TA, Jacobs MJ, Wazirie NA, Ocklenburg C, Spüntrup E. Unmasking pedal arteries in patients with critical ischemia using time-resolved contrast-enhanced 3D MRA. J Vasc Surg. 2009 49:1196-202.
In critical limb ischaemia, time-resolved MR angiography has proved more accurate than DSA for imaging distal calf and foot arteries]. The fast imaging of the arriving contrast bolus of time-resolved MR angiography comes at the cost of lower spatial resolution compared to the standard, static contrast-enhanced MR angiography.

NON-ENHANCED MR AND OTHER MR ANGIOGRAPHY TECHNIQUES

Good quality MR angiography is also possible without the use of MR contrast agents. In fact, one of the early promises of MR angiography was to obviate the use of contrast agent altogether. However, over time it has been shown that MR contrast agents added to the reliability and speed of imaging that non-contrast-enhanced MR angiography lacked [1111. Langer S, Krämer N, Mommertz G, Koeppel TA, Jacobs MJ, Wazirie NA, Ocklenburg C, Spüntrup E. Unmasking pedal arteries in patients with critical ischemia using time-resolved contrast-enhanced 3D MRA. J Vasc Surg. 2009 49:1196-202.
In critical limb ischaemia, time-resolved MR angiography has proved more accurate than DSA for imaging distal calf and foot arteries]. Nevertheless, patients with poor kidney function or other contraindications against MR contrast agents may still benefit from non-contrast-enhanced MR angiography. Especially if one is not interested in imaging lengthy vascular trajectories, e.g., if one is only interested in the below-the-knee arteries or the arteries of the foot, then this method may be of use. Tried-and-tested MR imaging methods are available for this purpose, including time-of-flight (inflow) and phase-contrast MR angiography.

Newer non-contrast-enhanced MR imaging methods are currently being developed which make full use of the enhanced technical possibilities of modern MR scanners [1212. Hodnett PA, Koktzoglou I, Davarpanah AH, Scanlon TG, Collins JD, Sheehan JJ, Dunkle EE, Gupta N, Carr JC, Edelman RR. Evaluation of peripheral arterial disease with nonenhanced quiescent-interval single shot MR angiography. Radiology. 2011 260:282-93. , 1313. Offerman EJ, Hodnett PA, Edelman RR, Koktzoglou I. Nonenhanced methods for lower-extremity MRA: a phantom study examining the effects of stenosis and pathologic flow waveforms at 1.5T. J Magn Reson Imaging. 2011 33:401-8. ].

Versatile phase-contrast MR sequences also form the basis for other vascular imaging techniques. Velocity-encoded MR angiography or quantitative MR flow mapping makes use of phase-contrast acquisition schemes to provide accurate information on blood flow velocities (in m/sec), to evaluate increased velocities at vascular stenoses, or decreased flow volume (in ml/sec) through obstructed vascular channels. Flow velocity maps may be used to show the regional variation in flow direction and flow velocity in post-stenotic areas, aneurysms, and vascular bifurcations and may be used as input for quantitative studies of wall stress [1414. Hope MD, Hope TA, Meadows AK, Ordovas KG, Urbania TH, Alley MT, Higgins CB. Bicuspid aortic valve: four-dimensional MR evaluation of ascending aortic systolic flow patterns. Radiology. 2010 255:53-61. ]. Whereas these methods are in frequent use in cardiac MR imaging, they have not yet found clinical use in peripheral vascular imaging.

GADOLINIUM-BASED MR CONTRAST AGENTS AND NEPHROGENIC SYSTEMIC FIBROSIS

The contrast agent in MR angiography is gadolinium (Gd, atomic number 64), which is a paramagnetic heavy lanthanide metal from the group of the rare earth elements. Its use is based on its T1-relaxation shortening properties. When used as a contrast agent a gadolinium-ion is firmly chelated to a ligand. The commercially available MR contrast agents differ with regard to the specific ligands which are used for chelation. Initially, gadolinium-based contrast agents were regarded as ultimately safe: the rate of allergic reactions was much lower than seen with iodinated contrast agents - though not zero- and nephrotoxicity seemed a non-issue. Yet, in 2000, rare but severe complications of gadolinium-based contrast were identified in patients with pre-existing renal insufficiency. Nephrogenic systemic fibrosis (NSF) has now been identified in several hundreds of patients worldwide. In most patients NSF is a purely debilitating disease and in some, a fatal one. NSF becomes manifest days to many months after exposure to the gadolinium-based contrast agent. The early symptoms of NSF occur within 2 months and consist of warm swellings, pain, discoloration and itching of the legs. Late or chronic NSF symptoms, 6 months or later after exposure, consist of fibrotic skin changes similar to scleromyxoedema and/or fibrosis of muscles, joints, eyes and internal organs. There is currently no treatment for NSF. Retrospective analysis indicates that NSF is related to those gadolinium-chelates that bind the gadolinium less tightly and allow dissociation of free Gd3+ ions. In response most radiologists have switched to gadolinium compounds which use tighter bonding chelates. However, it is advised to avoid the use of the “unsafe” MR contrast agents in patient with low renal function and in particular, a glomerular filtration rate under 30 ml/min/1.73 m2 [1515. Thomsen HS, Morcos SK, Almén T, Bellin MF, Bertolotto M, Bongartz G, Clement O, Leander P, Heinz-Peer G, Reimer P, Stacul F, Van der Molen A, Webb JA. Nephrogenic systemic fibrosis and gadolinium-based contrast media : updated ESUR Contrast Medium Safety Committee guidelines. Eur Radiol. 2013 23 :307-18.
NSF is a rare but potentially dangerous complication of gadolinium-based contrast agents in MR angiography. Only in patients with very low glomerular filtration rates of 30 ml/min/1.73m2 or less.]. In the US, the Food and Drug Administration specifically warned against the use of the MR contrast agents Magnevist, Omniscan and Optimark in such patients [1616. https://www.fda.gov/drugs/drug-safety-and-availability/fda-drug-safety-communication-new-warnings-using-gadolinium-based-contrast-agents-patients-kidney. Accessed 08/17/2020. ]. In Europe, Magnevist and Omniscan have been no longer authorized for use since 2018.

Computed tomography angiography

VALUE OF COMPUTED TOMOGRAPHY ANGIOGRAPHY FOR IMAGING PAD

CT angiography is a robust method providing highly detailed images of the peripheral arteries from the aorta to the plantar arch with high spatial resolution( Figure 9 ). It is an accurate modality to assess the presence and extent of PAD, especially in patients with intermittent claudication [1717. Met R, Bipat S, Legemate DA, Reekers JA, Koelemay MJ. Diagnostic performance of computed tomography angiography in peripheral arterial disease: a systematic review and meta-analysis. JAMA. 2009 301:415-24.
A meta-analysis showing the overall good accuracy of CT angiography in patients with intermittent claudication]. This method is widely available because the multichannel helical CT scanners required for CT angiography are a mainstay (in fact the workhorse imaging machines) in modern radiology departments. No additional hardware is required for CT angiography.

Vascular enhancement with iodinated contrast is standard. A bolus of iodinated contrast agent (typically 100 ml at a flow of 4 ml/sec) is injected into the antebrachial vein and CT images are made during first pass of the contrast agent at the peak concentration in the arterial lumen. In the CT image, the high-density (white) blood can be differentiated easily from the low density thrombus and vessel wall.

CT angiograms are mostly viewed using multiplanar reformats, that is relatively thin imaging slices and by scrolling through MPRs on the viewing station in the transverse, coronal, and/or sagittal planes. Curved planar reformats may be helpful in the parasagittal plane following the course of the iliac arteries, which in general are manually prepared by the CT technicians. It is of note that reformatting of the CT images can be done in any desired imaging plane without loss of image quality because isotropic voxels are obtained during image acquisition. Other than in MR angiography, no mask image is made in CT angiography. Computer algorithms exist to remove bone automatically from the images in order to obtain a clearer view of the arterial lumen, yet in clinical practice such post-processing is seldom needed. This technique is also susceptible to artefacts.

CT angiographic images show not only the arterial lumen but also the arterial wall. Assessment of the arterial wall with its calcifications provides useful information that may influence the treatment strategy in endovascular or surgical intervention.

CT ANGIOGRAPHY ARTEFACTS AND PROBLEM AREAS

The Achilles’ heel of CT angiography lies with dense vessel wall calcifications. Vessel calcifications are indistinguishable from contrast-enhanced blood. Evaluation of the vessel lumen is hampered and at times downright impossible in the presence of abundant calcifications. Below-the-knee vessels in diabetic patients are a case in point [1818. Ouwendijk R, Kock MC, van Dijk LC, van Sambeek MR, Stijnen T, Hunink MG. Vessel wall calcifications at multi-detector row CT angiography in patients with peripheral arterial disease: effect on clinical utility and clinical predictors. Radiology. 2006 241:603-8.
Vessel calcifications limit the accuracy of CT angiography in PAD] ( Figure 10 ). One may in some patients resolve these problems by using as thin imaging sections as possible and by carefully evaluating the axial source images - but in many patients the calcium load is just too high. Attempts to resolve the problem by using altered acquisition schemes (e.g., dual source CT imaging which employs radiation at two different kilovoltage settings) or by post-processing filtering schemes have proved disappointing ( Figure 11 ).

As in MR angiography, metal hip or knee prosthesis may distort the CT image. Imaging of stents tends to be less of a problem with CT angiography than with MR angiography though it remains an issue to look for in in-stent restenosis within balloon expandable stents made of dense stainless steel. The less radio-opaque Nitinol stents do not pose a major problem in CT angiography.

In general practice CT angiography, like static MR angiography, shows the arterial tree at equilibrium phase but lacks the dynamic inflow information provided by DSA or time-resolved MR angiography. However, experimental studies have shown the feasibility of time-resolved dynamic-CT angiography of the lower leg arteries, whereby the calf region is repetitively imaged using acquisition schemes such as continuous bidirectional table movement [1919. Sommer WH, Helck A, Bamberg F, Albrecht E, Becker CR, Weidenhagen R, Kramer H, Reiser MF, Nikolaou K. Diagnostic value of time-resolved CT angiography for the lower leg. Eur Radiol. 2010 20:2876-81. ]. This technique has never gained popularity, likely because of the drawbacks of a greatly increased contrast load.

The complications associated with the use of iodinated contrast agents are similar to those discussed in the DSA section. A typical CT angiography protocol uses 100 millilitres of contrast agent with 250 milligrams of iodine per millilitre, corresponding to a total contrast dose of 25 grams of iodine. A common injection rate is 4 ml/sec into the antebrachial vein.

Concerns have been raised about the proliferating use of modern helical CT in medical practice which has resulted in an overall increasing dose of ionising radiation to the population at large. Generally speaking this is a legitimate concern because of the two long-term hazards known to be associated with ionising radiation: the induction of neoplasms, and the induction of genetic changes that may be carried into future generations. However, in this author’s opinion, concern over radiation dose is of lesser relevance in this specific patient population with obstructive PAD. The typically older PAD patient is well beyond the procreational age and, in addition, his life expectancy is more likely to be limited by cardiovascular mortality than by mortality because of a potentially induced cancer through the use of CT angiography. That is to say, the risk of induced neoplasms is present but it would take 10-20 years before this might occur.

CT ANGIOGRAPHY IMAGE FORMATION AND IMAGE ACQUISITION

Image formation in CT angiography is based on the addition of contiguous thin slices into a large 3D-volume. Section thickness is typically 1 millimetre or less and the in-plane resolution, 1 mm or better (e.g., a 25 x 25 cm2 field of view covered by a 256 x 256 imaging matrix). The individual voxels that make up the imaged 3D-volume have therefore equal dimensions along the X, Y and Z-axes and are said to be isotropic. This means that, once the full 3D-volume has been acquired, reconstructions can be made in any desired cross-sectional plane, including curved ones, without perceptible loss of image quality.

CT imaging sections are produced by continuously rotating the x-ray tube in a circular motion in the gantry perpendicular to the CT-table, while simultaneously moving the CT-table with the patient on top through the gantry. From the patient’s perspective, the x-ray tube will then follow a spiralling path along his or her long axis, hence the name spiral or helical CT.

Each individual 1 mm-thick imaging section is acquired during a 180 degree rotation of the x-ray tube around the long-axis of the patient. The information collected in this way is then reconstructed into a 256 x 256 image by using a 180-degree Fourier transformation mathematical algorithm. Modern CT-scanners employ tube rotation times of 1/3 to 1 second, which means that the time to image each individual section is 1/6 - ½ second. This short imaging time means that motion artefacts play only a minor role in CT angiography.

The x-ray beam fan from the rotating x-ray tube is collected by not one but multiple x-ray detectors lying next to each other. When two detectors are used, each tube rotation will yield two instead of one imaging section. The acquisition time is shortened in proportion to the number of x-ray detectors (channels) built into the CT-scanner. Current multichannel CT-scanners typically use 16, 64, or 128 detector configurations (2011 data). To acquire a full peripheral run-off from distal aorta to the toes roughly corresponds to a 3D-volume of approximately 120 centimetres length along the long-axis of the patient, i.e., 1,200 imaging sections of 1 mm thickness. A 16-slice CT-scanner with tube rotation time of 1/2 second can do this in 20 seconds, a 128-slice CT-scanner with tube rotation time of 1/3 second, in less than 2 seconds. The speed of the CT-scanners therefore greatly exceeds the optimal speed for CT angiography which is dictated by the speed at which the contrast bolus is propelled along the peripheral arteries by the blood flow. This is why high-quality CT angiograms

can also be obtained with relatively modest 16-channel CT-systems and in addition there is little if any quality gain for peripheral CT angiography in using higher-end CT-scanners. The imaging protocol needs to be optimised for each individual CT-scanner type. If the imaging speed of the CT coincides with the speed of the contrast bolus, this means that optimal use can be made of the iodinated contrast agent and that the contrast dose can be minimised.

Duplex doppler ultrasound

VALUE OF DUPLEX DOPPLER ULTRASOUND FOR IMAGING PAD

Duplex Doppler ultrasound (DUS) has the advantage over other imaging modalities in that it directly detects the physiological effect of arterial obstruction, rather than merely the anatomical substrate of decreased lumen diameter. Arterial obstruction is detected by noticing a steep increase in blood flow velocity. This may in fact be clinically more meaningful than detecting lumen stenosis because an increase in flow velocity is inherently linked to a pressure drop over the obstruction. Arguably the most important role for DUS in the pre-treatment work-up of PAD is to help decide whether an ambiguous stenotic lesion at DSA, CT or MR angiography is haemodynamically significant or not.

The term Duplex refers to the fact that DUS combines the anatomical greyscale (B-mode)ultrasound image with a pulsed-wave Doppler ultrasound measurement. This combination allows one to sample the Doppler spectrum at any chosen point or area of interest within the anatomic image, i.e., exactly in the artery segment of interest. Pulsed wave Doppler has considerable advantages over continuous wave Doppler, which samples all Doppler spectra encountered along the path of the insonated ultrasound beam. An in-depth discussion of continuous wave versus pulsed-wave Doppler ultrasound is beyond the scope of this chapter and can be found in textbooks on Doppler ultrasound and also in the 2006 review article of the American Society of Echocardiography and the Society of Vascular Medicine and Biology [2020. Gerhard-Herman MG, Gardin JM, Jaff M, Mohler E, Roman M, Naqvi TZ. Guidelines for noninvasive vascular laboratory testing: a report from the American Society of Echocardiography and the Society of Vascular Medicine and Biology. J Am Soc Echocardiogr. 2006 19:955-72.
Review article of duplex ultrasound in PAD].

DUS makes use of the Doppler effect (named after the Austrian physicist Christian Doppler, 1803-1853) whereby particles in flowing blood reflect an insonated ultrasound beam with altered frequency. The Doppler frequency shift depends on the velocity of the flowing blood, the frequency of the insonated ultrasound beam, the speed of sound in the body tissues and the angle between the flow velocity and the insonated beam, according to the equation ( Figure 12 ).

The measurement depends on the cos (α) and small errors in angle measurements will result in relatively large errors in calculated Doppler shift at higher values of α nearing 90 degrees. This is why, in clinical practice, only DUS measurements obtained at α of 60 degrees or less are considered trustworthy. This can be a problem in, e.g., the common iliac artery which is usually insonated from the anterior abdomen at an angle near 90 degrees. It should also be noted that the correct angle to be taken relates to the direction of the blood flow (jet), which may differ from the direction of the long axis of the blood vessel. For all practical purposes, however, an abnormal DUS result obtained at an insonation angle of 60 degrees or less is convincing proof of a haemodynamically significant stenosis.

DUS CRITERIA FOR ARTERIAL STENOSIS

The physiological peak systolic velocity (PSV) of blood flow in healthy larger sized arteries does not exceed 100 cm/sec. DUS criteria for haemodynamically significant stenosis rely either on measuring elevated absolute PSV or an increased ratio of a high PSV at the site of stenosis relative to a lower PSV in the same vessel segment proximal (or far distal) to the stenosis. For example, a PSV value of 230 cm/sec is generally taken as evidence of a more than 70% carotid artery stenosis and a PSV ratio of more than 2 as evidence of haemodynamically significant stenosis in PAD. The precise cut-off values to distinguish between normal and abnormal may differ slightly between ultrasound machines and operators, and are therefore best determined individually in each vascular lab.

The normal arterial blood flow in peripheral arteries has a triphasic or biphasic waveform, showing a steep systolic upslope, a second diastolic phase whereby the flow velocity drops slightly below zero (i.e., the blood flows backwards for a short moment) and a third phase of small positive upstroke. This complex blood flow pattern is a result of the summation of the pressures on the blood by the forward pushing proximal blood column and modulating pressures exhibited by the arterial wall ( Figure 13 ). For clinical practice it is of interest that distal to a significant stenosis the waveform often changes to a monophasic signal, characterised by continuous antegrade flow with a slight arterial ripple (Thus, if one sees a monophasic waveform pattern at DUS, this is a reliable sign of a significant stenosis located more proximally. The reverse, however, is not true: A normal biphasic or even triphasic waveform does not exclude a more proximal stenosis [2121. Spronk S, Den Hoed PT, de Jonge LC, van Dijk LC, Pattynama PM. Value of the duplex waveform at the common femoral artery for diagnosing obstructive aortoiliac disease. J Vasc Surg. 2005 42:236-42. ]. Another tell-tale waveform pattern is the so-called sharp monophasic signal seen just proximal to an occlusion or near-occlusion: bursts of short and sharp monophasic antegrade flow separated by periods of zero flow, indicating a lack of net antegrade flow through the vessel.

DUS IMAGING PROBLEMS

DUS is an examination which is not easy to perform. A full DUS examination of the peripheral arteries of both legs is time-consuming and requires technical skill and dedication. A problem for a comprehensive DUS examination is that theoretically, all vessel segments should be interrogated with a pulsed-wave Doppler beam. In practice it can be a problem with DUS to find an unobstructed view of the common iliac arteries with an adequate insonation angle of less than 60 degrees. Obesity and/or bowel gas further hamper the reliability of DUS in this region. Another problem area for DUS is the lower leg arteries which cannot be reliably examined.

DUS is operator dependent and the results, especially if normal, cannot be reliably checked afterwards. In most institutions DUS is performed by vascular technicians and the interventionalist has to base his therapeutic decision purely on their report. As previously mentioned, a hard copy recording of an abnormally high PSV ratio obtained at an insonation angle of less than 60 degrees and abnormal waveforms are reliable proof of a haemodynamically significant stenosis. Yet on the other hand, when a normal DUS-examination is called by the ultrasound technician and the recorded imaging files only show screen prints of normal vessel segments, there is in fact no way to check if this represents a truly normal examination or that a stenosis may have been overlooked because of technical difficulties, time constraints etc.

DUS IN POST-INTERVENTION FOLLOW-UP

For this author, if one is particularly interested in evaluating small and well-circumscribed areas of interest, DUS has its greatest value as a follow-up examination after endovascular treatment or vascular surgery. For example: a femoral stent to search for in-stent restenosis, a vessel anastomosis after surgical bypass grafting to look for stenosis or the formation of a false aneurysm, and the endoluminal sac of an aortic aneurysm after endovascular repair to search for endoleaks. The main drawbacks of DUS have been obviated: the examiner knows where to look and what to look for and the advantages of DUS then come to full expression: a direct physiological measurement without ionising radiation and without contrast agent by using a quick and relatively cheap technique.

Comparison of imaging modalities

Both CT angiography and MR angiography have proved to be accurate imaging modalities in obstructive PAD. In a randomised comparison, we have found that in the overall cohort of patients with obstructive PAD, the use of CT angiography was more cost-effective than MR angiography or DSA [2222. Kock MC, Adriaensen ME, Pattynama PM, van Sambeek MR, van Urk H, Stijnen T, Hunink MG. DSA versus multi-detector row CT angiography in peripheral arterial disease: randomized controlled trial. Radiology. 2005 237:727-37. ]. In particular, CT angiography was associated with higher degree of diagnostic certainty and reduced the costs of extra imaging tests to resolve diagnostic uncertainties. However, the study was done as long ago as 2004 and since then there have been several developments which may prompt one to take a fresh look at the available imaging modalities:

1) Further technical developments in endovascular technique have expanded the options to treat patients with CLI. Many more patients with CLI are now being considered for revascularisation of below-the-knee arteries. These patients, in earlier times, would have undergone amputation of the leg and would not have had imaging of the lower limb arteries. These patients are now referred for pre-treatment planning and, because many of them have diabetes, have a higher incidence of severely calcified vessels and also decreased kidney function. Both factors limit the usefulness of CT angiography.

2) The diagnostic accuracy of MR angiography in below-the-knee vessels and pedal has substantially improved with the addition of time-resolved sequences not available to us in 2004.

3) Consequently, the balance has shifted somewhat over the past few years from CT towards MR-angiography.

In a relatively healthy patient with normal kidney function and without considerable vessel calcifications, CT angiography may be the ideal imaging modality: it is widely available, non-invasive, robust and fast. The obtained images are of exquisite spatial detail showing also the smaller arterial branches and without perceptible motion artefacts. Difficult to evaluate vessel segments may be evaluated by using multiple multiplanar reformatted images. Unfortunately, most patients with PAD have concomitant cardiovascular renal disease and vessel calcifications. In particular, diabetic patients with poor kidney function and CLI may be suboptimal candidates for CT angiography. The weak points of CT then become apparent: assessment of heavily calcified vessels becomes an issue which renders the examination non-diagnostic. In addition, contrast-induced nephrotoxicity is an important issue as this is a risk especially in azotemic diabetic patients.

MR angiography also provides a good diagnostic examination. MR imaging too is widely available and the resultant images, though lacking the fine crisp detail of the CT angiogram, have sufficiently high spatial resolution to steer the treatment decision. The fact that the imaging time is longer and that the patient must be able to lie still in a cramped MR-tunnel, though they will be tightly bound to the MR imaging-table during imaging, are minor limitations. Patients with pacemakers or claustrophobia cannot undergo MR angiography. A significant weak point of static MR angiography, i.e., venous contamination, may be obviated by using the newer, time-resolved MR angiography sequences. Other weak points of MR angiography remain, e.g., its unpredictability and non-robustness. Also in patients who have undergone previous stent placement or that have metallic prosthesis, MR angiography has a predictably low yield. In patients with severly impaired kidney function, the rare but severe complication of NSF should be considered.

DUS is used in many institutions as the initial test in obstructive PAD to steer subsequent diagnostic strategy. It is relatively cheap, readily available and does not require a nephrotoxic contrast agent. However, in the large multicentre randomised trial mentioned above, we found that whenever DUS was used as the initial imaging test in PAD, it resulted overall in less diagnostic confidence and more requests for additional imaging tests when compared with the situations when either CT or MR angiography were used as the initial imaging test. In the final analysis, the DUS-as-initial test strategy was expensive and not cost-effective [2323. Ouwendijk R, de Vries M, Pattynama PM, van Sambeek MR, de Haan MW, Stijnen T, van Engelshoven JM, Hunink MG. Imaging peripheral arterial disease: a randomized controlled trial comparing contrast-enhanced MR angiography and multi-detector row CT angiography. Radiology. 2005 236:1094-103. , 2424. Ouwendijk R, de Vries M, Stijnen T, Pattynama PM, van Sambeek MR, Buth J, Tielbeek AV, van der Vliet DA, SchutzeKool LJ, Kitslaar PJ, de Haan MW, van Engelshoven JM, Hunink MG Program for the Assessment of Radiological Technology. Multicenter randomized controlled trial of the costs and effects of noninvasive diagnostic imaging in patients with peripheral arterial diseas: the DIPAD trial. AJR Am J Roenthenol. 2008 190:1349-57.
Randomised trials indicating CT angiography as the most cost-effective initial imaging method in PAD, in comparison with MR angiography, DSA and DUS]. The previously mentioned trend of more patients presenting with CLI strengthens these results because DUS is less reliable in below-the-knee arteries.

An issue, as previously stated, is that the sensitivity for detecting stenoses is operator-dependent and that especially a normal DUS cannot be confirmed by examining the recordings of the examination. This author would therefore be hesitant to accept a normal DUS as the sole basis for treatment.

DSA is, from the diagnostic point of view, probably the superior method of the three. Standard midstream aorta injections will give a good general overview of the peripheral run-off and, if desired, selective injections combined with zoomed-in acquisition and/or invasive pressure measurements will resolve remaining uncertainties. Diagnostic problems are mainly in the non-opacification patent vessels distal to long segments of occlusion and insufficient collaterals due to contrast dilution and slow flow. However, on the downside, DSA is an invasive technique associated with higher costs and potential complications at the puncture site and as a result of catheter manipulations.

WHICH IMAGING MODALITY IS BEST IN PERIPHERAL ARTERY DISEASE?

The question to ask when considering an imaging study in patients with PAD is exactly what information is needed and how this information will affect treatment decisions. It is also important to consider not only the first imaging test but also the potential need for follow-up tests in case of indeterminate results from the first test. Overall, the aim is to get the required information in the quickest and most cost-effective manner. The answer will be different for different clinical scenarios:

Intermittent claudication

The a priori assumption in this patient group is that the vascular lesion will be located in the aorto-iliac or femoral region. Based on the history, the physical examination and the ankle-brachial indices there is an initial understanding of where the lesion will be located. If the clinical question is to decide between exercise training versus endovascular treatment, should an easy-to-treat central aorto-iliac lesion be identified, both MR angiography and CT angiography would seem a rational choice.

If the decision for revascularisation has been made, e.g., because exercise has failed to bring adequate relief, it is relevant whether endovascular treatment is considered the more likely option or if no such preference can be made on an a priori basis. If endovascular treatment can be combined with a full diagnostic DSA in the same session, this will obviate the need for a prior non-invasive imaging study. Combining the diagnostic and therapeutic sessions saves time and resources.

If more extensive work-up is needed to choose between endovascular treatment and open surgery, both MR angiography and CT angiography are suitable initial imaging methods. Any diagnostic uncertainties could be resolved next by DSA, perhaps done in the context of endovascular treatment of a central inflow lesion. In diabetic patients with claudication, this author would prefer MR angiography over CT angiography because of the high likelihood of extensive vessel calcifications. If there are contraindications against MR angiography, e.g., because the patient is claustrophobic or has a pacemaker, then CT angiography should be done.

In a patient with recurrent claudication after earlier revascularisation, there is probable failure of the earlier procedure, e.g., by the development of in-stent restenosis or of an anastomotic stenosis in a surgical bypass. If the specific area of interest is within easy reach of ultrasound, a dedicated DUS study might seem an appropriate first-line test. Otherwise, both MR angiography and CT angiography could be useful. However, in case a vascular stent has been placed earlier, CT angiography seems the more appropriate of the two.

Critical limb ischaemia

Here, the question to answer is whether revascularisation by surgical or endovascular means is possible in order to salvage the leg. The diagnostic approach could then be similar to that described in this scenario for intermittent claudication.

In the case of endovascular treatment being more likely, MR angiography or CT angiography will not have much additional value and we tend to rely more and more on conventional DSA immediately followed by a catheter-based treatment in the same session. Some endovascular treatments may require an antegrade femoral approach and in these instances too, it is adamant to rule out aorto-iliac artery obstruction on the affected side. This could be done by DUS or, and this is this author’s preference, with a diagnostic DSA with retrograde approach from the contralateral side. My common strategy is to do a contralateral DSA first and, based on these images, decide whether to proceed with an over-the-bifurcation procedure or to switch to an antegrade femoral approach.

If more extensive diagnostic work-up is needed before treatment then this author would have a preference for MR angiography, including a time-resolved MR sequence. This, given the higher incidence of diabetes in this population with heavily calcified vessels including the arteries below-the-knee and also considering the risk of iodinated contrast agents in azotemic patients.

Personal perspective - Peter M. Pattynama

If endovascular therapy on an a priori basis seems the likely therapy, then combine the treatment session with a full diagnostic DSA. This saves time and resources.

If endovascular therapy on an a priori basis seems the likely therapy AND an antegrade femoral approach seems preferable, then one may use DUS to rule out aorto-iliac stenotic disease or use DSA from the contralateral side. This author prefers the latter approach.

If more extensive work-up is needed for treatment planning or if open surgery seems the likely treatment option then one can use either CT angiography or MR angiography.

The choice between CT and MR angiography is based on considering specific patient characteristics (including: expected site of PAD lesion? Previous stent placement? Comorbidity? Kidney function? Claustrophobia? Pacemaker?). This author has a personal preference for MR angiography that includes, in critical limb ischaemia, time-resolved MR angiography of the lower leg and foot arteries.