ct angiography of peripheral arterial disease

CT Angiography of Peripheral Arterial DiseaseDominik Fleischmann, MD, Richard L. Hallett, MD, and Geoffrey D. Rubin, MD

Lower-extremity computed tomographic (CT) angiography (ie, peripheral CT angiography) is increasingly used toevaluate patients with peripheral arterial disease. It is therefore increasingly important for all vascular specialists tobecome familiar with the strengths and limitations of this new technique. The aims of this review are to explain theprinciples of scanning and injection technique for a wide range of CT scanners, to explain and illustrate the propertiesof current image postprocessing tools for effective visualization and treatment planning, and to provide an overviewof current clinical applications of peripheral CT angiography.

J Vasc Interv Radiol 2006; 17:3–26

Abbreviations: CPR � curved planar reformation, DSA � digital subtraction angiography, MIP � maximum intensity projection, MPR � multiplanar reforma-tion, 3D � three-dimensional, VR � volume reconstruction

ALTHOUGH computed tomographic(CT) imaging of lower-extremity arter-ies has been attempted with single–detector row scanners (1–5), it was notbefore the introduction of multiple–detector row CT that adequate resolu-tion imaging of the entire inflow andrunoff vessels became possible with asingle acquisition and a single intrave-nous contrast medium injection (6).With increasing availability of multi-ple–detector row CT scanners, periph-eral CT angiography has gradually en-tered clinical practice (7–11), and as aresult of the concomitant rapid evolu-tion of CT scanner technology, high-resolution imaging of the peripheralvasculature has become routinely pos-sible.

For the vascular and interventionalradiologist, but also for the vascularsurgeon and cardiovascular specialist,

it is increasingly important to be famil-iar with this latest vascular imagingtechnique, to know its strengths andlimitations, and, most importantly, tolearn how to read and interpret thelarge CT angiographic data sets andtheir reformatted images for clinicaldecision making and treatment plan-ning.

The organization of this review re-flects three interrelated objectives.First, we describe the technical princi-ples of image acquisition and contrastmedium injection parameters for awide range of currently available mul-tiple–detector row CT systems. Thelevel of detail should allow the diag-nostic imager to develop his or herown scanner-specific peripheral CTangiography acquisition and injectionprotocol. Next, we will review theproperties of different visualizationtechniques for extracting the relevantfindings and explain how they are in-terpreted. Finally, we will discuss theaccuracy and the practical applicationof CT angiography within the contextof various clinical situations.

SCANNING TECHNIQUE

Peripheral CT angiograms can beobtained with all current multiple–de-tector row CT scanners (ie, four ormore channels). No special hardwareis required. With a standardized scan-

ning protocol programmed into thescanner, peripheral CT angiography is avery robust technique for elective andemergency situations. When patientsare mobile, the study can easily be per-formed in 10–15 minutes of room time.

In general, peripheral CT angiogra-phy acquisition parameters followthose of abdominal CT angiography.Unless automated tube current modu-lation is available, a tube voltage of120 kV and a maximum tube amper-age of 300 mA (depending on the scan-ner) is used for peripheral CT angiog-raphy, which results in a similarradiation exposure and dose (12.97mGy, 9.3 mSv) as abdominal CT an-giography (7). Breath-holding is re-quired only at the begining of the CTacquisition through the abdomen andpelvis. Lower amperage (and/or volt-age) can and should be used in pa-tients with low body mass index. Inobese patients, tube voltage and tubecurrent often need to be increased. Amedium to small imaging field of view(with the greater trochanter used as abony landmark) and a medium to softreconstruction kernel are generallyused for image reconstruction.

SCANNING PROTOCOL

One or more dedicated peripheralCT angiographic acquisition and con-trast medium injection protocol(s)

From the Cardiovascular Imaging Section, Depart-ment of Radiology, Stanford University MedicalCenter, 300 Pasteur Drive, S-072, Stanford, Califor-nia 94305-5105 (D.F., R.L.H., G.D.R.); and Depart-ment of Angiography and Interventional Radiology,Medical University of Vienna, Waehringer Guertel18-20, A-1090, Vienna, Austria (D.F.). Received July27, 2005; accepted October 8. Address correspon-dence to D.F.; E-mail: [email protected]

None of the authors have identified a conflict ofinterest.

© SIR, 2006

DOI: 10.1097/01.RVI.0000191361.02857.DE

Review Articles

3

should be established for each scannerand programmed into the scanner. Afull scanning protocol consists of (i)the digital radiograph (“scout” imageor “topogram”), (ii) an optional non-enhanced acquisition, (iii) one seriesfor a test bolus or bolus triggering, (iv)the actual CT angiography acquisitionseries, and (v) a second optional “late-phase” CT angiography acquisition(initiated only on demand) in theevent of nonopacification of distal ves-sels (Fig 1).

Patient Positioning and ScanningRange

The patient is placed feet-first andsupine on the couch of the scanner. Tokeep the image reconstruction field ofview small, and also to avoid off-cen-ter stair-step artifacts (12), it is impor-tant to carefully align the patient’s legsand feet close to the isocenter of thescanner. Tape may be required to holda patient’s knees together. Althoughcushions can be used to stabilize theextremities for the patient’s comfort,large cushions under the knees shouldnot be used so the field of view can bekept small. Also, as in conventional an-giography, excessive plantar flexion ofthe feet is avoided to prevent an artifac-tual stenosis or occlusion of the dorsalispedis artery (ie, “ballerina sign” [13]).

The anatomic coverage extendsfrom the T12 vertebral body level (toinclude the renal artery origins) prox-imally through the patient’s feet dis-tally (Fig 1). The average scan length isbetween 110 cm and 130 cm. Smallerscan ranges or a smaller field of view(ie, one leg only) may be selected inspecific clinical situations.

Image Acquisition andReconstruction Parameters

The choice of acquisition parame-ters (ie, detector configuration/pitch)and the corresponding reconstructionparameters (ie, section thickness/re-construction interval) depends largelyon the type and model of the scanner.Table 1 provides an overview of pe-ripheral CT angiography acquisitionparameters for a wide range of mul-tiple–detector row CT scanners, andreflects our clinical experience withfour-, 16-, and 64- channel Siemens CTscanners (Siemens, Erlangen, Ger-many), and four-, eight,- and 16-chan-

nel General Electric CT scanners (GEMedical Systems, Milwaukee, WI). De-tector configuration is denoted as num-ber of channels times channel width (eg,4 � 2.5 mm); section thickness and re-construction interval are expressed as aratio, eg, 3 mm/1.5 mm.

Four-channel CT

To cover the entire peripheral arte-rial tree within acceptable scanning

times, the detector configuration isusually set to 4 � 2.5 mm. As a result,the thinnest effective section thicknessachievable with these scanners is ap-proximately 3 mm. With overlappingimage reconstruction every 1–2 mm,these “standard-resolution” data setsare adequate for visualizing the aor-toiliac and femoropopliteal vessels,and also provide enough detail to as-sess the patency of crural and pedalarteries if vessel calcification is absentor minimal. Four-channel CT thereforeprovides adequate imaging in patientswith intermittent claudication, inacute embolic disease (Fig 2), aneu-rysms, anatomic vascular mapping,and also in the setting of trauma. Suchstandard-resolution data sets may notbe fully diagnostic when visualizationof small arterial branches (ie, crural orpedal) is clinically relevant, such as inpatients with critical limb ischemiaand predominantly distal disease, no-tably in the presence of excessive arte-rial wall calcifications. If higher reso-lution imaging is required and if theclinical situation permits a smaller an-atomic coverage area (eg, limited tothe legs or to popliteal to pedal vesselsonly) exquisite high-resolution imag-ing can also be achieved with four-channel systems with 4 � 1-mm or 4 �1.25-mm detector collimation (7)within acceptable acquisition times.

Eight-channel CT

A detector configuration of 8 � 1.25mm permits anatomic coverage of theentire peripheral arterial tree withinthe same scan time as a 4 � 2.5-mmfour-channel CT acquisition at sub-stantially improved resolution. Imageswith a nominal section thickness of1.25 mm, reconstructed at 0.8 mm (ie,high resolution) can routinely be ac-quired. Crural and even pedal arteriescan be reliably identified with use ofthese settings. With faster gantry rota-tion speed (0.5 sec/360° rotation) andmaximum pitch, eight-channel CT al-lows for extended volume coverage,such as to visualize the entire thoracic,abdominal, and lower-extremity arter-ies within a single acquisition (Fig 3).

Sixteen-channel CT

When a detector configuration of 16� 1.25 mm or 16 � 1.5 mm is used,similar high-resolution data sets of the

Figure 1. Digital CT radiograph for pre-scribing peripheral CT angiography. Thepatient’s legs and feet are aligned with thelong axis of the scanner. Scanning range(from T12 through the feet) and reconstruc-tion field of view (determined by thegreater trochanters; arrows) are indicatedby the dotted line. A second optional CTangiography acquisition is prescribed forthe crural/pedal territory (dashed line).

4 • CT Angiography of Peripheral Arterial Disease January 2006 JVIR

peripheral arterial tree can be acquiredif 1.25–2-mm-thick sections are recon-structed at 0.8–1-mm intervals (Fig 4).However, the acquisition speed withthese parameter settings is substan-tially faster compared with four- andeight-channel systems. In fact, it maybe too fast for patients with alteredflow dynamics (as detailed later).Therefore, in contrast to four- andeight-channel CT, it is generally notnecessary and potentially detrimentalto choose the maximum pitch and/orthe fastest gantry rotation speed withthese scanners.

Sixteen-channel CT also allows, forthe first time, acquisition of submilli-meter “isotropic” data sets of the en-tire peripheral arterial tree. With de-tector configuration settings such as 16� 0.625 mm or 16 � 0.75 mm, it ispossible to reconstruct sections lessthan 1 mm in thickness spaced every0.5–0.8 mm. With these settings, theacquisition speed is somewhat slower,in the range of 40–50 seconds, which iscomparable to routine four-channel (4� 2.5 mm) and eight-channel (8 � 1.25mm) system acquisitions. These isotro-pic resolution data sets further im-prove the visualization of small cruralor pedal vessels; however, image noiseand increased dose and tube current

requirements may be problematic inthe abdomen unless automated tubecurrent modulation is available.

However, it is important to bear inmind that submilllimeter acquisition(16 � 0.625 mm, 16 � 0.75 mm) doesnot necessarily require submillimeterimage reconstruction (ie, reconstruct-ing the thinnest possible images). Forexample, one might routinely chooseto reconstruct thicker high-resolution(eg, 1.25 mm) data sets from a submil-limeter acquisition. This strategy stillallows the user to go back and recon-struct another isotropic submillimeterdata set at maximum spatial resolu-tion if clinically necessary. It is ourexperience that the reconstruction of asingle high-resolution data set (ap-proximately 1.25–1.5 mm sectionthickness) obtained with eight- and 16-channel CT and 1-mm section thick-ness with 64-channel CT provides ad-equate spatial resolution for the entireperipheral arterial tree while keepingthe noise level in the abdomen andpelvis within an acceptable range.

Sixty-four–channel CT

Data acquisition with 64-channelCT normally ensues on a submillime-ter scale (64 � 0.6 mm or 64 � 0.625

mm). Because of peripheral arterial en-hancement dynamics, it is importantto deliberately slow the acquisitionspeed with these scanners by selectinga low pitch and refraining from usingthe maximum gantry rotation speed,most notably in patients with steno-occlusive disease.

Again, one can use the raw datafrom the submillimeter acquisition toroutinely reconstruct data sets with asection thickness of 1.0–1.5 mm (ie,high-resolution; Fig 5). Submillimeterisotropic images with a section thick-ness as small as 0.6–1.0 mm, spacedevery 0.4–0.7 mm, may be recon-structed from the same acquisition.This maximum spatial resolution maytranslate into improved visualizationand treatment planning of patientswith advanced peripheral arterial oc-clusive disease.

Peripheral CT angiography datasets are generally large, ranging from900 to 2,500 images. We currently saveall images in our Picture Archivingand Communication System, but thismay not be feasible for all institutions.One potential solution to reduce filesizes is to permanently archive 1.5–2-mm-thick images, and use thinner sec-tions (eg, submillimeter) for dataviewing only and for the generation of

Table 1CT Acquisition Parameters for Peripheral CT Angiography

Equipment

GantryRotation

Time (sec)

DetectorConfiguration

(channels � mm) Pitch

TableIncrement(mm/360°)

TableSpeed

(mm/sec)

ScanTime(sec)†

InjectionProtocol(Type)

Four channelsGE 0.8 4 � 2.5 1.5 15 19 69 SlowSiemens 0.5 4 � 2.5 1.5 15 30 43 Slow

Eight channelsGE 0.5 8 � 1.25 1.35 13.5 27 48 SlowGE 0.5 8 � 2.5 1.35 27 54 24 Fast

Sixteen channelsGE 0.6 16 � 1.25 1.375 35 46 28 FastSiemens 0.5 16 � 1.5 1.2 33 58 23 FastSubmillimeterGE 0.5 16 � 0.625 1.375 17.5 28 47 SlowSiemens 0.5 16 � 0.75 1.2 18 29 45 Slow

Sixty-four channelsSiemens 0.5 64 � 0.6‡ 0.85 17 32 40 Slow§Siemens 0.33 64 � 0.6‡ 1.1 21.1 63 20 FastGE* 0.7 64 � 0.625 0.5625 22.5 32 40 SlowGE* 0.6 64 � 0.625 0.9375 37.5 63 20 Fast

* Sixty-four–channel GE scanners not been used in practice by the authors at time of writing.† Scan times shown for a scanning range of 130 cm.‡ Physical detector configuration is 32 detector rows.§ Scan time fixed to 40 sec.

Fleischmann et al • 5Volume 17 Number 1

Figure 2. Four-channel CT (4 � 2.5 mm, 3.0 mm/1.0 mm) peripheral CT angiogram of a 62-year-old man with abdominal and bilateralcommon iliac artery aneurysms and subacute onset of right foot pain. (a) Oblique (45° left anterior oblique) MIP image of entire dataset. Box indicates magnified views. (b–d) axial CT images through the right proximal calf show embolic filling defects in the anteriortibial artery (arrowheads) and the tibioperoneal trunk (arrows). (e,f) Corresponding CPR images from the popliteal artery through theanterior tibial artery (e) and posterior tibial artery (f) display intraluminal filling defects. DSA confirms CT angiography findings (g).


permanent high-quality reformattedimages.

Contrast Medium InjectionTechnique

Intravenous contrast medium is in-jected with a power injector into anantecubital vein with use of a 20-gauge intravenous cannula. The basicprinciples of contrast medium injec-tion for CT angiography, such as therelationship of the injection flow rateand the injection duration on arterialenhancement, also apply to peripheralCT angiography, at least for its aor-toiliac portion (14). However, periph-eral CT angiography is more complex

with respect to synchronizing the en-hancement of the entire lower-extrem-ity arterial tree with the CT data acqui-sition speed.

One to 1.5 g of iodine injected persecond usually achieves adequate ar-terial enhancement for an average (75kg) person. Body weight–based ad-justments of the injection flow rate andvolume are recommended, at least forthose subjects who weigh more than90 kg or less than 60 kg. The injectionduration also affects the time course ofarterial enhancement. With a continu-ous intravenous injection of contrastmedium over a prolonged period oftime (eg, 35 seconds), arterial enhance-ment continuously increases over

time (15). This explains why the at-tenuation values observed in periph-eral CT angiograms are usually low-est in the abdominal aorta and peakat the level of the infragenicular pop-liteal artery (7). In general, biphasicinjections result in more uniform en-hancement over time, notably withlong scan and injection times (�25–30seconds) (16).

Principles of Scan Timing

The time interval between the be-ginning of an intravenous contrast ma-terial injection and the arrival of thebolus in the aorta, referred to as thecontrast medium transit time, is very

Figure 3. Eight-channel CT angiogram (8 � 1.25 mm, 1.25 mm/0.9 mm) of the chest, abdomen, pelvis, and lower-extremity arteries(length, 150 cm) obtained in 45 seconds with use of a pitch of 1.675 (a). This 48-year-old woman with an occluded right axillofemoralbypass (not opacified) was referred for surveillance of a recent thoracic aortobifemoral bypass (b) and bilateral femoropopliteal bypassgrafts (c,d).


variable among patients with coexist-ing cardiocirculatory disease, and mayrange from 12 to 40 seconds. Individ-ualizing the scanning delay is there-fore generally recommended in pe-ripheral CT angiography. A patient’sindividual contrast medium transittime can be reliably determined with asmall test bolus injection or estimatedwith use of automated bolus trigger-ing techniques. The scanning delaymay then be chosen to equal the con-trast medium transit time (the scan istherefore initiated as soon as contrastmedium arrives in the aorta), or thescanning delay may be chosen at apredefined interval after the contrastmedium transit time. For example, thenotation “contrast medium transittime � 5 seconds” means that the scanstarts 5 seconds after contrast mediumhas arrived in the aorta.

Aortopopliteal Bolus Transit Times

The additional challenge in patientswith peripheral arterial occlusive dis-ease is related to the well-known factthat arterial stenoses, occlusions, oraneurysms anywhere between the in-frarenal abdominal aorta and thepedal arteries may substantially delaydownstream vascular opacification(17,18). More specifically, we found ina group of 20 patients with peripheralarterial occlusive disease that the tran-sit times of intravenously injected con-trast medium to travel from the aortato the popliteal arteries to range from 4seconds to 24 seconds (mean, 10 sec),which corresponds to bolus transitspeeds as fast as 177 mm/sec to asslow as 29 mm/sec, respectively (19).

The clinical implication for periph-eral CT angiography is that when atable speed of approximately 30 mm/sec is selected, it is very unlikely (al-though not impossible) that the dataacquisition is faster than the intravas-cular contrast medium bolus. How-ever, with increasing acquisitionspeeds, the scanner table may movefaster than the intravascular contrastmedium column, and the scanner maytherefore “outrun” the bolus. Of note,outrunning the bolus has been re-ported in only one study to our knowl-edge, which used a table speed of 37mm/sec (9), but it has not been re-ported in five other published studieson peripheral CT angiography, all of

Figure 4. Sixteen-channel CT (16 � 1.25 mm, 1.25 mm/0.8 mm) peripheral CT angio-gram in a 70-year-old man with right thigh claudication and right iliofemoral arteryocclusion demonstrates exquisite detail of small collateral vessels (a), as well as femoro-popliteal (b), pedal (c), and crural arteries (d).


Figure 5. Sixty-four–channel CT (64 � 0.6 mm, 1.0 mm/0.7 mm) peripheral CT angiogram in an 83-year-old man with rightside–dominant calf and foot claudication. Automated tube current modulation (CareDose 4D; Siemens) allows submillimeter acquisi-tion and reconstruction with acceptable image noise level in the abdomen and unprecedented spatial resolution down to the plantar archand metatarsal branches. MIP of the entire data set (a); VR views of the abdomen (b) and right leg runoff (c–f) show distal aortic calcificplaque and patent renal arteries (b) occlusion of the right popliteal trifurcation with reconstitution of the peroneal artery (c), whichreconstitutes the right posterior tibial artery above the ankle via a communicating branch (d), and supplies the foot through the commonand lateral plantar arteries (e) which fill the plantar arch (f).


which used table speeds of 19–30mm/sec (7,8,10,11,20).

For the following discussion, it istherefore useful to arbitrarily catego-

rize injection strategies for peripheralCT angiography into those for slowacquisitions (�30 mm/sec tablespeed), and those for fast acquisitions

(�30 mm/sec table speed). Sixty-four–channel CT injection strategies will bediscussed separately.

Injection Strategies for SlowAcquisitions

Detector configuration settings of 4� 2.5 mm, 8 � 1.25 mm, and 16 �0.625 mm all translate into acquisitionspeeds of approximately 30 mm/sec.Such a table speed usually translatesinto a scan time of approximately 40seconds for the entire peripheral arte-rial tree. Because the data acquisitionfollows the bolus from the aorta downto the feet, the injection duration canbe chosen to be approximately 5 sec-onds shorter than the scan time. Forexample, for a 40-second acquisition, a35-second injection duration is suffi-cient. This would translate into 140 mLof contrast medium if a constant injec-tion rate of 4 mL/sec was used. If thebeginning of the data acquisition istimed closely to the contrast mediumarrival time in the aorta (with use of atest bolus or bolus triggering), bipha-sic injections achieve more favorableenhancement profiles with improvedaortic enhancement. As an example,our current protocol for a submillime-ter acquisition with a Siemens 16-channel scanner is shown in (Table 2).A similar concept is used for the 64-channel Siemens scanner at our insti-tution as well (as described later).

Because the possibility of evenmore delayed arterial opacificationthan accounted for in the aforemen-tioned protocol cannot be excluded(19), a second CT angiography acqui-sition (covering the popliteal andinfrapopliteal vasculature) should bepreprogrammed into the scanningprotocol (Fig 1). This acquisition is ini-tiated by the CT technologist immedi-ately after the main CT angiographyacquisition only if he or she does notsee any contrast medium opacificationin the distal vessels.

Injection Strategies for FastAcquisitions

Detector configuration settings of 8� 2.5 mm, 16 � 1.25 mm, or 16 � 1.5mm translate into acquisition speedsof 45–65 mm/sec, which, in some in-dividuals, may be faster than the con-trast medium bolus travels through adiseased peripheral arterial tree. To

Table 2Peripheral CTA Injection Protocol for Slow Acquisitions

Parameter Specification

Table speed � 30 mm/secScanning time � 40 secondsExamples of acquisition

parameters4 � 2.5 mm, 8 � 1.25 mm, 16 � 0.75 mm, 16 � 0.625

mm, 64 � 0.6 mm, 64 � 0.625 mm (slow pitch andgantry rotation; see Table 1)

Injection duration Scan time minus 5 seconds (�35 sec)Scanning delay Equal to contrast medium transit time (contrast

material arrival in the aorta, as determined by atest bolus or automated bolus triggering)

Injection flow ratesBiphasic 5–6 mL/sec (1.8 gI/sec) for 5 seconds, plus

3–4 mL/sec (1.0 gI/sec) for remaining seconds (scantime minus 10)

Example: for anacquisition time of 45

30 mL at 6 mL/sec plus 115 mL at 3.3 mL/sec (300mgI/mL concentration contrast medium);

seconds, inject 25 mL at 4.5 mL/sec plus 85 mL at 2.5 mL/sec (400mgI/mL concentration contrast medium);

Total injection duration is 35 sec (5 sec plus 30 sec)Total CM volume is 145 mL or 110 mL, respectively

Note.—Injection flow rates vary depending on the iodine concentration of thecontrast agent used. Iodine injection rate is adequate for a 75-kg individual andshould be increased or decreased for subjects who weigh more than 90 kg or lessthan 60 kg.

Table 3Peripheral CT Angiographic Injection Protocol for Fast Acquisitions in Patientswith Peripheral Arterial Occlusive Disease


Table speed � 30 mm/secScanning time � 40 secExamples of acquisition 8 � 2.5 mm

parameters 16 � 1.25 mm, 16 � 1.5 mm, 64 � 0.6 mm, 64 � 0.625mm (for pitch and rotation; see Table 1)

Injection duration 35 secScanning delay: Contrast medium transit time plus (40 sec minus

scanning time)*Injection flow rates: 1.5 gI /sec (4–5 mL/sec)Examples: Always inject for 35 seconds, eg 140 mL at 4 mL/sec

For a scan time of 30 sec, start scanning 10 sec aftercontrast medium arrived in aorta;

For a scan time of 25 sec, start scanning 15 sec aftercontrast medium arrived in aorta;

For a scan time of 20 sec, start scanning 20 sec aftercontrast medium arrived in aorta

Note.—The term “40 sec minus scanning time” is also referred to as the diagnosticdelay. Iodine injection rate of 1.5 g/sec is adequate for a 75-kg individual, andshould be increased/decreased for subjects who weigh more than 90 kg or less than60 kg (20 mgI/kg body weight/sec). Injection flow rates will vary depending oniodine concentration of contrast medium.


prevent the CT acquisition to outrunthe bolus, it is necessary to allow thebolus a “head start.” This is accom-plished by combining a fixed injectionduration of 35 seconds to fill the arte-rial tree with a delay of the start of theCT acquisition relative to the time ofcontrast medium arrival in the aorta.The faster the acquisition, the longerthis diagnostic delay should be. Weuse such a strategy with a GE 16-chan-nel scanner with a 16 � 1.25 mm pro-tocol, pitch of 1.375, and 0.6 secondsgantry rotation period (table speed, 45mm/sec). The protocol, as well as itsgeneralization to other fast acquisi-tions, is shown in Table 3. Examplesof arterial opacification with use ofthis approach on fast scans are shownin Figure 4. The limitation of the use ofa very long diagnostic delay (�15 sec-onds between contrast medium arrivaland scan initiation) is undesirableopacification of the renal and portalveins and the inferior vena cava.

Injection Strategy for 64-ChannelCT

Whereas 32-, 40-, and 64-channelCT systems theoretically allow acqui-sition speeds of 80 mm/sec and more,we deliberately acquire our peripheralCT angiograms at a much slower paceby prescribing a fixed scan time of 40seconds for each individual. For a 40-

second scan time, we select a gantryrotation of 0.5 seconds and a pitch ofless than 1. Automated tube currentmodulation is used in this setting toavoid increased radiation dose. Theadvantage of always selecting thesame 40-second scan time is that it canbe combined with fixed (biphasic) in-jections for 35 seconds. The injectionflow rates and volumes are then indi-vidualized to patient weight (Table 4).Examples of image quality and opaci-fication are shown in Figure 5.

All these acquisition and contrastmedium injection strategies reduce therisk of outrunning the bolus; however,this is not entirely excluded. Figure 6shows an example of a patient withextremely delayed flow, most likelycaused by decreased cardiac outputand altered flow dynamics resultingfrom bilateral iliac and popliteal arteryaneurysms and diffuse arteriomegaly.

Other Injection Strategies

The aforementioned injection strat-egies combine individual scan timing(with use of bolus tracking/test bolus)in the aorta, with empiric (ie, nonindi-vidualized) parameter selection to ad-just for a broad range of possible bolustransit speeds down to the feet. Otherapproaches with use of more or lessindividualization are also possible. Forexample, protocols with a fixed, long

scanning delay (eg, 28 seconds) havebeen used successfully in the past, no-tably before automated bolus trackingtechnique was available on the firstfour-channel scanners (11). On theother end of the spectrum are attemptsto individualize the scanning delayand the scanning speed based on aor-tic and popliteal transit times, deter-mined individually with two test bo-luses (21).

Venous Enhancement

Opacification of deep and superfi-cial veins can be observed in periph-eral CT angiography (7) and is morelikely to occur with longer scan timesand in patients with active inflamma-tion, such as that from infected ornonhealing ulcers. Given the rapid ar-teriovenous transit times observed an-giographically in some patients (22),venous opacification cannot be com-pletely avoided. Because arterial en-hancement is always stronger than ve-nous enhancement when the injectionis timed correctly (7), and with ade-quate anatomic knowledge and post-processing tools, venous enhancementrarely poses a diagnostic problem.

Visualization and ImageInterpretation

Visualization, image interpretation,and effective communication of find-ings are demanding in peripheral CTangiography. A powerful medical im-age postprocessing workstation, aswell as standardized workflow and vi-sualization protocols, are requiredwhen peripheral CT angiograms areobtained on a regular basis.

Various two- and three-dimen-sional (3D) postprocessing techniquesare available on today’s state-of-the-art workstations. For patients withatherosclerotic disease, a combinationof 3D overview techniques with atleast one two-dimensional cross-sec-tional technique is generally required.The selection of postprocessing tech-niques not only depends on the type ofdisease, but also on the specific visu-alization goal: interpretation versusdocumentation. In the context of inter-active exploration of the data sets,usually done by the radiologist inter-preting the study during readout, fastnavigation and flexible visualizationtools are preferred. In the setting of

Table 4Weight-based Biphasic Injection Protocol for 64-Channel Peripheral CTAngiography for Patients with Peripheral Arterial Occlusive Disease


Table speed Variable (approximately 30 mm/sec) depending onlongitudinal coverage

Scanning time 40 sec (all patients)Acquisition parameters 64 � 0.6 mm

Gantry rotation period: 0.5 sec; pitch usually � 1120 kV, 250 quality reference mA (CareDose 4D)

Injection duration 35 sec (all patients)Scanning delay Contrast medium transit time � 3 sec (minimum

delay with CareBolus, including breath-holdcommand)

Biphasic injection Maximum flow rate for first 5 seconds of injection,continued with 80% of this flow rate for 30seconds and a saline solution flush

�55 kg body weight 20 mL (4.0 mL/sec) � 96 mL (3.2 mL/sec)�65 kg body weight 23 mL (4.5 mL/sec) � 108 mL (3.6 mL/sec)75 kg body weight* 25 mL (5.0 mL/sec) � 120 mL (4.0 mL/sec)�85 kg body weight 28 mL (5.5 mL/sec) � 132 mL (4.4 mL/sec)�95 kg body weight 30 mL (6.0 mL/sec) � 144 mL (4.8 mL/sec)

* Average.


Figure 6. Peripheral CT angiogram (16 � 1.25 mm, 1.25 mm/0.9 mm) inan 82-year-old man with arteriomegaly, bilateral common iliac artery aneu-rysms, and diffuse ectasia of the femoropopliteal arteries. Frontal VR image(a) demonstrates good aortoiliac opacification. Posterior-view VR image (b)shows gradually decreasing enhancement of the femoropopliteal arteriesbilaterally, with complete lack of enhancement of the popliteal trifurcationand crural arteries. A second phase of CT angiography acquisition was ob-tained from above the knees down to the feet immediately after the firstacquisition. Posterior view of second phase of CT angiography (c) showsinterval opacification of the bilateral popliteal and crural arteries down tothe feet (d).


creating standardized sets of static im-ages and measurements for documen-tation and communication, techniquesthat allow a protocol-driven genera-tion of predefined views are preferred.The typical protocol for postprocess-ing of peripheral CT angiography inour 3D laboratory consists of curvedplanar reformations (CPRs); thin-slabmaximum-intensity projections (MIPs)through the renal and visceral arteries;bone removal; full-volume MIPs andvolume renderings (VRs) of the abdo-men, pelvis, and each leg; and filmingand archiving of these views. Al-though the creation of such sets of im-ages may be time-consuming even foran experienced technologist, the clini-cal interpretation is fast and straight-forward for the image recipient.

Transverse Source Image Viewing

Reviewing of transverse CT slices ismandatory for the assessment of ex-travascular abdominal or pelvicpathologic processes. This is facilitatedby reconstructing an additional seriesof contiguous 5-mm-thick sectionsthrough the abdomen and pelvis.Browsing through the stack of sourceimages is also helpful to gain a firstimpression of vascular abnormalities,and in select cases, such as in patientswith minimal or absent disease or pa-tients with trauma or suspected acuteocclusions, source image viewing maybe sufficient to completely address theconcern (Fig 2). Axial images also dis-play relevant extravascular anatomy,such as the course and position of themedial head of the gastrocnemiusmuscle in popliteal entrapment syn-drome, which may not be apparent onimages such as MIPs. Source imagesmay also serve as a reference whentwo-dimensional or 3D reformattedimages suggest artifactual lesions.However, for the majority of caseswith vascular disease, transverse im-age viewing is inefficient and less ac-curate than viewing reformatted im-ages (10).

Postprocessing Techniques

MIP.—Assessment of vascular ab-normalities is facilitated when the ar-terial tree is displayed in an angio-graphic fashion. This can be accom-plished with MIP or VR techniques.MIP provides the most “angiogra-

Figure 7. Peripheral CT angiography (16 � 1.25 mm, 1.25 mm/0.9 mm) in a 72-year-oldwoman with nonhealing left forefoot ulcer. (a) VR image of the left superficial femoralartery shows excessive vessel wall calcifications, precluding the assessment of the flowchannel. Cross-sectional views were required to visualize the vessel lumen. Axial CTimages (b,c) through the mid-superficial femoral artery (dotted line in a and d) withviewing window settings (level/width) of 200 HU/600 HU (b) does not allow us todistinguish between opacified vessel lumen and vessel calcification, which can be distin-guished only when an adequately wide window width (300 HU/1,200 HU) is used (c).Similar wide window settings are also used for a CPR through the same vessel (d),displaying several areas of wall calcification with and without stenosis.


phy-like” display of the vasculature,particularly when no or minimal ves-sel calcifications are present. MIP istherefore ideal for communicatingfindings to the referring services andfor creation of a “road map” fortreatment planning within the an-giography suite or the operatingroom (Figs 2,5). The disadvantage ofMIP is that it requires that bones areremoved from the data set, and evenwith the help of automated or semi-automated computer algorithms onmodern workstations, this remains atime-consuming task. Additionally,inadvertent removal of vessels inclose vicinity to bony structures maylead to spurious lesions.

Volume rendering.—Because VRpreserves 3D depth information, un-like MIP, bone editing may not be re-quired. Rather than editing the bonesout of the data set, one can use clipplanes or volume slabs together withinteractive selection of the appropri-ate viewing angles to expose the rele-vant vascular segment. Interactiveadjustment of the opacity transferfunction allows the user to blend inor carve out exquisite vascular detailwhen necessary. VR is the ideal toolfor fast interactive exploration of pe-ripheral CT angiography data sets.Although “snapshot” views obtainedduring data exploration can be in-valuable for communicating a spe-cific finding or detail, VR is some-what less suited for standardizeddocumentation of images. In part,this is because most Picture Archiv-ing and Communication Systemworkstations do not display the colorinformation on high-resolution gray-scale monitors.

The main limitation of MIP andVR is that vessel calcifications andstents may completely obscure thevascular flow channel (Fig 7). Thisfundamental limitation precludes theexclusive use of these techniques in asubstantial proportion (approxi-mately 60%) of patients with periph-eral arterial occlusive disease (23).

Multiplanar reformation and CPR.—In the presence of calcified plaque,diffuse vessel wall calcification, orendoluminal stents, cross-sectionalviews are essential to assess the vas-cular flow channel (Fig 7). Trans-verse source images, sagittal, coronal,or oblique multiplanar reformations(MPRs) are useful in an interactive

setting, such as in conjunction withVR.

Alternatively, longitudinal cross-sections along a predefined vascularcenter line, ie, CPR, can be created(Fig 8). CPRs provide the most com-prehensive cross-sectional display ofluminal pathologic processes, but re-quire manual or (semi-)automatedtracing of the vessel center lines(24,25). CPR does not require boneediting, but at least two CPRs pervessel segment (eg, sagittal and coro-nal views) have to be created to fullyevaluate eccentric disease.

One problem of (single) CPR im-ages in the context of visualizing theperipheral arterial tree is their lim-ited spatial perception. With bonylandmarks out of the curved recon-struction plane, the anatomic contextof a vascular lesion may be ambigu-ous unless clear annotations arepresent. This limitation has recentlybeen alleviated by an extension ofstandard CPR to so-called multipathCPR images (24). Multipath CPR im-ages provide simultaneous longitudi-nal cross-sectional views through themajor conducting vessels (Fig 9)without obscuring vessel wall calcifi-cations and stents, while maintainingspatial perception (26). We are cur-rently evaluating this technique as aroutine clinical visualization, docu-mentation, and endovascular treat-ment-planning tool for peripheral CTangiography.

Thin-slab MIP, thin-slab VR, andthick MPR/CPR.—If applied to sub-volumes of the data sets with use ofclip planes or slabs of the volume tofocus on a particular area of interest,the resulting images are referred toas thin-slab MIPs, thin-slab VRs, orthick MPRs (or CPRs) (27,28). Inter-active real-time variation of the view-ing direction, the thickness of therendered volume, adjustments towindow-level settings or opacitytransfer functions, etc, is ideal for in-teractively exploring the data sets.

Automated Techniques forSegmentation and Visualization

Although fully automated detec-tion of vessel centerlines, automatedsegmentation of bony structures, anddetection (and subtraction) of vesselwall calcification are highly desirable,no such algorithms have yet been de-

veloped. This is not surprising whenconsidering the complex manifesta-tions of vascular disease; the widerange of vessel sizes; the wide overlapin CT density values of opacifiedblood, plaque, and low-attenuationbone; and the inherently limited spa-tial resolution and image noise inperipheral CT angiography data sets.Although traditional density- and gra-dient-based algorithms are unlikely tocompletely solve the problem in thenear future, several software tools forimproved and faster editing and forcreating center lines through the arter-ies have been developed and are avail-able on modern 3D workstations.

At this point it is therefore reason-able to expect further improvementsin computer-assisted segmentationand visualization in the near future,but it would be naı̈ve to expect thatexpert user interaction, such as by aradiologist or 3D imaging technolo-gist, can be completely avoided in thecreation of clinically relevant and rep-resentative images.

Interpretation and Pitfalls

Vascular abnormalities have to beinterpreted in the context of a patient’ssymptoms and stage of disease, andwith respect to the available treatmentoptions. For those familiar with read-ing conventional angiograms, the clin-ical aspect of interpreting a peripheralCT angiogram is usually straightfor-ward. However, visual perception andinterpretation of well-known abnor-malities in a new and different format(such as VR or CPR images) requiresadaptation and familiarity with thespecific techniques used.

Probably the most important pitfallrelated to the interpretation of periph-eral CT angiograms is related to theuse of narrow viewing window set-tings in the presence of arterial wallcalcifications or stents. Even at wider-than-normal CT angiographic windowsettings (window level/width,150/600 HU), high-attenuation objects(eg, calcified plaque, stents) appearlarger than they really are (“bloom-ing” caused by the point-spread func-tion of the scanner), which may lead toan overestimation of a vascular steno-sis or suggest a spurious occlusion.When scrutinizing a calcified lesion ora stent-implanted segment with use ofany of the cross-sectional grayscale


Figure 8. Peripheral CT angiography (16 � 0.75 mm, 2.0 mm/1.0 mm) in a 73-year-old woman with intermittent claudicationbilaterally. MIP (a) shows long right femoropopliteal occlusion (curved arrow) and diffuse disease of the left superficial femoral arterywith a short distal near-occlusion. CPR (b) through left iliofemoral arteries demonstrates multiple mild stenoses of the external iliacartery (arrowheads), a diffusely diseased left superficial femoral artery, and short (�3 cm) distal left superficial femoral artery occlusion.Corresponding selective DSA images of the left external iliac artery (c) and the distal left superficial femoral artery (d) were obtainedimmediately before angioplasty/stent implantation.


Figure 9. Peripheral CT angiography (16 � 0.75 mm, 2.0 mm/1.0 mm) of a diabetic male patient with bilateral claudication. MIP (a)shows arterial calcifications near the aortic bifurcation (arrow), as well as in the right (arrowheads) and left common femoral arteries,in the right femoropopliteal region, and in the crural vessels. A long stent is seen in the left femoropopliteal segment (curved arrow).Frontal view (b) and magnified 45° left anterior oblique (c) multipath CPR images provide simultaneous CPRs through the aorta andbilateral iliac through crural arteries. Note that prominent calcifications cause luminal narrowing in the proximal left common iliacartery (arrow) and in the right common femoral artery (arrowheads). The left common femoral artery is normal; the long femoropop-liteal stent is patent (curved arrow). Mixed calcified and noncalcified occlusion of the right distal femoral artery is also seen (openarrow).


images (eg, transverse source images,MPR, or CPR), a viewing windowwidth of at least 1,500 HU may berequired (Fig 7). Interactive windowadjustment on a Picture Archiving andCommunication System viewing sta-tion or on a 3D workstation are mosteffective because, when printed onfilm, window settings are usually toonarrow. In the setting of extensive ath-erosclerotic or media calcificationwithin small crural or pedal arteries,such as those found in diabetic pa-tients and in patients with end-stagerenal disease, the lumen may not beresolved regardless of the window/level selection. In these circumstances,other imaging techniques, notablymagnetic resonance (MR) imaging,may be preferable to CT angiography.

Other interpretation pitfalls resultfrom misinterpretation of editing arti-facts (eg, inadvertent vessel removal)in MIP images and pseudostenosisand/or occlusions in CPRs resultingfrom inaccurate center-line definition.Most of these artifacts are obvious oreasily identified when additional

views, complimentary viewing modal-ities, or source images are reviewed.

Accuracy and Clinical Perspective

For clinical conditions involvingthe lower-extremity vascular struc-tures, a variety of imaging choices areavailable. Ultrasound (US), CT an-giography, and MR angiography allcan provide useful information aboutlower-extremity arteries and veinsnoninvasively. The particular modal-ity of choice depends on many factors,including patient demographics andcomorbidities, availability and moder-nity of respective imaging equipment,level of training and confidence of theoperating technologists, and local in-terest and expertise of the radiologist.Each of these factors must be consid-ered when choosing the best imagingexamination for an individual patient.Multiple–detector row CT angiogra-phy has the advantages of widespread(and increasing) availability, high spa-tial resolution, and relative freedomfrom operator dependence.

Lower-extremity CT angiography

(introduced in 1998–1999) is the new-est technique for peripheral arterialimaging. Therefore, it is not surprisingthat only sparse original data on itsaccuracy in patients with peripheralarterial occlusive disease are available(Table 5) when compared with thebody of published data on US or MRangiography. The majority of pub-lished articles on peripheral CT an-giography report results with four-channel CT scanners (7–11,29). Thesesmall early series do not allow for clin-ically meaningful stratification of pa-tients into those with claudication andthose with limb-threatening ischemia.Other difficulties in interpreting theresults of the published literature arethe inconsistent thresholds used forgrading of significant stenoses, vari-able categorization of anatomic vesselsegments, and use of different visual-ization and postprocessing tools. Allbut one group of authors (29) reportgood overall sensitivities and specific-ities for the detection of hemodynam-ically relevant steno-occlusive lesionswith four-channel CT relative to in-traarterial digital subtraction angiog-

Table 5Lower-extremity CT Angiography in Comparison with DSA in Patients with Peripheral Arterial Occlusive Disease(8–11,19,20,30)

Study, YearCollimation

(channels � mm)STh /RI

(mm/mm)No. of

Pts.Sensitivity

(%)Specificity

(%)Accuracy

(%)

Ofer et al., 2003 (8)�50% stenosis 4 � 2.5 3.2/1.6 18 91 92 92Occlusion – – –

Martin et al, 2003 (9)�75% stenosis 4 � 5 5/2.5 41 92 97 –Occlusion 89 98 –

Ota et al, 2004 (10)�50% stenosis 4 � 2 2/1 24 99 99 99Iliac/femoral/crural 97/100/100 100/96/100 99/97/100Occlusion 96 98 98

Catalano et al, 2004 (11)�50% steno-occlusion 4 � 2.5 3/3 50 96 93 94Aortoiliac/femoral/popliteocrural 95/98/96 90/96/93 92/97/94

Portugaller et al., 2004 (20)�50% steno-occlusion 4 � 2.5 2.5/1.5 50 92 83 86Aortoiliac/femoropopliteal/crural 92/98/90 95/70/74 94/88/81

Edwards et al., 2005 (29)�50% stenosis 4 � 2.5 3.2/2.0 44 79/72 93/93 –Occlusion 75/71 82/81 –

Willmann et al., 2005 (30)�50% steno-occlusion 16 � 0.75 0.75/0.4 39 96/97 96/97 96/97Aortoiliac 95/99 98/98 97/98Femoral 98/97 94/96 96/96Popliteocrural 96/97 95/96 96/96

Note.—STh/RI � section thickness/reconstruction interval.


Figure 10. Intermittent left leg claudication in a 62-year-old woman with a history of tobacco use and aortobifemoral bypass grafting.The ankle-brachial index was 0.65. (a–d) MIP images with bone segmentation; (e–h) DSA images obtained before treatment. (a) ObliqueMIP image shows high-grade stenosis (arrows) at the origin of the left profunda femoris artery (P); a previously placed aortobifemoralgraft (G) is noted, as is a patent superficial femoral artery (SFA). (b) Coronal MIP of the left thigh demonstrates multifocal moderate tosevere stenosis in the SFA (arrowheads). The SFA is small in caliber with soft and calcified plaque present. (c) Coronal MIP of the calfshows a one-vessel runoff (peroneal; PER) to the left foot. Mild venous contamination (V) is present. (d) Sagittal MIP image of the leftfoot shows collateral vessel reconstitution (arrowheads) of the dorsalis pedis (DP) above the ankle from the peroneal artery. (e) DSAimage from selective catheterization of the left profunda femoris artery corroborates the CT angiographic finding of high-gradeprofunda femoris artery (P) stenosis (arrows). (f) DSA image of left superficial femoral artery shows multiple focal stenoses (arrow-heads) in the same segment of the superficial femoral artery as demonstrated by CT angiography. Note the lack of calcium visualizationon the subtracted image from DSA. (g) DSA image of the calf confirms single peroneal vessel runoff. (h) DSA image of left foot confirmsreconstitution of the dorsalis pedis artery (DP) from the peroneal artery (arrowheads).


Figure 11. Perianastomotic pseudoaneurysms with bilateral infrainguinal disease in an 87-year-old man with claudication and historyof aortobifemoral bypass grafting. (a) VR image demonstrates bilateral perianastomotic pseudoaneurysms at the distal attachment sitesof the aortobifemoral graft and common femoral arteries (arrows). There is a long-segment superficial femoral artery occlusion on theleft (arrowheads) with collateral vessels from the profunda femoris artery (c). (b) Thick-slab VR image of the right groin (left lateral view)shows the profile of the pseudoaneurysm (arrow), as well as the adjacent native external iliac artery (arrowheads). The proximal andmid-right superficial femoral artery is patent (SFA). T, ischial tuberosity. (c) VR image of the left thigh demonstrates abundant profundacollateral supply (c), which reconstitutes the distal superficial femoral artery. The length of the occluded segment was approximately11 cm (arrowheads). (d) MIP of both knees shows occlusion of the right anterior tibial artery at its origin (arrowheads). There is a focalsignificant stenosis in the midportion of the left anterior tibial artery (arrow). Heavy popliteal artery calcific plaque is present bilaterally.(e,f) MIP of both feet with bone segmentation demonstrates patent posterior tibial arteries (PT) at both ankles into the feet. Peronealarteries (PER) are patent to the lower calves. (e) There is reconstitution of the right dorsalis pedis (DP) via peroneal collaterals(arrowheads). The left dorsalis pedis is patent. (f) The focal left anterior tibial stenosis is again identified (arrowhead).


raphy (DSA; Table 5) (8–11). In gen-eral, sensitivity and specificity aregreater for arterial occlusions than forthe detection of stenoses. Accuracyand interobserver agreement are alsogreater for femoropopliteal and iliacvessels compared with infrapoplitealarteries when four-channel CT is used.In the first study of submillimter col-limated 16-channel CT (30), the overallsensitivities, specificities, and accura-cies were all greater than 96% andthere was no evident decrease of per-formance down to the popliteocruralbranches. Pedal arteries have not beenspecifically analyzed in any publishedseries to our knowledge. The presenceof vessel calcifications apparently re-duces diagnostic performance of mul-tiple–detector row CT in general.

Ouwendijk and coworkers (31) re-cently evaluated the clinical utility, pa-tient outcomes and costs of peripheralMR angiography and 16-channel CTangiography for initial imaging anddiagnostic workup of 157 patientswith peripheral arterial disease. In thisrandomized prospective study, confi-dence was slightly greater with CTand patients in the CT group under-went less additional imaging studiesand had greater improvement of clin-ical outcomes after treatment, butnone of these differences were statisti-cally significant. However, average di-agnostic imaging costs were signifi-cantly lower with CT compared withMR angiography.

CURRENT CLINICALAPPLICATIONS

CT angiography is increasinglyused at many institutions, such as ourown, for imaging the lower-extremityvasculature over a wide range of clin-ical indications. Evaluation of athero-sclerotic steno-occlusive disease andits complications is the main applica-tion of CT angiography at our institu-tion. However, congenital abnormali-ties, traumatic and iatrogenic injuries,inflammatory conditions, drug toxic-ity, embolic phenomena, and aneurys-mal changes can also affect the arteriesof the lower extremities. CT angiogra-phy may be used in many of theseconditions.

Figure 12. Acute thrombosis of the femoropopliteal and trifurcation vessels in an84-year-old woman with an acutely cool right leg. The patient refused angiography. (a)Transverse CT angiographic image at the level of the adductor canal shows roundedfilling defect in the right popliteal artery (arrow). The contralateral left popliteal artery (P)is patent at this level. (b) Large field of view multipath CPR and (c) enlarged image ofpopliteal region show the extent of right-sided thrombus (arrowheads). In addition to thepopliteal artery (POP), the anterior tibial artery (AT), tibioperoneal trunk (TPT), andposterior tibial artery (PT) are occluded. F, femoral condyle. High-grade left poplitealstenosis (asterisk) is also noted.


Figure 13. Utility of peripheral CT angiography in monitoring bypass graft patency in a 74-year-old man with a surgically placedfemorofemoral bypass graft for chronic left iliac arterial occlusion. The patient presented with rest pain after recent surgical bypassprocedure. Pain and bandages prevented adequate Doppler imaging and physical examination. (a) Axial CT angiography shows lackof contrast material opacification of the femorofemoral bypass graft (BPG). Only the right common femoral artery is patent (arrow). (b)Volume-rendered image demonstrates only a short area of flow at the right bypass graft anastomosis (arrow). There is completeocclusion of the left iliac arterial system (asterisks) and reconstitution of the left profunda femoris artery by collateral vessels to thelateral femoral circumflex artery (arrowheads). (c) MIP image at the level of the thighs shows bilateral long-segment superficial femoralartery occlusions (arrowheads), with reconstitution of the popliteal arteries (P) via collateral vessels (C) from the profunda femorisarteries (arrows). (d) MIP image of calf arteries demonstrates three-vessel runoff in the left lower extremity. There is occlusion of theright posterior tibial artery (arrowhead) in the mid-calf. Segmentation artifact from automated bone removal is noted (asterisk). (e) MIPimage of the abdomen and bilateral groin region 3 days later demonstrates interval surgical revision of femorofemoral bypass graft(BPG), with restored patency. Left iliac occlusion is again noted (arrowheads).


Intermittent Claudication

In patients with claudication, treat-ment aims at improving blood supplyto the femoral and calf muscles, and istherefore usually limited to the aor-toiliac and femoropopliteal vascularterritories. Medical management in-cluding exercise regimens, tobaccocessation programs, and drug therapymay improve symptoms (32–34), butpatient compliance is often poor. Sur-gical or endovascular revasculariza-tion is then considered. For many pa-tients, angioplasty is more cost-effective than surgical intervention(35). The original American Heart As-sociation task force guidelines (36) andthe TransAtlantic Inter-Society Con-sensus report (37) have providedguidelines for appropriate treatmentof femoropopliteal disease. Conven-tional treatment of aortoiliac lesionsand short (�5 cm) femoropoplitealstenoses or occlusions (TransAtlanticInter-Society Consensus A or B) typi-

cally has been endovascular, whereaslonger femoral occlusions (Trans-Atlantic Inter-Society Consensus C orD) and occlusions involving the com-mon femoral artery may require sur-gical revascularization. Factors influ-encing the method of treatmentdepend on the lesion’s morphology(degree of stenosis/occlusion and le-sion length) (38) and location, andmost importantly, the status of runoffvessels. The STAR Registry report andothers have shown distal runoff vesselpatency (and diabetes) as the most im-portant lesion characteristic with re-gard to expected long-term patency(39). Additional improvements incatheter and guide wire design andtechniques such as subintimal recana-lization (40–43) have expanded therange of lesions that can be adequatelytreated percutaneously. The calf arter-ies are not a primary target for endo-vascular or surgical treatment in pa-tients with claudication, but are

important as a predictor of short- andlong-term patency rates after femoro-popliteal intervention.

CT angiography can provide com-plete delineation of the femoropopli-teal segment and inflow and outflowarteries, including lesion number,lengths, stenosis diameter and mor-phology, adjacent normal arterial cali-ber, degree of calcification, and statusof distal runoff vessels (Fig 10). Thesefactors allow precise preproceduralplanning with regard to route of ac-cess (Figs 8–10), balloon selection, andexpected long-term patency after in-tervention. Effects of eccentric steno-ses on luminal diameter reduction arealso well-defined by multiple–detec-tor row CT angiography, ie, a moreaccurate determination of diameter re-duction may be possible with multi-ple–detector row CT angiography ver-sus DSA alone (44). Status of collateralvessels is well-assessed on MIP andVR images, and arterial segments dis-tal to long-segment occlusions arewell-visualized (Fig 11). From a finan-cial perspective, preprocedural US orMR angiography has been shown tobe cost-effective relative to catheterangiography (45); it is expected thatCT angiography is also cost-effectivein this regard (45,46).

Chronic Limb-threatening Ischemia

Patients with chronic limb-threat-ening ischemia also benefit from treat-ment of aortoiliac and femoropopliteallesions, which are the initial target forintervention if present. Because thetreatment goal in this patient group isthe prevention of tissue loss and needfor amputation, assessment and pro-motion of blood flow through the calfarteries is of much greater importancethan in patients with intermittent clau-dication. A variety of endovasculartechniques have shown efficacy in thetreatment of infrapopliteal lesions(47,48). Creation of a road map tolesions amenable to angioplasty orother endovascular techniques and de-lineation of patent, acceptable targetvessels for distal bypass are the chal-lenges of vessel analysis in this ad-vanced disease setting. Use of verythin collimation (�1 mm) and optimi-zation of contrast medium administra-tion will provide improved visualiza-tion of these small vessels. The newestgeneration of 64-channel CT machines

Figure 14. Utility of peripheral CT angiography for preoperative vascular mapping in a24-year-old woman after a motor vehicle accident with tibial/fibular fractures and de-gloving injury. Preoperative CT angiography was requested to evaluate patency andposition of calf runoff vessels. (a) VR image with opacity transfer functions adjusted forskin detail. There is extensive soft-tissue injury including exposed bone (yellow arrow).An external fixator for tibial/fibular fracture is also noted. (b) Oblique MIP image of leftcalf shows bowing of the peroneal artery secondary to mass effect from adjacent displacedfibular fracture (white arrows). Popliteal (P) and posterior tibial arteries are patent(arrowheads). F, external fixator.


allows for further increase in spatialresolution and should allow improvedvisualization of small crural and pedalvessels. It is hoped these technical im-

provements translate to improved di-agnostic accuracy and treatment plan-ning; results from ongoing clinicalstudies are eagerly anticipated.

Aneurysms

Aneurysmal disease can affect anyportion of the lower-extremity arterialsystem. Popliteal artery aneurysms areof particular interest as a source ofdistal embolic material and as amarker for the presence of abdominalaortic aneurysm. Peripheral CT an-giography provides a rapid, noninva-sive, cost-effective alternative to cath-eter angiography for detection andcharacterization of lower-extremityaneurysms. CT angiography yields ro-bust data on aneurysm size, presenceand amount of thrombus, presence ofdistal embolic disease, associated sig-nificant proximal and distal steno-oc-clusive disease, and detection of coex-istent abdominal or iliac aneurysms(Fig 6). Three-dimensional volumetricanalysis provides accurate measure-ment of aneurysm volume and lumi-nal dimension.

Acute Ischemia

Few published data are availableregarding the use of CT angiographyin the setting of acute lower-extremityischemia. If immediate percutaneousintervention (eg, thrombolysis) isplanned, catheter angiography may bethe most appropriate choice. If urgentsurgical intervention is warrantedbased on the clinical status of the leg(ie, Rutherford criteria) (49), CT an-giography may add an unnecessarydelay to definitive therapy. CT angiog-raphy may be helpful in certain situa-tions to guide the choice of percutane-ous or surgical intervention and to aidin preprocedural planning. Determi-nation of the extent and location ofthrombosis can be accomplished byCT angiography. If thrombus or em-boli involves all trifurcation vessels ora previously patent bypass graft or re-sides within a popliteal aneurysm,thrombolytic therapy may be most ef-ficacious option (50). Demonstrationof thrombus in locations not accessibleto embolectomy could also direct ther-apy to catheter-based techniques.Large vessel thrombus or a recentlythrombosed bypass graft could be besttreated surgically (50). In the subacuteischemic setting, in which surgerymay be best, CT angiography offers acomplete overview of the affected vas-cular territories for optimum surgical

Figure 15. (a,b) Peripheral CT angiography of penetrating trauma in a 16-year-old malepatient with a gunshot wound to the left leg. (a) Posterior VR image and (b) obliquethin-slab MIP image of the left calf show a lobulated pseudoaneurysm arising from theperoneal artery (arrow). The immediate distal peroneal artery shows luminal narrowing,most likely spasm (arrowhead). The remainder of the trifurcation vessels and adjacentbony structures are intact. (c,d) Images in a 45-year-old man with groin hematoma aftercardiac catheterization. (c) VR image of the pelvis shows a rounded mass at the leftcommon femoral bifurcation (arrow), consistent with pseudoaneurysm. The contralateralright common femoral artery is normal (arrowhead). F, femoral head. (d) Sagittal thin-slab VR image with opacity transfer function set to render translucent vessel interiordemonstrates a narrow neck (arrowheads) of the pseudoaneurysm (�), which arises fromthe common femoral artery (c) immediately proximal to the bifurcation of the superficialfemoral artery (S) and profunda femoris artery (P). F, femoral head.


treatment. Patients who refuse cathe-ter angiography and/or thrombolysismay also be rapidly and adequatelyevaluated by CT angiography (Fig 12).A delayed phase CT angiography ac-quisition initiated immediately afterthe initial arterial phase is often help-ful to differentiate patent but slowlyflowing vessels from thrombus.

Follow-up and Surveillance afterSurgical or PercutaneousRevascularization

CT angiography is accurate and re-liable in the assessment of peripheralarterial bypass grafts and detection ofgraft-related complications, includingstenosis, aneurysmal changes, and ar-teriovenous fistulas (Figs 11,13). Thisis helpful in the immediate postoper-ative period, when US is limited by thepresence of bandages and wounds

(51). CT angiography can also demon-strate the results of percutaneous in-terventions, reveal residual disease,and diagnose vascular and extravas-cular complications.

CT angiography is not the firstchoice for routine bypass graft surveil-lance or for serial follow-up evaluationafter intervention (eg, in research pro-tocols) because US is easy to performand reliable in this setting (52,53).However, CT is a very convenientproblem-solving tool for the workupof patients with nondiagnostic USstudies (limited access as a result ofskin lesions, draping, or obesity),when US results are equivocal, orwhen ankle-brachial measurementsand US yield conflicting results. In thissetting, CT angiography has replacedcatheter DSA completely at our insti-tution and is used to decide furthermanagement.

Vascular Trauma

CT angiography, even with single–detector row CT, has been shown to bean accurate and useful test in the set-ting of suspected vascular trauma (5),and can replace diagnostic angiogra-phy. CT is easily accessible (oftenwithin or adjacent to the emergencydepartment), and examination time isshort. Moreover, CT angiography canbe performed in combination with CTof other organ systems (eg, abdomen,chest) for complete delineation of thedistribution and severity of injuries ineach individual organ system (54).Traumatic arterial injuries and rela-tionship of arterial segments to adja-cent fractures and soft-tissue injuriesare well-depicted (Fig 14). CT angiog-raphy reliably depicts hematoma andassociated vascular compression orpseudoaneurysm, and can simulta-neously show bone fragments. Bulletsor other metal artifacts, if present, maylimit small portions of the acquiredvolume.

Extremity CT angiography is alsoused for pediatric trauma patients(54). Particular attention to radiationdose and contrast medium is manda-tory in this instance. Limited scanrange, reduced dose settings, and alimit of 2 mL/kg body weight of con-trast medium are employed to maxi-mize patient safety.

Visualization is straightforward.For initial diagnosis, transverse im-ages are usually sufficient, althoughMPR image creation and viewing mayimprove rapidity of analysis (Fig 15).The addition of VR images improvesdepiction of the anatomic relationshipbetween arteries and adjacent bony/soft tissue injuries and foreign bodies.Generation of real-time 3D and/or VRimages also promotes rapid communi-cation to, and understanding by, refer-ring clinical services.

Vascular Mapping

Data obtained from peripheral CTangiography in the setting of traumaare also of use when subsequent sur-gical reconstruction is performed (Fig14). Additionally, precise knowledgeof arterial anatomy is paramountwhen plastic surgical reconstructionfor various diseases is considered. Fib-ular free flap procurement requirespreoperative assessment of the limb to

Figure 16. Images of adventitial cystic disease in a 55-year-old patient. (a) Axial sourceimage from peripheral CT angiography shows marked compression of the poplitealartery (P) by an ovoid fluid-density lesion (arrowheads). Note the predominant trans-verse compression of the flow lumen. F, fabella. (b) Corresponding sagittal thin-slab MIPimage shows craniocaudal extent of fluid-density lesion in the popliteal arterial wall(arrowheads). (c) VR image viewed obliquely from the posterior direction demonstratessevere narrowing of the right popliteal artery (arrows).


prevent ischemic complications andfailure of the flap and to exclude vari-ant peroneal artery anatomy and oc-clusive disease, which could alter thesurgical procedure (55). Previously,catheter angiography, with its atten-dant costs and risks, was required forthis assessment. CT angiography al-lows high-resolution 3D evaluation ofarteries, veins, and soft tissues (56–58)with less risk and at lower cost thancatheter angiography (56). Other po-tential uses for CT angiography in-clude evaluation of character and vas-cular supply of musculoskeletaltumors (54) and evaluation of suitabil-ity of the thoracodorsal and internalmammary arteries before transverserectus abdominis muscle flap recon-struction.

Other Indications

Multiple–detector row CT angiog-raphy can provide exquisite noninva-sive characterization of many othervascular conditions affecting the lowerextremity. Vascular malformations, ar-terial compression by adjacent masses,adventitial cystic disease (Fig 16), andpopliteal entrapment syndrome (59)can be fully evaluated. In the lattercase, image acquisition at rest andwith provocative maneuvers (eg, ac-tive plantar flexion against resistance)allows determination of the anatomyof the medial head of the gastrocne-mius and the dynamic degree of ar-terial obstruction. Vasculitides andinflammatory processes affecting ad-jacent vessels can be well-character-ized and the relationship of soft tissueand bone infection to the arterial struc-tures can be delineated (54).

SUMMARY

Peripheral CT angiography has ex-cellent spatial resolution and can showexquisite detail of peripheral vascula-ture. The current generation of 16- to64–detector-row CT scanners and thedevelopment of refined 3D renderingtechniques have made peripheral CTangiography a powerful tool for non-invasive imaging and treatment plan-ning of peripheral arterial disease.

References1. Lawrence JA, Kim D, Kent KC, et al.

Lower extremity spiral CT angiogra-

phy versus catheter angiography. Ra-diology 1995; 194:903–908.

2. Rieker O, Duber C, Schmiedt W, et al.Prospective comparison of CT angiog-raphy of the legs with intraarterial dig-ital subtraction angiography. AJR Am JRoentgenol 1996; 166:269–276.

3. Beregi JP, Elkohen M, Deklunder G, etal. Helical CT angiography comparedwith arteriography in the detection ofrenal artery stenosis. AJR Am J Roent-genol 1996; 167:495–501.

4. Kramer SC, Gorich J, Aschoff AJ, et al.Diagnostic value of spiral-CT angiog-raphy in comparison with digital sub-traction angiography before and afterperipheral vascular intervention. Angi-ology 1998; 49:599–606.

5. Soto JA, Munera F, Morales C, et al.Focal arterial injuries of the proximalextremities: helical CT arteriography asthe initial method of diagnosis. Radiol-ogy 2001; 218:188–194.

6. Rubin GD, Schmidt AJ, Logan LJ, et al.Multidetector-row CT angiography oflower extremity occlusive disease: anew application for CT scanning. Radi-ology 1999; 210:588.

7. Rubin GD, Schmidt AJ, Logan LJ, et al.Multi-detector row CT angiography oflower extremity arterial inflow andrunoff: initial experience. Radiology2001; 221:146–158.

8. Ofer A, Nitecki SS, Linn S, et al. Mul-tidetector CT angiography of periph-eral vascular disease: a prospectivecomparison with intraarterial digitalsubtraction angiography. AJR Am JRoentgenol 2003; 180:719–724.

9. Martin ML, Tay KH, Flak B, et al.Multidetector CT angiography of theaortoiliac system and lower extremi-ties: a prospective comparison withdigital subtraction angiography. AJRAm J Roentgenol 2003; 180:1085–1091.

10. Ota H, Takase K, Igarashi K, et al.MDCT compared with digital subtrac-tion angiography for assessment oflower extremity arterial occlusive dis-ease: importance of reviewing cross-sectional images. AJR Am J Roentgenol2004; 182:201–209.

11. Catalano C, Fraioli F, Laghi A, et al.Infrarenal aortic and lower-extremityarterial disease: diagnostic perfor-mance of multi-detector row CT an-giography. Radiology 2004; 231:555–563.

12. Fleischmann D, Rubin GD, Paik DS, etal. Stair-step artifacts with single ver-sus multiple detector-row helical CT.Radiology 2000; 216:185–196.

13. Hartnell GG. Contrast angiographyand MR angiography: still not opti-mum. J Vasc Interv Radiol 1999; 10:99–100.

14. Fleischmann D. Use of high-concen-tration contrast media in multiple-de-

tector-row CT: principles and ratio-nale. Eur Radiol 2003; 13(suppl 5):M14–M20.

15. Fleischmann D, Hittmair K. Mathe-matical analysis of arterial enhance-ment and optimization of bolus geom-etry for CT angiography using thediscrete Fourier transform. J ComputAssist Tomogr 1999; 23:474–484.

16. Fleischmann D, Rubin GD, BankierAA, et al. Improved uniformity ofaortic enhancement with customizedcontrast medium injection protocols atCT angiography. Radiology 2000; 214:363–371.

17. Bron KM. Femoral arteriography. In:Abrams HL, ed. Abrams angiography:vascular and interventional radiology,3rd ed. Boston: LittleBrown, 1983:1835–1875.

18. Versteylen RJ, Lampmann LE. Kneetime in femoral arteriography. AJRAm J Roentgenol 1989; 152:203.

19. Fleischmann D, Rubin GD. Quantifi-cation of intravenously administeredcontrast medium transit through theperipheral arteries: implications for CTangiography. Radiology 2005; 236:1076–1082.

20. Portugaller HR, Schoellnast H, Hau-segger KA, et al. Multislice spiral CTangiography in peripheral arterial oc-clusive disease: a valuable tool in de-tecting significant arterial lumen nar-rowing? Eur Radiol 2004; 14:1681–1687.

21. Qanadli SD, Chiappori V, Kelekis A, etal. Multislice computed tomographyof peripheral arterial disease: New ap-proach to optimize vascular opacifica-tion with 16-row platform. Eur Radiol2004; 14 Suppl. 2:S304.

22. Milne EN. The significance of earlyvenous filling during femoral arteriog-raphy. Radiology 1967; 88:513–8.

23. Koechl A, Kanitsar A, Lomoschitz E, etal. Comprehensive assessment of pe-ripheral arteries using multi-pathcurved planar reformation of CTAdatasets. Eur Radiol 2003; 13:268–269.

24. Kanitsar A, Fleischmann D, WegenkittlR, et al. CPR—curved planar refor-mation. In: IEEE Visualization. Boston:IEEE Computer Society; 2002;37–44.

25. Raman R, Napel S, Beaulieu CF, et al.Automated generation of curved pla-nar reformations from volume data:method and evaluation. Radiology2002; 223:275–280.

26. Fleischmann D, Kanitsar A, Lomos-chitz E, et al. Multi-path curved pla-nar reformation of the peripheral arte-rial tree. Radiology 2002; 225:363.

27. Napel S, Rubin GD, Jeffrey RB Jr.STS-MIP: a new reconstruction tech-nique for CT of the chest. J ComputAssist Tomogr 1993; 17:832–838.

28. Raman R, Napel S, Rubin GD.Curved-slab maximum intensity pro-


jection: method and evaluation. Radi-ology 2003; 229:255–260.

29. Edwards AJ, Wells IP, Roobottom CA.Multidetector row CT angiography ofthe lower limb arteries: a prospectivecomparison of volume-rendered tech-niques and intra-arterial digital sub-traction angiography. Clin Radiol 2005;60:85–95.

30. Willmann JK, Baumert B, Schertler T, etal. Aortoiliac and lower extremity ar-teries assessed with 16-detector rowCT angiography: prospective compari-son with digital subtraction angiogra-phy. Radiology 2005; 236:1083–1093.

31. Ouwendijk R, de Vries M, PattynamaPM, et al. Imaging peripheral arterialdisease: a randomized controlled trialcomparing contrast-enhanced MR an-giography and multi-detector row CTangiography. Radiology 2005; 236:1094–1103.

32. Hiatt WR. Medical treatment of pe-ripheral arterial disease and claudica-tion. N Engl J Med 2001; 344:1608–1621.

33. Hiatt WR. Pharmacologic therapy forperipheral arterial disease and claudi-cation. J Vasc Surg 2002; 36:1283–1291.

34. Stewart KJ, Hiatt WR, Regensteiner JG,et al. Exercise training for claudica-tion. N Engl J Med 2002; 347:1941–1951.

35. de Vries SO, Visser K, de Vries JA, et al.Intermittent claudication: cost-effec-tiveness of revascularization versus ex-ercise therapy. Radiology 2002; 222:25–36.

36. Pentecost MJ, Criqui MH, Dorros G, etal. Guidelines for peripheral percuta-neous transluminal angioplasty of theabdominal aorta and lower extremityvessels. A statement for health profes-sionals from a special writing group ofthe Councils on Cardiovascular Radi-ology, Arteriosclerosis, Cardio-Tho-racic and Vascular Surgery, ClinicalCardiology, and Epidemiology andPrevention, the American Heart Asso-ciation. Circulation 1994; 89:511–531.

37. Dormandy JA, Rutherford RB. Man-agement of peripheral arterial disease(PAD). TASC Working Group. Trans-Atlantic Inter-Society Consensus(TASC). J Vasc Surg 2000; 31(suppl):S1–S296.

38. Surowiec SM, Davies MG, Eberly SW,et al. Percutaneous angioplasty and

stenting of the superficial femoral ar-tery. J Vasc Surg 2005; 41:269–278.

39. Clark TW, Groffsky JL, Soulen MC.Predictors of long-term patency afterfemoropopliteal angioplasty: resultsfrom the STAR registry. J Vasc IntervRadiol 2001; 12:923–933.

40. Reekers JA, Bolia A. Percutaneous in-tentional extraluminal (subintimal) re-canalization: how to do it yourself. EurJ Radiol 1998; 28:192–198.

41. Tisi PV, Mirnezami A, Baker S, et al.Role of subintimal angioplasty in thetreatment of chronic lower limb isch-aemia. Eur J Vasc Endovasc Surg 2002;24:417–422.

42. Saketkhoo RR, Razavi MK, Padidar A,et al. Percutaneous bypass: subinti-mal recanalization of peripheral occlu-sive disease with IVUS guided luminalre-entry. Tech Vasc Interv Radiol 2004;7:23–27.

43. Nadal LL, Cynamon J, Lipsitz EC, et al.Subintimal angioplasty for chronic ar-terial occlusions. Tech Vasc Interv Ra-diol 2004; 7:16–22.

44. Hirai T, Korogi Y, Ono K, et al. Max-imum stenosis of extracranial internalcarotid artery: effect of luminal mor-phology on stenosis measurement byusing CT angiography and conven-tional DSA. Radiology 2001; 221:802–809.

45. Visser K, de Vries SO, Kitslaar PJ, et al.Cost-effectiveness of diagnostic imag-ing work-up and treatment for patientswith intermittent claudication in TheNetherlands. Eur J Vasc Endovasc Surg2003; 25:213–223.

46. Rubin GD, Armerding MD, Dake MD,et al. Cost identification of abdominalaortic aneurysm imaging by using timeand motion analyses. Radiology 2000;215:63–70.

47. Clair DG, Dayal R, Faries PL, et al.Tibial angioplasty as an alternativestrategy in patients with limb-threaten-ing ischemia. Ann Vasc Surg 2005; 19:63–68.

48. Dorros G, Jaff MR, Murphy KJ, et al.The acute outcome of tibioperonealvessel angioplasty in 417 cases withclaudication and critical limb ischemia.Cathet Cardiovasc Diagn 1998; 45:251–256.

49. Rutherford RB, Baker JD, Ernst C, et al.

Recommended standards for reportsdealing with lower extremity ischemia:revised version. J Vasc Surg 1997; 26:517–538.

50. Costantini V, Lenti M. Treatment ofacute occlusion of peripheral arteries.Thromb Res 2002; 106:V285–V294.

51. Willmann JK, Mayer D, Banyai M, etal. Evaluation of peripheral arterialbypass grafts with multi-detector rowCT angiography: comparison withduplex US and digital subtraction an-giography. Radiology 2003; 229:465–474.

52. Mills JL, Harris EJ, Taylor LM Jr, et al.The importance of routine surveillanceof distal bypass grafts with duplexscanning: a study of 379 reversed veingrafts. J Vasc Surg 1990; 12:379–386.

53. Moody P, Gould DA, Harris PL. Veingraft surveillance improves patency infemoro-popliteal bypass. Eur J VascSurg 1990; 4:117–21.

54. Karcaaltincaba M, Akata D, AydingozU, et al. Three-dimensional MDCTangiography of the extremities: clinicalapplications with emphasis on muscu-loskeletal uses. AJR Am J Roentgenol2004; 183:113–117.

55. Whitley SP, Sandhu S, Cardozo A.Preoperative vascular assessment ofthe lower limb for harvest of a fibularflap: the views of vascular surgeons inthe United Kingdom. Br J Oral Maxil-lofac Surg 2004; 42:307–310.

56. Klein MB, Karanas YL, Chow LC, et al.Early experience with computed tomo-graphic angiography in microsurgicalreconstruction. Plast Reconstr Surg2003; 112:498–503.

57. Karanas YL, Antony A, Rubin G, et al.Preoperative CT angiography for freefibula transfer. Microsurgery 2004; 24:125–127.

58. Chow LC, Napoli A, Klein MB, et al.Vascular mapping of the leg withmulti-detector row CT angiographyprior to free-flap transplantation. Radi-ology 2005; 237:353–360.

59. Takase K, Imakita S, Kuribayashi S, etal. Popliteal artery entrapmentsyndrome: aberrant origin of gastroc-nemius muscle shown by 3D CT. JComput Assist Tomogr 1997; 21:523–528.


ct angiography of peripheral arterial disease

Documents