Você está na página 1de 7

576094

research-article2015
JETXXX10.1177/1526602815576094Journal of Endovascular TherapyMacharzina et al

Clinical Investigation

Journal of Endovascular Therapy

Duplex Ultrasound Assessment of Native


2015, Vol. 22(2) 254­–260
© The Author(s) 2015
Reprints and permissions:
Stenoses in the Superficial Femoral and sagepub.com/journalsPermissions.nav
DOI: 10.1177/1526602815576094

Popliteal Arteries: A Comparative Study www.jevt.org

Examining the Influence of Multisegment


Lesions

Richard-Roland Macharzina, MD1*, Simon F. Schmid, MD1*, Ulrich Beschorner,


MD1, Elias Noory, MD1, Aljoscha Rastan, MD1, Werner Vach, PhD2,
Uwe Schwarzwälder, MD1, Sebastian Sixt, MD1, Karlheinz Bürgelin, MD1,
Franz-Josef Neumann, MD1, and Thomas Zeller, MD1

Abstract
Purpose: To evaluate duplex ultrasonography (DUS) in the assessment of femoropopliteal stenoses comparing a
single native stenosis (SNS) to multisegmental native stenoses (MNS). Methods: Among the 1284 patients treated for
atherosclerotic occlusive disease involving the femoropopliteal segment between November 2002 and November 2012,
139 patients (97 men; mean age 68±8 years) with 142 SNS or 143 MNS in 79 and 60 patients, respectively, were eligible
for this retrospective analysis. The peak systolic velocity ratios with proximal (PSVRprox) and distal (PSVRdist) reference for
the 285 lesions were compared with their respective angiographic stenosis grade as measured by 2 independent readers
using quantitative vascular analysis to ensure objectivity. Receiver operating characteristic curve analysis was used to
evaluate sensitivity, specificity, and the optimal thresholds of PSV and PSVR for detection of stenoses by grade (>50%,
>70%, or >80% diameter stenosis). The area under the curve (AUC) values of dependent and independent receiver
operating characteristic curves were compared. Results: For SNS, correlation of PSVRprox to diameter stenosis (R=0.88)
was higher (p<0.001) than the correlation for MNS (R=0.78). In the SNS group, the AUC for detecting a >50% (0.99±0.01),
>70% (0.98±0.01), and >80% (0.96±0.01) stenosis with PSVRprox was significantly higher than in the MNS group [AUC50%
0.93±0.02 (p=0.01), AUC70% 0.92±0.02 (p=0.02), and AUC80% 0.87±0.03 (p=0.003)]. The optimal thresholds for detecting
>50%, >70%, and >80% stenoses for SNS using PSVRprox were 2.6, 3.3, and 3.9, respectively. For MNS, the optimal
thresholds of PSVRprox were 2.6, 3.4, and 3.9, respectively, with respective sensitivities of 87%, 81%, and 75%; respective
specificities of 93%, 90%, and 82%; negative predictive values of 45%, 64%, and 74%; and positive predictive values of 99%,
95%, and 83%. Conclusion: DUS is an optimal tool for quantification of SNS. However, a multisegment setting has a
significant negative impact on the quantification of femoropopliteal artery stenosis.

Keywords
peripheral occlusive disease, superficial femoral artery, popliteal artery, stenosis, duplex ultrasonography, proximal systolic
velocity ratio, angiography, quantitative vessel analysis

Introduction 1
University Heart Center Freiburg-Bad Krozingen, Bad Krozingen,
Duplex ultrasonography (DUS) has become the diagnostic Germany
2
Center for Medical Biometry and Medical Informatics, University
method of choice for identification of stenosis in the super- Freiburg, Germany
ficial femoral (SFA) and popliteal arteries (PA) because of
*Richard-Roland Macharzina and Simon F. Schmid contributed equally to
its noninvasive nature, lack of side effects, and acceptable this work and have shared first authorship.
reproducibility.1 Peak systolic velocity (PSV) and peak
systolic velocity ratio (PSVR) serve as DUS markers for Corresponding Author:
Prof Dr Thomas Zeller, Abteilung Angiologie, Universitäts-Herzzentrum
grading stenoses.2 However, multisegment occlusive dis- Freiburg-Bad Krozingen, Abteilung Angiologie, Südring 15, Bad
ease, particularly in the femoropopliteal segment, repre- Krozingen, D-79189, Germany.
sents a major challenge for investigators using DUS to Email: thomas.zeller@universitaets-herzzentrum.de
Macharzina et al 255

Figure 1.  Flowchart illustrating the selection of study patients from the entire cohort of patients treated between 2002 and 2012.

characterize a stenotic lesion and to decide further steps in subtotal occlusions or severely calcified stenoses were
the treatment of this complex atherosclerotic disease.3 excluded from this retrospective analysis.
Different studies have described a negative influence of Among the 1284 patients identified in the database
multisegment disease on the quality of DUS,4,5 whereas search, 1084 were ineligible (Figure 1) owing to amputation
some studies did not assess the effect on the accuracy of of the target limb, prior surgery on the target segment, no
diagnosis.6–8 Moreover, there are different opinions on the stenosis, or severe calcification. Of the 200 eligible patients
thresholds for detecting a hemodynamically significant with 461 lesions, a further 61 were excluded owing to SFA
>50% diameter stenosis.9 Using gold standard angiography ostial or bifurcation lesions (n=21, 56 lesions) and in-stent
as an objective benchmark, the present single-center study restenosis (n=40, 120 lesions). This left 139 patients (97
analyzed velocities for a single native stenosis (SNS) and men; mean age 68±8 years) with 142 SNS (116 de novo
multisegment native stenoses (MNS) of the femoropopli- stenoses and 26 restenoses in 142 limbs) or 143 MNS (98
teal segment to determine any influence of MNS on the limbs) in 79 and 60 patients, respectively. Patient and lesion
thresholds for stenosis grading and the test quality, as well characteristics are given in Table 1. Diagnostic DUS scans
as to define thresholds for SNS. from the 240 limbs were reviewed retrospectively and com-
pared with the follow-up angiogram performed an average
of 3±23 days later.
Methods
Study Design and Patient Sample Duplex Ultrasonography
The databases for 10 prospective interventional studies con- Each DUS was performed by experienced physicians (vas-
ducted at our institution were searched to identify patients cular specialists, angiologists) using a Philips IU 22 scanner
who underwent endovascular treatment for atherosclerotic (Philips, Amsterdam, the Netherlands) or a GE L9 system
occlusive disease involving the SFA and/or PA between (General Electric Healthcare, Milwaukee, WI, USA) with
November 2002 and November 2012. Patients with total or dynamic range linear transducers operating at 7.5 or 8 MHz.
256 Journal of Endovascular Therapy 22(2)

Table 1.  Characteristics of Study Patients and Lesions Stratified by Single Stenoses (SNS) vs. Multisegmental Native Stenoses (MNS).a

SNS (n=79) MNS (n=60) p


Age, y 67±8 (47–81) 70±8 (50–85) 0.11
Men 58 (73) 39 (65) 0.28
Body mass index, kg/m2 27.5±3.9 (18.6–46.9) 27.9±5.5 (19.8–47.6) 0.90
Hypertension 68 (86) 50 (83) 0.66
Dyslipidemia 73 (92) 52 (87) 0.27
Diabetes mellitus 27 (34) 25 (42) 0.37
Smoking history 28 (35) 16 (50) 0.27
Active smoking 17 (22) 23 (38) 0.03
Family history of cardiovascular disease 19 (24) 9 (15) 0.19
Heart failure 12 (17) 10 (18) 0.91
Valvular disease 9 (13) 3 (6) 0.16
Limbs / lesions 142 / 142 98 / 143  
  Right / left 71 / 71 71 / 72 0.95
  Ankle-brachial index 0.73±0.24 0.65±0.21 0.02
  Rutherford category 0 / 1 / 2 / 3 / 4 / 5 / 6 2 / 1 / 23 / 106 / 2 / 8 / 0 2 / 3 / 21 / 63 / 6 / 3 / 0 0.19
  Fontaine stage I / IIa / IIb / III / IV 3 / 20 / 109 / 2 / 8 2 / 23 / 64 / 6 / 3 0.16
  Morphologic characteristics
  TASC category A: 139, B:3 B: 135, C:8 <0.0001
   Proximal SFA / mid SFA / distal SFA / PA 22 / 53 / 56 / 11 25 / 46 / 47 / 25 0.08
   Percent diameter stenosis, % 69±25 (6–97) 75±17 (15–97) 0.30
    0% –49% / 50% –69% / 70% –99% 35 / 17 / 90 15 / 25 / 103 0.006
  Eccentricity index 0.50±0.25 0.54±0.25 0.18
   Proximal collateralization vessel 74 (52) 62 (44) 0.14
   Stenosis length, mm 21±20 (5–140) 20±19b (5–156) 0.98
  Functional characteristics
   Intrastenotic PSV, cm/s 355±138 (62–889) 360±137 (58–1047) 0.87
  PSVprox, cm/s 98±42 (28–359) 96±45 (28–360) 0.57
  PSVdist, cm/s 89±38 (22–210) 84±47 (18–289) 0.04
  PSVRprox 4.3±2.7 (1.0–21.3) 4.6±2.9 (1.0–17.5) 0.54
  PSVRdist 5.0±3.4 (1.0–24.5) 5.6±4.0 (1.1–31.0) 0.16

Abbreviations: TASC, TransAtlantic Inter-Society Consensus; SFA, superficial femoral artery; PA, popliteal artery, PSV, peak systolic velocity; PSVR,
peak systolic velocity ratio.
a
 Continuous data are presented as the means ± standard deviation (range); categorical data are given as the counts (percentage).
b
 Length for each isolated lesion within a multisegment lesion complex.

The Doppler signal was measured at a 60° insonation angle (16 bit, 512×512 pixels). To ensure objectivity and
to the vessel centerline beginning in the iliac artery and examiner-independent results, quantitative vascular
extending to the popliteal artery. A stenotic lesion was doc- analysis was employed with an automated vessel detec-
umented in a prospective and anonymized fashion. Three tion algorithm (QAngio XA 7.1; Medis, Leiden, the
measurement points were defined (Figure 2): at the largest Netherlands), which allowed calibration of each angio-
degree of stenosis and proximal and distal within a disease- gram (using the length of the catheter tip) and automatic
free segment. For purposes of this study, PSVs were measurement of the percent diameter stenosis by 2 inde-
extracted from the Doppler frequency spectrum analysis at pendent readers blinded to the patient and DUS data
each defined measurement point. Ratios to proximal (Figure 3). The highest degree of stenosis for each lesion
(PSVRprox) and distal reference points (PSVRdist) were cal- in different planes was documented and used for com-
culated to determine which ratio correlated best with the parison with the DUS findings. For multisegment lesions,
angiographic findings. stenoses in the femoropopliteal and aortoiliac segments
were assessed independently.
Angiography and Quantitative Vessel Analysis
Statistical Analysis
Using the Axiom Artis dTC angiography system (Siemens,
Erlangen, Germany), different planes were electronically Continuous variables are described as the mean ± standard
documented as anonymized pictures in DICOM format deviation (range). The Mann-Whitney U test was used to
Macharzina et al 257

Figure 2.  Measurements for duplex ultrasonography of a single native stenosis of a right superficial femoral artery. (A) The proximal
measurement is ~2 cm proximal to the lesion. (B) In middle of the artery, there is a moderate increase of intrastenotic peak systolic
velocity (PSV). (C) The distal measurement is ~2 cm distal to the lesion.

Figure 3.  (A, C) A single native stenosis in the distal part of the right superficial femoral artery. (B, D) Quantitative vascular analysis
of the lesion. (A) and (B) are 2 planes of the same stenosis at 90° to each other. An automated vessel detection algorithm traces
vessel contours (yellow lines) and assumed lumens (red lines). Minimal lumen diameter as well as the proximal and the distal reference
diameters are depicted as yellow strokes.

compare the means of unpaired variables, while the Wilcoxon- frequencies. The Kruskal-Wallis test examined the impact of
signed rank test was employed to compare the means of paired outflow disease on flow velocities in the femoropopliteal seg-
variables. The chi-square test was applied for comparison of ment. The threshold of statistical significance was p<0.05.
258 Journal of Endovascular Therapy 22(2)

Spearman’s rank correlation coefficients (R) were calculated


to determine any association of PSV and PSVR to percent
diameter stenosis and to compare reader accuracy for the
angiographic results. Receiver operating characteristic
(ROC) curve analysis was used to evaluate sensitivity, speci-
ficity, and the optimal thresholds of PSV and PSVR for
detection of stenoses by grade (>50%, >70%, or >80% diam-
eter stenosis). Outcomes of the ROC analysis are presented
with the 95% confidence intervals as [lower bound; upper
bound]. The area under the curve (AUC) values of dependent
and independent ROC curves were compared to each other.
The analyses were performed using STATA software (version
12.1; StataCorp LP, College Station, TX, USA).

Results
Baseline characteristics of the 2 patient groups were similar,
differing only for active smoking, which was significantly
higher for MNS patients (p=0.03). Each group had only 4
limbs with an iliac inflow lesion, whereas no patent outflow
vessel was found in 3/142 (2.1%) in the SNS group and
7/98 (7.1%) in the MNS cohort. Characteristics of limbs
and stenoses are presented for both groups in Table 1. Most
of the morphological lesion characteristics did not differ
significantly, including diameter stenosis, lesion eccentric-
ity, lesion length, and the presence of collateral vessels.
There were more TransAtlantic Inter-Society Consensus II
type B and C lesions in the MNS group (p<0.001), as well
as more high-grade (70%–99%) stenoses (p=0.006).
Accuracy of angiographic measurements between the 2 Figure 4.  Receiver operating characteristic curve analysis
readers was high (p<0.001), with a Spearman correlation computed to diagnose (A, B) a >50% diameter stenosis (%DS), (C,
coefficient of 0.94. D) a >70% DS, and (E, F) a >80%DS for a single native stenosis (A,
Regarding functional characteristics, only the PSVdist C, E) and for multisegment native stenoses (B, D, F). SNS, single
was on average 4 cm/s lower for the MNS group than for native stenosis; MNS, multisegment native stenoses; PSV, peak
the SNS limbs (p=0.04), but the PSVRprox and PSVRdist did systolic velocity; PSVR, peak systolic velocity ratio.
not differ between groups. The strongest correlation to per-
cent diameter stenosis for the SNS group was the PSVRprox 94%, 92%, and 93%. For MNS, the corresponding optimal
(R=0.88); for the MNS group, the correlation coefficient for thresholds of PSVRprox were 2.6, 3.4, and 3.9 respectively,
PSVRprox dropped significantly to 0.78 (p<0.001). For with respective sensitivities of 87%, 81%, and 75%; respec-
MNS, the best correlation for comparison was the PSVRdist tive specificities of 93%, 90%, and 82%; negative predictive
(R=0.81). The PSV was less accurate, resulting in signifi- values of 45%, 64%, and 74%; and positive predictive val-
cantly lower Spearman correlation coefficients (RSNS=0.73 ues of 99%, 95%, and 83%.
and RMNS=0.51). Distal runoff disease had no impact on velocities in the
The ROC curves (Figure 4) illustrate the sensitivity and femoropopliteal segment. Mean distal, intrastenotic, or
specificity to diagnose >50%, >70%, and >80% stenoses in proximal velocities were equal for 0 to 3-vessel runoff in
both groups. In the SNS group, the AUC for detecting a the lower leg.
>50% (0.99±0.01), >70% (0.98±0.01), and >80%
(0.96±0.01) stenosis with PSVRprox was significantly higher
Discussion
than in the MNS group [AUC50% 0.93±0.02 (p=0.01),
AUC70% 0.92±0.02 (p=0.02), and AUC80% 0.87±0.03 The impact of multisegment femoropopliteal stenosis on
(p=0.003)]. The optimal thresholds for detecting >50%, DUS identification and quantification is still a matter of
>70%, and >80% stenoses for SNS using PSVRprox were 2.6, debate because some investigators have reported an
3.3, and 3.9 (Table 2), respectively, with respective sensitivi- effect3–5,13–15 and others did not.6–8 The present study con-
ties of 98%, 92%, and 89% and respective specificities of firmed a negative effect of MNS compared to SNS in terms
Macharzina et al 259

Table 2.  Receiver Operating Characteristic Curve Analysis Outcomes for PSVRs According to the Stenosis Grade for a Single
Stenosis in the Entire Cohort.a

Stenosis Grade PSVR Criterion Sensitivity, % Specificity, % PPV, % NPV, %


≥50% 2.1 99 [97; 100] 83 [77; 89] 95 [91; 98] 97 [94; 100]
  2.2 98 [96; 100] 86 [80; 92] 96 [92; 99] 94 [90; 98]
  2.3 98 [96; 100] 91 [87; 96] 97 [95; 100] 94 [90; 98]
  2.6 98 [96; 100] 94 [91; 98] 98 [96; 100] 94 [91; 98]
  2.7 95 [92; 99] 94 [91; 98] 98 [96; 100] 87 [81; 92]
  2.8 92 [87; 96] 97 [94; 100] 99 [97; 100] 79 [72; 86]
≥70% 3.0 94 [91; 98] 89 [83; 94] 93 [89; 98] 90 [85; 95]
  3.1 93 [89; 97] 90 [86; 95] 94[91; 98] 89 [84; 94]
  3.3 92 [88; 97] 92 [88; 97] 95 [92; 99] 87 [82; 93]
  3.4 90 [85; 95] 94 [90; 98] 96 [93; 100] 85 [79; 90]
  3.5 88 [82; 93] 96 [93; 99] 98 [95; 100] 82 [76; 88]
  3.6 83 [77; 90] 98 [96; 100] 99 [97; 100] 77 [70; 84]
≥80% 3.6 92 [87; 96] 87 [81; 93] 88 [83; 94] 91 [86; 96]
  3.7 92 [87; 96] 90 [85; 95] 91 [86; 95] 91 [87; 96]
  3.8 89 [84; 94] 92 [87; 96] 92 [87; 96] 89 [84; 94]
  3.9 89 [84; 94] 93 [89; 97] 93 [89; 97] 89 [84; 94]
  4.0 86 [81; 92] 96 [92; 99] 96 [92; 99] 87 [81; 92]
  4.1 84 [78; 90] 79 [94; 100] 97 [94; 100] 85 [79; 91]

Abbreviations: PSVR, peak systolic velocity ratio; PPV, positive predictive value; NPV, negative predictive value.
a
 Percentages are presented with the 95% confidence interval [lower bound; upper bound].

of accuracy, with a 10% decrease in sensitivity and a 25% to present study, only 4 patients in each cohort had significant
45% decrease in the negative predictive value. The analysis iliac inflow lesions. As opposed to previous studies, we
resulted in similar cutoff values for SNS and MNS in the excluded other potential confounders from analysis, such as
stratification of femoropopliteal stenosis grade and defined bifurcation and ostial lesions, severe calcification, and pre-
new thresholds for single femoropopliteal lesions without vious surgical reconstruction of the index lesion.
excluding inflow and outflow lesions. The present study rejected the hypothesis that multiseg-
The strength of the present study compared with previ- ment femoropopliteal disease has an impact on the thresh-
ous studies was the use of quantitative vascular analysis for olds for detecting various grades of femoropopliteal
computing the percent diameter stenosis in an automatic stenosis, with PSVRprox >50% thresholds of 2.6 and 2.5 for
and standardized fashion, whereas all previous studies the SNS and MNS groups, respectively. Even if the accu-
except one16 have used visual assessment for semi-quantitative racy of PSVRprox were reduced as compared with PSVRdist
angiographic analysis. Therefore, in the present study, angi- in MNS, the differences in the threshold and its accuracy
ographic quantification was independent of the readers’ (R=0.78 vs. 0.81) were not significant. Potential confound-
individual approach. ing factors affecting the functional parameters, such as
Over more than 2 decades, the ratio of the intrastenotic lesion eccentricity and the presence of large prestenotic col-
to the proximal velocities (PSVRprox) has served as a more lateral vessels, were equally distributed in both study
stable marker than the intrastenotic velocity alone for grad- cohorts.19
ing stenoses by DUS, with an acceptable sensitivity of 88% The present study includes almost exclusively claudi-
and specificity of 96%.9 A 50% stenosis is considered cants (~96%), which is underlined by the low number of
hemodynamically significant, and authors have calculated patients having no patent outflow artery (2.1% and 7.1% for
from ROC analyses different thresholds for PSVRprox to SNS and MNS, respectively). The impact of outflow dis-
diagnose a 50% stenosis that range from 1.4 to 3.0.2,7,8,16–18 ease in the lower leg on functional DUS parameters has not
The big range implies the presence of inhomogeneous study been discussed in the past. The presence of tissue loss in
populations including multisegment lesion complexes com- other cohorts (37%16) may imply conditions of reduced
prising not only single femoropopliteal lesions but also flow resulting in a lower PSVR, as seen in the present study
inflow and outflow lesions potentially affecting the femoro- with a smaller proportion of occluded outflow vessels. For
popliteal segment flow pattern. Unfortunately, most previ- detection of a 50% diameter stenosis, Khan et al16 proposed
ous publications did not precisely describe the impact of a PSVRprox of 1.5 compared with the 2.6 and 2.5 thresholds,
inflow and outflow lesions on their study cohorts. In the respectively, in our study.
260 Journal of Endovascular Therapy 22(2)

Of note is the significant difference in lesion length in 3. Allard L, Guy C, Zhenyu G, et al. Review of the assess-
the present cohort compared to former studies. Our average ment of single level and Multilevel arterial occlusive disease
lesion length for SNS and each isolated MNS was only ~21 in lower limbs by duplex ultrasound. Ultrasound Med Biol.
mm, much shorter compared to Aly et al7 (42% of lesions 1999;25:495–502.
4. Allard L, Cloutier G, Durand LG, et al. Limitations of ultra-
>50 mm) and Khan et al16 (mean lesion length 203±100
sonic duplex scanning for diagnosing lower limb arterial
mm). In both these former studies, the PSVRprox for a 50%
stenoses in the presence of adjacent segment disease. J Vasc
diameter stenosis was lower (2.0 and 1.5, respectively) than Surg. 1994;19:650–657.
in the present study. Whether this difference is related to 5. Allard L, Cloutier G, Durand LG. Doppler velocity ratio mea-
lesion length remains uncertain. surements evaluated in a phantom model of multiple arterial
The present study proposes for the first time cutoff val- disease. Ultrasound Med Biol. 1995;21:471–480.
ues for isolated single femoropopliteal lesions excluding 6. Kohler TR, Nance DR, Cramer MM, et al. Duplex scanning
multisegment and ostial lesions. The study results confirm for diagnosis of aortoiliac and femoropopliteal disease: a pro-
previous findings that the PSVRprox is a good hemodynamic spective study. Circulation. 1987;76:1074–1180.
parameter for DUS in the evaluation of a single native ste- 7. Aly S, Sommerville K, Adiseshiah M, et al. Comparison of
nosis. With balanced sensitivity and specificity, the duplex imaging and arteriography in the evaluation of lower
limb arteries. Br J Surg. 1998;85:1099–1102.
PSVRprox values of 2.6, 3.3, and 3.9 are good hemodynamic
8. Legemate D, Teeuwen C, Hoeneveld H, et al. Spectral
discriminators to detect a >50%, >70%, and >80% femoro- analysis criteria in duplex scanning of aortoiliac and femo-
popliteal stenoses. ropopliteal arterial disease. Ultrasound Med Biol. 1991;17:
769–776.
Conclusion 9. Collins R, Burch J, Cranny G, et al. Duplex ultrasonography,
magnetic resonance angiography, and computed tomography
Multisegment femoropopliteal occlusive disease has a neg- angiography for diagnosis and assessment of symptomatic,
ative influence on the accuracy of DUS to quantify stenosis. lower limb peripheral arterial disease: systematic review.
Whereas the thresholds for detecting the 3 grades of steno- BMJ. 2007;334:1257.
sis (>50%, >70%, and >80%) are nearly identical for MNS 10. Hanley J, McNeil B. The meaning and use of the area under
and SNS, sensitivity as well the negative predictive values a receiver operating characteristic (ROC) curve. Radiology.
1982;143:29–36.
decrease significantly for MNS. This implies less reliable
11. Hanley J, McNeil B. A method of comparing the areas under
detection of femoropopliteal stenosis in a multisegment receiver operating characteristic curves derived from the
lesion setting. To overcome the low negative predictive same cases. Radiology. 1983;148:839–843.
value in MNS, concurrent mapping of the aortoiliac inflow 12. DeLong ER, DeLong DM, Clarke-Person DL. Comparing
as well as of the infrapopliteal outflow vessels is advised. the areas under two or more correlated receiver operating
For SNS, the PSVRprox thresholds of 2.6 and 3.3 are recom- characteristic curves: a nonparametric approach. Biometrics.
mended for the detection of >50% and >70% stenosis, 1988;44:837–845.
respectively, with excellent test quality. 13. Ranke C, Creutzig A, Klaus A. Duplex scanning of the

peripheral arteries: correlation of the peak velocity ratio
with angiographic diameter reduction. Ultrasound Med Biol.
Declaration of Conflicting Interests 1992;18:433–440.
The author(s) declared no potential conflicts of interest with respect 14. Bergamini TM, Tatum CM Jr, Marhall C, et al. Effect of mul-
to the research, authorship, and/or publication of this article. tilevel sequential stenosis on lower extremity arterial duplex
scanning. Am J Surg. 1995;169:564–566.
Funding 15. Sensier Y, Hartshorne T, Thrush A, et al. The effect of adja-
cent segment disease on the accuracy of color duplex scan-
The author(s) received no financial support for the research,
ning for the diagnosis of lower limb arterial disease. Eur J
authorship, and/or publication of this article.
Vasc Endovasc Surg. 1996;12:238–242.
16. Khan S, Khan M, Bradley B, et al. Utility of duplex ultra-
References sound in detecting and grading de novo femoropopliteal
1. Tendera M, Aboyans V, Bartelink M, et al. ESC guidelines lesions. J Vasc Surg. 2011;54:1067–1073.
on the diagnosis and treatment of peripheral artery diseases. 17. Leng GC, Whyman MR, Donnan PT, et al. Accuracy and
Document covering atherosclerotic disease of extracranial reproducibility of duplex ultrasonography in grading femoro-
carotid and vertebral, mesenteric, renal, upper and lower popliteal stenoses. J Vasc Surg. 1993;17:510–517.
extremity arteries. The Task Force on the Diagnosis and 18. Sacks D, Robinson M, Marinelli D, et al. Peripheral arterial
Treatment of Peripheral Artery Diseases of the European Doppler ultrasonography: diagnostic criteria. J Ultrasound
Society of Cardiology. Eur Heart J. 2011;32:2851–2906. Med. 1992;11:95–103.
2. Jager K, Philipps D, Martin R, et al. Noninvasive map- 19. Traupe T, Ortmann J, Stoller M, et al. Direct quantitative assess-
ping of lower limb arterial lesions. Ultrasound Med Biol. ment of the peripheral artery collateral circulation in patients
1985;11:515–521. undergoing angiography. Circulation. 2013;128:737–744.