Donate Help Contact The AHA Sign In Home
American Heart Association
Circulation
Search: search_blue_button Advanced Search
« Previous Article | Table of Contents | Next Article »
Circulation. 1996;93:1919-1927

This Article
Right arrow Abstract Freely available
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow Request Permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Molloi, S.
Right arrow Articles by Leung, C. Y.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Molloi, S.
Right arrow Articles by Leung, C. Y.

(Circulation. 1996;93:1919-1927.)
© 1996 American Heart Association, Inc.

Articles

Quantification of Volumetric Coronary Blood Flow With Dual-Energy Digital Subtraction Angiography

Sabee Molloi, PhD; Atila Ersahin, PhD; Jerry Tang, PhD; James Hicks, PhD; Cyril Y. Leung, MD

From the Departments of Radiological Sciences (S.M., A.E., J.T.), Medicine (Cardiology) (S.M., C.Y.L.), and Ecology and Evolutionary Biology (J.H.), University of California (Irvine).


*    Abstract
up arrowTop
 *Abstract
down arrowIntroduction
 down arrowMethods
 down arrowResults
 down arrowDiscussion
 down arrowAppendix 1
 down arrowReferences
 
Background As a solution to the well-documented problems associated with visual interpretation of coronary arteriograms, more physiological methods of assessing coronary artery stenosis are being investigated. Volumetric coronary blood flow (BF) can be a valuable aid in the analysis of functional significance of arterial obstruction.

Methods and Results The left anterior descending coronary artery (LAD) of 15 anesthetized pigs (40 to 50 kg) was dissected free from the epicardium in its proximal portion, and a transit-time ultrasound flow probe of the appropriate size was applied. A vascular occluder was positioned distal to the flow probe for flow adjustments. Contrast injections (2 to 4 mL/s for 3 seconds) were made into the left main coronary artery during image acquisition with a motion-immune dual-energy digital subtraction angiography (DE DSA) system. Tissue-suppressed energy-subtracted images were used to generate time-density curves. BF measurements were made in the LAD vascular bed with use of the time-density curve, with consideration that blood was momentarily replaced with contrast during the injection. In 19 comparisons, the mean BF, measured with the use of DE DSA, correlated extremely well with the mean ultrasound flow (DE DSA=0.90 ultrasound+3.10 mL/min, r=.96). Also, contrast injection increased the BF by an average of only 15% during the image-acquisition time interval.

Conclusions Accurate BF measurements can be made with motion-immune DE DSA. The BF measurements can be completed before the onset of significant changes in BF due to contrast injection. Furthermore, it is possible to make the BF measurements during routine coronary arteriography.


Key Words: blood flow • angiography • coronary disease • atherosclerosis • imaging


*    Introduction
up arrowTop
 up arrowAbstract
 *Introduction
down arrowMethods
 down arrowResults
 down arrowDiscussion
 down arrowAppendix 1
 down arrowReferences
 
Coronary arteriography remains the standard method for defining coronary anatomy and assessing coronary atherosclerosis. The limitations of visual estimation of lumen narrowing have been identified.1 2 3 4 There is significant intraobserver and interobserver variability in the interpretation of coronary angiograms, and several pathological studies have shown poor correlation with postmortem anatomic findings.5 6 7 Furthermore, poor correlation with physiological measures of lesion severity has been reported.8 9

Because of the major limitations of standard cineangiography, a functional measure of stenosis severity, such as measurement of coronary blood flow, that is obtainable during cardiac catheterization is desirable. The introduction of digital angiographic techniques provides the potential for obtaining such data during routine coronary angiography through its ability to provide physiological information in addition to the anatomic information present in the angiogram. For example, analysis of the propagation of contrast material through the circulatory system can be used to obtain measurements of physiological variables such as coronary blood flow and flow reserve.10 11

Various angiographic methods of assessing coronary arterial blood flow have been investigated. After contrast injection, blood flow is determined with the contrast pass curve data derived from the epicardial arteries or the myocardial vascular bed. The contrast pass curve data are obtained from epicardial arteries with the use of contrast dilution12 13 14 15 and contrast transit-time methods.16 17 18 However, these techniques have serious limitations, including the need to accurately determine arterial dimensions, misregistration artifacts, and the need for a high frame rate.

The flow measurement techniques based on contrast pass curve data measured in the myocardial vascular bed involve the use of mean contrast transit-time19 20 and parametric imaging methods.21 22 23 24 However, these techniques can be used only to measure relative blood flow because they quantify parameters, such as time of arrival, that are correlated only to coronary blood flow.

Techniques based on the first-pass distribution analysis (FPA) have previously been investigated for the measurement of coronary blood flow.25 26 These techniques combine videodensitometric analysis of spatial and temporal aspects concerning the contrast propagation through the myocardium. The flow measurements are made by summing pixel values in large regions of interest with the use of temporal subtraction images. The major limitations of these studies include the determination of blood iodine concentration, motion misregistration artifacts in temporal subtraction images, and nonlinearities due to physical degradation factors, such as x-ray scatter and veiling glare.

As a solution to the motion misregistration artifacts, we developed a dual-energy subtraction mode that eliminates these misregistration artifacts.27 We also developed a method to correct for nonlinearities due to x-ray scatter and veiling glare.28 29 An in vitro validation of flow measurements with the FPA algorithm and the dual-energy subtraction technique has been previously reported.30

The purpose of the present study was to achieve in vivo validation of the coronary blood flow measurement technique with FPA and a motion-immune dual-energy digital subtraction technique. In addition, we investigated the effects of contrast injection on coronary blood flow and a method for the determination of blood iodine concentration.


*    Methods
up arrowTop
 up arrowAbstract
 up arrowIntroduction
 *Methods
down arrowResults
 down arrowDiscussion
 down arrowAppendix 1
 down arrowReferences
 
Theory
Indicator dilution methods have commonly been used for determination of blood flow. However, the limitations of indicator dilution methods have hampered the routine clinical application of angiographic coronary blood flow measurement techniques. Blood flow into the myocardium can also be modeled with the first-pass distribution theory (Fig 1Down), which overcomes some of the limitations of indicator dilution methods. In the first-pass distribution theory, the volume of the vascular bed (Vp) supplied by a major coronary artery is modeled as a reservoir with a single input. The model does not require any assumptions regarding the internal structure of Vp or the nature of exit conduits.



View larger version (10K):
[in this window]
[in a new window]
  		Figure 1. Schematic of the flow model.

The system input supplies the vascular bed with blood flow at rate Q(t). Flow into the vascular bed is measured by injecting contrast material into the input conduit. The injection of the iodinated contrast material results in concentration C(t) in the blood entering the vascular bed. The concentration C(t) is a dimensionless volume concentration parameter that is equal to the ratio of volume of contrast material in a volume element divided by the total volume of blood and contrast in that element. It is assumed that there is a minimum transit time (tmin) through the vascular bed before which none of the contrast material exits. For times less than this minimum system transit time, all of the contrast material entering the vascular bed accumulates there. Therefore, the volume of contrast that accumulates in time period {Delta}t is calculated as

(1)

for t <= tmin.

If the input arterial concentration is approximately constant [C(t)=C] during the flow measurement, Equation 1Up becomes:

(2)
{smtexp}The average flow for a time period {Delta}t is defined as

(3)

With Equations 2 and 3 combined, the following expression for average volumetric blood flow can be obtained:

(4)

This model makes two necessary assumptions: (1) the flow measurement is performed before the contrast begins to exit the vascular bed, and (2) the input contrast concentration is approximated by the average value C and is known during the flow measurement period.

In Equation 4Up, average flow depends on the difference in contrast material volume ({Delta}V) between two images separated by a time period of {Delta}t and the average iodine concentration (C). Therefore, if absolute videodensitometric measurements of {Delta}V and C are made, the flow rate into the vascular bed can be calculated. It is well known that, in the ideal case, the pixel value in a logarithmically subtracted image is directly proportional to the thickness of contrast material that is present at that spatial location. The constant of proportionality is the effective attenuation coefficient of iodine µI. The volume of contrast material can be calculated by multiplying the thickness with the pixel area. Therefore, the contrast volume over any region is proportional to the integrated videodensitometric iodine signal over that region of the image. Also, the change of contrast volume in the vascular bed can be determined with use of the change in the integrated videodensitometric iodine signal ({Delta}DI) with the expression:

(5)

where {Delta}DI=DI(t+{Delta}t)-DI(t), and it represents the difference in the videodensitometric iodine signal over the region corresponding to the vascular bed for the time period {Delta}t. With Equations 4 and 5 combined, the expression for the average flow rate (Q) for a time period {Delta}t can be rewritten as:

(6)

The attenuation coefficient of iodine µI and the pixel area A can be measured with the use of calibration images. The time period {Delta}t may be measured with the known acquisition frame rate. The iodine concentration C of the bolus entering the myocardial vascular bed is assumed to be the same as the iodine concentration of the contrast agent. This assumption is based on the fact that contrast is injected at a sufficiently high injection rate to momentarily replace blood with contrast.

Animal Preparation
Fasted swine (40 to 55 kg, n=15) were sedated with acepromazine (1.0 mg/kg IM), atropine (0.11 mL/kg IM), ketamine (20 mg/kg IM), and zylazyne (1.0 mg/kg IM). The animals were anesthetized with halothane (0.5% to 2.0%) and nitrous oxide (70% N2O/30% O2) and ventilated with a Harvard Apparatus ventilator. Ventilator settings were adjusted during the experiments to maintain PO2 and PCO2 within normal ranges. A peripheral vein was used for the administration of medication and intravenous fluid. The right femoral artery was isolated and was used to measure the blood pressure with a calibrated Stratham P23ID pressure transducer. A modified left thoracotomy was performed, and the heart was suspended in a pericardial cradle. The left anterior descending coronary artery (LAD) was dissected free from the epicardium in its proximal portion, and a transit-time ultrasound flow probe (Transonic Systems) of appropriate size was applied. A vascular occluder was positioned distal to the ultrasound probe for flow adjustments. Standard techniques were used to cannulate the left main ostium with a 7F multipurpose diagnostic catheter inserted through the femoral artery under fluoroscopic guidance. An intracoronary administration of adenosine (1 to 3 mg at 0.5 mg/mL) and the vascular occluder were used to produce a range of coronary blood flow. ECG, femoral artery blood pressure, LAD blood flow, and x-ray tube voltage (kVp) were recorded continuously (Biopac Systems). The procedures used in this study were approved by the University of California-Irvine Animal Research Committee.

Each animal was positioned under the image intensifier, and the projection was optimized for separation of the LAD and circumflex vascular beds ({approx}30° right anterior oblique). Dual-energy coronary angiograms were then obtained after power injection of 6 to 10 mL of a nonionic iodinated contrast material (Hexabrix, Mallinckrodt Medical) with an injection rate of 2 and 4 mL/s for baseline and maximum hyperemia, respectively. To minimize the effect of ventilation on angiographic acquisitions, the respirator was turned off at the end of full expiration.

Image Acquisition and Processing
All images were acquired with the use of a conventional x-ray tube (Dynamax 79-45/120, Machlett Laboratories), a constant potential x-ray generator (Optimus M200, Philips Medical Systems), a 23/15-cm CsI image intensifier, a focused grid (8:1 grid ratio, 36 lines/cm), and a charge-coupled device camera (Multicam MC-1134GN, Texas Instruments). Light intensity in front of the camera was controlled with an adjustable aperture. The video signal was linearly digitized to 512x512x8-bit precision, and the image processing was performed with a Fischer DA-100 image processor (Fischer Imaging).

The images were acquired with the 15-cm image intensifier mode and the large (1.2 mm nominal) focal spot. A dual-energy mode was used for image acquisition. In this mode, the low-energy beam was formed with a 60- to 70-kVp beam with 2-mm aluminum filtration. The high-energy beam was produced with an x-ray tube voltage of 110 to 120 kVp with an additional 2.5 mm of copper filtration.27 The kVp and beam filtration were switched at 30 Hz. The low- and high-energy images were separated by only 5 milliseconds.

Dual-energy images were formed with a weighted subtraction of logarithmically processed high- and low-energy images.27 Corrections were made for the spatially varying scatter- and veiling glare before logarithmic transformation. A convolution filtering technique was used to estimate scatter-glare distribution in images without the need to sample the scatter-glare intensity for each experiment.28 29 This technique involves the use of the exposure parameters and detected intensity distribution to estimate scatter-glare intensity by predicting the total thickness at every pixel in the image. The thickness information was used to estimate scatter-glare on a pixel-by-pixel basis. Corrections were also made for variation of scatter-glare fraction with beam energy and field size. The detector field nonuniformity and Heel effect were corrected by dividing the detected image with a template image, acquired without scattering material in the x-ray beam.

Flow Measurements
Flow measurements were made on tissue-suppressed energy-subtracted images. A region of interest (ROI) was drawn around the LAD vascular bed (Fig 2Down). The ROI was drawn sufficiently large to ensure that the vascular bed stays within the ROI during the entire cardiac cycle. The average background signal was estimated by drawing a narrow-strip ROI around the vascular bed. The background-corrected videodensitometric iodine signal was used for contrast volume calculations.



View larger version (194K):
[in this window]
[in a new window]
  		Figure 2. Dual-energy coronary arteriogram after injection of contrast into left coronary ostium of a pig. Region of interest used for left anterior descending coronary artery flow measurement is outlined.

An iodine calibration phantom consisting of an array of cylindrical cells 7 mm in diameter and 10 mm in depth was used to calibrate videodensitometric iodine signal for a range of iodine thicknesses (0 to 125 mg/cm2). A typical system calibration curve is shown in Fig 3Down. As previously shown (Equation 4Up), the calibration curve and the known iodine concentration of the contrast material can be used to convert the integrated videodensitometric iodine signal to contrast volume. Therefore, the difference in integrated videodensitometric iodine signal in the vascular bed can be converted to the volume of contrast bolus entering the vascular bed between successive images. A typical curve showing the contrast volume accumulation in the LAD vascular bed after injection of contrast material is shown in Fig 4Down. The contrast volume accumulation in the vascular bed and the known time period between dual-energy images were used to generate phasic flow of contrast into the LAD vascular bed. Mean coronary blood flow was calculated by averaging instantaneous flow measurements over two cardiac cycles. An example of mean coronary blood flow through the LAD vascular bed is shown in Fig 5Down. The initial underestimation is due to the mixture of contrast and blood in the initial segment of the contrast bolus. The underestimation at the end is mainly due to the entrance of diluted contrast material into the vascular bed after the end of contrast injection (injection time of 2.5 seconds). The simultaneous recording of ultrasound flow probe data along with the kVp signal from the x-ray generator, during image acquisition, was used to compare the flow measurements during the same time period.



View larger version (15K):
[in this window]
[in a new window]
  		Figure 3. Calibration curve relating integrated dual-energy iodine signal to total iodine thickness (mg/cm2). Line represents a linear curve fit with the use of regression analysis.



View larger version (16K):
[in this window]
[in a new window]
  		Figure 4. Example of a contrast accumulation curve obtained from left anterior descending coronary artery vascular bed. Decline of contrast volume in vascular bed, toward end of contrast accumulation curve, is an indication of contrast leakage through coronary sinus vein.



View larger version (12K):
[in this window]
[in a new window]
  		Figure 5. Measured mean coronary blood flow calculated with contrast pass curve data shown in Fig 4Up. Actual blood flow measured with ultrasound flow probe was 34.7 mL/min, and heart rate was 112 bpm.

Effect of Contrast Injection on Blood Flow
The intracoronary injection of contrast agents has long been recognized to produce hemodynamic effects, which are mainly due to the physical and chemical properties of the contrast material. The high viscosity and osmolarity relative to blood, sodium content, and pH are some of the properties that are thought to cause these hemodynamic effects. Therefore, it is important to be certain that measured changes in flow are due to the presence or absence of stenosis rather than to perturbation caused by contrast media. Typical mean ultrasound coronary blood flow after an injection of contrast material is shown in Fig 6Down. The immediate effects of intracoronary contrast injection on coronary blood flow have been reported previously.31 These flow alterations occur in three distinct phases. The first phase is the slight increase in flow during contrast injection. The second phase is a reduction in flow, which begins immediately after injection of the contrast material. The third phase consists of the contrast-induced hyperemic response that increases the flow above preinjection levels. In "Appendix," a physical model was used to evaluate the effects of contrast injection on coronary artery blood flow.



View larger version (17K):
[in this window]
[in a new window]
  		Figure 6. Example of mean ultrasound (US) coronary blood flow showing effect of contrast injection on blood flow. Contrast material was power-injected at 3 mL/s for 3 seconds. Contrast injection time and time interval used for first-pass distribution analysis (FPA) flow measurements are shown.

In the present study, the measurements of blood flow were made during the contrast injection time interval. Therefore, only the initial slight increase, characterizing phase 1, affected our flow measurements. During this phase, there are two main factors that need to be considered; the minimum injection rate required for complete replacement of blood with contrast material and the change in blood flow due to contrast injection during the flow measurement time interval.

To determine the injection rate required for complete replacement of blood with contrast material, the catheter was positioned in the left coronary ostium and the injection rate was varied over the range of 0.75 to 4.0 mL/s. A small ROI was positioned on the proximal segment of the LAD in phase-matched subtracted images. The background-corrected videodensitometric signal in the ROI was converted to contrast volume with the use of the system calibration curve. The volume of contrast material enclosed in the ROI was used to generate a contrast pass curve. The maximum of the contrast pass curve was then compared for different injection rates. The minimum injection rate that produced the maximum contrast volume in the ROI was assumed to be the minimum required injection rate for complete replacement of blood with contrast material.

The perturbation in blood flow during the time period used for the FPA flow measurements was calculated with the ultrasound flow probe data. The preinjection blood flow was calculated by averaging LAD blood flow for several cardiac cycles before contrast injection. The change in blood flow was determined by comparing the preinjection blood flow with the average blood flow for the two cycles used for flow measurements with use of the FPA algorithm. Also, in vitro flow measurement was used to establish the fact that accurate flow measurements can be made with the transit-time ultrasound flow probe in the presence of pure contrast material.


*    Results
up arrowTop
 up arrowAbstract
 up arrowIntroduction
 up arrowMethods
 *Results
down arrowDiscussion
 down arrowAppendix 1
 down arrowReferences
 
Flow Measurements
The accumulation of contrast volume in the vascular bed and the known time period between dual-energy images were used to generate phasic flow of contrast into the LAD vascular bed. An example of phasic coronary blood flow measurement with the FPA algorithm and the ultrasound flow probe is shown in Fig 7Down. Mean coronary blood flow was calculated using the instantaneous flow measurements over two cardiac cycles. A comparison of mean coronary blood flow with the FPA algorithm and with the ultrasound flow probe is shown in Fig 8Down. The mean coronary flow was calculated by averaging blood flow during two cardiac cycles ({approx}30 instantaneous flow data points). The two flow measurements were made during the same time period; therefore, the effect of contrast injection on coronary blood flow is eliminated. The measured FPA algorithm (DE) and known ultrasound (US) mean coronary blood flow rates were related by DE=0.90 US-3.10 mL/s (r=.96).



View larger version (28K):
[in this window]
[in a new window]
  		Figure 7. Example of measured coronary blood flow rate in left anterior descending coronary artery of a pig with use of an ultrasonic transit time flow probe (US) and dual-energy first-pass distribution analysis technique (DE). Diastolic (D) and systolic (S) segments of coronary blood flow have been marked. Raw instantaneous first-pass distribution analysis flow data were smoothed with three-point averaging to reduce the noise.



View larger version (16K):
[in this window]
[in a new window]
  		Figure 8. Comparison of mean coronary blood flow measurement with use of the first-pass distribution analysis (FPA) technique and ultrasonic (US) transit time flow probe during the same time interval. Mean coronary blood flow was calculated by averaging instantaneous flow measurements over two heart cycles (30 data points). Solid line represents a linear curve fit with the use of regression analysis, and dashed line represents line of identity.

Effect of Contrast Injection on Blood Flow
An important factor in this flow-measurement technique is the accurate determination of iodine concentration of the contrast bolus entering the vascular bed. Therefore, it is essential to determine the minimum required injection rate to momentarily replace blood with contrast material. It is also necessary to determine the change in coronary blood flow due to contrast injection during the time period that FPA flow measurements were made.

To determine the required injection rate for complete replacement of blood with contrast material, an ROI was positioned on the proximal segment of the LAD. The volume of the contrast material enclosed in the ROI was used to generate contrast pass curves for different injection rates. An example of a contrast pass curve is shown in Fig 9Down. The maximum of the contrast pass curve for different injection rates is shown in Fig 10Down. The contrast volume for different injection rates indicates that an injection rate of {approx}2 mL/s is adequate for complete replacement of blood with contrast material for baseline.



View larger version (13K):
[in this window]
[in a new window]
  		Figure 9. Example of contrast pass curve through left anterior descending coronary artery obtained from a region of interest on a proximal segment of the artery. Diameter and length of the artery segment were {approx}2.5 mm and {approx}2.2 mm, respectively. A total of 9 mL of contrast material was injected into the left coronary ostium at an injection rate of 3 mL/s. Mean coronary blood flow, before contrast injection, was 13.7 mL/min.



View larger version (12K):
[in this window]
[in a new window]
  		Figure 10. Maximum contrast volume in a proximal segment of left anterior descending coronary artery for different injection rates of contrast material in left coronary ostium. Diameter and length of the arterial segment were {approx}2.5 mm and {approx}2.2 mm, respectively. Injection time was kept fixed at 3 seconds. Mean arterial blood flow before contrast injection, measured with use of the transit time ultrasonic flow probe, was 10.4±2.2 mL/min.

The change in coronary blood flow for different injection rates is shown in Fig 11Down; there is only a small change in coronary blood flow during the time period used for the FPA flow measurement. The overall effect of contrast injection on coronary blood flow was determined with ultrasound flow probe data (Fig 12Down). The mean coronary blood flow rates before contrast injection (B) and during the FPA flow measurement (D) time interval were related by D=1.15B+1.82 mL/min (r=.94). These measurements were made with standard injection rates of 2 and 4 mL/s for baseline and maximum hyperemia, respectively.



View larger version (13K):
[in this window]
[in a new window]
  		Figure 11. Increase in mean left anterior descending coronary artery blood flow, during time period used in first-pass distribution analysis measurement, for different injection rates of contrast material in left coronary ostium. Injection time was kept fixed at 3 seconds. Mean arterial blood flow before contrast injection, measured with use of the transit time ultrasonic flow probe, was 10.4±2.2 mL/min.



View larger version (18K):
[in this window]
[in a new window]
  		Figure 12. Comparison of ultrasound (US) coronary blood flow in left anterior descending coronary artery before and during first-pass distribution analysis (FPA) flow measurement time interval. Solid line represents a linear curve fit with the use of regression analysis, and dashed line represents line of identity.


*    Discussion
up arrowTop
 up arrowAbstract
 up arrowIntroduction
 up arrowMethods
 up arrowResults
 *Discussion
down arrowAppendix 1
 down arrowReferences
 
The limitations in the assessment of coronary stenosis from subjective interpretation of standard coronary cineangiograms are well recognized.1 2 3 9 32 Coronary flow reserve has previously been suggested as a reliable measure of the physiological severity of coronary artery stenosis11 ; however, several fundamental limitations exist.33 34 Therefore, a technique capable of measuring the actual coronary blood flow rate would be valuable in quantifying the severity of coronary lesions. Recently, angiographic methods for coronary blood flow measurement with three-dimensional reconstruction of the arterial tree have been reported.35 36 These flow measurements require determination of the volume of an arterial segment and the contrast transit time through that segment. These techniques are limited by the pulsatile nature of coronary blood flow and the change in the contrast bolus profile. In addition, the volume of the arterial segment is calculated with use of diameter measurements with an edge-detection algorithm and the arterial length. These algorithms require an assumption of circular cross section that can produce significant errors in the cases where lesions are present.37

Angiographic methods for measuring coronary blood flow, using the FPA algorithm, have been reported.26 38 Marinus et al26 used the measured volume of the entrance arterial segment for blood iodine calibration. Hangiandreou et al38 used a pump to directly infuse contrast with known iodine concentration into the coronary artery. The limitations of these two techniques include x-ray scatter- and veiling glare, videocamera lag, blood iodine concentration measurement, determination of arterial segment volume, change in coronary blood flow due to contrast injection, and motion misregistration artifacts. These limitations have been addressed in the present study with the use of FPA algorithm in conjunction with dual-energy digital subtraction angiography.

It is well known that x-ray scatter- and veiling glare are two of the major sources of nonlinearities in videodensitometric iodine quantification using image-intensifier/television systems. A convolution filtering method to estimate x-ray scatter- and veiling glare has been reported.39 It was shown that scatter- and veiling glare in humanoid phantom and animal images can be estimated with a root-means-squared error of {approx}8%. This technique does not require sampling of scatter- and veiling glare intensity for each study. It is estimated with image gray level and the readily available x-ray beam parameters.

Previous reports have indicated that videocamera temporal lag causes a systematic underestimation in flow measurements38 and a significant reduction in dual-energy signal.40 As a solution to the significant limitations of the videocamera, with Plumbicon pickup tube, it was replaced with a charge-coupled device camera.40 This camera eliminates the systematic underestimation in flow measurement because it does not have any temporal lag.30 Furthermore, it reduces the possibility of misregistration artifacts in dual-energy images because the low- and high-energy images can be acquired at only 5 milliseconds apart.40

The lack of knowledge regarding blood iodine concentration, after contrast injection, has been another limitation in volumetric coronary blood flow measurement. Two different methods of estimating blood iodine concentration have been reported.26 30 In the first method, referred to as the entrance vessel technique, measured volume of a segment of the entrance vessel was used for calibration purposes.26 In the second method, referred to as videodensitometric calibration technique, it is assumed that blood is momentarily replaced with contrast material.30 The advantage of the entrance vessel technique is that it does not require power injection of contrast material. Its disadvantages are the uncertainties involved in measuring the volume of the entrance arterial segment. The volume is calculated from measured diameter with an edge-detection algorithm and the measured length of the arterial segment. There are significant limitations in quantifying cross-sectional area along the length of an arterial segment.37 Also, orthogonal projections are required for measurement of the arterial length. Another complicating factor is the three-dimensional motion of coronary arteries. The advantages of the videodensitometric calibration, used in this study, include the fact that it does not require the measurement of the entrance vessel volume. Its disadvantage is the requirement of power injection to ensure the complete replacement of blood with contrast material.

Power injection of contrast into a coronary artery produces a back pressure that momentarily prevents blood from entering the coronary artery.37 The magnitude of the generated back pressure depends on the injection rate, the ratio of vascular and aortic resistances, and vessel compliance. The results based on the required injection rate for complete replacement of blood with contrast material indicate that a standard injection rate of 2 and 4 mL/s was adequate for baseline and maximum hyperemia, respectively (see Figs 10Up and 12Up).

Another important factor in coronary blood flow measurement is the significant perturbation in coronary blood flow after contrast injection.31 With this technique, flow measurements were made during contrast injection and completed within 3 seconds after the start of contrast injection. During this time interval, there is only a slight elevation in coronary blood flow, which is caused by the local increase in pressure (see Fig 11Up). Therefore, the flow measurement is completed before the onset of significant changes in coronary blood flow. The magnitude of the elevation in coronary blood flow during contrast injection is influenced by the position of the catheter and whether it has side holes.41 42 The increase in flow is more significant during maximum hyperemia, where the vascular resistance is at its minimum. On the other hand, the elevation in flow is eliminated in the presence of a tight stenosis. The average overestimation in coronary blood flow for different flow conditions was 15% (see Fig 12Up). As the results in Fig 11Up show, the overestimation in blood flow can be minimized when the catheter is carefully positioned in the left coronary ostium. In cases where the tip of the catheter is inside the left main coronary artery, the flow is further elevated during contrast injection.

With the technique that we present, which is the combination of motion-immune dual-energy digital subtraction angiography with the FPA algorithm, absolute volumetric coronary blood flow can be measured accurately. Fig 7Up is an example of phasic coronary blood flow measurement with use of this technique. The noise in instantaneous blood flow measurement limits the measurement of phasic coronary blood flow. However, it is possible to accurately measure mean coronary blood flow with this technique (see Fig 8Up).

The present results demonstrated the feasibility and potential utility of the FPA algorithm in conjunction with motion-immune dual-energy digital subtraction angiography for measuring volumetric coronary blood flow. Quantification of coronary blood flow will be useful for evaluation of the functional significance of stenosis and the effects of interventional procedures on absolute blood flow. The technique can be implemented in available angiographic systems. Flow measurements can be made in conjunction with routine angiography that is performed for assessment of vessel anatomy. Therefore, both anatomic and physiological information can be derived from the same images.

Clinical Implications and Limitations
For clinical use, our technique has some important limitations. One potential limitation is that flow measurements have to be made before contrast starts exiting the vascular bed. Also, diluted contrast material cannot enter the vascular bed during this time interval. However, all of the flow measurements in the present study were made before coronary sinus opacification, which indicates that the LAD vascular bed empties sufficiently slowly to allow the flow measurements to be performed. Also, an injection time of 3 seconds was adequate to ensure that only undiluted contrast material was entering the vascular bed during the flow measurement time interval.

Overlapping of vascular beds from multiple vessels is another potential limitation of this technique. This problem is an inherent limitation of the projection nature of radiographic images. However, it is possible to choose an imaging projection that will minimize potential errors due to overlapping vascular beds.

The application of this flow measurement technique for determination of stenosis severity is limited when collateral blood flow is present. However, clinically, the situations most likely to involve collateral flow are those in which lesion severity is reliably assessed by angiography (eg, >90%). For less severe stenosis, for which angiographic assessment is less reliable, the frequency of appreciable collateral flow is markedly reduced.

Our current implementation of this technique uses dual-energy subtraction angiography that would require some modification for use with the existing cardiac catheterization x-ray equipment. However, this technique can also be implemented with phase-matched diastolic time subtraction images during a breath hold. The flow measurements can be completed within approximately four cardiac cycles, which reduces the possibility of motion misregistration artifacts compared with other techniques requiring temporal subtraction images during approximately 15 to 20 heart beats.20

Our technique offers the ability to compare resting or maximal myocardial blood flow before and after an appropriate intervention such as angioplasty or bypass surgery. However, a limitation of volumetric blood flow measurement in a coronary artery is that it does not account for the size of the vascular bed that is being perfused; that is, it is necessary to know the mass of the perfused muscle to determine whether a single measured maximal myocardial blood flow is normal. A previous study has shown that the regional myocardial mass was linearly related to the sum of coronary artery branch lengths distal to any point in the coronary artery tree.43 44 Therefore, it is possible to use the branch length information from the arteriogram for determination of myocardial mass. The volumetric blood flow measurement can be combined with the estimated myocardial mass to calculate myocardial perfusion. Thus, despite potential clinical limitations, results of the present study contribute to the understanding of the angiographic methods used for coronary blood flow measurement and show that this videodensitometric technique can potentially be used for quantification of myocardial perfusion during routine coronary arteriography.


*    Acknowledgments
 
This work was supported in part by NHLBI grant R01-HL48030. We thank Dr Xian Lao and Guy Shepard for their technical assistance.


*    Footnotes
 
Reprint requests to Sabee Molloi, PhD, Department of Radiological Sciences, Medical Sciences I, B-140, University of California, Irvine, CA 92717. E-mail symolloi@uci.edu.


*    Appendix 1
up arrowTop
 up arrowAbstract
 up arrowIntroduction
 up arrowMethods
 up arrowResults
 up arrowDiscussion
 *Appendix 1
down arrowReferences
 
Effects of Contrast Injection on Coronary Blood Flow
The hemodynamic effects of intracoronary injection of contrast agents due to their physical and chemical properties have long been recognized. The high viscosity and osmolarity relative to blood, sodium content, and pH are some of the properties that are thought to cause the hemodynamic effects. The observed flow alterations can be understood by considering the high viscosity of contrast agents relative to blood and using Poiseuille's law for the resistance Rv of a cylindrical vessel under viscous conditions:

(1A)

where µ is the viscosity of the fluid, r is the inner radius of the vessel, and L is vessel length. With an assumed flow rate of Q (cm3/s), the pressure drop across the cylindrical vessel can be expressed as the following:

(2A)

After the simplifying assumption is made that the single value Rbed can be used to reflect the vascular resistance of a nonstenotic vessel, flow can be expressed as the following:

(3A)

where Pa is aortic pressure, Pv is venous pressure, and RBbed is the vascular resistance for blood flow. After contrast injection, blood is temporarily displaced by contrast. This results in an increase in Rbed because µcontrast>µblood.

A single proximal stenosis adds a further complication because there is an additional pressure drop across the stenosis, {Delta}Ps. The pressure drop across the stenosis, {Delta}Ps, can be expressed in terms of flow, Q, with the following quadratic expression11 :

(4A)

where:

and

In the above equation, a is the viscous resistance previously defined and b is the pressure drop due to exit separation losses. The expression for b contains the fluid density {rho}, the normal cross-sectional area An, and stenotic cross-sectional area As of the vessel. The pressure drop across the stenosis, {Delta}Ps, through the a and b terms, also depends on the physical characteristics of the fluid. Therefore, an increased pressure drop occurs across a stenosis for a higher viscosity fluid such as contrast media. This leads to decreased coronary blood flow.

Although the above simple model neglects several important factors, it is a useful aid in understanding the changes in coronary blood flow after injection of contrast media. During contrast injection, there is a local increase in arterial pressure that is dependent on many parameters, including the injection rate and viscosity of the contrast material. Proximal to the injection site, flow decreases and may actually reverse. The elevated downstream pressure causes an increase in blood flow into the peripheral vascular bed until contrast material has reached the arterioles, which provide the main resistance to flow (see Equations A1Up and A3Up). At the end of the injection, pressure in the artery rapidly drops to the preinjection levels. However, the presence of the high viscosity contrast material in the artery and resistance vessels causes a drop in flow below control levels until the contrast material is washed out of the resistance vessels. The degree of this flow reduction is primarily influenced by the volume and concentration of contrast material.31 Next, as the distal bed dilates in response to the reduction of flow, Rbed reduces and the flow increases. After the contrast has passed through and blood flow resumes, the flow peaks before returning to the baseline flow as Rbed returns to its resting value.

This discussion does not include other physiological factors, such as osmolarity, that have been reported to cause a reduction in flow. However, the general behavior of coronary blood flow after contrast injection appears to be predicted based on viscosity alone. Furthermore, in vitro and in vivo studies have shown that when the injected solution viscosity exceeds the viscosity of the perfusion media, the flow occurring immediately after the injection decreases below control levels.45 Also, it has been found that the use of nonionic contrast material decreases the magnitude of the later hyperemic response but has little effect on the severity of the flow reduction compared with ionic contrast material.46 This is consistent with the fact that both ionic and nonionic contrast materials also have high viscosity.

Received September 5, 1995; revision received October 30, 1995; accepted November 5, 1995.


*    References
up arrowTop
 up arrowAbstract
 up arrowIntroduction
 up arrowMethods
 up arrowResults
 up arrowDiscussion
 up arrowAppendix 1
 *References
 

  1. Detre KM, Wright E, Murphy ML, Takaro T. Observer agreement in evaluating coronary angiograms. Circulation. 1975;52:979-986. [Abstract/Free Full Text]
  2. Zir LM, Miller SW, Dinsmore RE, Gilbert JP, Hawthorne JW. Interobserver variability in coronary angiography. Circulation. 1976;53:627-632. [Abstract/Free Full Text]
  3. Galbraith JE, Murphy ML, de Soyza N. Coronary angiogram interpretation: interobserver variability. JAMA. 1978;240:2053-2056. [Abstract]
  4. Fisher LD, Judkins MP, Lesperance J, Cameron A, Swaye P, Ryan T, Maynard C, Bourassa MG, Kennedy JW, Cosselin A, Kemp H, Faxon D, Wexler L, Davis K. Reproducibility of coronary angiographic reading in the Coronary Artery Surgery Study (CASS). Cathet Cardiovasc Diagn. 1982;8:656-675.
  5. Robbins SL, Rodriquez FL, Wragy AL, Fish SJ. Problems in quantitation of coronary atherosclerosis. Am J Cardiol. 1966;18:153-159. [Medline] [Order article via Infotrieve]
  6. Vlodaver Z, Frech R, Van Tassel RA, Edwards JE. Correlation of the antemortem coronary angiogram and the postmortem specimen. Circulation. 1973;47:162-169. [Abstract/Free Full Text]
  7. Grondin CM, Dyrda I, Pasternac A, Campeau L, Bourassa MG, Lesperance J. Discrepancies between cineangiographic and post-mortem findings in patients with coronary artery disease and recent myocardial revascularization. Circulation. 1974;49:703-708. [Abstract/Free Full Text]
  8. Harrison DG, White CW, Hiratzka LF, Doty DB, Barnes DH, Eastham CL, Marcus ML. The value of lesion cross-sectional area determined by quantitative coronary angiography in assessing the physiological significance of proximal left anterior descending coronary arterial stenoses. Circulation. 1984;69:1111-1119. [Abstract/Free Full Text]
  9. White CW, Wright CB, Doty DB, Hiratza LF, Eastham CL, Harrison DG, Marcus ML. Does visual interpretation of the coronary arteriogram predict the physiologic importance of a coronary stenosis? N Engl J Med. 1984;310:819-824. [Abstract]
  10. Gould KL, Lipscomb K, Hamilton GW. Physiologic basis for assessing critical coronary senosis. Am J Cardiol. 1974;33:87-94. [Medline] [Order article via Infotrieve]
  11. Gould KL, Lipscomb K. Effect of coronary stenoses on coronary flow reserve and resistance. Am J Cardiol. 1974;34:48-55.[Medline] [Order article via Infotrieve]
  12. Foerster JM, Lantz BM, Holcroft JW, Link DP, Mason DT. Angiographic measurement of coronary blood flow by videodilution technique. Acta Radiol Diagn. 1981;22:121-127.
  13. Foerster JM, Link DP, Lantz BM, Lee G, Holcroft JW, Mason DT. Measurement of coronary reactive hyperemia during clinical angiography by videodilution technique. Acta Radiol Diagn. 1981;22:209-216.
  14. Bursch JH. Use of digitized functional angiography to evaluate arterial blood flow. Cardiovasc Intervent Radiol. 1983;6:303-310. [Medline] [Order article via Infotrieve]
  15. Nissen SE, Elion JL, Booth DC, Evans J, DeMaria AN. Value and limitations of computer analysis of digital subtraction angiography in the assessment of coronary flow reserve. Circulation. 1986;73:562-571. [Abstract/Free Full Text]
  16. Rutishauser W, Bussman WD, Noseda G, Meier W, Wellauer J. Blood flow measurement through single coronary arteries by roentgen densitometry. Am J Roentgenol. 1970;109:12-20. [Abstract]
  17. Smith HC, Sturm RE, Wood EH. Videodensitometric system for measurement of vessel blood flow, particularly in the coronary arteries in man. Am J Cardiol. 1973;32:144-150. [Medline] [Order article via Infotrieve]
  18. Spiller P, Schmiel FK, Politz B, Block M, Fermor U, Hackbarth W, Jehle J, Korfer R, Pannek H. Measurement of systolic and diastolic flow rates in the coronary artery system by x-ray densitometry. Circulation. 1983;68:337-347. [Abstract/Free Full Text]
  19. Eigler NL, Pfaff JM, Zeiher A, Whiting JS, Forrester JS. Digital angiographic impulse response analysis of regional myocardial perfusion: linearity, reproducibility, accuracy and comparison with conventional indicator dilution curve parameters in phantom and canine models. Circ Res. 1989;64:853-866. [Abstract/Free Full Text]
  20. Pijls NH, Uijen GJ, Hoevelaken A, Arts T, Aengevaeren WR, Ros HS, Fast JH, Van Leeuwen KL, Van der Werf T. Mean transit time for the assessment of myocardial perfusion by videodensitometry. Circulation. 1990;81:1331-1340. [Abstract/Free Full Text]
  21. Vogel R, LeFree MT, Bates ER, O'Neill W, Foster R, Kirlin P, Smith D, Pitt B. Application of digital techniques to selective coronary arteriography: use of myocardial contrast appearance time to measure coronary flow reserve. Am Heart J. 1984;107:153-164. [Medline] [Order article via Infotrieve]
  22. Hodgson JM, LeGrand V, Bates ER, Mancini GB, Aueron FM, O'Neill WW, Simon SB, Beauman GJ, LeFree MT, Vogel RA. Validation in dogs of a rapid digital angiographic technique to measure relative coronary blood flow during routine cardiac catheterization. Am J Cardiol. 1985;55:188-193. [Medline] [Order article via Infotrieve]
  23. Cusma JT, Toggart EJ, Folts JD, Peppler WW, Hagiandreou NJ, Lee CS, Mistretta CA. Digital subtraction angiographic imaging of coronary flow reserve. Circulation. 1987;75:461-472. [Abstract/Free Full Text]
  24. Zijlstra F, Van Ommeren J, Reiber JH, Serruys PW. Does the quantitative assessment of coronary artery dimensions predict the physiologic significance of a coronary stenosis? Circulation. 1987;75:1154-1161. [Abstract/Free Full Text]
  25. Hangiandreou NJ. Coronary Blood Flow Measurement Using Digital Subtraction Angiography and First Pass Distribution Analysis. Madison, Wis: University of Wisconsin-Madison; 1990. Thesis.
  26. Marinus H, Buis B, Van Benthem A. Pulsatile coronary flow determination by digital angiography. Int J Cardiac Imaging. 1990;5:173-182. [Medline] [Order article via Infotrieve]
  27. Molloi SY, Mistretta CA. Quantification techniques for dual-energy cardiac imaging. Med Phys. 1989;16:209-217. [Medline] [Order article via Infotrieve]
  28. Molloi SY, Mistretta CA. Scatter-glare corrections in quantitative dual-energy fluoroscopy. Med Phys. 1988;15:289-297. [Medline] [Order article via Infotrieve]
  29. Ersahin A. Instantaneous Brain Blood Flow Measurements Using First Pass Distribution Analysis and Digital Subtraction Angiography. Irvine, Calif: University of California-Irvine; 1993. Thesis.
  30. Molloi S, Qian Y, Ersahin A. Absolute volumetric blood flow measurements using dual-energy digital subtraction angiography. Med Phys. 1993;20:85-91. [Medline] [Order article via Infotrieve]
  31. Hodgson J, Mancini J, Legrand V, Vogel R. Characterization of changes in coronary blood flow during the first six seconds after intracoronary contrast injection. Invest Radiol. 1985;20:246-252. [Medline] [Order article via Infotrieve]
  32. DeRouen TA, Murray JA, Owen W. Variability in the analysis of coronary arteriogram. Circulation. 1977;55:324-328. [Abstract/Free Full Text]
  33. Hoffman JI. Maximal coronary flow and the concept of coronary vascular reserve. Circulation. 1984;70:153-159. [Free Full Text]
  34. Klocke FJ. Measurements of coronary flow reserve: defining pathophysiology versus making decisions about patient care. Circulation. 1987;76:1183-1189. [Free Full Text]
  35. Parker DL, Pope DL, Van Bree RE, Marshal H. Blood flow measurements in digital cardiac angiography. In: Reiber JHC, Serruys PW, eds. Developments in Quantitative Coronary Arteriography. Dordrecht, Kluwer Academic Publishers; 1988:215-220.
  36. Guggenheim N, Dorsaz PA, Doriot PA, Suilen C, Chappuis F, Rutishauser W. 3D determination of the intravascular volume and flow of coronary arteries. Int J Bio-Med Comput. 1994;35:13-23.
  37. Molloi SY, Ersahin A, Roeck WW, Nalcioglu O. Absolute cross-sectional area measurement in quantitative coronary arteriography by dual-energy DSA. Invest Radiol. 1991;26:119-127. [Medline] [Order article via Infotrieve]
  38. Hangiandreou N, Folts J, Peppler W, Mistretta C. Coronary blood flow measurement using an angiographic first pass distribution technique: a feasibility study. Med Phys. 1991;18:947-954. [Medline] [Order article via Infotrieve]
  39. Ersahin A, Molloi S, Qian Y. A digital filtering technique for scatter-glare correction based on thickness estimation. IEEE Trans Med Imaging. 1995;14:587-595. [Medline] [Order article via Infotrieve]
  40. Molloi S, Ersahin A, Qian Y. CCD camera for dual-energy digital subtraction angiography. IEEE Trans Med Imaging. 1995;14:747-752. [Medline] [Order article via Infotrieve]
  41. Levin DC, Philips DA, Stanley L, Maroko PR. Hemodynamic changes distal to selective arterial injections. Invest Radiol. 1977;12:116-120. [Medline] [Order article via Infotrieve]
  42. Wolf GL, Shaw DD, Baltaxe HA. A proposed mechanism for transient increases in arterial pressure and flow during angiographic injections. Invest Radiol. 1978;13:195-199. [Medline] [Order article via Infotrieve]
  43. Seiler C, Kirkeeide RL, Gould KL. Basic structure-function relations of the coronary vascular tree: the basis of quantitative coronary arteriography for diffuse coronary artery disease. Circulation. 1992;85:1987-2003. [Abstract/Free Full Text]
  44. Seiler C, Kirkeeide RL, Gould KL. Measurement from arteriograms of regional myocardial bed size distal to any point in the coronary vascular tree for assessing anatomic area at risk. J Am Coll Cardiol. 1993;21:783-797. [Abstract]
  45. Morris T, Kern M, Katzberg R. The effects of media viscosity on the hemodynamics in the selective arteriograph. Invest Radiol. 1982;17:70-76. [Medline] [Order article via Infotrieve]
  46. Vogel RA. Digital coronary arteriography: assessment of coronary flow reserve by parametric imaging. In: Wasserman A, Ross A, eds. Cardiac Application of Digital Angiography. Mount Kisco, NY: Futura Publishing Company; 1989:123-141.



This article has been cited by other articles:


Home page
 CirculationHome page
S. Molloi, G. S. Kassab, and Y. Zhou
Quantification of Coronary Artery Lumen Volume by Digital Angiography: In Vivo Validation
Circulation, November 6, 2001; 104(19): 2351 - 2357.
[Abstract] [Full Text] [PDF]

Home page
 CirculationHome page
A. Abaci, A. Oguzhan, N. K. Eryol, and A. Ergin
Effect of Potential Confounding Factors on the Thrombolysis in Myocardial Infarction (TIMI) Trial Frame Count and Its Reproducibility
Circulation, November 30, 1999; 100(22): 2219 - 2223.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow