Article Text

Download PDFPDF

Aortic arch injection with C-arm cone beam CT for radiosurgery treatment planning of cerebral arteriovenous malformations: technical note
  1. Martin G Radvany1,
  2. Tina Ehtiati2,
  3. Judy Huang3,
  4. Mahadevappa Mahesh4,
  5. Philippe Gailloud1
  1. 1The Johns Hopkins University School of Medicine, Division of Interventional Neuroradiology, Baltimore, Maryland, USA
  2. 2Siemens Corporate Research, Princeton, New Jersey, USA
  3. 3The Johns Hopkins University School of Medicine, Department of Neurosurgery, Baltimore, Maryland, USA
  4. 4The Johns Hopkins University School of Medicine, Department of Radiology, Baltimore, Maryland, USA
  1. Correspondence to Dr Martin G Radvany, The Johns Hopkins University School of Medicine, Division of Interventional Neuroradiology, 600 N Wolfe Street, Radiology B-100, Baltimore, MD 21287, USA; mradvan2{at}jhmi.edu

Abstract

Background C-arm cone beam CT (CBCT) with selective intra-arterial contrast injection combined with digital subtraction angiography (DSA) is currently used for the evaluation and treatment planning of cerebral arteriovenous malformations (AVMs). In some instances an AVM will derive its blood supply from more than one main cervical artery (carotid and/or vertebral artery) and a single-vessel injection will not adequately demonstrate the entire AVM nidus.

Methods Three patients with cerebral AVM in whom the entire nidus could not be visualized by injection of a single cervical artery are reported. CBCT dataset acquisition was performed by intra-arterial contrast injection in the ascending thoracic aorta through a 5 F pigtail catheter. The injection of diluted iodinated contrast agent (35%) lasted 22 s at a rate of 8 ml/s for a total volume of 176 ml (61.6 ml of contrast agent). The dataset was then processed using standard reconstruction methods.

Results Contrast injection in the ascending aorta during a single CBCT acquisition provided a volumetric dataset adequate for subsequent radiosurgical treatment planning.

Conclusion This is a safe and effective angiographic technique for the acquisition of volumetric datasets using CBCT that are suitable for treatment planning of intracranial AVMs deriving their blood supply from more than one major cervical artery. This technique allows imaging of the entire AVM nidus during a single CBCT acquisition.

  • Angiography
  • coil
  • arteriovenous malformation
  • ultrasound
  • hydrocephalus
  • aneurysm
View Full Text

Statistics from Altmetric.com

Introduction

Stereotactic radiosurgery is a well-established method of treatment for cerebral arteriovenous malformations (AVMs).1 Prior to treatment, patients undergo multimodality imaging for treatment planning which usually includes digital subtraction angiography (DSA) and MRI/MRA imaging. Recent advances in C-arm cone beam CT (CBCT) technology now permit the acquisition of high-resolution volumetric vascular datasets in the angiography suite that can be used for radiosurgery planning.2

We previously studied CBCT for target delineation in stereotactic radiosurgery of brain AVMs under an IRB approved protocol (unpublished data). After establishing submillimeter registration accuracy with a phantom, CBCT imaging was incorporated into the routine radiotherapy treatment planning workflow as follows. After the placement of the external frame, patients underwent MRI/MRA followed by DSA with a catheter in the common carotid or vertebral artery, depending upon the blood supply to the AVM. This was immediately followed by CBCT imaging while injecting dilute contrast through the catheter. The patients were then transferred for radiotherapy with treatment planning.

However, when an AVM derives its blood supply from more than one major cervical artery, a single-vessel injection will only offer partial opacification of the nidus and the three-dimensional dataset will not be adequate for treatment planning. We describe three cases in which CBCT obtained by injecting the contrast agent in the ascending thoracic aorta resulted in complete nidus opacification from a single rotational acquisition.

Methods

Patient 1

The first patient was a man in his late 20s diagnosed at an outside institution with a left hemispheric parasagittal AVM (Spetzler–Martin grade 4). He underwent several embolization procedures at that time. He was admitted to our center 7 years later after suffering an intraventricular hemorrhage. DSA demonstrated an aneurysm on one of the pericallosal branches feeding the AVM. The aneurysm was embolized along with approximately 20% of the AVM nidus without complication. The patient was then scheduled for radiosurgical treatment. Our standard stereotactic angiography protocol includes CBCT for nidus delimitation. However, the residual nidus could not be adequately demonstrated without injection of both carotid arteries and the left vertebral artery (figure 1A–C).

Figure 1

Patient with left hemispheric parasagittal arteriovenous malformation (AVM): early arterial images for stereotactic radiosurgery planning. (A) Selective left common carotid injection. (B) Selective left vertebral injection. (C) Selective right common carotid injection. (D). C-arm cone beam CT image from aortic contrast injection demonstrating diffuse nature of AVM nidus.

A modified imaging protocol was used for CBCT in which a 5 F pigtail catheter (Cook Medical, Bloomington, Indiana, USA) was advanced into the ascending thoracic aorta. A standard 20 s CBCT acquisition was performed using a flat-panel angiography system (Artis zee VC13D; Siemens AG Healthcare Sector, Forchheim, Germany) while injecting the contrast agent through the pigtail catheter. The injector was filled with 70 ml iohexol-300 (GE Healthcare, Milwaukee, Wisconsin, USA) and 130 ml normal saline. The diluted contrast mixture was administered using a power injector with the following parameters: rate rise 0.5 s; flow rate 8 ml/s; total volume 176 ml; contrast volume 61.6 ml; injection duration 22 s; x-ray delay 5 s. Dataset reconstruction was performed on a dedicated workstation (syngo DynaCT, Siemens AG). The reconstructed dataset included 394 slices through the acquired volume, with a 512×512 matrix and isotropic voxels of 0.46 mm3 (figure 1D).

Patient 2

The second patient was a woman in her late 60s who presented to the emergency room with a severe headache. A CT scan of the head demonstrated intraventricular hemorrhage. MRI revealed an enlarged draining vein without clear AVM nidus (figure 2A,B). DSA showed a small left choroidal AVM (1.5×1.3 cm) vascularized by pericallosal branches of the left anterior cerebral artery (ACA) and parietal branches of the left posterior cerebral artery (PCA), with a solitary deep draining vein with outflow stenosis (Spetzler–Martin grade 2) (figure 2C). She remained neurologically intact and was discharged home with plans to return for stereotactic radiosurgery. DSA performed for treatment planning again noted that the nidus could not be fully visualized without injection of both the left carotid artery and the left vertebral artery (figure 2D). Stereotactic localization was obtained with CBCT performed using the previously described protocol (figure 2E).

Figure 2

Patient with left choroidal arteriovenous malformation (AVM). (A) T2-weighted MRI demonstrating enlarged draining vein. (B) T1-weighted MRI with contrast without evidence of AVM. (C) Selective left vertebral artery injection demonstrating majority of nidus and single deep draining vein. (D) Selective left carotid artery injection demonstrating additional nidal component of choroidal AVM supplied by anterior cerebral circulation (arrow). (E) C-arm cone beam CT image with aortic contrast injection demonstrating AVM nidus. (F) C-arm cone beam CT image demonstrating draining vein.

Patient 3

The third patient was a man in his early 20s who underwent MRI as a volunteer and was found to have an AVM. DSA showed a left frontal-basal AVM (Spetzler-Martin grade 4) supplied by lenticulostriate arteries of the left A1 and M1 segments and by the left frontopolar and anterior temporal arteries with drainage via deep and superficial veins. DSA was performed for treatment planning. Owing to the flow dynamics of the AVM, the entire AVM was not visualized on the left common carotid artery (CCA) injection; segments supplied by the left A1 branches only filled from the right CCA injection. Stereotactic localization was obtained with CBCT performed using the previously described protocol.

Results

The high-resolution volumetric datasets obtained from the single aortic injection of diluted contrast agent were adequate for radiosurgery planning when compared with the MRA images. The images reconstructed from these aortic injections were equivalent in quality to CBCT images obtained from the routinely used selective injection of a single cervical artery. Aortic arch CBCT offered a similar ability to distinguish between blood vessel and cerebral parenchyma, and between nidus and feeding arteries or draining veins (figure 2F). No complications were associated with the angiographic procedures.

Air kerma measurements were conducted to compare the relative radiation dose exposure from conventional DSA and CBCT acquisitions. The DSA protocol utilizes biplane imaging in anteroposterior and lateral projections at 3 frames per second (fps) for a total of 10 s per acquisition (3.6 μGy system dose per frame). The CBCT protocol utilizes a 20 s scan time of a single plane (plane A), 200 degree rotation, 0.4 degree angulation step, 500 projection images and 1.2 μGy system dose per pulse. Air kerma measurements were performed for the two types of acquisitions; radiation exposure was 415±1 mR for CBCT acquisition and 88.5±1 mR for DSA acquisition at 3 fps for a 10 s acquisition. Based on these air kerma measurements, the radiation exposure for CBCT is approximately 4.7 more than for DSA acquisition (3 fps, 10 s), but it should also be noted that, during DSA acquisition, the radiation dose is accumulated to an identical part of the body while, during CBCT acquisition, the dose is spread out over a rotation with varying projection angles.

Discussion

Our experience shows that three-dimensional CBCT angiography with contrast injection in the aortic arch provides high-resolution volumetric datasets that can be used for planning cerebral AVM radiosurgery. In cases in which the supply is from multiple arteries, the entire nidus can be visualized without having to overlay separate acquisitions. In our experience, it has proved particularly useful for the stereotactic evaluation of cerebral AVMs deriving their blood supply from several major cervical vessels (carotid or vertebral arteries).

This technique can be performed without excessive contrast administration and without the higher radiation dose that would be associated with multiple CBCT dataset acquisitions. Further reduction of radiation in the CBCT acquisitions may be feasible by reducing the radiation dose per frame or by using the 8 s CBCT protocol with fewer projection images acquired. Because the CBCT data used for this study involve high contrast images, greater degrees of noise may be tolerated, but this will require further investigation to determine if these datasets provide images acceptable for radiotherapy treatment planning.

Conclusion

We describe a safe and effective angiographic technique for the acquisition of volumetric datasets using CBCT that are suitable for treatment planning of intracranial AVMs deriving their blood supply from more than one major cervical artery. This technique limits the radiation exposure by reducing the number of volumetric CBCT acquisitions and simplifies treatment planning by demonstrating the AVM on a single acquisition.

References

View Abstract

Footnotes

  • Funding Siemens Corporate Research provided research support.

  • Competing interests None.

  • Patient consent Local IRB approved consent forms were used for all patients in the study along with standard institutional consent forms for the procedures.

  • Ethics approval Ethics approval was provided by the Johns Hopkins Investigational Review Board.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.