Transvenous Hemodynamic Assessment of Experimental Arteriovenous Malformations
Doppler Guidewire Monitoring of Embolotherapy in a Swine Model
Background and Purpose A Doppler guidewire was used to monitor progressive changes in draining vein flow parameters during experimental embolotherapy in a swine arteriovenous malformation (AVM) model.
Methods A microcatheter was positioned superselectively in the main arterial feeder and main draining vein in each of 10 AVM models in swine. With use of the Doppler guidewire, preembolization arterial and venous average peak velocities (APVs) and pulsatility indices were recorded. The device was left in the draining vein during transarterial particulate (in 8 swine) or liquid adhesive (in 2 swine) embolization, and continuous transvenous flow during and after treatment was monitored. Periembolization Doppler flow parameters were correlated qualitatively with angiographic changes in the nidus.
Results Preembolization draining vein flow was pulsatile, with a mean APV of 38.9±13.7 cm/s. After embolization, this changed significantly to a less pulsatile or nonpulsatile pattern, with a lower mean APV of 9.2±4.9 cm/s (P=.0001). A novel expression, the maximum minus the minimum peak velocity (MxPV−MnPV), was used in evaluating the transvenous Doppler spectra. This was reduced significantly after embolization from a mean of 11.1±3.5 cm/s to 6.7±2.5 cm/s (P=.0025). Objective periembolization hemodynamic changes were detected in the draining veins earlier than the visually subjective angiographic changes within the nidus.
Conclusions Transvenous Doppler guidewire assessment of two parameters, APV and MxPV−MnPV, is useful in the hemodynamic evaluation of experimental arteriovenous shunting and may be used for future objective and quantitative monitoring during endovascular AVM embolotherapy.
Intracranial AVMs possess two fundamental features, one morphological (the presence of an abnormal tangle of nidus vessels interposed between feeding arteries and draining veins) and one hemodynamic (the consequent presence of arteriovenous shunting through the low-resistance nidus vessels). Therefore, in addition to morphological characterization of these lesions, the assessment of their hemodynamics remains an integral step in planning their treatment, understanding the effects of treatment, and preventing treatment complications. Despite the recognition of the importance of AVM hemodynamics, clinical monitoring during endovascular embolotherapy remains difficult. Several methods for hemodynamic evaluation of AVMs, eg, TCD1 2 3 4 5 6 7 or intravascular pressure measurements,8 9 10 11 12 13 14 15 16 have been attempted previously. Most of these studies, however, have focused on investigations of feeding artery hemodynamics. Despite draining veins being an integral and often quite accessible component of intracranial AVMs, the procurement of hemodynamic information from draining veins remains relatively uncommon.13 14 15 16 17 To enhance our knowledge of the hemodynamic behavior of AVMs during endovascular treatment, we used an experimental AVM model in laboratory swine to simulate embolotherapy and evaluate the feasibility and usefulness of transvenous blood flow velocity monitoring of treatment using an intravascular Doppler guidewire.
Materials and Methods
To date, in vivo laboratory research of AVM embolotherapy has been hindered by the lack of a suitable animal model. Recently, Massoud et al18 developed an experimental AVM model with a nidus fashioned from bilateral carotid retia mirabilia of swine. In this study, we used this model for investigating transvenous hemodynamic monitoring during simulated endovascular embolotherapy.
All animal experimentation was conducted in accordance with policies set by the local University Chancellor's Animal Research Committee and National Institutes of Health guidelines. Ten Red Duroc swine were used in this study. The animals were 3 to 4 months old, weighed 30 to 40 kg, were of mixed sex, and were maintained on a standard laboratory diet. After an overnight fast, each swine was premedicated with 20 mg/kg ketamine and 2 mg/kg xylazine IM. General anesthesia was maintained with mechanical ventilation and inhalation of 1% to 2% halothane after endotracheal intubation.
The carotid rete mirabile of the swine is a fine network of vessels (with connections across the midline to the contralateral rete) situated at the termination of each ascending pharyngeal artery as it perforates the skull base. This vascular structure has some morphological similarities to a plexiform AVM nidus but lacks its inherent arteriovenous shunting. The AVM model reported previously18 entails principally the simple surgical formation of a carotid-jugular fistula in the neck of swine. This model offers two main benefits that represent a significant advance in experimental in vivo modeling of these vascular malformations: simplicity of creation and clear visualization of all components of the “AVM,” including an intact “nidus.”
The relevant vascular anatomy of the swine head and neck and details of constructing the AVM model have been described previously.18 In the present study, a minor modification to the original model consisted of omitting the performance of preoperative endovascular occlusion of the right occipital artery, the muscular branch of the right ascending pharyngeal artery, and the right external carotid artery. We chose to not perform these maneuvers to simplify even further the construction of this AVM model, accepting the likelihood that this also resulted in marginal reduction of the postoperative blood shunting through and channeling from both retia to the fistula. Another modification consisted of maintaining a patent (ie, not tied at surgery) proximal right CCA, caudad to the arteriovenous fistula. This still resulted in an effective sump effect (although probably to a lesser degree compared with the original model), which reversed blood flow cephalad to the fistula while allowing easy transarterial access (through a patent right CCA) and microcatheterization of the “draining vein” portion of the AVM model (see below), particularly the right ascending pharyngeal artery adjacent to the “nidus.”
A side-to-side arteriovenous fistula between the right CCA and the right external jugular vein was fashioned surgically in the necks of the swine. Under general anesthesia, the right side of the neck was shaved and scrubbed with betadine solution and then draped in a sterile fashion. With sterile techniques, a 10-cm incision was made in the neck parallel to the sternocleidomastoid muscle. Reflecting this muscle medially, a 3-cm segment of the external jugular vein, free of tributaries, was dissected, isolated, and cleaned of adventitia. Next, an adjacent 3-cm segment of the CCA also was dissected, isolated, and cleaned of adventitia. Four microvascular clamps were placed at each end of the isolated artery and vein segments to achieve temporary occlusion during fistula construction. A 2-cm-long elliptical arteriotomy and a similar venotomy were performed microsurgically. With 7-0 prolene suture, a side-to-side arteriovenous fistula was fashioned. The vascular clamps were removed subsequently to reveal the marked dilatation and pulsation of the external jugular vein due to the presence of arterialized blood. Subcutaneous tissues and skin were sutured in layers.
Preembolization Angiography and Hemodynamics
Immediate postoperative angiography was performed in all swine to demonstrate the “AVM” (Fig 1⇓). Via the transfemoral route, a 6F Fasguide guiding catheter (Target Therapeutics) was positioned in the left CCA. Nonionic contrast medium (6 mL) was injected to outline the three “feeding arteries” (left ascending pharyngeal artery, left ramus anastomoticus, and left arteria anastomotica), the “nidus” (bilateral retia mirabilia), and the main “draining vein” (right ascending pharyngeal artery and right CCA down to the fistula). The rapid transit of blood from the left to the right side of the neck, across bilateral retia, was seen on rapid-sequence (up to 30 frames per second) digital subtraction imaging. A bolus intra-arterial dose of 3000 U heparin was given. Next, superselective angiograms were obtained via the main terminal “feeding artery” (the left ascending pharyngeal artery) to demonstrate the anatomy of the “nidus” and document subjectively the induced fast flow through it before its embolization. For performance of the left ascending pharyngeal arteriogram and the subsequent hemodynamic evaluation and embolization, a Tracker 18 microcatheter/Seeker 14 microguidewire combination (Target Therapeutics) was navigated coaxially through the guiding catheter, and the microcatheter tip was positioned in the ascending pharyngeal artery just proximal to the left rete. Via a second transfemoral route, an 8F guiding catheter (Balt) was positioned in the right CCA just caudad to the carotid-jugular fistula. A larger guiding catheter was used on this side of the neck to reduce as much as possible the contribution of blood flow (around the catheter) from the right CCA to the fistula, thus allowing for a greater contribution to the fistula from the cephalad aspect of the artery (and thus increasing the flow reversal down the “draining vein” portion of the AVM model). A second Tracker 18 microcatheter/Seeker 14 microguidewire combination was navigated coaxially through the 8F guiding catheter, and the microcatheter tip was positioned in the right ascending pharyngeal artery just proximal to the right rete.
After construction, bilateral microcatheterization, and angiographic demonstration of the AVM model, blood flow parameters were measured in the terminal “feeder” (the left ascending pharyngeal artery) and the main “draining vein” (the right ascending pharyngeal artery) immediately adjacent to the “nidus.” An intravascular Doppler guidewire (SmartWire, Cardiometrics) was used. This consisted of a 12-MHz piezoelectric ultrasonic transducer mounted on the tip of a 175-cm-long flexible and steerable 0.014-in guidewire. The distal flexible coil portion of the guidewire is 45 cm in length. The ultrasonic beam diverges approximately 14° from the axis of the transducer, with a range gate depth of 4.3 mm and gate integration distance of 1 mm. These specifications produce an estimated axial sample diameter of 2.5 mm and a sample volume of 5 mm3. The pulsed Doppler signals were processed by a real-time spectrum analyzer (Flomap, Cardiometrics), using a fast Fourier transform at a rate of 100 spectra per second. The Doppler guidewire was navigated first through the “transarterial” microcatheter, and its tip was advanced as close as possible to the left rete (ie, just proximal to the “nidus”) for preembolization sampling. This was then withdrawn and advanced in turn through the “transvenous” microcatheter, and its tip was advanced as close as possible to the right rete (ie, just distal to the “nidus”). This was left in place for pre-, peri-, and postembolization sampling of “venous” APV and PI. The PI value was used to calculate the maximum minus the minimum peak velocity (MxPV−MnPV=APV×PI). This novel expression was used in analyzing the available Doppler spectra. In our analysis, attention was paid to two parameters in particular, APV and MxPV−MnPV, in evaluating the subsequent hemodynamic changes resulting from transarterial embolization of the AVM model.
AVM Embolization and Hemodynamics
Eight of the 10 swine AVM models were embolized with embolic particles (6 with collagen microbeads [≈300 μm in size, Imedex] and 2 with polyvinyl alcohol [200 μm in size, Biodyne Inc]). Two AVM models were embolized with liquid glue (NBCA). Embolization of the normal swine rete and the swine AVM model using collagen microbeads was described previously by Turjman et al19 and Massoud et al,20 respectively. A suspension of either type of particles was diluted by mixing with contrast agent and saline. Embolization was performed through the microcatheter placed superselectively in the arterial feeder. A slow injection of 0.5 mL of the dilute particle suspension was performed under fluoroscopic digital subtraction imaging. This was repeated until complete contrast stasis was achieved in the distal feeder and proximal nidus. Careful angiographic monitoring of the extent of nidus or feeder occlusion was necessary, and the speed of injection was adjusted to produce satisfactory embolization without particle reflux. Pre-, peri-, and postembolization angiograms were obtained along with simultaneous transvenous Doppler guidewire flow measurements.
Embolization of the swine AVM model with NBCA has also been described previously.20 Iodized oil (Ethiodol, Savage) was mixed with NBCA and tantalum powder in a clean container. NBCA (0.5 mL) was mixed with Ethiodol in a 1:2 volume ratio. A preembolization flush injection of 5% dextrose was performed; then 0.2 mL of embolic mixture was injected as a bolus and flushed gently into the “nidus” with a chaser bolus of 5% dextrose. Preembolization and postembolization angiograms were obtained along with simultaneous transvenous Doppler guidewire flow measurements.
The preembolization and postembolization blood flow data (arterial APV; venous APV, MxPV−MnPV, and PI) were examined quantitatively. Statistical analysis was performed using Student's t test (paired one-tail). Results were considered significant at P<.05 and reported as mean±SD. The periembolization flow velocity parameters and angiograms were also correlated qualitatively.
All swine tolerated the general anesthesia and surgical and endovascular procedures with no ill effects. All AVM models were created successfully, resulting in a clear angiographic demonstration of the “feeding arteries,” “nidus,” and a main “draining vein” in each swine. All embolizations and hemodynamic studies were performed on AVM models immediately after creation.
It was possible to obtain all the necessary blood flow data from 9 of the 10 swine AVM models. In 1 of the swine embolized with NBCA, it was not possible to measure the preembolization “feeder” APV due to technical difficulties resulting from a damaged Doppler guidewire. The results from this swine were excluded from the statistical analysis of the entire data. Nevertheless, the angiographic and transvenous blood flow velocity changes in this swine were interesting and instructive; therefore, these findings are described below.
The preembolization mean APV in the terminal “feeder” was 63.4±24.6 cm/s. In all models, the flow pattern in the main “draining vein” was reversed in direction (away from the nidus and toward the fistula) and was pulsatile (ie, exhibited an arterial-like waveform), with a mean APV of 38.9±13.7 cm/s. Thus, the preembolization transnidal mean APV gradient was 61% (ie, the ratio of “feeder” APV to “draining vein” APV). This reduction in blood flow values across the “nidus” was thought to be similar to that present in intracranial AVMs. There was a positive linear correlation (y=0.5x+7.8, r=.775) between the “feeder” APV and the “draining vein” APV, as shown graphically in Fig 2⇓. The baseline PI across the nidus before embolization was 0.6±0.2 in the “feeder” and 0.3±0.1 in the “draining vein” (P=.0001), a reflection on the proximity of the simulated draining vein (the right ascending pharyngeal artery) to the ipsilateral carotid-jugular fistula and the consequent lower vascular resistance in this vessel.
Embolization of the “nidus” with particles proceeded in a stepwise fashion, using low concentrations and small volumes of injection. However, despite the often imperceptible or only small incremental changes in occlusion of the “nidus” as judged subjectively on successive superselective angiograms, the corresponding objective changes in transvenous blood flow parameters were always significant and clearly identifiable by the change in their numerical values (Fig 3⇓). This discrepancy in the ability of angiography and transvenous Doppler guidewire monitoring to detect the effects of embolization was most pronounced in AVM models exhibiting higher transnidal flow gradients and particularly in the early stages of embolization of the “nidus.”
Embolization of the “nidus” with NBCA was performed in two AVM models. In one of these (Fig 4⇓) (in which it was not possible to measure the preembolization “feeder” APV), glue was injected via the main “feeder” (the left ascending pharyngeal artery), and the “nidus” appeared to be embolized successfully on angiography via the CCA. The “draining vein” blood flow, however, was reduced only marginally (APV from 31 to 24 cm/s) and remained pulsatile, as demonstrated by the spectra obtained from the Doppler guidewire (Fig 4D⇓). The suspicion arose that unsatisfactory proximal occlusion of the “nidus” had occurred while leaving more distal aspects of the “nidus” to be supplied by patent “en passage feeders.” Therefore, a left external carotid arteriogram was performed to delineate the extent of “nidus” supply by the two en passage feeders, the left ramus anastomoticus and the left arteria anastomotica. These feeders indeed were shown to supply the “nidus” beyond its proximally occluded portion (Fig 4C⇓). The left external carotid artery was therefore embolized with a large 2.0-mL bolus of NBCA mixture to occlude the artery and its branches (including the supply to the “nidus”) in toto (Fig 4E⇓). This resulted in an immediate significant reduction in “draining vein” flow velocity and pulsatility (Fig 4F⇓).
After embolization, all transvenous flow velocities were reduced and nonpulsatile waveforms were obtained. The “draining vein” postembolization mean APV was reduced significantly to 9.2±4.9 cm/s (P=.0001) (Fig 5⇓, left). On embolization, the mean preembolization MxPV−MnPV was reduced significantly from 11.1±3.5 to 6.7±2.5 cm/s (P=.0025) (Fig 5⇓, middle). Similarly, the PI increased significantly from 0.3±0.1 to 0.9±0.5 (P=.0013) (Fig 5⇓, right).
Doppler Guidewire in AVM Draining Veins
Despite past extensive hemodynamic studies of intracranial AVMs (eg, using TCD,1 2 3 4 5 6 7 TCCD,21 22 direct intravascular pressure measurements,8 9 10 11 12 13 14 15 16 or MR angiography23 24 ), it has proved difficult to use these modalities and the data they provide in a clinically practical manner to monitor critical hemodynamic changes during endovascular embolotherapy. In this study, we hypothesized that the transvenous placement of a Doppler guidewire for continuous measurement of draining vein flow parameters might offer several advantages as a clinical monitoring technique for conventional transarterial embolotherapy of AVMs. These potential advantages could be categorized broadly as technical, practical, and hemodynamic and would include the following: (1) There would be no dependence on size and thickness of the temporal bone window, as is the case for flow measurements using TCD25 and TCCD.26 27 28 (2) It might offer the possibility of continuous monitoring of blood flow during transarterial treatment, which is difficult in practice or impossible with most other modalities, including TCD, TCCD, MR angiography, and intravascular pressure measurements. (3) There would be no physical interference with arterial feeders used for the embolization process itself. (4) In many instances, the navigation/manipulation/positioning of a transvenous microcatheter and Doppler guidewire might be easier and safer in AVM draining veins than in arterial feeders, particularly in large ones or within the dural venous sinuses draining the AVM. (5) It would provide an objective and representative blood flow wave pattern from which qualitative hemodynamic information could be obtained. (This is unlike intravascular pressure measurements, in which the pressure waveform is strongly dependent on the caliber of the microcatheter used because of attenuation of values at the systolic peak and the diastolic trough.9 ) (6) It might be possible, if the caliber of the vein is established beforehand (eg, at angiography or ultrasonography), to determine true volumetric blood flow (in milliliters per minute) draining an AVM (and the changes induced by embolization of the nidus), especially if the Doppler guidewire is positioned in a dural venous sinus, because these veins exhibit less variation in caliber than arteries during each vascular pulsation. Although many of these potential advantages could be appreciated only in the clinical setting, we have set out in the present study to test and study the preliminary experimental feasibility of transvenous flow monitoring of AVM embolotherapy using an appropriate AVM model in laboratory swine.
Swine AVM Model
To date, experimental research and training in endovascular embolization of intracranial AVMs have been hampered by the lack of a suitable laboratory animal model. To address this shortfall, Massoud et al18 20 have developed a swine model of an AVM with closer resemblance to human lesions than was available previously. The model makes use of the carotid rete mirabile of swine but with the added experimentally induced feature of faster blood flow through bilateral retia. Shunting of blood across both retia, from one side of the neck to the other, is produced by surgical formation of a large unilateral carotid-jugular fistula. This novel swine AVM model possesses the mandatory morphological and hemodynamic requirements of rapid shunting through a “nidus.”
Of prime relevance to our study was consideration of the validity of using this particular experimental AVM model to investigate the transvenous hemodynamic changes that occur during transarterial treatment of AVMs. In this respect, the suitability of this swine AVM model for laboratory simulations of endovascular embolotherapy had been established previously.20 Previous intravascular blood pressure measurements within components of this model have confirmed the presence of a mean pressure gradient of 23 mm Hg between the terminal “feeder” and the main “draining vein” (across the “nidus”), a hemodynamic feature akin to that of some human AVMs, particularly those exhibiting distant stenoses in their draining veins, and a trait unavailable previously in experimental AVM simulations using the normal rete mirabile. Further hemodynamic characterization of the AVM model in the present study has demonstrated arterial-like pulsatility of the “draining vein,” a baseline PI (which reflects the peripheral vascular resistance) reduction across the nidus of 50%, a baseline 61% drop in transnidal blood flow velocity, and the existence of a positive linear correlation between “feeder” APV and “draining vein” APV, a correlation not unlike that established previously by Young et al15 for transnidal pressure changes in human AVMs (intravascular pressure is proportional to flow velocity6 ). Previously reported simulations of endovascular embolization using the swine AVM model were performed with acrylic glue and particles.20 In one model, the intravascular mean blood pressure in the terminal feeder rose from 48 to 74 mm Hg after nidus embolization, consistent with observations in feeders of human AVMs. Although this model lacks the angioarchitectural complexity of most large and giant intracranial AVMs, it provides an adequate experimental replica of simpler plexiform malformations. Therefore, the gross morphological, histopathological,29 angiographic, and hemodynamic features of this AVM model lend support to its suitability as a laboratory simulator for endovascular embolotherapy and for the purpose of our study. Experience in the cognitive and technical aspects of endovascular embolotherapy can also be gained by repeated performances using this model.
Transvenous Flow Monitoring of Embolotherapy
Our results have shown that all “transvenous” flow velocities were reduced after embolization of the “nidus” in the swine AVM model. The exact values of two quantifiable parameters in particular, “venous” APV and MxPV−MnPV, appear to provide measurements that objectively reflect the degree of “nidus” embolization. Transvenous monitoring of embolotherapy also resulted in a measurable increase in PI and a decrease in the pulsatile wave pattern within the “draining vein.”
This study confirms the usefulness of the carotid-jugular fistula–type AVM model in swine as a valuable research tool for basic hemodynamic investigations of simulated simple AVMs. An acknowledged limitation of the present experimental feasibility study is the inability to test the significance of the above hemodynamic findings in a more realistic setting of complex (morphological and hemodynamic) arteriovenous shunting, as would be present in most human AVMs referred for endovascular treatment. However, the experimental results using the swine AVM model serve the purpose of demonstrating the underlying general principles and usefulness of transvenous flow monitoring of embolotherapy. In reality, the extent of transvenous reduction in APV and MxPV−MnPV during treatment of human AVMs is likely to depend on a complex interplay of numerous morphological and hemodynamic factors in the nidus, arterial feeders, and draining veins (eg, presence or absence of intranidal fistulae, size of nidus, diffuse or compact nidus, compartmentalized or noncompartmentalized nidus, number of and pressure within arterial feeders, number and size of draining veins, stenoses within draining veins, etc) and possibly on the presence of a compressive hematoma adjacent to an AVM. Many of these factors are known to affect the hemodynamic stresses within AVMs by affecting their propensity to rupture spontaneously. Therefore, it is conceivable that these factors also may affect hemodynamic measurements within draining veins during nidus embolization. A true understanding of these complex issues might be possible from future clinical implementation and accumulation of morphological and hemodynamic data from patients harboring AVMs.
Two additional findings emerging from this study were deemed particularly important. First, it was possible to detect hemodynamic changes within the “draining vein” before angiographic attenuation of the “nidus,” during transarterial embolization of the “AVM.” The reduction in transvenous flow velocities was noted objectively by the corresponding reduction in numerical values displayed on a monitor or recorded on the Doppler spectra. On the other hand, the corresponding angiographic occlusion of “nidus” microvessels was always subtle and often imperceptible. In our study, this discrepancy between the two modalities was observed when the “AVM” was embolized with particles because this afforded a convenient method for small incremental occlusion of the “nidus.” We speculate that this observation of early changes in transvenous flow parameters would similarly occur in human AVMs during embolization with particles or when a prolonged injection of the nidus is performed with slowly polymerizing NBCA.30 The availability of flow measurement within draining veins would provide a method to objectively monitor changes during embolization of an AVM that would represent a significant improvement on current practices relying mostly on angiographic assessment of the degree of nidus occlusion. Furthermore, it may provide a future method of objectively quantifying and accurately grading the extent of embolization of an intracranial AVM. With the accumulation of future clinical data, this may in turn provide objective indicators of treatment effectiveness and impending complications.
Another important result from this study was the demonstration of the ability of transvenous monitoring of blood flow to detect and quantify persistent shunting through the “AVM” despite an apparent successful (ie, complete) embolization demonstrated angiographically. This is exemplified by one of the AVM models embolized with NBCA. The embolization through the terminal “feeder” resulted unsuspectedly in too proximal an occlusion of the “nidus” and the distal “feeder.” A superselective angiogram through the same “feeder” after embolization showed an apparent obliteration of the “nidus.” However, there was minimal reduction of “draining vein” APV, indicating persistent flow through the “AVM” resulting from an alternative feeder source. A less selective CCA arteriogram was then performed, showing significant filling of the two “en passage feeders” that supply the “AVM nidus.” The recruitment of these two minor feeders on proximal occlusion of the “AVM nidus” in this model has been described previously.20 Further embolization of these two feeders resulted in dramatic reduction of the “draining vein” APV, indicating abolition of the shunt. In this case, the transvenous monitor proved invaluable in detecting and recording the presence and extent of draining vein flow due to persistent shunting across an incompletely embolized nidus. Thus, it is possible that future use of this modality may provide an objective way of determining the end point of embolization within intracranial AVMs, especially in complex lesions supplied by numerous arterial feeders or when angiography is equivocal regarding the presence of a persistent nidus after an initial embolization. In this regard, a recent report has indicated the usefulness of intraoperative TCCD in the detection of persistent supply through feeding arteries to what was thought to be “angiographically completely embolized” AVMs.31 The future clinical use of transvenous (even if positioned in adjacent dural venous sinuses) Doppler guidewire detection of blood flow might provide a similar but more advantageous noninvasive technique to monitor the success of endovascular embolotherapy of intracranial AVMs.
In conclusion, transvenous Doppler guidewire assessment of two parameters, APV and MxPV−MnPV, is useful in the hemodynamic evaluation of experimental arteriovenous shunting and may be used for future objective and quantitative monitoring during endovascular AVM embolotherapy. Further experimental and clinical investigations of this technique will be necessary to establish its true role in the management of intracranial AVMs.
Selected Abbreviations and Acronyms
|APV||=||average peak velocity|
|CCA||=||common carotid artery|
|MxPV−MnPV||=||maximum minus minimum peak velocity|
|TCCD||=||transcranial color-coded Doppler ultrasonography|
|TCD||=||transcranial Doppler ultrasonography|
This study was supported in part by National Institutes of Health grant RO1 HL/NS52352-01A1. The authors gratefully acknowledge the technical assistance of John Robert, Christopher Carangi, and Roger McGath in the laboratories of the Leo G. Rigler Radiological Research Center at University of California at Los Angeles. We also wish to thank James Sayre, PhD, for his assistance with the statistical analysis.
Reprint requests to Yuichi Murayama, MD, Department of Neurosurgery, The Jikei University School of Medicine, 3-25-8 Nishishinbashi, Minato-ku 105 Tokyo, Japan.
- Received March 18, 1996.
- Revision received May 3, 1996.
- Accepted May 3, 1996.
- Copyright © 1996 by American Heart Association
Petty GW, Massaro AR, Tatemichi TK, Mohr JP, Hilal SK, Stein BM, Solomon RA, Duterte DI, Sacco RL. Transcranial Doppler ultrasonographic changes after treatment for arteriovenous malformations. Stroke. 1990;21:260-266.
Fleischer L, Young WL, Pile-Spellman J, terPenning B, Kader A, Stein BM, Mohr JP. Relationship of transcranial Doppler flow velocities and arteriovenous malformation feeding artery pressures. Stroke. 1993;24:1897-1902.
Jungreis CA, Horton JA, Hecht ST. Blood pressure changes in feeders to cerebral arteriovenous malformations during therapeutic embolization. AJNR Am J Neuroradiol. 1989;10:575-577.
Duckwiler G, Dion J, Viñuela F, Jabour B, Martin N, Bentson J. Intravascular microcatheter pressure monitoring: experimental results and clinical evaluation. AJNR Am J Neuroradiol. 1990;11:169-175.
Quisling RG, Mickle JP. Venous pressure measurements in vein of Galen aneurysms. AJNR Am J Neuroradiol. 1989;10:411-417.
Benndorf G, Podrabsky P, Wellnhofer E, Biamino G, Molsen HP, Fleck E, Lanksch W, Felix R. Measurement of blood flow velocity by Doppler-tipped guidewires in patients with arteriovenous malformations and fistulae. Neuroradiology. 1995;37(suppl):489-491.
Massoud TF, Ji C, Viñuela F, Guglielmi G, Robert J, Duckwiler GR, Gobin YP. An experimental arteriovenous malformation model in swine: anatomic basis and construction technique. AJNR Am J Neuroradiol. 1994;15:1537-1545.
Turjman F, Massoud TF, Vinters HV, Ji C, Tardy M, Guglielmi G, Viñuela F. Collagen microbeads: experimental evaluation of an embolic agent in the rete mirabile of the swine. AJNR Am J Neuroradiol. 1995;16:1031-1036.
Massoud TF, Ji C, Viñuela F, Turjman F, Guglielmi G, Duckwiler GR, Gobin YP. Laboratory simulations and training in endovascular embolotherapy using a swine arteriovenous malformation model. AJNR Am J Neuroradiol. 1996;17:271-279.
Klötzsch C, Henkes H, Nahser HC, Kühne D, Berlit P. Transcranial color-coded duplex sonography in cerebral arteriovenous malformations. Stroke. 1995;26:2298-2301.
Seidel G, Kaps M, Gerriets T. Potential and limitations of transcranial color-coded sonography in stroke patients. Stroke. 1995;26:2061-2066.
Massoud TF, Vinters HV, Ji C, Viñuela F, Chao KH, Guglielmi G. Histopathologic and histometric characterization of nidus microvessels in a chronic arteriovenous malformation model in swine. In: Proceedings of the 33rd annual meeting of the American Society of Neuroradiology; April 23-27, 1995; Chicago, Ill. Abstract:255-256.
Debrun G. Persistent feeding arteries to angiographically completely embolized arteriovenous malformation demonstrated by intraoperative color-flow Doppler testing: report of two cases. Surg Neurol. 1995;44:332-333. Comment on: Surg Neurol. 1995;44:326-333.