Download Assessment of Coronary Microcirculation in a Swine Animal Model

Articles in PresS. Am J Physiol Heart Circ Physiol (May 27, 2011). doi:10.1152/ajpheart.00213.2011
1
Assessment of Coronary Microcirculation in a Swine Animal Model
2
3
Authors:
4
Zhang Zhang, M.S., 1, 2 Shigeho Takarada, M.D., 3 Sabee Molloi, Ph.D., 1, 3, *
5
6
1. Department of Radiological Sciences, University of California-Irvine, Irvine,
7
California 92697, United States
8
2. Department of Radiology, Tianjin Medical University General Hospital, Tianjin
9
300052, China
10
3. Department of Medicine (Cardiology), University of California-Irvine, Irvine,
11
California 92697, United States
12
13
Short running title: assessment of coronary microcirculation in a swine model
14
15
*Corresponding Author: Sabee Molloi, Ph.D.
16
Department of Radiological Sciences, University of California-Irvine
17
Medical Sciences B, B-140, Irvine, CA 92697
18
Telephone: 949-824-5904, Fax: 949-824-8115
19
E-mail: [email protected]
20
21
1
Copyright © 2011 by the American Physiological Society.
22
ABSTRACT
23
24
Background: Coronary microvascular dysfunction has important prognostic
25
implications. Several hemodynamic indices, such as coronary flow reserve (CFR),
26
microvascular resistance (MR) and zero-flow pressure (Pzf) were used to establish the
27
most reliable index to assess coronary microcirculation.
28
29
Methods: Fifteen swine were instrumented with a flow probe, and a pressure wire
30
was advanced into the distal LAD. Adenosine was used to produce maximum
31
hyperemia. Microspheres were used to create microvascular dysfunction. An occluder
32
was used to produce stenosis. Blood flow from the probe (Qp), aortic pressure (Pa),
33
distal coronary pressure (Pd) and right atrium pressure (Pv) were recorded.
34
Angiographic flow (Qa) was calculated using a time-density curve. Flow probe based
35
CFR (CFRq) and angiographic CFR (CFRa) were calculated using Qp and Qa,
36
respectively. Flow probe based normalized MR (NMRqh) and angiographic
37
normalized MR (NMRah) were determined using Qp and Qa, respectively, during
38
hyperemia. Pzf was calculated using Qp and Pd.
39
40
Results: Two series of ROC curves were generated: normal epicardial artery model
41
(N-model) and stenosis model (S-model). The areas under the ROC curves for CFRq,
42
CFRa, NMRqh, NMRah, Pzf were 0.855, 0.836, 0.976, 0.956, 0.855 in N-model and
43
0.737, 0.700, 0.935, 0.889, 0.698 in S-model. Both NMRqh and NMRah were
44
significantly more reliable than CFR and Pzf in detecting the microvascular
45
deterioration.
46
2
47
Conclusions: Compared to CFR and Pzf, NMR provided a more accurate assessment
48
of microcirculation. This improved accuracy was more prevalent when stenosis
49
existed. Moreover, NMRah is potentially a less invasive method for assessing coronary
50
microcirculation.
51
52
Keywords:
Angiography;
53
microvascular function.
Blood
flow;
54
55
Total word count: 5,780
56
3
Cardiovascular
imaging; Coronary
57
INTRODUCTION
58
59
Coronary angiography, a routine examination for patients with heart disease, provides
60
an assessment of stenosis severity by visualizing the opacified arterial lumen.
61
However, in many cases(12, 40), the severity of myocardium ischemia correlates
62
poorly with epicardial stenosis. Therefore, intracoronary physiological techniques
63
have been developed to provide physiological evaluations for patients who undergo
64
coronary angiography(8, 34). Recently, the medical community’s interest in coronary
65
microcirculation has grown with the increasing
66
dysfunction occurs in a number of myocardial disease states and has important
67
prognostic implications(19, 20). With this recognition comes the need for a method to
68
accurately assess the functional capacity of coronary microcirculation for diagnostic
69
purposes and to monitor the effects of therapeutic interventions targeted at reversing
70
the extent of coronary microvascular dysfunction. To assess changes in the
71
microcirculatory bed, several indices have been proposed, such as coronary flow
72
reserve (CFR), microvascular resistance (MR) and zero-flow pressure (Pzf). In the
73
early 1990s, Gould et al.(14) introduced CFR as the fist clinically applicable
74
functional index for assessing epicardial vessel stenosis. It serves as a marker for the
75
integrity of epicardial coronary circulation and microcirculation(36). MR is
76
theoretically defined as the perfusion to coronary back pressure (vide infra) gradient
77
divided by absolute coronary blood flow at hyperemia(9). In 1974, Gould et al.(13, 15)
78
postulated that the minimum microvascular resistance is independent of epicardial
79
stenosis severity. Pzf is the arterial pressure at which blood flow would cease. In 2003,
80
Shimada et al.(31) speculated that Pzf increases with the severity of injured coronary
81
microvasculature, particularly the capillaries(28). However, previous studies have not
4 / 28
awareness that microvascular
82
compared these indices.
83
84
In a previous report, an angiographic technique for MR measurement, using a first-pass
85
distribution analysis (FPA) technique in an in vivo coronary microvascular disruption
86
model, was validated (43). The purpose of the current study is to compare the CFR,
87
MR, and Pzf at various stages of severity in microvascular disruption and epicardial
88
artery conditions by using both a flow probe as the gold standard and the angiographic
89
method.
90
91
METHODS
92
93
Protocol
94
This study was conducted according to guidelines of the National Institutes of Health
95
(NIH) and approved by the University of California, Irvine Institutional Animal Care
96
and Use Committee. In an open-chest swine model, CFR, MR and Pzf measurements
97
were made at various stages of severity of microvascular disruption in the left anterior
98
descending artery (LAD). Microcirculation was disrupted by embolized microspheres.
99
An external vascular occluder was used to produce a moderate epicardial stenosis.
100
Coronary angiograms were acquired for each data set.
101
102
Animal preparation
103
Fifteen fasted Yorkshire swine (37.6±5.7kg, male, S&S Farms, CA) were sedated and
104
pre-medicated with Telazol-Ketamine-Xylazine (TKX, 4.4mg T/kg, 2.2mg K/kg,
5 / 28
105
2.2mg X/kg) and atropine (0.05 mg/kg). Anesthesia was maintained with
106
approximately 1~2% isoflurane (Highland Medical Equipment Vaporizer; Temecula,
107
CA). Carotid artery and jugular vein were surgically prepared for sheath placement.
108
Adenosine (400µg/kg/min)(35) was used to induce maximum hyperemia.
109
110
Surgery and catheterization
111
A lateral thoracotomy was performed using standard surgical techniques. The 4th and
112
5th ribs were spread apart and the heart was fully exposed. The pericardium was
113
opened and the proximal segment of the LAD was dissected free. A transit-time
114
ultrasound flow probe (Transonic systems, Inc. Ithaca, NY) was placed on the
115
proximal segment of the LAD. An extravascular occluder (IVM In Vivo Metric,
116
Healdsburg, CA) was placed around the LAD just distal to the flow probe to produce
117
a moderate epicardial stenosis. The position of the flow probe and the occluder were
118
adjusted to ensure that there were no side branches between them. The occluder was
119
also used to produce a 100% occlusion for reactive maximum hyperemia.
120
121
The left main ostium was cannulated with a 6F hockey-stick catheter through the left
122
carotid artery under fluoroscopic guidance. Another 4F hockey-stick catheter was
123
placed in the right atrium to measure the coronary back pressure. Microcirculation
124
was disrupted by gradually injecting 50~100µm (1.8×104 microspheres/ml)
125
microspheres (Polysciences Inc. Warrington, PA)(2, 9) down the LAD through the
126
catheter. This procedure was repeated for different degrees of severity of
127
microcirculatory embolism. An intracoronary pressure wire (Radi Medical System,
128
0.014") was advanced into the distal segment of the LAD to measure the distal
129
coronary pressure. Aortic pressure (Pa), distal coronary pressure (Pd) and right atrium
6 / 28
130
pressure (Pv) were measured continuously with pressure transducers and the pressure
131
wire, respectively.
132
133
Image acquisition and processing
134
All images were acquired using a conventional x-ray tube with a constant potential
135
x-ray generator (Optimus M200, Philips Medical Systems, Shelton, CT). A
136
cesium-iodide-based flat panel detector (PaxScan 4030A, Varian Medical Inc., Palo
137
Alto, CA) was used for image acquisition. Images were acquired at 30 frames per
138
second. All images were corrected for X-ray scatter before logarithmic
139
transformation(7). A publicly available software (Image J, NIH, Bethesda, MD) was
140
used for image analysis.
141
142
Each swine was positioned on its right side under the flat panel detector. The
143
projection angle was optimized for separating the LAD and the left circumflex artery
144
(LCX) perfusion beds. Pancuronium (0.1 mg/kg) was administered intravenously.
145
Electrocardiogram, arterial blood pressure, coronary blood flow were continuously
146
recorded (MP100, Biopac Systems, Inc. Santa Barbara, CA) to establish that the
147
animals were sufficiently sedated under this condition. Coronary angiograms were
148
acquired when the blood flow reached maximum hyperemia. The ventilator was
149
turned off at the end of a full expiration to minimize respiratory motion. Images of at
150
least one full heart cycle were acquired prior to contrast injection for selecting a
151
cardiac phase-matched mask image for temporal subtraction. Contrast material
152
(Omnipaque-350; Princeton, NJ) was power injected (Leibel-Flarsheim Angiomat
153
6000; Cincinnati, OH) at 3 ml/sec for 3 seconds. An image of a calibration phantom
154
positioned over the heart was also acquired to determine the correlation between the
7 / 28
155
image’s gray level and iodine mass.
156
157
Angiographic blood flow
158
Coronary blood flow was determined from the change in contrast volume in the
159
arterial tree during one cardiac cycle. A region of interest (ROI) for flow measurement
160
was drawn around the LAD vascular bed, which encompassed both the visible arteries
161
and the microcirculatory blush(41). Power injection of contrast material was assumed
162
to momentarily replace blood with contrast material. The known iodine concentration
163
in the contrast material and a linear regression analysis between the measured
164
integrated gray levels in the calibration phantom were used to convert the gray level
165
to contrast volume. The time period of the cardiac cycle was calculated from the
166
image acquisition rate of 30 frames per second. Therefore, the ratio of the change in
167
volume to the time it takes to complete one cardiac cycle yielded the value of
168
volumetric coronary blood flow (Figure 1). Our previous report has validated that the
169
angiographic flow based on the FPA technique have strong correlations with flow
170
from the flow probe in the swine microvascular disruption model (43).
171
172
Calculation of CFR, NMR and Pzf
173
CFR is defined as the ratio of coronary artery flow during maximal hyperemia to the
174
flow at rest. The gold standard flow probe based CFR (CFRq) and angiography based
175
CFR (CFRa) were calculated using the blood flow from flow probe (Qq) and the
176
angiographic flow based on the FPA technique (Qa), respectively.
177
178
MR is defined as myocardial perfusion pressure divided by volumetric coronary flow
179
at hyperemia. However, volumetric coronary blood flow and the calculated resistance
8 / 28
180
are dependent on the arterial perfusion bed size. Previous reports (21, 25) have shown
181
that the arterial lumen volume is related to the myocardial mass and can be used to
182
account for the dependent arterial bed size. Therefore, the normalized MR (NMR) (43)
183
can be calculated as:
184
NMR =
ΔP
Q
 V

V
 ref




(1)
3/ 4
185
All the NMR measurements were calculated using equation 1 where ΔP is the
186
coronary pressure gradient (mmHg), Q is the hyperemic blood flow (ml/min) through
187
a stem, V is the corresponding crown arterial volume (ml) and Vref is the reference
188
arterial lumen volume (1 ml). Therefore, the gold standard normalized MR (NMRqh)
189
was calculated as (Pd-Pv) divided by Qq during hyperemia. The angiography based
190
normalized MR (NMRah) was calculated as (Pa-Pv) divided by hyperemic Qa.
191
192
To measure Pzf, the instantaneous Qq was plotted with respect to the simultaneously
193
measured Pd. The linear regression between Qq and Pd was used to calculate the slope
194
of the curve in diastole, and the X-intercept of the line was used to calculate the Pzf
195
(Figure 2). Every data point was calculated as the average of three consecutive cardiac
196
cycles without recording artifacts.
197
198
Comparison of hemodynamic indices without epicardial stenosis
199
CFRq, CFRa, NMRqh, NMRah and Pzf were used to evaluate the different severities of
200
coronary microvascular disruption. The measurements without epicardial stenoses
201
were divided into four groups: Group A (normal condition, N=18), Group B (mild
202
microvascular disruption, less than 1.5×105 microspheres, N=63), Group C (moderate
9 / 28
203
microvascular disruption, 1.5×105~3.0×105 microspheres, N=51), and Group D
204
(severe microvascular disruption, more than 3.0×105 microspheres, N=35).
205
206
Diagnosis of microcirculation disruption involving epicardial stenosis
207
Diagnostic abilities for microcirculation disruption using CFR, NMR and Pzf were
208
tested in two models: (1) normal epicardial artery model (N model), which included
209
the normal condition and different severities of microvascular disruption with normal
210
epicardial arteries, and (2) moderate epicardial stenosis model (S model), which
211
included moderate coronary epicardial stenoses (50% diameter stenosis) and different
212
severities of microvascular disruption with the same moderate epicardial stenoses.
213
The gold standard for detecting microvascular disease is determined by whether
214
microspheres were injected, regardless of the amount of microspheres.
215
216
Statistical Analysis
217
Linear regression analysis was performed between (1) CFRq and CFRa, and (2)
218
NMRqh and NMRah to determine the coefficient in the regression equation. The
219
correlation coefficient (r) and standard error of estimate (SEE) were determined by a
220
linear regression analysis. The degree of agreement between the different methods
221
was assessed using the Bland-Altman analysis. One-way analysis of variance
222
(ANOVA) and Student Newman-Keuls Test (SNK) were used to compare the CFRq,
223
CFRa, NMRqh, NMRah and Pzf in the four groups of different coronary microvascular
224
conditions. Two series of receiver operating characteristic (ROC) curves were made
225
for both the N model and the S model. The areas under each curve (AUC) and the best
226
cut-off values were calculated to compare the diagnostic abilities of the five dynamic
10 / 28
227
indices. A p<0.05 was considered to be statistically significant for all statistical
228
analyses.
229
230
RESULTS
231
From a total of 337 measurements, there were 167 measurements without epicardial
232
stenosis and 170 measurements involving stenoses. And 251 measurements, excluding
233
the conditions with only epicardial stenoses and without microspheres, were used to
234
compare the angiographic and gold standard flow measurements.
235
236
Comparison of angiographic and flow probe measurements
237
A total of 251 CFR and NMR measurements were made. CFRa was linearly related to
238
the gold standard CFRq as CFRa=0.91CFRq+0.30 (p<0.001) with a good correlation
239
coefficient (r=0.901, SEE=0.535). NMRah correlated linearly with NMRqh as
240
NMRah=0.89NMRqh+0.04 mmHg/ml/min (p<0.001) with a good correlation
241
coefficient (r=0.955, SEE=0.208 mmHg/ml/min). Additionally, in the Bland-Altman
242
plot, the mean differences between the two measurements were 0.12 for CFR and 0.10
243
mmHg/ml/min for NMR. The limits of agreement were 1.19 and -0.95 for CFR and
244
0.35 and -0.55 mmHg/ml/min for NMR. None of the values had a statistically
245
significant difference from zero, implying a lack of bias between the two techniques.
246
Additionally, there were two observers measuring the angiographic flow using the
247
FPA method for the 55 measurements taken from the three animals. The mean error
248
for variability between two different observers was 4.95ml/min (error relative to the
249
mean was 7.9%) and the error for reproducibility was 4.83 ml/min (error relative to
250
the mean was 7.7%).
251
11 / 28
252
253
Comparison of hemodynamic indices without epicardial stenosis
254
Results from ANOVA showed significant differences (p<0.001) amongst the four
255
groups for each hemodynamic index in a total of 167 measurements. Figure 3 shows
256
the SNK multiple comparisons between the four groups for each index. For all the
257
indices (CFRq, CFRa, NMRqh, NMRah and Pzf), there were significant differences
258
between each group at various degrees of severity of microvascular disruption
259
(p<0.05).
260
261
262
Diagnosis of microcirculation disruption involving epicardial stenosis
263
In 167 measurements of the N model, the AUCs for CFRq, CFRa, NMRqh, NMRah and
264
Pzf were 0.855, 0.836, 0.976, 0.956, and 0.855, respectively. In 170 measurements of
265
the S model, the AUCs were 0.737, 0.700, 0.935, 0.889 and 0.698, respectively
266
(Figure 4). There were significant area reductions for CFRq, CFRa and Pzf in the S
267
model (p<0.05), but no significant difference for NMRqh and NMRah (p>0.05). The
268
percent area stenosis was measured to be 75 ±10% (50% diameter stenosis). The CFRq
269
was 2.82±1.17 in the N model compared to 2.20±1.12 in the S model (p<0.05), and
270
FFR was 0.74±0.14 in the S model .Table 1 also shows the sensitivity and specificity
271
of the best cut-off value for each index.
272
273
DISCUSSION
274
275
The present study demonstrated that NMR, which appeared to be independent of
276
moderate epicardial artery disease, was the most reliable diagnostic index to detect
12 / 28
277
microvascular deterioration. As compared with CFR and Pzf, NMR showed
278
improvement especially in cases of moderate epicardial artery disease. Furthermore,
279
the best cut-off value of NMR had the highest sensitivity and specificity as compared
280
to CFR and Pzf. Therefore, even in the presence of a moderate epicardial stenosis, the
281
increase of NMR can predict coronary microvascular dysfunction.
282
283
Comparison of CFR, Pzf and NMR
284
With the growing awareness that coronary microcirculatory dysfunction is an
285
important pathophysiological component in many cardiac conditions, several different
286
physiological dynamic indices, such as CFR, Pzf and NMR , have been introduced.
287
From the current study’s results, CFR, Pzf and NMR can all be used to evaluate the
288
severities of microvascular disruption.
289
290
Some previous studies have used the concept of CFR as the theoretical framework to
291
study microcirculation invasively(6, 23). CFR can provide an integral assessment of
292
both epicardial coronary circulation and microcirculation. Quantitative assessment of
293
CFR can be easily performed by intracoronary Doppler wires in the cardiac
294
catheterization laboratory. However, the cutoff value for CFR is highly affected by
295
epicardial stenosis and is very sensitive to hemodynamic changes(10, 29). Fractional
296
flow reserve (FFR) has been established as a useful physiological index of the
297
severity of an epicardial stenosis in the catheterization laboratory. FFR has previously
298
been shown to be a reliable technique to functionally assess a given coronary
299
intermediate stenosis with unclear hemodynamic significance (1, 5). A previous report
300
has suggested that coronary outflow pressure corrected FFR cannot only be
13 / 28
301
considered a specific index for the epicardial stenosis alone (32). FFR is also a
302
function of myocardial bed resistance, which vary as a consequence of variable
303
hemodynamic conditions (32). Therefore, FFR can’t be used as an independent index
304
to assess microcirculation since it is highly affected by the presence of an epicardial
305
stenosis. However, in the absence of microvascular disease, it is possible to estimate
306
FFR using only angiographic image data (41, 42) .
307
308
A previous study had reported that increased Pzf may reflect accentuation of
309
microvascular tone and decreased perfusion bed mass, which is caused by non-viable
310
tissue(8, 38, 39). Tanaka et al. (37) described this relationship during hyperemia. The
311
slope of the curve in diastole and the X-intercept of changed significantly during
312
maximal hyperemia induced by papaverine administration. Terai et al.(16) suggested
313
that increased Pzf is responsible for the no reflow phenomenon in patients. Pzf appears
314
to be an accurate index for evaluating microcirculation because only Qp and Pd are
315
used during the diastolic phase to calculate Pzf. In diastole, coronary flow depends on
316
coronary pressure without regards to cardiac contraction, and the diastolic relationship
317
between coronary flow and pressure permits analysis of coronary artery resistance(17,
318
28, 31). However, sometimes it is hard to perform Pzf on patients using Doppler wires
319
because of the large variance. Pzf is easily influenced by individual cardiac cycle and
320
diastolic phase.
321
322
Previous studies have reported that in comparison to other indices, MR’s advantages
323
allow it to be reliably applied for the interrogation of microcirculatory resistance in
14 / 28
324
the catheterization laboratory(4, 9, 18, 22). First, MR is reproducible and largely
325
independent of variation in the hemodynamic state(29). Second, MR appears to be
326
independent of epicardial artery disease. Previous reports (3, 30, 33) have documented
327
an increase in microvascular resistance in the presence of an epicardial artery stenosis.
328
From the current study’s results, even in the presence of a moderate epicardial
329
stenosis, the increase of NMR could still be used to evaluate the microvascular
330
disruption. Additionally, MR is an independent predictor of acute and short term
331
myocardial damage in patients undergoing primary percutaneous coronary
332
intervention (PCI). It may allow us to determine the efficacy of therapeutic strategies
333
for microvascular protection in patients with ST-segment elevation myocardial
334
infarction (STEMI)(11). Therefore, measurements of coronary microvascular
335
resistance play a pivotal role in diagnosing and monitoring the effects of therapeutic
336
interventions. However, MR calculation requires an accurate measurement of absolute
337
blood flow. A relatively simple and minimally invasive method for quantitatively
338
assessing the status of the coronary microcirculation is still required.
339
340
FPA and angiographic NMR
341
Several invasive and noninvasive techniques have been developed for measuring
342
blood flow to evaluate the status of coronary microcirculation (6, 18, 20). FPA
343
technique, using angiographic image data, has previously been used to measure
344
absolute coronary blood flow through a stenotic artery(26, 27). Our previous
345
report(43) focused on validation of the FPA technique in a swine microvascular
346
disruption model. It showed that angiographic NMR measurement based on the FPA
347
technique has good accuracy and reproducibility. In the current study, different indices
348
for assessment of microvascular disruption were compared. As compared with CFR
15 / 28
349
and Pzf, NMR was shown to be the most reliable diagnostic index to detect
350
microvascular deterioration, especially in the presence of moderate epicardial artery
351
stenosis. The results indicate that NMR can provide an accurate and minimally
352
invasive assessment of coronary microcirculation in the catheterization laboratory.
353
Furthermore, angiographic NMR based on the FPA technique has important
354
advantages over the previously reported methods for MR measurement (24, 41). The
355
angiographic NMR measurement, which requires no wires, reduces the cost and
356
procedure time associated with the other techniques. All angiographic NMR
357
measurements can be done using image data acquired during routine diagnostic
358
cardiac catheterization. Therefore, both anatomical and physiological information can
359
be derived from the same angiographic images. Therefore, the angiographic method
360
for NMR measurement could potentially be used to evaluate the microcirculatory
361
system of patients with stable chest pain in the cardiac catheterization laboratory.
362
363
The current study used (Pd-Pv) to calculate NMRqh as the gold standard, and (Pa-Pv) to
364
calculate NMRah for both normal and stenosis conditions. As discussed above, one of
365
the most important advantages of the angiographic technique for NMR measurement
366
is the elimination of risks associated with advancing a pressure wire across a stenotic
367
lesion and the reduction of procedure’s cost. Our preliminary study compared pairs of
368
NMR calculated from (Pa-Pv) and (Pd-Pv) during various stages of severity of
369
microvascular disease involving mild epicardial stenosis of up to approximately 50%
370
area stenosis by using both flow probe and the proposed angiographic method. In 45
371
pairs of measurements, Paired Samples t Tests showed no significant difference
372
between Pa and Pd. The SEE of NMR measurements between using Pa and Pd relative
373
to the mean NMR measured was 10.7% for NMRqh and 10.5% for NMRah.
16 / 28
374
Additionally, a Bland-Altman plot identified the mean differences between NMR
375
measurements using Pa and Pd as being only -0.08±0.24 for NMRqh and -0.08±0.20
376
for NMRah. Therefore, NMRah calculated from Pa can be potentially used to screen the
377
microvascular condition of patients with normal and mild epicardial stenosis (up to
378
approximately 50% area stenosis). In cases of severe epicardial stenosis, NMR
379
calculated from Pa will be overestimated because Pd may be significantly lower than
380
Pa. However, in the cases of this study, the use of wires was inevitable because
381
intervention of the epicardial stenosis is clinically indicated.
382
383
Study Limitations
384
In this study, several injections of microspheres were made. This may cause
385
heterogeneous microinfarcts, which is not the same pathological change as normal
386
myocardial infarction or diffused microvascular disease(2). Moreover, this
387
experimental animal model only introduced a 50% diameter stenosis in addition to the
388
microcirculatory disruption. More severe epicardial stenosis along with microvascular
389
disease should be studied in future. Furthermore, other disease conditions, such as
390
ventricular hypertrophy, diabetes mellitus, flow limiting diffuse coronary artery
391
disease, and previous myocardial infarction, should be considered. The impact of
392
other disease conditions on NMR requires additional study.
393
394
In the case of severe stenosis, collateral flow might lead to an overestimation of NMR
395
because myocardial perfusion is a combination of both coronary and collateral flows.
396
For our current angiographic flow measurement, an ROI large enough to encompass
397
the LAD vascular bed was drawn. This technique ensured that any visible potential
17 / 28
398
collateral flow perfusion would be included in the angiographic NMR measurement.
399
However, in cases of severe stenosis, coronary angiography still has only a limited
400
sensitivity for quantifying collateral circulation capacity. Additionally, in a clinical
401
setting, it is difficult to avoid using wires in cases of severe epicardial stenosis
402
because the lesion has to undergo intervention. Therefore, our angiographic method is
403
more suitable for the patients with less severe stenoses for which intervention might
404
not be required. In future studies, it will be necessary to measure the coronary wedge
405
pressure (Pw) to assess the collateral flow and incorporate it into the calculation of
406
microvascular resistance in the presence of a severe epicardial stenosis.
407
408
Another limitation of the angiographic method for flow measurement is motion
409
misregistration. Respiratory motion can introduce misregistration artifacts in
410
phase-matched subtracted images and increase measurement error in coronary flow.
411
However, motion artifacts can be minimized with breath-holding, since only a short
412
time interval is required for blood flow measurement (3~5 seconds).
413
414
CONCLUSION
415
416
Compared to CFR and Pzf, NMR is a more reliable index to diagnose microcirculation
417
disruption and to provide a more accurate assessment of microcirculation, especially
418
when epicardial stenosis is present. Moreover, angiography based NMR can
419
potentially be a simpler and less invasive method for specific assessment of the
420
coronary microvascular condition of patients with stable chest pain during routine
421
coronary arteriography.
18 / 28
422
423
FUNDING
424
This work was supported by the National Heart, Lung and Blood Institute and the
425
Department of Health and Human Services [R01 HL89941].
426
427
ACKNOWLEDGMENTS
428
The authors would like to thank Drs. Jerry Wong and Charles Dang for their technical
429
support. We would like to acknowledge partial funding for Zhang Zhang from China
430
National Natural Science Foundation grant 30870698, and Tianjin Medical University
431
Graduate Innovation Fund 2009GSLI13.
432
433
DISCLOSURES: None
434
435
REFERENCES
436
437
1.
Bishop AH, Samady H. Fractional flow reserve: critical review of an important
438
physiologic adjunct to angiography. Am Heart J 147: 792-802, 2004.
439
2.
440
microinfarcts caused by coronary microemboli: evaluation with multidetector CT and
441
MR imaging in a swine model. Radiology 254: 718-728, 2010.
442
3.
443
Association between coronary lesion severity and distal microvascular resistance in
444
patients with coronary artery disease. Am J Physiol Heart Circ Physiol
Carlsson M, Saloner D, Martin AJ, Ursell PC, Saeed M. Heterogeneous
Chamuleau SA, Siebes M, Meuwissen M, Koch KT, Spaan JA, Piek JJ.
19 / 28
285:
445
H2194-2200, 2003.
446
4.
447
takotsubo cardiomyopathy. Catheter Cardiovasc Interv, 2010. [Epub ahead of print]
448
5.
449
Heart 94: 949-959, 2008.
450
6.
451
thermodilution to assess flow reserve: experimental validation. Circulation
452
2003-2006, 2001.
453
7.
454
deconvolution. Phys Med Biol 55: 1295-1309, 2010.
455
8.
456
Salas C, Alfonso F, Hernandez R, Angiolillo DJ, Jimenez-Quevedo P, Banuelos C,
457
Alonso-Pulpon L, Macaya C. Assessment of microcirculatory remodeling with
458
intracoronary
459
endomyocardial sampling in cardiac allografts. Circulation
460
9.
461
Fitzgerald PJ, Yock PG, Yeung AC. Novel index for invasively assessing the
462
coronary microcirculation. Circulation 107: 3129-3132, 2003.
463
10. Fearon WF, Farouque HM, Balsam LB, Caffarelli AD, Cooke DT, Robbins
464
RC, Fitzgerald PJ, Yeung AC, Yock PG. Comparison of coronary thermodilution
465
and Doppler velocity for assessing coronary flow reserve. Circulation
466
2198-2200, 2003.
Daniels DV, Fearon WF. The index of microcirculatory resistance (IMR) in
De Bruyne B, Sarma J. Fractional flow reserve: a review: invasive imaging.
De Bruyne B, Pijls NH, Smith L, Wievegg M, Heyndrickx GR. Coronary
104:
Ducote JL, Molloi S. Scatter correction in digital mammography based on image
Escaned J, Flores A, Garcia-Pavia P, Segovia J, Jimenez J, Aragoncillo P,
flow
velocity
and
pressure
measurements:
validation
with
120: 1561-1568, 2009.
Fearon WF, Balsam LB, Farouque HM, Caffarelli AD, Robbins RC,
20 / 28
108:
467
11. Fearon WF, Shah M, Ng M, Brinton T, Wilson A, Tremmel JA, Schnittger I,
468
Lee DP, Vagelos RH, Fitzgerald PJ, Yock PG, Yeung AC. Predictive value of the
469
index of microcirculatory resistance in patients with ST-segment elevation myocardial
470
infarction. J Am Coll Cardiol 51: 560-565, 2008.
471
12. Folland ED, Vogel RA, Hartigan P, Bates ER, Beauman GJ, Fortin T,
472
Boucher C, Parisi AF. Relation between coronary artery stenosis assessed by visual,
473
caliper, and computer methods and exercise capacity in patients with single-vessel
474
coronary artery disease. The Veterans Affairs ACME Investigators. Circulation 89:
475
2005-2014, 1994.
476
13. Gould KL, Lipscomb K. Effects of coronary stenoses on coronary flow reserve
477
and resistance. Am J Cardiol 34: 48-55, 1974.
478
14. Gould KL, Lipscomb K, Hamilton GW. Physiologic basis for assessing critical
479
coronary stenosis. Instantaneous flow response and regional distribution during
480
coronary hyperemia as measures of coronary flow reserve. Am J Cardiol 33: 87-94,
481
1974.
482
15. Gould KL, Lipscomb K, Calvert C. Compensatory changes of the distal
483
coronary vascular bed during progressive coronary constriction. Circulation
484
1085-1094, 1975.
485
16. Ito H. No-reflow phenomenon in patients with acute myocardial infarction: its
486
pathophysiology and clinical implications. Acta Med Okayama 63: 161-168, 2009.
487
17. Ito H, Terai K, Iwakura K, Kawase I, Fujii K. Hemodynamics of
488
microvascular dysfunction in patients with anterior wall acute myocardial infarction.
21 / 28
51:
489
Am J Cardiol 94: 209-212, 2004.
490
18. Kern MJ, Lerman A, Bech JW, De Bruyne B, Eeckhout E, Fearon WF,
491
Higano ST, Lim MJ, Meuwissen M, Piek JJ, Pijls NH, Siebes M, Spaan JA.
492
Physiological assessment of coronary artery disease in the cardiac catheterization
493
laboratory: a scientific statement from the American Heart Association Committee on
494
Diagnostic and Interventional Cardiac Catheterization, Council on Clinical
495
Cardiology. Circulation
496
19. Kitabata H, Imanishi T, Kubo T, Takarada S, Kashiwagi M, Matsumoto H,
497
Tsujioka H, Ikejima H, Arita Y, Okochi K, Kuroi A, Ueno S, Kataiwa H,
498
Tanimoto T, Yamano T, Hirata K, Nakamura N, Tanaka A, Mizukoshi M,
499
Akasaka T. Coronary microvascular resistance index immediately after primary
500
percutaneous coronary intervention as a predictor of the transmural extent of
501
infarction in patients with ST-segment elevation anterior acute myocardial infarction.
502
JACC Cardiovasc Imaging 2: 263-272, 2009.
503
20. Knaapen P, Camici PG, Marques KM, Nijveldt R, Bax JJ, Westerhof N,
504
Gotte MJ, Jerosch-Herold M, Schelbert HR, Lammertsma AA, van Rossum AC.
505
Coronary microvascular resistance: methods for its quantification in humans. Basic
506
Res Cardiol
507
21. Le HQ, Wong JT, Molloi S. Allometric scaling in the coronary arterial system.
508
Int J Cardiovasc Imaging
509
22. Marques KM, Knaapen P, Boellaard R, Lammertsma AA, Westerhof N,
510
Visser FC. Microvascular function in viable myocardium after chronic infarction
114: 1321-1341, 2006.
104: 485-498, 2009.
24: 771-781, 2008.
22 / 28
511
does not influence fractional flow reserve measurements. J Nucl Med 48: 1987-1992,
512
2007.
513
23. Meuwissen M, Chamuleau SA, Siebes M, Schotborgh CE, Koch KT, de
514
Winter RJ, Bax M, de Jong A, Spaan JA, Piek JJ. Role of variability in
515
microvascular resistance on fractional flow reserve and coronary blood flow velocity
516
reserve in intermediate coronary lesions. Circulation 103: 184-187, 2001.
517
24. Molloi S, Kassab GS, Zhou Y. Quantification of coronary artery lumen volume
518
by digital angiography: in vivo validation. Circulation
519
25. Molloi S, Chalyan D, Le H, Wong JT. Estimation of coronary artery hyperemic
520
blood flow based on arterial lumen volume using angiographic images. Int J
521
Cardiovasc Imaging, 2011. [Epub ahead of print]
522
26. Molloi S, Ersahin A, Tang J, Hicks J, Leung CY. Quantification of volumetric
523
coronary blood flow with dual-energy digital subtraction angiography. Circulation
524
93: 1919-1927, 1996.
525
27. Molloi S, Bednarz G, Tang J, Zhou Y, Mathur T. Absolute volumetric
526
coronary blood flow measurement with digital subtraction angiography. Int J Card
527
Imaging 14: 137-145, 1998.
528
28. Nanto S, Masuyama T, Hori M, Shimonagata T, Ohara T, Kubori S. Zero
529
flow pressure in human coronary circulation. Angiology 47: 115-122, 1996.
530
29. Ng MK, Yeung AC, Fearon WF. Invasive assessment of the coronary
531
microcirculation: superior reproducibility and less hemodynamic dependence of index
532
of microcirculatory resistance compared with coronary flow reserve. Circulation
23 / 28
104: 2351-2357, 2001.
533
113: 2054-2061, 2006.
534
30. Sambuceti G, Marzilli M, Fedele S, Marini C, L'Abbate A. Paradoxical
535
increase in microvascular resistance during tachycardia downstream from a severe
536
stenosis in patients with coronary artery disease : reversal by angioplasty. Circulation
537
103: 2352-2360, 2001.
538
31. Shimada K, Sakanoue Y, Kobayashi Y, Ehara S, Hirose M, Nakamura Y,
539
Fukuda D, Yamagishi H, Yoshiyama M, Takeuchi K, Yoshikawa J. Assessment of
540
myocardial viability using coronary zero flow pressure after successful angioplasty in
541
patients with acute anterior myocardial infarction. Heart 89: 71-76, 2003.
542
32. Siebes M, Chamuleau SA, Meuwissen M, Piek JJ, Spaan JA. Influence of
543
hemodynamic conditions on fractional flow reserve: parametric analysis of underlying
544
model. Am J Physiol Heart Circ Physiol 283: H1462-1470, 2002.
545
33. Siebes M, Verhoeff BJ, Meuwissen M, de Winter RJ, Spaan JA, Piek JJ.
546
Single-wire pressure and flow velocity measurement to quantify coronary stenosis
547
hemodynamics and effects of percutaneous interventions. Circulation 109: 756-762,
548
2004.
549
34. Spaan JA, Piek JJ, Hoffman JI, Siebes M. Physiological basis of clinically
550
used coronary hemodynamic indices. Circulation
551
35. Spinale FG, Tanaka R, Crawford FA, Zile MR. Changes in myocardial blood
552
flow during development of and recovery from tachycardia-induced cardiomyopathy.
553
Circulation
554
36. Struijker-Boudier HA, Rosei AE, Bruneval P, Camici PG, Christ F, Henrion
113: 446-455, 2006.
85: 717-729, 1992.
24 / 28
555
D, Levy BI, Pries A, Vanoverschelde JL. Evaluation of the microcirculation in
556
hypertension and cardiovascular disease. Eur Heart J 28: 2834-2840, 2007.
557
37. Tanaka N, Takazawa K, Takeda K, Aikawa M, Shindo N, Amaya K, Kobori
558
Y, Yamashina A. Coronary flow--pressure relationship distal to epicardial stenosis.
559
Circ J 67: 525-529, 2003.
560
38. Van Herck PL, Carlier SG, Claeys MJ, Haine SE, Gorissen P, Miljoen H,
561
Bosmans JM, Vrints CJ. Coronary microvascular dysfunction after myocardial
562
infarction: increased coronary zero flow pressure both in the infarcted and in the
563
remote myocardium is mainly related to left ventricular filling pressure. Heart 93:
564
1231-1237, 2007.
565
39. Versluis JP, Heslinga JW, Sipkema P, Westerhof N. Microvascular pressure
566
measurement reveals a coronary vascular waterfall in arterioles larger than 110
567
microm. Am J Physiol Heart Circ Physiol 281: H1913-1918, 2001.
568
40. White CW, Wright CB, Doty DB, Hiratza LF, Eastham CL, Harrison DG,
569
Marcus ML. Does visual interpretation of the coronary arteriogram predict the
570
physiologic importance of a coronary stenosis? N Engl J Med 310: 819-824, 1984.
571
41. Wong JT, Molloi S. Determination of fractional flow reserve (FFR) based on
572
scaling laws: a simulation study. Phys Med Biol 53: 3995-4011, 2008.
573
42. Wong JT, Le H, Suh WM, Chalyan DA, Mehraien T, Kern MJ, Kassab GS,
574
Molloi S. Quantification of fractional flow reserve based on angiographic image data.
575
Int J Cardiovasc Imaging, 2011. [Epub ahead of print]
576
43. Zhang Z, Takarada S, Molloi S. Quantification of coronary microvascular
25 / 28
577
resistance using angiographic images for volumetric blood flow measurement: In-vivo
578
validation. Am J Physiol Heart Circ Physiol, 2011. [Epub ahead of print]
579
580
26 / 28
581
FIGURE LEGENDS
582
583
Figure 1. Angiographic blood flow measurement using FPA method
584
The calibration phantom was imaged over the heart to determine the correlation
585
between image’s gray level and iodine mass. Correction was made for differential
586
magnification of the phantom and the heart. Then the measured integrated grey levels
587
from the arteries was converted to volume using the calibration curve. The coronary
588
blood flow was determined from the change in volume during one cardiac cycle.
589
590
Figure 2. Pzf measurement
591
The instantaneous blood flow from flow probe (Qq) was plotted against the
592
simultaneously measured coronary distal pressure (Pd) displayed in an X-Y scatter
593
plot. Linear regression between Qq and Pd was used to calculate the slope of the curve
594
in diastole, and the x-intercept of the slope was calculated as Pzf (a). The diastolic
595
interval analyzed was selected from the maximal diastolic blood flow point to the
596
beginning of the phase of rapid decrease of blood flow induced by ventricular
597
contraction (b).
598
599
Figure 3. Comparison of Hemodynamic Indices at different degrees of severity of
600
microvascular disruption
601
Results from ANOVA showed significant differences (p<0.001) in the four groups for
602
each hemodynamic index in a total of 167 measurements. The SNK multiple
603
comparisons were made between the four groups for each index. For all the indices
27 / 28
604
(CFRq, CFRa, NMRqh, NMRah and Pzf), there were significant differences between
605
each group at different degrees of severity of microvascular disruption (p<0.05).
606
607
Figure 4. Diagnosis for Microcirculation Disruption Involving Epicardial
608
Stenosis
609
The figures showed the ROC curves of both normal (N) model (figure a, N=167) and
610
stenotic (S) model (Figure b, N=170). There were significant drops of AUCs for CFRq,
611
CFRa and Pzf in the S model (p<0.05) compared to N model, but no significant
612
differences for NMRqh and NMRah (p>0.05).
28 / 28
TABLE
Table 1. Diagnosis for Microcirculation Disruption
The table showed the areas under the ROC curves (AUC) and the best cut-off values of al; the indices in both normal (N) and stenotic
(S) models. The units of cut-off value for NMR and Pzf are mmHg/ml/min and mmHg, respectively.
N model (N=167)
S model (N=170)
AUC
Best Cut-off Value
Sensitivity
Specificity
AUC
Best Cut-off Value
Sensitivity
Specificity
CFRq
0.855
3.389
0.772
0.833
0.737
3.066
0.705
0.667
CFRa
0.836
3.709
0.698
0.889
0.700
3.139
0.557
0.810
NMRqh
0.976
0.641
0.893
0.944
0.935
0.713
0.826
0.905
NMRah
0.956
0.632
0.886
0.889
0.889
0.761
0.792
0.857
Pzf
0.855
26.563
0.899
0.722
0.698
32.014
0.591
0.857
1/1