Niranjan Balu1, Li Chen2, Thoetphum Benyakorn1, Daniel S Hippe1, Henrik Haraldsson3, Warren Gasper3, David Saloner3, Chun Yuan1, and Thomas Hatsukami4
1Radiology, University of Washington, Seattle, WA, United States, 2Electrical Engineering, University of Washington, Seattle, WA, United States, 3University of California SanFrancisco, San Francisco, CA, United States, 4Surgery, University of Washington, Seattle, WA, United States
Synopsis
The extent of collateralization/branching
(CB) of lower limb vessels in peripheral artery disease (PAD) can inform risk
of ischemia and response to revascularization. However quantitative imaging
metrics of CB have not been assessed in the setting of severe PAD. We developed
automated quantitative measurements of lower limb vascular morphology (pCafe)
and compared CB in patients undergoing revascularization for PAD. Assessment of
pCafe metrics in severe PAD suggests CB is increased with occlusion compared to
stenosis indicating compensatory CB development.
Background
PAD due to atherosclerosis of
the superficial femoral artery (SFA) can cause limb threating ischemia. However,
the extent of collateralization/branching (CB) can ameliorate the
symptoms [1] and also determine the success of revascularization procedures.
Therefore, knowledge of the individual patient’s CB can help stratify ischemia risk
and guide revascularization/treatments. However quantitative imaging measurements
of CB and its effects in severe PAD are limited [1]. The aim of this study is
to compare CB in different degrees of SFA patency using an automated
quantitative measurement of lower limb vessel morphology.Methods
Study subjects and procedures: Ten patients
with peripheral artery disease (PAD) scheduled for lower extremity endovascular
reconstruction were recruited at two different institutions. Imaging procedures
followed institutional IRB guidelines and informed consent was obtained from
subjects prior to the scan.
MRI scans: Subjects
were scanned on either 3T Siemens Skyra or Philips 3T Ingenia scanners. Single
station first pass 3D CEMRA was obtained after single dose gadolinium contrast
injection (Prohance or Gadavist) covering the lower thigh and knee bilaterally
using surface phased array coils. Similar scan parameters were used on both
platforms, namely: TR/TE = 4.56/2.195 ms, flip angle = 20°, in-plane resolution
= 0.81 mm×0.81 mm, slice thickness = 3 mm, field of view = 430 mm*430 mm.
Collateral and branch artery
quantification: One subject was excluded due to insufficient image
quality. Peripheral arterial feature extraction (pCafe) was based on centerline
tracing [2]. Briefly, CEMRA was
resampled to 0.81 mm isotropic resolution followed by Nyul intensity
normalization [3] and arterial centerline tracing [4]. An expert reader then
labeled key landmark points of major arteries namely superficial femoral artery
(SFA) and profunda femoris (PA). In cases of SFA stenosis, the site of stenosis
was labeled. In cases of SFA occlusion, proximal and distal SFA segments were
labeled (distal labeled if present). Arterial branches besides the SFA and PA
segments were then automatically labeled as collaterals/branches (CB) (figure
1). Both collaterals and branches were included in this category since the
intent was to examine whether total branches distal to occlusion/stenosis was
increased. Based on the centerlines and labels, length, radius and number of
branches of each segment were automatically calculated (list provided in table
1).
Statistical analysis: Arteries
from each leg (18 legs in total) were considered separately. Based on SFA patency
on the original CEMRA, lower limbs were classified into those with no-stenosis,
stenosed or occluded SFA. Representative examples of each type are shown in
figure 2. Each of the pCafe metrics was compared using Spearman’s rank
correlation coefficient between ordered stenosis group (None, Stenosis,
Occluded). Wilcoxon rank-sum test was used to compare the no-stenosis and
occluded arteries. P-values less than 0.05 were considered statistically
significant.Results
Collateral/branch
quantification was achieved in all cases. There was marked individual variation
in collateral/branch appearances. There were statistically significant
differences in SFA length and average radius between the three groups with
decreased length and radius in occluded arteries compared to arteries with no
stenosis (length 335 ± 49 mm vs 252 ± 82 mm, p<0.05; radius 2.9 ± 0.7 vs 2.4
± 0.2 mm, p<0.05). Comparing arteries with stenosis and occlusions, we found
a trend towards increased volume, length and number of branches in CB in both
SFA and PA when the artery was occluded as opposed to those with only stenosis
(table 2). Discussion
pCafe metrics showed decreased
SFA length and radius as expected in patent SFA compared to SFA with stenosis. There
was a trend for increased CB (length, radius and number of branches) in legs
with occluded SFA when compared to limbs with stenosis. With chronic occlusion,
collateralization is known to increase. pCafe metrics also show this to be the
case and the increased CB in these cases suggest that these limbs may be at
lower risk of functional impairment.
This is the first study to
compare objective quantitative measurements of lower limb CB for stratification
of stenosis and occlusion. Our sample size was restricted due to the specific
type of patients recruited (PAD subjects scheduled for revascularization). Due
to the small sample sizes, standard statistical methods which treat all
observations as independent were used (e.g., Wilcoxon rank-sum test, Spearman’s
rank correlation). Additional subject inclusion will also help to study patient
level differences in collateralization/perfusion using iCafe metrics and
relation to patient symptoms.Conclusions
In PAD with compromised SFA
(stenosis/occlusion) there was a trend towards increased CB with occlusion
compared to stenosis. Our results suggest that CB quantification by pCafe may
be useful in stratification of perfusion to lower limb tissues in PAD.Acknowledgements
This work was partially supported by grants from R01HL128816, R01HL103609 and R01NS092207References
[1] McDermott MM, Carr J, Liu K, Kramer
CM, Yuan C, Tian L, Criqui MH, Guralnik JM, Ferrucci L, Zhao L, Xu D, Kibbe M,
Berry J, Carroll TJ., Collateral vessel number, plaque burden, and functional
decline in peripheral artery disease., Vasc Med. 2014 Aug;19(4):281-288. Epub
2014 Jul 21.
[2] Chen L, Mossa-Basha M, Balu N, et al.
Development of a quantitative intracranial vascular features extraction tool on
3D MRA using semiautomated open-curve active contour vessel tracing. Magn
Reson Med. 2018;79(6):3229-3238. doi:10.1002/mrm.26961.
[3] Nyul LG, Udupa JK, Zhang X. New
variants of a method of MRI scale standardization. IEEE Trans Med Imaging.
2000;19(2):143-150. doi:10.1109/42.836373.
[4] Wang Y, Narayanaswamy A, Tsai CL,
Roysam B. A broadly applicable 3-D neuron tracing method based on open-curve
snake. Neuroinformatics. 2011;9(2-3):193-217.
doi:10.1007/s12021-011-9110-5.