- Email: [email protected]
Design and evaluation of a flow phantom
Design and Evaluation of a Flow Phantom Jeffrey L. Creasy, MD 1, Daniel B. Crump 2, Kimberly Knox 3, Charles W. Kerber, MD 4, Ronald R. Price, PhD t
R a t i o n a l e a n d Objectives. We constructed a nearanatomically correct large-vessel phantom to perform repeatable flow dynamics research examinations by angiography, magnetic resonance (MR) angiography, and computed tomography (CT) angiography. M e t h o d s . An internal carotid artery was constructed within a head phantom. The internal carotid artery branches into a middle and an anterior cerebral artery; the former trifurcates and ends in the superior sagittal sinus, and the latter ends in the inferior sagittal sinus. A transverse and sigmoid sinus drains the model. All four vessels connecting the arterial and venous vessels have variable flow-constricting ligatures placed around them. These ligatures are accessible on the skull surface. The skull cavity is filled with a silicone polymer that is isodense to brain on CT scans and isointense on most MR images. Results. The flow in the phantom's vessels may be varied in a repeatable manner. Multiple scan sequences may be performed without the image degradation caused by patient motion. The homogeneity of the filler polymer allows visualization of flow-related artifacts that may be hidden by complex human anatomy. C o n c l u s i o n . Preliminary images of each modality show promise for use of the phantom in imaging research on large-vessel flow dynamics. K e y Words. Flow phantom; magnetic resonance angiography; computed tomography angiography.
method for the rigorous scientific assessment of new imaging techniques for the visualization of flowing blood has been difficult to optimize. The traditional approach has been to use either fixed-flow phantoms, which are not anatomically correct, or to use human volunteers. In the former case, the wide anatomic variation of vessel diameters, orientations, and flow velocity has been difficult to reproduce within a single phantom. In the latter case, it is generally not possible for normal volunteers to lie on a scanner for the many hours required to fully compare new techniques. Therefore, we conceived, designed, and constructed a more anatomically correct large-vessel phantom and performed initial scan evaluations to determine its value and, more important, its limitations. This phantom has near-anatomic representations of the anterior and middle cerebral arteries as well as the sagittal, transverse, and sigmoid sinus within one hemisphere. Initial evaluation of this model with magnetic resonance (MR) angiography, computed tomography (CT) angiography, and projection angiography revealed that the phantom is a promising tool to help assess new vascular imaging methods.
A
MATERIALS AND METHODS
The cranial blood flow phantom (Golden Pacific Arts, Encinitas, CA) consists of an acrylic skull filled with a
From the 1Del0artment of Radiology and Radiological Sciences, Vanderbilt University Medical Center, Nashville, TN; 2VanderbiR University School of Medicine, Nashville, TN; 3Golden Pacific Arts, Encinitas, CA; and 4 University of California, San Diego, CA. Address reprint requests to J. L. Creasy, MD, Department of Radiology and RadJological Sciences, Vanderbilt University Medical Center, 21st Ave. South & Garland, Nashville, TN 37232-2675. Received February 3, 1995, and accepted for publication after revision June 6, 1995. Acad Radio11995;2:902-904 © 1995, Association of University Radiology
902
Vol. 2, No. 10, October 1995
visually translucent silicone polymer that has MR imaging characteristics similar to those of the human brain. The right hemisphere contains an internal carotid artery with curved carotid siphon and a middle and anterior cerebral artery (Fig. 1). The anterior cerebral artery does not branch; it ends in the straight sinus. The middle cerebral artery trifurcates, passing over the insula, and ends without further branching in the superior sagittal sinus. Both sinuses meet at the torcular herophil and exit the skull by the right transverse and sigmoid sinuses. All arteries were modeled from actual human arteries by injecting fresh human cadavers with acrylic resin;
DESIGN AND EVALUATION
OF A FLOW PHANTOM
the details of this method are reported elsewhere [1]. Flow in the four main arteries may be controlled by monofilament tourniquets. All veins were hand constructed in wax using a resin cast human model, duplicating all dimensions, variations, and shapes. Once the vein and artery models were arranged in the skull, they were coated with a thin elastic silicone. Next, the entire skull was filled with the clear silicone polymer. Finally, the wax was removed thermally and chemically [1]. Silicone tubing was attached to the vessels at the base of the skull for fluid access. Imaging of the phantom was performed in a variety of ways. Routine angiographic plain films were performed following infusion of 25% by volume of ioxaglate meglumine (Hexabrix; Mallinckrodt Medical, St. Louis, MO) contrast material (Fig. 2). CT angiography was performed on a Somatom Plus scanner (Siemens Medical Systems, Iselin, NJ), using a similar concentration of ioxaglate meglumine (Fig. 3). MR angiography was performed on a Siemens GBS2 1.5-T system (Fig. 4). RESULTS AND DISCUSSION
Prior efforts at producing an anatomically correct vascular phantom have been limited to a short segment of the carotid artery that includes the bifurcation [2]. The need for a more complete vascular phantom of the head prompted us to build a near-anatomic human
FIGURE 1. Lateral view of the flow phantom demonstrating the anatomic appearance of the acrylic skull model. Connecting tubes at the skull base for the connection of the bidirectional flow are present. Over the convexity, the small screw adjustments for adjusting the constriction of the anterior cerebral artery and middle cerebral artery branches are indicated (arrows).
FIGURE 2. Lateral plane film view following contrast injection into the phantom. All three branches of the middle cerebral artery, as well as the single anterior cerebral artery branch continuing in the superior sagittal sinus, are visible.
903
CREASY
ET AL.
vol. 2, No. 10, October 1995
A
B
FIGURE 3. Computed tomography angiography was performed with a table speed of 1 mm, a table increment of 1 mm per revolution, and a 2-mm slice thickness reconstruction. Lateral view (A) and anteroposterior view (B) demonstrate good visualization of the vessels within the phantom, Because of placement of the phantom, an air bubble causes incomplete visualization of the most anterior portion of the anterior cerebral artery.
FIGURE 4. Lateral maximum intensity projection reconstruction from a dualslab axially acquired time-of-flight magnetic resonance angiogram. Flow artifacts can be seen in the region of the torcular herophil, the junction of the most anterior middle cerebral artery branch, the superior sagittal sinus, and the midanterior cerebral artery.
intracranial vascular phantom that was conceived, designed, built, and used for preliminary imaging research. Use of this anatomic phantom for pulse sequence and parameter refinement improves the accuracy of observations compared with less anatomically
904
correct phantoms or human volunteers, w h o tend to tire and move. Buxton et al. [3] and Frank et al. [4] began to study the complex artifacts that are produced by most MR sequences. These artifacts are displaced in three-dimensional space and by wrap-around. The complexity of anatomy found in humans obscures these confusing shadows, but the model, which is filled with a homogenous material, allows the clear appreciation of these artifacts. The major disadvantage of the model is the lack of third- and fourth-order vessels and the lack of a capillary bed. However, for our goals, such complexity was deliberately deleted to keep the model simple and to control costs. Although our initial applications were in the area of optimizing MR imaging sequences, we think that this model also will be helpful in CT angiography and rotational angiography research. REFERENCES 1. Kerber CW, Heilman CB, Zanetti PH. Transparent elastic arterial models: I. A brief technical note. Biorheology1989;26:1041-1049. 2. Frayne R, Gowman LM, Rickey DW, et al. A geometrically accurate vascular phantom for comparative studies of X-ray, ultrasound, and magnetic resonance vascular imaging: construction and geometrical verification. Med Phya 1993;20:415-425. 3. Buxton RB, Kerber CW, Frank LR. Pulsatile flow artifacts in two-dimensional time-of-flight MR angiography: initial studies in elastic models of human carotid arteries. J Magn Reson Imaging 1993;3:625-636. 4. Frank LR, Buxton RB, Kerber CW. Pulsatile flow artifacts in 3D MR imaging. Magn Reson Med 1993;30:296-304.
R a t i o n a l e a n d Objectives. We constructed a nearanatomically correct large-vessel phantom to perform repeatable flow dynamics research examinations by angiography, magnetic resonance (MR) angiography, and computed tomography (CT) angiography. M e t h o d s . An internal carotid artery was constructed within a head phantom. The internal carotid artery branches into a middle and an anterior cerebral artery; the former trifurcates and ends in the superior sagittal sinus, and the latter ends in the inferior sagittal sinus. A transverse and sigmoid sinus drains the model. All four vessels connecting the arterial and venous vessels have variable flow-constricting ligatures placed around them. These ligatures are accessible on the skull surface. The skull cavity is filled with a silicone polymer that is isodense to brain on CT scans and isointense on most MR images. Results. The flow in the phantom's vessels may be varied in a repeatable manner. Multiple scan sequences may be performed without the image degradation caused by patient motion. The homogeneity of the filler polymer allows visualization of flow-related artifacts that may be hidden by complex human anatomy. C o n c l u s i o n . Preliminary images of each modality show promise for use of the phantom in imaging research on large-vessel flow dynamics. K e y Words. Flow phantom; magnetic resonance angiography; computed tomography angiography.
method for the rigorous scientific assessment of new imaging techniques for the visualization of flowing blood has been difficult to optimize. The traditional approach has been to use either fixed-flow phantoms, which are not anatomically correct, or to use human volunteers. In the former case, the wide anatomic variation of vessel diameters, orientations, and flow velocity has been difficult to reproduce within a single phantom. In the latter case, it is generally not possible for normal volunteers to lie on a scanner for the many hours required to fully compare new techniques. Therefore, we conceived, designed, and constructed a more anatomically correct large-vessel phantom and performed initial scan evaluations to determine its value and, more important, its limitations. This phantom has near-anatomic representations of the anterior and middle cerebral arteries as well as the sagittal, transverse, and sigmoid sinus within one hemisphere. Initial evaluation of this model with magnetic resonance (MR) angiography, computed tomography (CT) angiography, and projection angiography revealed that the phantom is a promising tool to help assess new vascular imaging methods.
A
MATERIALS AND METHODS
The cranial blood flow phantom (Golden Pacific Arts, Encinitas, CA) consists of an acrylic skull filled with a
From the 1Del0artment of Radiology and Radiological Sciences, Vanderbilt University Medical Center, Nashville, TN; 2VanderbiR University School of Medicine, Nashville, TN; 3Golden Pacific Arts, Encinitas, CA; and 4 University of California, San Diego, CA. Address reprint requests to J. L. Creasy, MD, Department of Radiology and RadJological Sciences, Vanderbilt University Medical Center, 21st Ave. South & Garland, Nashville, TN 37232-2675. Received February 3, 1995, and accepted for publication after revision June 6, 1995. Acad Radio11995;2:902-904 © 1995, Association of University Radiology
902
Vol. 2, No. 10, October 1995
visually translucent silicone polymer that has MR imaging characteristics similar to those of the human brain. The right hemisphere contains an internal carotid artery with curved carotid siphon and a middle and anterior cerebral artery (Fig. 1). The anterior cerebral artery does not branch; it ends in the straight sinus. The middle cerebral artery trifurcates, passing over the insula, and ends without further branching in the superior sagittal sinus. Both sinuses meet at the torcular herophil and exit the skull by the right transverse and sigmoid sinuses. All arteries were modeled from actual human arteries by injecting fresh human cadavers with acrylic resin;
DESIGN AND EVALUATION
OF A FLOW PHANTOM
the details of this method are reported elsewhere [1]. Flow in the four main arteries may be controlled by monofilament tourniquets. All veins were hand constructed in wax using a resin cast human model, duplicating all dimensions, variations, and shapes. Once the vein and artery models were arranged in the skull, they were coated with a thin elastic silicone. Next, the entire skull was filled with the clear silicone polymer. Finally, the wax was removed thermally and chemically [1]. Silicone tubing was attached to the vessels at the base of the skull for fluid access. Imaging of the phantom was performed in a variety of ways. Routine angiographic plain films were performed following infusion of 25% by volume of ioxaglate meglumine (Hexabrix; Mallinckrodt Medical, St. Louis, MO) contrast material (Fig. 2). CT angiography was performed on a Somatom Plus scanner (Siemens Medical Systems, Iselin, NJ), using a similar concentration of ioxaglate meglumine (Fig. 3). MR angiography was performed on a Siemens GBS2 1.5-T system (Fig. 4). RESULTS AND DISCUSSION
Prior efforts at producing an anatomically correct vascular phantom have been limited to a short segment of the carotid artery that includes the bifurcation [2]. The need for a more complete vascular phantom of the head prompted us to build a near-anatomic human
FIGURE 1. Lateral view of the flow phantom demonstrating the anatomic appearance of the acrylic skull model. Connecting tubes at the skull base for the connection of the bidirectional flow are present. Over the convexity, the small screw adjustments for adjusting the constriction of the anterior cerebral artery and middle cerebral artery branches are indicated (arrows).
FIGURE 2. Lateral plane film view following contrast injection into the phantom. All three branches of the middle cerebral artery, as well as the single anterior cerebral artery branch continuing in the superior sagittal sinus, are visible.
903
CREASY
ET AL.
vol. 2, No. 10, October 1995
A
B
FIGURE 3. Computed tomography angiography was performed with a table speed of 1 mm, a table increment of 1 mm per revolution, and a 2-mm slice thickness reconstruction. Lateral view (A) and anteroposterior view (B) demonstrate good visualization of the vessels within the phantom, Because of placement of the phantom, an air bubble causes incomplete visualization of the most anterior portion of the anterior cerebral artery.
FIGURE 4. Lateral maximum intensity projection reconstruction from a dualslab axially acquired time-of-flight magnetic resonance angiogram. Flow artifacts can be seen in the region of the torcular herophil, the junction of the most anterior middle cerebral artery branch, the superior sagittal sinus, and the midanterior cerebral artery.
intracranial vascular phantom that was conceived, designed, built, and used for preliminary imaging research. Use of this anatomic phantom for pulse sequence and parameter refinement improves the accuracy of observations compared with less anatomically
904
correct phantoms or human volunteers, w h o tend to tire and move. Buxton et al. [3] and Frank et al. [4] began to study the complex artifacts that are produced by most MR sequences. These artifacts are displaced in three-dimensional space and by wrap-around. The complexity of anatomy found in humans obscures these confusing shadows, but the model, which is filled with a homogenous material, allows the clear appreciation of these artifacts. The major disadvantage of the model is the lack of third- and fourth-order vessels and the lack of a capillary bed. However, for our goals, such complexity was deliberately deleted to keep the model simple and to control costs. Although our initial applications were in the area of optimizing MR imaging sequences, we think that this model also will be helpful in CT angiography and rotational angiography research. REFERENCES 1. Kerber CW, Heilman CB, Zanetti PH. Transparent elastic arterial models: I. A brief technical note. Biorheology1989;26:1041-1049. 2. Frayne R, Gowman LM, Rickey DW, et al. A geometrically accurate vascular phantom for comparative studies of X-ray, ultrasound, and magnetic resonance vascular imaging: construction and geometrical verification. Med Phya 1993;20:415-425. 3. Buxton RB, Kerber CW, Frank LR. Pulsatile flow artifacts in two-dimensional time-of-flight MR angiography: initial studies in elastic models of human carotid arteries. J Magn Reson Imaging 1993;3:625-636. 4. Frank LR, Buxton RB, Kerber CW. Pulsatile flow artifacts in 3D MR imaging. Magn Reson Med 1993;30:296-304.