Radiology
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF)
Right arrow Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
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 reprints & permissions
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Bude, R. O.
Right arrow Articles by Rubin, J. M.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Bude, R. O.
Right arrow Articles by Rubin, J. M.
(Radiology. 1999;212:79-87.)
© RSNA, 1999

Experimental Studies

Stenosis of the Main Artery Supplying an Organ: Effect of End-Organ Vascular Resistance on the Poststenotic Peak Systolic Velocity in an in Vitro Hydraulic Model at Doppler US1

Ronald O. Bude, MD, Ronald G. Larson, PhD, Wilmer W. Nichols, PhD and Jonathan M. Rubin, MD, PhD

1 From the Department of Radiology, TC 2910K, University of Michigan Medical Center, 1500 E Medical Center Dr, Ann Arbor, MI 48109-0326 (R.O.B., J.M.R.); the Department of Chemical Engineering, University of Michigan, Ann Arbor (R.G.L.); and the Division of Cardiology, Department of Medicine, University of Florida, College of Medicine, Gainesville (W.W.N.). Received June 3, 1998; revision requested July 15; revision received August 25; accepted November 25. Address reprint requests to R.O.B. (e-mail: ronbude@umich.edu).


   Abstract
TOP
 Abstract
Introduction
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 References
 
PURPOSE: To test the hypothesis that increased end-organ vascular resistance reduces blood flow to the kidney, thus reducing the mean velocity in the renal artery and secondarily lowering the peak systolic velocity (PSV).

MATERIALS AND METHODS: An in vitro hydraulic model with a pulsatile pump, blood-mimicking fluid, interchangeable stenoses, and variable compliance and resistance was used to investigate the relationship between end-organ vascular resistance and poststenotic PSV.

RESULTS: Poststenotic PSV was mildly dependent on end-organ vascular resistance and decreased with increasing resistance.

CONCLUSION: The results help explain some of the reported variability from using poststenotic PSV to detect hemodynamically significant renal arterial stenoses, but the effect is not great enough to completely explain the variability. Other factors not investigated in this study must be at work as well.

Index terms: Blood, flow dynamics, 961.72 • Renal arteries, flow dynamics, 961.72 • Renal arteries, stenosis or obstruction, 961.72 • Renal arteries, US, 961.12983, 961.12984 • Test objects, 961.12983, 961.12984, 961.72


   Introduction
TOP
 Abstract
 Introduction
MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 References
 
Variable results have been reported for the detection of hemodynamically significant renal arterial stenosis through evaluation of the poststenotic peak systolic velocity (PSV) (1–19), with sensitivities ranging from 0% (2) to 98% (11). We postulated that renal end-organ vascular resistance might be responsible for some of this variability by causing the poststenotic PSV to decrease as increasing end-organ resistance reduces the blood flow to the kidney, thus reducing the mean velocity in the renal artery.

If this postulate is true, since renal vascular resistance varies from individual to individual (especially in those with renal disease), depending on the end-organ resistances, stenoses of the same geometric severity might produce different PSVs in different individuals. An in vitro hydraulic model was developed, and experiments were performed to evaluate the effect of end-organ vascular resistance on the poststenotic PSV.


   MATERIALS AND METHODS
TOP
 Abstract
 Introduction
 MATERIALS AND METHODS
RESULTS
 DISCUSSION
 References
 
Hydraulic Model
The model (Fig 1) was constructed to study the effect of end-organ vascular resistance on the poststenotic PSV of a stenotic artery that supplies an organ, such as the kidney. A pulsatile pump supplied flow through a single tube that branched, which represented the descending aorta; one branch simulated blood flow to the kidney, and the other branch simulated blood flow to the rest of the body. (These components are referred to as "aorta," "renal" or "kidney," and "body" for clarity and brevity, although we acknowledge this is an in vitro and not an in vivo study.)



View larger version (20K):
[in this window]
[in a new window]
  		Figure 1. Schematic diagram of the experimental design.

 
This setup allowed the flow to be diverted from the kidney to the rest of the body as end-organ vascular resistance increased, as this happens in vivo. Four experimental trials were performed; two were performed with a low-grade stenosis, and two were performed with a high-grade stenosis (discussed later) that was inserted into the renal branch.

For each stenosis, PSVs were obtained immediately downstream of the stenosis with end-organ resistances that ranged from "normal" to infinite (complete occlusion of the distal artery) and with two degrees of distal vascular compliance (vascular compliance downstream of the stenosis; described later): compliance that simulated normal in vivo compliance and compliance lower than the compliance that is possible in vivo (referred to as very low compliance).

Pump and pump output tubing.—A constant-volume pump (model 1421; Harvard Apparatus, Holliston, Mass) supplied pulsatile flow through a single 3/8-inch (9.53-mm)–inside diameter, 9/16-inch (1.43-cm)–outside diameter vinyl tube at fixed settings throughout the experiment: 60 strokes per minute; 10 mL per stroke; with a duty cycle, or systolic fraction of the cardiac cycle, of 0.3.

Because the pump output was very pulsatile without diastolic flow, the fluid-flow analogue of a resistive-capacitive circuit (20) was installed immediately downstream of the pump to modify the pump output waveform. By moving the clamps up or down the tubing to adjust the position and by tightening or loosening the clamp to adjust the degree of the stenoses on the two parallel lengths of gum rubber tubing (5/16-inch [7.94-mm] inner diameter, 7/16-inch [1.11-cm] outer diameter; 24 and 45 cm long) (Fig 1) that composed this network, it was possible to produce diastolic flow in the renal arterial waveform, so each experimental trial started with as nearly the same prestenotic renal arterial resistive index (RI) as possible (actual RIs measured 0.52 or 0.53).

Aorta, renal branch, and body branch.—After the resistive-capacitive network, the descending aorta was simulated with 1/4-inch (6.35-mm)–inner diameter, 3/8-inch (9.53-mm)–outer diameter vinyl tubing that bifurcated into the body branch (and that was entirely composed of the same tubing that formed the aorta) and the renal branch (discussed later).

Needle valves (catalog no. 6393-60; Cole-Parmer Instrument, Vernon Hills, Ill) near the ends of these branches simulated the total vascular resistances of the body and kidney. Initially, these valves were set so the mean pressure was as close to physiologic pressure as the model allowed (pressure ranged from 58 to 90 mm Hg for all experimental trials and increased with increasing renal end-organ resistance), and so approximately 10% of pump output flowed through the renal branch, since the normal kidney receives approximately 10% of the cardiac output (21). Renal volume flows for the four trials were as follows: for high-grade stenosis with normal compliance, 9.68% (0.70 mL/sec flow to the kidney ÷ [0.70 mL/sec flow to the kidney + 6.53 mL/sec flow to the body]); for low-grade stenosis with very low compliance, 9.92% (0.72 mL/sec ÷ [0.72 mL/sec + 6.54 mL/sec]); for low-grade stenosis with normal compliance, 9.94% (0.77 mL/sec ÷ [0.77 mL/sec + 6.98 mL/sec]); and for high-grade stenosis with very low compliance, 10.1% (0.68 mL/sec ÷ [0.68 mL/sec + 6.08 mL/sec]) (Table).


View this table:
[in this window]
[in a new window]
  		Experimental Data
 
Once set at the beginning of the first trial, the valve on the body side was not readjusted throughout the remainder of the experiment. When stenoses were interchanged for a new experimental trial, the renal resistance valve was readjusted until renal flow was approximately 10% of pump output at the beginning of that trial.

The proximal renal branch, which extended from the aorta through the stenotic region, was composed of the same vinyl tubing that formed the aorta and was the branch through which Doppler ultrasonography (US) was successfully performed. The remainder of the renal branch was composed of stiff polypropylene tubing to simulate as closely as possible zero compliance in this portion; the tubing was so stiff that intraluminal signals could not be obtained because the ultrasound beam could not sufficiently penetrate the tubing. The polypropylene tubing had a 1/4-inch (6.35-mm) inner diameter and a 3/8-inch (9.53-mm) outer diameter, except in the following regions (discussed later): the stenoses; the compliance region (Fig 1); a 27-cm segment of vinyl tubing immediately proximal to the distal resistance valve that provided an insonation port, when necessary; and the short lengths of vinyl tubing that allowed leak-free placement of pressure monitor needles (silicone caulk prevented leakage from the vinyl tubing, but no method could be found to prevent leakage with the polypropylene tubing) and that allowed the clamps to occlude flow, when flow was diverted from the compliance region, because the polypropylene tubing was too stiff to clamp effectively (the polypropylene tubing was so stiff that it could not be perceptibly compressed by hand). The tubing was bent to fit the model by softening it in nearly boiling water and by allowing it to cool and to set in place.

Flow through the body and renal branches returned to a common reservoir that was placed on a magnetic stirrer (Magnestir; A.S. Aloe, St Louis, Mo). Renal arterial pressures upstream of the stenosis, downstream of the stenosis but upstream of the renal vascular resistance valve, and downstream of the renal vascular resistance valve were measured with pressure transducers (Transpac IV; Abbott Laboratories, Hospital Products, North Chicago, Ill) (Fig 1) that were connected to a monitor (model 78354A; Hewlett-Packard, Bad Homburg, Germany).

Interchangeable renal arterial stenoses.—Two renal arterial stenoses were made with 23-mm-long segments of copper tubing with 0.7- and 2.5-mm inner diameters, which produced 61% diameter ([6.35 mm - 2.5 mm]/6.35 mm) and 89% diameter ([6.35 mm - 0.7 mm]/6.35 mm) stenoses in the 1/4-inch (6.35-mm)–inner diameter vinyl tubing. These stenoses were glued into centrally drilled holes in 2-cm lengths of 5/16-inch (7.94-mm)-diameter wooden dowel, which was carefully painted to preclude water absorption or exudation, so that the tubing edges extended from the cut surfaces of the dowel plugs approximately 1.5 mm on each side.

Spectral Doppler US was performed immediately distal to the copper-tubing edges, which could be clearly seen at US. Color or power Doppler US of the poststenotic jet was performed to correctly determine the insonation angle. The 5/16-inch (7.94-mm)–outer diameter wooden dowel segments slid into the 1/4-inch (6.35-mm)–inner diameter vinyl tubing, which facilitated stenosis interchanges, to produce a watertight fit.

At the beginning of the experiment, pressure gradients across the stenoses and across the renal resistance valve were measured to determine the relative hemodynamic significance of the stenoses. The pressure gradients across the 61% diameter stenosis and across the renal resistance valve were 6 and 53 mm Hg, respectively, which gave a ratio of the stenotic pressure gradient to the total renal arterial pressure gradient of 10% (6 mm Hg/[6 mm Hg + 53 mm Hg]). The pressure gradients across the 89% diameter stenosis and across the renal resistance were 39 and 30 mm Hg, respectively, which gave a ratio of the stenotic pressure gradient to the total renal arterial pressure gradient of 57% (39 mm Hg/[39 mm Hg + 30 mm Hg]).

In the typical human, the mean arterial pressure is approximately 100 mm Hg (22). Therefore, a pressure gradient of 10% in the human renal artery would give a 10 mm Hg pressure gradient for our 61% diameter stenosis (10% x 100 mm Hg = 10 mm Hg), and a pressure gradient of 57% in the human renal artery would give a 57 mm Hg pressure gradient for our 89% diameter stenosis (57% x 100 mm Hg = 57 mm Hg).

In our institution, University of Michigan, Ann Arbor, a 10 mm Hg transstenotic renal arterial pressure gradient is the threshold level for angioplasty or the placement of stents, whereas a 57 mm Hg transstenotic pressure gradient is so high that it is not often encountered. On the basis of these results, we refer to these stenoses as low- and high-grade stenoses throughout this article.

Renal arterial compliance.—A pulse dampener (catalog no. 07596-20; Cole-Parmer Instrument), which is a hollow plastic container roughly the shape of a half sphere that rests on its flat surface (Fig 1), was used to provide finite compliance downstream of the stenosis, which simulated the normal compliance of the renal vasculature. Air that is introduced through a valve that is inserted into the top of this device collects at the top, while liquid flows, unmixed with air, through the base. The entrapped air compresses during systole and expands during the lower pressure of diastole, thus absorbing some of the flow pulsatility. If the quantity of air is large enough, highly pulsatile flow can be completely converted to steady, nonpulsatile flow. This device is a windkessel, which is well known in the physiology literature and which was first described by Hales near the turn of the 18th century (23).

Trial-and-error adjustment of the amount of air in the pulse dampener revealed that 1.3 mL of air for the low-grade stenosis and 3.0 cm3 of air for the high-grade stenosis were sufficient to damp the input renal arterial RI from the initial value of 0.52 or 0.53 to 0.20, or near nonpulsatility (Fig 2). An RI of 0.20 was chosen rather than an RI of 0 because at very low RIs it is extremely difficult to measure the RI accurately and because an RI of 0.20 was so nearly nonpulsatile that it was considered adequate to simulate the degree of pulsatility damping that happens when flow progresses from the pulsatile renal arteries to the essentially nonpulsatile capillaries. Yet, the flow was still pulsatile enough to be measured and reproduced accurately.



View larger version (109K):
[in this window]
[in a new window]
  		Figure 2. Waveforms used to validate normal and very low compliances used with low-grade stenosis. 1, Input waveforms in the renal artery upstream of the stenosis, without compliance. Mean RI is 0.53 ± 0.02. 2, Nearly nonpulsatile waveforms downstream of the stenosis, with compliance. Mean RI was 0.20, which indicates a large portion of the pulsatility was damped by the compliance. 3, Waveforms downstream of the stenosis, with flow diverted past the compliance through a diverting tube (Fig 1). Mean RI is 0.48 ± 0.03. Very slight change in RIs between the waveforms in 1 and 3 suggests our model had very little compliance in the very low compliance mode.

 
These amounts of air were used in the pulse dampener for the normal compliance trials, since they produce approximately the same amount of pulsatility damping that occurs in vivo. For the very low compliance trials, clamps diverted flow away from the pulse dampener through polypropylene tubing (Fig 1).

Inlet lengths of tubing.—A flow perturbation (eg, bifurcation, stenosis, tight turn) transiently alters the flow profile for a variable distance, up to a maximal length known as the inlet length (24). Since the nonpulsatile laminar–flow inlet length is longer than the inlet lengths for other types of flow, it was used to ensure that the Doppler measurements were obtained in areas of stable flow. All US and pressure measurements were obtained at least one inlet length from any flow perturbation, except for the poststenotic PSV, which was obtained in the poststenotic jet at the distal orifice of the stenosis. The nonpulsatile laminar–flow inlet length is defined as follows: L = 2kvr2/{upsilon}, where L is the inlet length in centimeters, k is an experimentally derived constant with a value of 0.08, v is the mean velocity in centimeters per second, r is the tube radius in centimeters, and {upsilon} is the kinematic viscosity in stokes (centimeters squared per second) (24).

Blood-mimicking liquid.—The model was filled with 1,500 mL of a solution with the mean viscosity of blood (2.5 centipoise [25]) that was composed of 35 mL of glycerol and 0.67 g of microparticles (Sephadex G-50; Sigma Chemical, St Louis, Mo) as ultrasound scatterers for every 100 mL of water (26). Green food coloring tinted the fluid, so air bubbles could be seen and eliminated from the nearly opaque white tubing, since bubbles cyclically changed in volume with the cardiac cycle and caused undesirable and unmeasurable compliance.

Measuring Renal Vascular Resistance
The fluid-flow analogue of the Ohm law (P = QR, where P is the transstenotic pressure gradient, Q is the volume flow rate, and R is the resistance) was not used to calculate vascular resistance. This is because the Ohm law applies to steady flow, but it might not apply to pulsatile flow (27,28). Since the only variable altered, once the compliance and stenosis were selected during each experimental trial, was the degree of stenosis of the renal resistance valve (described in a subsequent section), renal resistance was assumed to be related to the ratio of the flow rate at increased resistance to the baseline flow rate, with this ratio decreasing with increasing renal resistance because the flow to the kidney decreased with increasing renal resistance. This is the same method used by Norris and Barnes (29) in an in vivo dog study in which they investigated the relationship between RI and vascular resistance.

Doppler US
A freestanding aluminum frame supported the US transducers (Spectra; Diasonics Vingmed Ultrasound, Santa Clara, Calif). Doppler waveforms were obtained with 5.0-MHz curvilinear transducers (Doppler frequency, 4.0 MHz) at gains at which noise first became apparent and at pulse repetition frequencies that were sufficient to prevent aliasing. Waveforms were obtained at the aorta (48° insonation angle, 2.9–4.0-kHz pulse repetition frequency, 75–105-Hz wall filter), in the renal artery proximal to the stenosis (48°–53° insonation angle, 1.4-kHz pulse repetition frequency, 35-Hz wall filter), at the distal orifice of the low-grade stenosis (43°–51° insonation angle, 2.9–6.0-kHz pulse repetition frequency, 75–160-Hz wall filter), and at the distal orifice of the high-grade stenosis (59°–70° insonation angle, 14.2–22.2-kHz pulse repetition frequency, 375–590-Hz wall filter).

Color and power Doppler US ensured accurate angle correction for spectral Doppler US in the poststenotic jet. The large insonation angles for the high-grade stenosis were needed to prevent aliasing in the very high velocity jet. Renal arterial US was performed in a water bath (wallpaper soaking tray); water-filled plastic bags (acting as interfaces between the transducer and tubing) and US gel were used for US elsewhere. A sound-absorbent material (70-durometer Sorbothane; Sorbothane, Kent, Ohio) was interposed between the tubing and its resting surfaces to reduce ultrasound reverberations.

PSVs and RIs were measured by hand using calipers, with RIs calculated according to the following formula: RI = (S - D)/S, where S is the height of the systolic peak and D is the height of the end-diastolic trough. All reported PSVs and RIs were means and were obtained by averaging the values for four to five consecutive waveforms (RIs) or five consecutive waveforms (PSVs). The Doppler scale that produced the largest possible waveform without aliasing was used to decrease measurement error.

Validation That the Model, Downstream of the Stenosis and without the Windkessel, Had Very Little Compliance
Vascular compliance is defined as dV/dP, where V is the volume and P is the pressure. It is a dynamic phenomenon that could not be measured in our model. Therefore, an indirect method was used to determine if our model had appreciable compliance in the very low compliance mode. Vascular compliance, in conjunction with vascular resistance, causes the progressive dampening of pulsatility that occurs as flow progresses from the highly pulsatile central arteries to the essentially nonpulsatile capillaries (30–33).

Therefore, our model, in the very low compliance mode (in the absence of the pulse dampener) can be assumed to have very low compliance, which is lower than that which can occur in vivo if the RI in the distal renal artery just proximal to the resistance valve (which, in terms of the distal RI, corresponds to the small arteriolar or capillary region) is nearly the same as the RI in the proximal renal artery during maximal flow.

Waveforms at the input and output portions of the renal segment of our model, with the low-grade stenosis in the very low compliance mode at a 10% renal volume flow, had RIs of 0.53 ± 0.02 (SD) and 0.48 ± 0.03, respectively (Fig 2), which indicated the compliance in this mode was very low, much lower than that which occurs in vivo.

Experimental Trials
Four experimental trials were performed; two trials each were performed with 61% and 89% diameter stenoses. For each stenosis, trials were performed with normal and very low compliances by varying the flow into the renal branch from a baseline of approximately 10% of pump output to complete occlusion of flow in five increments, for a total of six measurements per trial (Table). This was done by progressively tightening the distal renal resistance valve, which produced increasing renal end-organ resistances. Since the purpose of the experiment was to show the relationship of PSV to end-organ resistance, initial input flows were adjusted to be very nearly the same; they ranged from 9.68% to 10.1% (Table), but they were not adjusted until they were all exactly 10.0%.

The upstream compliance assembly was adjusted so that each trial commenced with a renal arterial RI, as measured on the monitor of the US unit, of 0.53 (which measured 0.53 for three of the trials and 0.52 for the trial with high-grade stenosis and normal compliance at subsequent measurement with hand calipers on the hard-copy images at the time of data analysis). This RI value was chosen because it is within the in vivo range and because it was the maximal value the model could achieve. PSVs immediately distal to the stenoses and in the aorta were obtained at each flow increment (Table).

Data Analysis
Since the aortic PSV varied slightly during the experiment (Table), and since this in turn affected the poststenotic PSV, all poststenotic PSVs were normalized to the aortic PSV by dividing the poststenotic PSV by the aortic PSV to alleviate this effect. Normalized poststenotic PSVs were plotted against the renal flow percentage for each trial (Fig 3).



View larger version (13K):
[in this window]
[in a new window]
  		Figure 3a. Normalized PSV versus volume flow percentage to the kidney for (a) high-grade stenosis and normal compliance, (b) high-grade stenosis and very low compliance, (c) low-grade stenosis and normal compliance, and (d) low-grade stenosis and very low compliance. Error bars (± 1 SD) are not visible at the highest volume flow percentage in a because the SD is very small (0.08). Linear regression lines have been plotted. PSV decreases mildly with increasing vascular resistance for all four modes of the experiment.

 


View larger version (13K):
[in this window]
[in a new window]
  		Figure 3b. Normalized PSV versus volume flow percentage to the kidney for (a) high-grade stenosis and normal compliance, (b) high-grade stenosis and very low compliance, (c) low-grade stenosis and normal compliance, and (d) low-grade stenosis and very low compliance. Error bars (± 1 SD) are not visible at the highest volume flow percentage in a because the SD is very small (0.08). Linear regression lines have been plotted. PSV decreases mildly with increasing vascular resistance for all four modes of the experiment.

 


View larger version (14K):
[in this window]
[in a new window]
  		Figure 3c. Normalized PSV versus volume flow percentage to the kidney for (a) high-grade stenosis and normal compliance, (b) high-grade stenosis and very low compliance, (c) low-grade stenosis and normal compliance, and (d) low-grade stenosis and very low compliance. Error bars (± 1 SD) are not visible at the highest volume flow percentage in a because the SD is very small (0.08). Linear regression lines have been plotted. PSV decreases mildly with increasing vascular resistance for all four modes of the experiment.

 


View larger version (15K):
[in this window]
[in a new window]
  		Figure 3d. Normalized PSV versus volume flow percentage to the kidney for (a) high-grade stenosis and normal compliance, (b) high-grade stenosis and very low compliance, (c) low-grade stenosis and normal compliance, and (d) low-grade stenosis and very low compliance. Error bars (± 1 SD) are not visible at the highest volume flow percentage in a because the SD is very small (0.08). Linear regression lines have been plotted. PSV decreases mildly with increasing vascular resistance for all four modes of the experiment.

 

   RESULTS
TOP
 Abstract
 Introduction
 MATERIALS AND METHODS
 RESULTS
DISCUSSION
 References
 
The data for the four experimental trials are presented in the Table and in Figures 3–5. Note that the poststenotic PSV decreased with increasing end-organ vascular resistance for both stenoses, with both normal and very low vascular compliances. Poststenotic waveforms for the high-grade stenosis are shown in Figures 4 and 5.



View larger version (57K):
[in this window]
[in a new window]
  		Figure 4a. Sequence of poststenotic Doppler waveforms, with the high-grade stenosis and normal vascular compliance, from lowest (part 1) to highest (part 6) vascular resistance. The fifth set of waveforms was inadvertently not obtained due to experimental error. Note the progressive narrowing of the systolic peak (arrowhead in 1 and 6) and the progressive loss of antegrade diastolic flow (arrow in 1 and 6) as vascular resistance increases, with relative preservation of the PSV and with reversed diastolic flow (aliased) at complete distal vascular occlusion (part 6). (Reversed diastolic flow is analogous to the reversed flow that occurs in vivo in renal vein thrombosis.) The amplitudes of the waveforms in 3 appear artifactually lower than those of the other waveforms owing to a larger velocity scale. In 6, diastolic aliasing could not be eliminated, because displaying the baseline higher than the bottom of the image would result in aliasing of the systolic peak and also because the waveform scale could not be made any larger due to the insonation angle required by the stenotic jet.

 


View larger version (54K):
[in this window]
[in a new window]
  		Figure 4b. Sequence of poststenotic Doppler waveforms, with the high-grade stenosis and normal vascular compliance, from lowest (part 1) to highest (part 6) vascular resistance. The fifth set of waveforms was inadvertently not obtained due to experimental error. Note the progressive narrowing of the systolic peak (arrowhead in 1 and 6) and the progressive loss of antegrade diastolic flow (arrow in 1 and 6) as vascular resistance increases, with relative preservation of the PSV and with reversed diastolic flow (aliased) at complete distal vascular occlusion (part 6). (Reversed diastolic flow is analogous to the reversed flow that occurs in vivo in renal vein thrombosis.) The amplitudes of the waveforms in 3 appear artifactually lower than those of the other waveforms owing to a larger velocity scale. In 6, diastolic aliasing could not be eliminated, because displaying the baseline higher than the bottom of the image would result in aliasing of the systolic peak and also because the waveform scale could not be made any larger due to the insonation angle required by the stenotic jet.

 


View larger version (44K):
[in this window]
[in a new window]
  		Figure 4c. Sequence of poststenotic Doppler waveforms, with the high-grade stenosis and normal vascular compliance, from lowest (part 1) to highest (part 6) vascular resistance. The fifth set of waveforms was inadvertently not obtained due to experimental error. Note the progressive narrowing of the systolic peak (arrowhead in 1 and 6) and the progressive loss of antegrade diastolic flow (arrow in 1 and 6) as vascular resistance increases, with relative preservation of the PSV and with reversed diastolic flow (aliased) at complete distal vascular occlusion (part 6). (Reversed diastolic flow is analogous to the reversed flow that occurs in vivo in renal vein thrombosis.) The amplitudes of the waveforms in 3 appear artifactually lower than those of the other waveforms owing to a larger velocity scale. In 6, diastolic aliasing could not be eliminated, because displaying the baseline higher than the bottom of the image would result in aliasing of the systolic peak and also because the waveform scale could not be made any larger due to the insonation angle required by the stenotic jet.

 


View larger version (60K):
[in this window]
[in a new window]
  		Figure 4d. Sequence of poststenotic Doppler waveforms, with the high-grade stenosis and normal vascular compliance, from lowest (part 1) to highest (part 6) vascular resistance. The fifth set of waveforms was inadvertently not obtained due to experimental error. Note the progressive narrowing of the systolic peak (arrowhead in 1 and 6) and the progressive loss of antegrade diastolic flow (arrow in 1 and 6) as vascular resistance increases, with relative preservation of the PSV and with reversed diastolic flow (aliased) at complete distal vascular occlusion (part 6). (Reversed diastolic flow is analogous to the reversed flow that occurs in vivo in renal vein thrombosis.) The amplitudes of the waveforms in 3 appear artifactually lower than those of the other waveforms owing to a larger velocity scale. In 6, diastolic aliasing could not be eliminated, because displaying the baseline higher than the bottom of the image would result in aliasing of the systolic peak and also because the waveform scale could not be made any larger due to the insonation angle required by the stenotic jet.

 


View larger version (57K):
[in this window]
[in a new window]
  		Figure 4e. Sequence of poststenotic Doppler waveforms, with the high-grade stenosis and normal vascular compliance, from lowest (part 1) to highest (part 6) vascular resistance. The fifth set of waveforms was inadvertently not obtained due to experimental error. Note the progressive narrowing of the systolic peak (arrowhead in 1 and 6) and the progressive loss of antegrade diastolic flow (arrow in 1 and 6) as vascular resistance increases, with relative preservation of the PSV and with reversed diastolic flow (aliased) at complete distal vascular occlusion (part 6). (Reversed diastolic flow is analogous to the reversed flow that occurs in vivo in renal vein thrombosis.) The amplitudes of the waveforms in 3 appear artifactually lower than those of the other waveforms owing to a larger velocity scale. In 6, diastolic aliasing could not be eliminated, because displaying the baseline higher than the bottom of the image would result in aliasing of the systolic peak and also because the waveform scale could not be made any larger due to the insonation angle required by the stenotic jet.

 


View larger version (56K):
[in this window]
[in a new window]
  		Figure 5a. Sequence of poststenotic Doppler waveforms, with the high-grade stenosis and very low vascular compliance, from lowest (part 1) to highest (part 6) vascular resistance. Note the waveforms in 6 have a smaller velocity scale than that of the other waveforms. PSV (arrowhead in 1 and 6) of the waveforms in 6 was the lowest of the series. (Retrograde diastolic flow is not depicted at infinite resistance probably because the flow entering the nearly noncompliant vessel during systole is so small that retrograde diastolic flow is lost in the wall filter.)

 


View larger version (55K):
[in this window]
[in a new window]
  		Figure 5b. Sequence of poststenotic Doppler waveforms, with the high-grade stenosis and very low vascular compliance, from lowest (part 1) to highest (part 6) vascular resistance. Note the waveforms in 6 have a smaller velocity scale than that of the other waveforms. PSV (arrowhead in 1 and 6) of the waveforms in 6 was the lowest of the series. (Retrograde diastolic flow is not depicted at infinite resistance probably because the flow entering the nearly noncompliant vessel during systole is so small that retrograde diastolic flow is lost in the wall filter.)

 


View larger version (52K):
[in this window]
[in a new window]
  		Figure 5c. Sequence of poststenotic Doppler waveforms, with the high-grade stenosis and very low vascular compliance, from lowest (part 1) to highest (part 6) vascular resistance. Note the waveforms in 6 have a smaller velocity scale than that of the other waveforms. PSV (arrowhead in 1 and 6) of the waveforms in 6 was the lowest of the series. (Retrograde diastolic flow is not depicted at infinite resistance probably because the flow entering the nearly noncompliant vessel during systole is so small that retrograde diastolic flow is lost in the wall filter.)

 


View larger version (51K):
[in this window]
[in a new window]
  		Figure 5d. Sequence of poststenotic Doppler waveforms, with the high-grade stenosis and very low vascular compliance, from lowest (part 1) to highest (part 6) vascular resistance. Note the waveforms in 6 have a smaller velocity scale than that of the other waveforms. PSV (arrowhead in 1 and 6) of the waveforms in 6 was the lowest of the series. (Retrograde diastolic flow is not depicted at infinite resistance probably because the flow entering the nearly noncompliant vessel during systole is so small that retrograde diastolic flow is lost in the wall filter.)

 


View larger version (54K):
[in this window]
[in a new window]
  		Figure 5e. Sequence of poststenotic Doppler waveforms, with the high-grade stenosis and very low vascular compliance, from lowest (part 1) to highest (part 6) vascular resistance. Note the waveforms in 6 have a smaller velocity scale than that of the other waveforms. PSV (arrowhead in 1 and 6) of the waveforms in 6 was the lowest of the series. (Retrograde diastolic flow is not depicted at infinite resistance probably because the flow entering the nearly noncompliant vessel during systole is so small that retrograde diastolic flow is lost in the wall filter.)

 


View larger version (46K):
[in this window]
[in a new window]
  		Figure 5f. Sequence of poststenotic Doppler waveforms, with the high-grade stenosis and very low vascular compliance, from lowest (part 1) to highest (part 6) vascular resistance. Note the waveforms in 6 have a smaller velocity scale than that of the other waveforms. PSV (arrowhead in 1 and 6) of the waveforms in 6 was the lowest of the series. (Retrograde diastolic flow is not depicted at infinite resistance probably because the flow entering the nearly noncompliant vessel during systole is so small that retrograde diastolic flow is lost in the wall filter.)

 
From the linear regressions, we determined that the calculated poststenotic PSVs decreased from normal renal end-organ resistance (10% volume flow through the kidney) to infinite resistance (complete occlusion of distal renal vasculature) as follows: 42% decrease ([{1.60 calculated normalized PSV - 0.93 calculated normalized PSV at infinite resistance} ÷ 1.60 calculated normalized PSV] x 100) for the low-grade stenosis and normal compliance, 44% decrease ([{1.00 - 0.56} ÷ 1.00] x 100) for the low-grade stenosis and very low compliance, 8% decrease ([{9.69 - 8.96} ÷ 9.69] x 100) for the high-grade stenosis and normal compliance, and 29% decrease ([{9.04 - 6.41} ÷ 9.04] x 100) for the high-grade stenosis and very low compliance.

For the high-grade stenosis with normal compliance, the very slight PSV decrease with increasing end-organ resistance is better depicted on the graph (Fig 3a) than in the waveform illustrations (Fig 4). The major visual alterations of the waveforms are a decrease in the width of the systolic pulse with increasing end-organ vascular resistance and a progressive loss of antegrade diastolic flow, which culminated in reversed diastolic flow at high (infinite) end-organ vascular resistance (Fig 4). A similar pattern (not illustrated) was seen with the low-grade stenosis.

For the high-grade stenosis with very low compliance, the poststenotic PSV progressively decreased with increasing end-organ vascular resistance. The progressive narrowing of the systolic pulse and the loss of antegrade diastolic flow that were noted at normal compliance were again present, but they were more pronounced (Fig 5). At infinite end-organ vascular resistance (complete occlusion of flow through the renal artery) (Fig 5, part 6), the width of the systolic pulse was extremely narrow. Even at infinite resistance, however, with a system as noncompliant as possible (the model at very low compliance had a compliance that is much lower than that which occurs in vivo), the poststenotic PSV was still relatively preserved when compared with the marked overall change in waveform morphology that occurred. Similar results were again seen with the low-grade stenosis.


   DISCUSSION
TOP
 Abstract
 Introduction
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
References
 
Hemodynamically significant renal arterial stenoses can be detected by measuring the jet of increased velocity that the stenosis causes. Many investigators (1–14) have evaluated the PSV of the stenotic jet for this purpose. Unfortunately, widely disparate results have been reported in the detection of hemodynamically significant stenoses; results range from 0% sensitivity and 37% specificity (2) to 98% sensitivity and 98% specificity (11).

Technical factors, such as obesity and bowel gas, preclude visualization of the renal artery. Also, accessory arteries, which may be stenotic, occur in up to approximately 20% of kidneys and are often not appreciated during US (2,3,34). These factors limit the usefulness of this technique, but they do not completely explain the discrepant results. Furthermore, a large range of PSV thresholds has been suggested for stenosis detection, with values ranging from 100 to 350 cm/sec advocated (1–14).

As renal end-organ vascular resistance increases, flow is shunted from the kidney to the aorta, thus reducing the volume of flow passing through the stenosis. We believed this reduced flow might decrease the stenotic PSV and might help explain the variable thresholds that have been reported. Since end-organ renal vascular resistance varies from individual to individual due to physiologic variation and renal disease, stenoses of the same degree of hemodynamic severity might, therefore, produce different PSVs in different individuals, if our hypothesis is correct.

To test this hypothesis in vitro, we studied the effect of varying the renal resistance from normal to infinite. We did this using two values of vascular compliance, normal and very low, which span a range of compliances that are greater than that possible in vivo. We did this not to completely study the effect of compliance, but to determine if compliance substantially modifies the effect of end-organ vascular resistance. (To study completely the effect of compliance alone, an experiment in which end-organ vascular resistance is held constant while the compliance is varied through a large range is required.)

Our results show poststenotic PSV decreases only mildly with increasing vascular resistance distal to the stenosis for stenoses at both ends of the hemodynamic spectrum (low- and high-grade stenoses). This is true for both normal and very low vascular compliance. We found it surprising that the decrease in PSV for both stenotic grades was not much more substantial, especially at infinite end-organ vascular resistance (absence of net flow through the kidney), not only with normal compliance, but especially with very low compliance.

It is likely this occurs because the compliance (the normal ability of the artery to respond to the pressure gradient between systole and diastole by expanding in systole and by relaxing in diastole) allows the artery to accommodate enough of the systolic jet so PSV is only mildly affected, even when the compliance is much smaller than that which occurs in vivo. Rather than substantially reducing the PSV, the main effect of increasing end-organ resistance is both progressively reducing diastolic flow and progressively narrowing the width of the systolic pulse (Fig 4). Halpern et al (35) have shown similar results regarding the PSV in vessels without proximal stenosis.

Our results also likely apply to another parameter that is used to detect hemodynamically significant stenoses and that incorporates the poststenotic PSV: the ratio of the renal arterial PSV to the aortic PSV (RAR, or its closely related variant in renal transplants, the ratio of the renal arterial PSV to the iliac arterial PSV), where RAR is stenotic PSV/aortic PSV (15–19).

Variable results have been reported for the detection of hemodynamically significant renal arterial stenoses with this parameter, as well; results range from 71% sensitivity and 91% specificity (19) to 91% sensitivity and 95% specificity (18). Our results help explain some of this variability because variations in end-organ resistance change the poststenotic PSV, but it is known experimentally that the aortic PSV is essentially independent of increases in peripheral vascular resistance in vivo (36). Thus, the ratio of the renal arterial PSV to the aortic PSV should change with changes in end-organ vascular resistance, as the PSV does.

Even though PSV is only mildly dependent on end-organ vascular resistance, the dependence is still important. Consider a stenosis at the threshold of hemodynamic significance (10 mm Hg pressure gradient in vivo), as modeled by our low-grade stenosis. At normal compliance in our model, a vascular resistance high enough to decrease the volume of blood flowing to the kidney by 50% (reduction from 10% to 5% volume flow) decreases the normalized PSV from 1.60 to 1.27 (Fig 3c), a decrease of 21%.

If a velocity threshold of 100 cm/sec is used for stenosis detection at normal compliance, any stenosis that produces a PSV of up to 127 cm/sec at normal end-organ resistance produces a PSV of less than 100 cm/sec at the higher resistance (127 cm/sec x [100% - 21%] = 100 cm/sec). Therefore, in our model, any stenosis at normal renal resistance that gives a PSV in the 100–127 cm/sec range will give a spuriously normal PSV at the higher resistance.

These results may, therefore, help to explain some of the false-negative results that occur from using PSV to detect hemodynamically significant stenoses. They do not, however, provide a complete explanation, since the reported PSV thresholds for detecting hemodynamically significant stenoses vary substantially more than this (eg, 100–350 cm/sec). Other factors not investigated in our study must be at work, as well.

Although end-organ vascular resistance exerts much the same effect on high-grade and low-grade stenoses, the effect on a high-grade stenosis is inconsequential. This is because in vivo, a high-grade stenosis often has a very high PSV that far exceeds the threshold for stenosis detection. Even if this PSV is lowered by the maximal amount demonstrated in our experiment (a resistance increase from normal to infinite), it is reduced by only 8% at normal compliance and by 29% at very low compliance (see Results) and should still exceed the PSV threshold and not escape detection on that basis alone.

Another feature of the waveforms in our experiment needs to be addressed: the small component of reversed flow during systole in many of the waveforms (Figs 4, 5). The cause is uncertain, but it may be related to the following. The Doppler sample gate included the entire width of the systolic jet, as determined with color Doppler US. Since blooming causes overestimation of vessel size at color Doppler US, the sample volume probably included areas of fluid adjacent to but not actually in the jet. Since eddy currents occur at substantial flow boundaries, it is possible that the eddy currents adjacent to the jet were included in the Doppler sample volume and caused the low-amplitude reversed systolic flow.

A limitation of our model is it only mimics in vivo conditions. It does not simulate the autoregulatory ability of the kidney. However, autoregulation changes renal vascular resistance, and even if a kidney autoregulates, the relationship we have shown between PSV and end-organ resistance still applies.

Practical application: A greater understanding of the effect of end-organ resistance on the stenotic PSV may enable future studies in which end-organ resistance is taken into account to better use the PSV to detect hemodynamically significant stenoses. The effect of end-organ resistance also helps explain the morphology of the stenotic jet Doppler waveform.


   Acknowledgments
 
The authors acknowledge the help of the two research assistants in this project, Lois Otte Bude, BS, and Doris Otte Tomlin, BS. Without their assistance, this project could not have been completed.


   Footnotes
 
Abbreviations: PSV = peak systolic velocity RI = resistive index

Author contributions: Guarantor of integrity of entire study, R.O.B.; study concepts, R.O.B., R.G.L., W.W.N.; study design, R.O.B.; definition of intellectual content, R.O.B., J.M.R.; literature research, R.O.B.; experimental studies, R.O.B.; data acquisition, R.O.B.; data analysis, R.O.B., J.M.R.; manuscript preparation, R.O.B.; manuscript editing, R.O.B., J.M.R., W.W.N.; manuscript review, all authors.


   References
TOP
 Abstract
 Introduction
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 References
 

  1. Avasthi PS, Voyles WF, Greene ER. Noninvasive diagnosis of renal artery stenosis by echo-Doppler velocimetry. Kidney Int 1984; 25:824-829.[Medline]
  2. Berland LL, Koslin DB, Routh WD, Keller FS. Renal artery stenosis: prospective evaluation of diagnosis with color duplex US compared with angiography—work in progress. Radiology 1990; 174:421-423.[Abstract/Free Full Text]
  3. Desberg AL, Paushter DM, Lammert GK, et al. Renal artery stenosis: evaluation with color Doppler flow imaging. Radiology 1990; 177:749-753.[Abstract/Free Full Text]
  4. Duda SH, Erley CM, Wakat JP, et al. Posttransplant renal artery stenosis: outpatient intraarterial DSA versus color aided duplex Doppler sonography. Eur J Radiol 1993; 16:95-101.[Medline]
  5. Kliewer MA, Tupler RH, Carroll BA, et al. Renal artery stenosis: analysis of Doppler waveform parameters and tardus-parvus pattern. Radiology 1993; 189:779-787.[Abstract/Free Full Text]
  6. Saarinen O, Salmela K, Edgren J. Doppler ultrasound in the diagnosis of renal transplant artery stenosis: value of resistive index. Acta Radiol 1994; 35:586-589.[Medline]
  7. Baxter GM, Ireland H, Moss JG, et al. Colour Doppler ultrasound in renal transplant artery stenosis: which Doppler index?. Clin Radiol 1995; 50:618-622.[Medline]
  8. Gottlieb RH, Lieberman JL, Pabico RC, Waldman DL. Diagnosis of renal artery stenosis in transplanted kidneys: value of Doppler waveform analysis of the intrarenal arteries. AJR 1995; 165:1441-1446.[Abstract/Free Full Text]
  9. Hélénon O, El Rody F, Correas JM, et al. Color Doppler US of renovascular disease in native kidneys. RadioGraphics 1995; 15:833-854.[Abstract]
  10. Hoffmann U, Edwards JM, Carter S, et al. Role of duplex scanning for the detection of atherosclerotic renal artery disease. Kidney Int 1991; 39:1232-1239.[Medline]
  11. Olin JW, Piedmonte MR, Young JR, DeAnna S, Grubb M, Childs MB. The utility of duplex ultrasound scanning of the renal arteries for diagnosing significant renal artery stenosis. Ann Intern Med 1995; 122:833-838.[Abstract/Free Full Text]
  12. Strotzer M, Fellner CM, Geissler A, et al. Noninvasive assessment of renal artery stenosis: a comparison of MR angiography, color Doppler sonography, and intraarterial angiography. Acta Radiol 1995; 36:243-247.[Medline]
  13. Miralles M, Cairols C, Cotillas J, Gimenez A, Santiso A. Value of Doppler parameters in the diagnosis of renal artery stenosis. J Vasc Surg 1996; 23:428-435.[Medline]
  14. van der Hulst VPM, van Baalen J, Kool LS, et al. Renal artery stenosis: endovascular flow wire study for validation of Doppler US. Radiology 1996; 200:165-168.[Abstract/Free Full Text]
  15. Merkus JW, Hoitsma AJ, van Asten WN, et al. Echo-Doppler diagnosis of renal allograft artery stenosis. Clin Transplant 1995; 9:383-389.[Medline]
  16. Taylor DC, Kettler MD, Moneta GL, et al. Duplex ultrasound scanning in the diagnosis of renal artery stenosis: a prospective evaluation. J Vasc Surg 1988; 7:363-369.[Medline]
  17. Hansen KJ, Tribble RW, Reavis SW, et al. Renal duplex sonography: evaluation of clinical utility. J Vasc Surg 1990; 12:227-236.[Medline]
  18. Kohler TR, Zierler RE, Martin RL, et al. Noninvasive diagnosis of renal artery stenosis by ultrasonic duplex scanning. J Vasc Surg 1986; 4:450-456.[Medline]
  19. Halpern EJ, Needleman L, Nack TL, East SA. Renal artery stenosis: should we study the main renal artery or segmental vessels?. Radiology 1995; 195:799-804.[Abstract/Free Full Text]
  20. Bude RO, Rubin JM, Platt JF, Fechner KP, Adler RS. Pulsus tardus: its cause and potential limitations in detection of arterial stenosis. Radiology 1994; 190:779-784.[Abstract/Free Full Text]
  21. Folkow B, Neil E. General principles In: Circulation. New York, NY: Oxford University Press, 1971; 10.
  22. Nichols WW, O'Rourke MF. The nature of flow of a fluid. In: Nichols WW, O'Rourke MF, eds. McDonald's blood flow in arteries: theoretical, experimental and clinical principles. 4th ed. New York, NY: Oxford University Press, 1998; 11.
  23. Nichols WW, O'Rourke MF. Introduction. In: Nichols WW, O'Rourke MF, eds. McDonald's blood flow in arteries: theoretical, experimental and clinical principles. 4th ed. New York, NY: Oxford University Press, 1998; 1.
  24. Nichols WW, O'Rourke MF. The nature of flow of a fluid. In: Nichols WW, O'Rourke MF, eds. McDonald's blood flow in arteries: theoretical, experimental and clinical principles. 4th ed. New York, NY: Oxford University Press, 1998; 40-47.
  25. Duck FA. Mechanical properties of tissue In: Physical properties of tissue. San Diego, Calif: Academic Press, 1990; 161.
  26. Weast RC. Handbook of chemistry and physics: a ready-reference book of chemical and physical data 48th ed. Cleveland, Ohio: Chemical Rubber, 1967; D191-D192.
  27. Gould KL. Dynamic coronary stenosis. Am J Cardiol 1980; 45:286-292.[Medline]
  28. Nichols WW, O'Rourke MF. The nature of flow of a fluid. In: Nichols WW, O'Rourke MF, eds. McDonald's blood flow in arteries: theoretical, experimental and clinical principles. 3rd ed. Philadelphia, Pa: Lea & Febiger, 1990; 26.
  29. Norris CS, Barnes RW. Renal artery flow velocity analysis: a sensitive measure of experimental and clinical renovascular resistance. J Surg Res 1984; 36:230-236.[Medline]
  30. Nichols WW, O'Rourke MF. Introduction. In: Nichols WW, O'Rourke MF, eds. McDonald's blood flow in arteries: theoretical, experimental and clinical principles. 4th ed. New York, NY: Oxford University Press, 1998; 3.
  31. Nichols WW, O'Rourke MF. Coupling of the left ventricle with the systemic circulation: implications to cardiac failure. In: Nichols WW, O'Rourke MF, eds. McDonald's blood flow in arteries: theoretical, experimental and clinical principles. 4th ed. New York, NY: Oxford University Press, 1998; 295-296.
  32. McDonald DA, Taylor MG. The hydrodynamics of the arterial circulation. Prog Biophys Chem 1959; 9:107-173.
  33. Halpern EJ, Deane CR, Needleman L, Merton DA, East SA. Normal renal artery spectral Doppler waveform: a closer look. Radiology 1995; 196:667-673.[Abstract/Free Full Text]
  34. Merklin RJ, Michels NA. The variant renal and suprarenal blood supply with data on the inferior phrenic, ureteral, and gonadal arteries. J Int Coll Surg 1958; 29:41-76.[Medline]
  35. Halpern EJ, Merton DA, Forsberg F. Effect of distal resistance on Doppler US flow patterns. Radiology 1998; 206:761-766.[Abstract/Free Full Text]
  36. Westerhof N, Sipkema P, Bos GC, van den Elzinga Z. Forward and backward waves in the arterial system. Cardiovasc Res 1972; 6:648-656.[Medline]




This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF)
Right arrow Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
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 reprints & permissions
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Bude, R. O.
Right arrow Articles by Rubin, J. M.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Bude, R. O.
Right arrow Articles by Rubin, J. M.

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
RADIOLOGY RADIOGRAPHICS RSNA JOURNALS ONLINE