Formulate your best endograft plan with a combination of preprocedure imaging and postimaging processing.
To view the figures that correspond to this article, please refer to the print version of our November/December issue, page 34.
Interventionalists must rely on accurate imaging in order to select appropriate candidates for endovascular aneurysm repair (EVAR). The limitations of the available endograft devices require the assessment of several anatomic features of an abdominal aortic aneurysm (AAA) prior to embarking on an endoluminal repair.1 We will discuss these pertinent anatomic features as well as the imaging techniques that are used in preprocedural evaluation of an AAA patient. The specific techniques to be discussed are digital subtraction angiography (DSA), computed tomography angiography (CTA), MRI and magnetic resonance angiography (MRA), and intravascular ultrasound (IVUS). We will address the advantages and limitations of each of these techniques, and we will include a discussion of the postimaging processing used for CTA and MRA, highlighting their unique advantages for preprocedural imaging. Finally, we will outline our protocol for preprocedural imaging of our patients using these techniques.
The combination of DSA and CTA initially served as the gold standard for accurate pre-EVAR measurements. Advances in CTA and MRA imaging techniques have substantially lessened DSA’s role in preprocedural imaging. In fact, many endovascular centers (including ours in Utrecht) have eliminated DSA from their imaging arsenal and use CTA alone.2 The main reasons for eliminating DSA stem from the combined risks and potential complications of arterial puncture and iodinated contrast that the procedure requires. DSA also does not provide as much information when compared with a well-performed CTA.
DSA does offer some advantages, including the ability to visualize the true vessel lumen and pertinent branches. Physicians can obtain images in a lateral and AP view in order to evaluate the aorta, the renal arteries, the iliac system, and both femoral arteries. A calibrated pigtail catheter is commonly used to gauge the required length of the endovascular graft. Visualization of the aorta and its branches allows the interventionalist to assess several anatomic features important in the selection of appropriate EVAR candidates, including (1) the length and angulation of the aneurysm neck, (2) any associated renal artery stenoses, (3) iliac artery tortuosity or stenosis, (4) patency of internal iliac arteries, and (5) luminal diameter of the iliac vessels, especially the external iliac arteries. Such an assessment may even make it possible for the physician to intervene before EVAR to deal with significant problems detected on the arteriogram, including renal artery or iliac artery stenoses. These interventions may allow easier insertion of the AAA device and decrease the incidence of periprocedural complications.
In addition, DSA can detect patent lumbar arteries and a patent inferior mesenteric artery. Although the value of preprocedural detection of these patent vessels in the development of type II endoleaks remains controversial, some researchers have reported that successful preprocedural embolization of patent side branches decreases the incidence of type II endoleaks.3 In centers (such as ours) where DSA is not performed routinely, this strategy does not appear to be a useful option.
Despite its potential usefulness, DSA has significant limitations. DSA is unable to detect actual aneurysm size, as it only reveals the true lumen and not the associated thrombus (Figure 1). In addition, this imaging method does not accurately detect calcification at the aortic bifurcation and within the iliac vessels—a very important aspect of preprocedure imaging that needs to be assessed. CTA allows for the accurate assessment of all these important issues in preprocedure imaging, leading many centers to discontinue their use of DSA. As this article discusses, performing a combination of MRI and MRA may provide even more accurate preprocedural imaging than CTA in the near future.
In most centers, spiral CTA has taken DSA’s place as the gold standard for pre-EVAR imaging.4,5 As mentioned earlier, high-quality preprocedural imaging is essential for patient selection, endograft sizing, and procedure planning. Spiral CTA with postimaging processing provides all of the necessary information and has made DSA superfluous. Postimaging processing of spiral CTA data has proven to be superior over the two-dimensional DSA projection for both patient selection and endograft sizing.6,7
In our center, we use a standardized CTA acquisition protocol. CT scans are performed on a Philips AV-EP spiral CT scan (Philips Medical Systems, Best, The Netherlands). We use a 5 mm/sec table speed with a reconstruction index of 2 mm; we perform one unenhanced scan and another using 140 mL Ultravist (Berlex Laboratories, Montville, NJ) at a rate of 3 mL/sec with a 30-second delay. The scan starts directly above the level of the celiac trunk and continues until the bilateral femoral bifurcation. At 140 kV/225 mA and 70 continuous rotations, the entire volume can be scanned in one 50-second breath hold. The raw data are sent to a postimaging workstation for exact analysis.
We employ the single-breath-hold technique to avoid respiratory artifact and optimize image quality. The patient must hold his breath during scanning for as long as possible; most patients can do this for approximately 30 seconds. By that time, the aortic bifurcation has passed through the scan and motion artifacts from this level down will be minimal. Modern multislice helical CT scanners are able to cover the entire abdomen within 25 seconds, reducing motion artifacts even farther.8 Multislice scanners can also produce thinner slices, leading to a more isometric voxel size. This technology will slightly increase the inplane resolution because of the reduced partial-volume effect, but the resolution of reconstructed planes will improve dramatically compared with normal spiral CTA.
In order to evaluate a patient’s suitably for EVAR, it is essential that the interventionalist measure the diameter and length of the infrarenal aortic neck and both common iliac arteries. CTA provides tissue contrast that easily distinguishes between lumen mural thrombus and aortic wall; in combination with postimaging processing, this provides the basis for exact measurements. Furthermore, it is important to depict the amount of thrombus and calcium in the infrarenal neck, the common iliac arteries, and the external iliac arteries (Figure 2). Unlike DSA, CTA gives an excellent depiction of thrombus as well as calcium. The interventionalist will also measure the aorta in a patient who is suitable for EVAR.
Aortic length and diameter measurements allow the interventionalist to choose an endograft of the right configuration (monoiliac or bifurcated). The physician can perform these measurements using commercially available postimaging processing software or workstations.
After the physician has selected both the appropriate patient and the correct endograft, planning the intervention is the last step in preprocedure imaging. In most AAAs, the aorta is elongated and tortuous. This anatomy may result in a more difficult introduction through one of the iliac systems or in a very angulated infrarenal neck complication graft delivery. Using a three-dimensional model of the aortic lumen, the physician can assess the correct angulation and rotation of the C-arm during the intervention (Figure 3). The origin of the most distal renal artery is essential to the exact placement of the proximal attachment system.9 The hypogastric artery origin is important for proper placement of the distal attachment systems. In our experience, a rotation and angulation of the C-arm based upon preprocedural measurements on the three-dimensional model results in precise placement of the endograft without losing precious millimeters in the proximal and distal seal zones.
CTA has drawbacks as well, but they are only significant in postprocedural follow-up. Because of the lifelong CT scan surveillance that patients require after EVAR, their potential accumulated dose of radiation is high. In this group of patients, renal insufficiency is common, so the repeated load of iodine-containing contrast agent used in CTA is also undesirable.
MRI and MRA
MRI and MRA techniques have taken a tremendous leap forward in the last decade. New developments in MR hardware have made faster scanning at a higher resolution possible. MRA techniques make use of gadolinium-based contrast agents and allow fast three-dimensional imaging of the abdominal aorta and branches, as well as imaging of the iliac arteries. Gadolinium is non-nephrotoxic, so the risk of inducing further renal insufficiency is absent in MRA. MR has other theoretical advantages over CTA, including inherent three-dimensionality, excellent soft tissue contrast, and lack of ionizing radiation.
In order to extract all the necessary information from MR for patient selection, sizing, and procedure planning, however, several scans must be obtained.
We use a standardized MR protocol at our center, and perform MRI/MRA scans on a Philips Intera 1.5-T scanner (Philips Medical Systems). For all scans, we use a quadrature wraparound synergy body coil as a receiver coil. Patients enter the scanner headfirst. After the initial survey, we obtain the following scans (total examination time is approximately 20 minutes):
• T1-weighted spin echo (transverse orientation)
• T2-weighted turbo spin echo (transverse orientation)
• High-resolution, three-dimensional, contrast-enhanced (CE) MRA (sagittal orientation)
• Postcontrast T1-weighted spin echo (transverse orientation, as precontrast)
For the three-dimensional, CE-MRA, we administer 25 mL of gadolinium contrast agent at a rate of 2 mL/sec followed by 20 mL of saline solution at a rate of 1.5 mL/sec. A simple high-resolution three-dimensional, CE-MRA, however, is insufficient. This scan can provide the physician with an excellent depiction of the aortic lumen and all patent side branches, but because of the pronounced T1-weighting of this scan, the intrinsic tissue contrast is minimal. Exact measurements of the aortic infrarenal neck cannot be performed on this scan unless a part of the T1-weighting is given up. The vessel wall and thrombus are not clearly distinguishable and measuring the lumen alone is insufficient. Additional noncontrast-enhanced, transverse T2-weighted turbo spin echo and T1-weighted spin echo MRI scans are necessary. A T1-weighted postcontrast scan is particularly well-suited for patient selection and sizing because it provides a good contrast between lumen, thrombus, and vessel wall.
The MIP of the CE-MRA is ideally suited for procedure planning and predicting the best rotation of the fluoroscope during the intervention (Figure 4). The MIP images of the three-dimensional CE-MRA are comparable to a three-dimensional DSA. These images allow the same assessment as DSA: the length and angulation of the aneurysm neck, any associated renal artery stenoses, iliac artery tortuosity or stenosis, patency of internal iliac arteries, and luminal diameter of the iliac vessels and the external iliac arteries. However, neither MRI nor MRA will depict calcium, which can be problem in a tortuous iliac system or a short aortic neck.
Raw MRI and MRA data can be sent to a postimaging workstation for exact measurements and detailed analysis.
Physicians commonly employ intraprocedural IVUS to ensure accurate endograft deployment (especially with a difficult aneurysm neck). The technique is less-commonly used for preprocedural imaging. Some investigators have advocated the use of IVUS in patients with renal insufficiency, proposing that this subset of patients should undergo an initial noncontrast CT scan.10 If their anatomy remains unclear from the study, the interventionalist can then use IVUS as an adjunct technique to determine if a patient is an EVAR candidate.
Some researchers are enthusiastic about the routine use of IVUS. White et al have called IVUS, “the ultimate tool for AAA assessment and endovascular graft delivery.”11 IVUS can accurately measure the aortic neck diameter, and can also determine the quality and extent of calcification in the arterial wall of the iliac vessels. This technique can also accurately determine the required endograft length by using a pullback method that measures the distance from the distal renal artery to the aortic bifurcation, and then to the internal iliac arteries. This technique is considered an accurate method of length measurement.
The limitations of IVUS include the required arterial puncture with its inherent complications (as discussed with DSA). The advantage of using IVUS over DSA is that contrast is not required for this technique. The learning curve associated with IVUS presents another disadvantage. Because it is not a technique that is as widely applied as other imaging modalities, new users require training that is specific to IVUS. Finally, IVUS is not currently available in many endovascular centers. Its application would require a significant investment in equipment and training in order to be routinely used as a preprocedural imaging technique for EVAR.
Processing of the acquired images (whether they are gathered through CTA, MRI/MRA, or even IVUS) is essential in the preprocedural analysis of a patient. Taking measurements on the hard copies of a scan can be inaccurate. For instance, an angulated neck will look wider on an axial slice than when measurements are taken perpendicular to the central lumen line (CLL) (Figure 5). Furthermore, the window level and width of an axial slice on a hard copy is fixed, making it difficult to distinguish lumen from calcium. Using a higher window width, the amount of calcium is easily assessed.
In order to minimize preprocedural as well as postprocedural complications, the physician must make no compromises concerning workup. Inaccurate measurements can lead to inappropriate graft placement, or use of the wrong size endograft. These mistakes can lead to an endoleak, or to the covering of one of the renal or hypogastric arteries.
Postimaging processing can be performed on several commercially available programs. We currently process our images on a Philips EasyVision workstation, release 4 (Philips Medical Systems). The following postimaging processing options are useful in the assessment of an AAA: cine-mode, multiplanar reconstruction (MPR) formats, MIP, and three-dimensional shaded-surface reconstructions (Figure 6).
Postprocessing begins by stacking the individual image slices. The interventionalist can then display the resulting series in a movie-like fashion. This cine-mode is interactive and allows the viewer to move the cutplane perpendicular to the slice direction. In CTA, T1-weighted, and T2-weighted MRI scans, the movements will follow the vertical body axis. In the CE-MRA, the movements will be perpendicular to the coronal plane. The up-and-down movement facilitates the physician’s interpretation of sidebranches and tortuosity.
Using the cine-mode display, a large proportion of aneurysms can be excluded from EVAR at an early stage, based on rough estimates of the diameter, length, and quality of the infrarenal aortic neck. Extension of the aneurysm into the iliac arteries and the extent of calcification of the access arteries can also render the aneurysm unsuitable for EVAR without further measurement.
For aneurysms that are not excluded on the basis of the axial slices in cine-mode, a more detailed analysis is necessary for exact diameter and length measurements. The physician can use MPR to simultaneously display the aneurysm in axial, sagittal, and coronal planes. Using MPR, a curved line can be drawn following the center of the CLL. Reformats perpendicular to this line (curved linear reformats) provide the physician with the exact measurement of the vessel lumen in a plane perpendicular to the vessel axis. In addition, the relative positions of the perpendicular reformats allow measurement of the vessel segment length along the CLL. This method is comparable to length measurements using DSA, however, postprocessing CTA data allows the operator to position the CLL and can be considered equivalent to the calibrated catheter in any desired position. For instance, the physician can choose a line that estimates the placement of the endograft within the aorta.
MIP uses the density of the pixels in the reconstructed slices to compose two-dimensional projections of the stacked slices. In this manner, angiograph-like images can be produced from any desirable viewpoint. MIP images are particularly useful for quick evaluation of the iliac arteries.
Three-dimensional reconstructions provide virtual images of the vessels and selected other structures. The basic principle in the generation of these three-dimensional reconstructions is the selection of voxels with a density above a certain threshold, in a process called segmentation. The complete spiral data set, with a wide range of densities, is reduced to a binary set of voxels that can be displayed or hidden. Mathematical manipulation of these selected voxels adds the illusion of depth by generating shade surfaces on the selected object. The physician can obtain three-dimensional shaded surface reconstructions by performing semiautomatic and manual segmentation of lumen and thrombus, however, this is a labor-intensive method and not really necessary for endograft sizing.
Once a patient is accepted for endovascular grafting, these three-dimensional images are useful for determining the surgical plan. By viewing and rotating the virtual aneurysm in any direction, the tortuosity of the access arteries and angulation of the aortic lumen can be appreciated.
The interventionalist can submit raw CTA data commercially using the Preview system (Medical Media Systems, West Lebanon, NH) and obtain prefabricated, three-dimensional, shaded-surface reconstructions of lumen, thrombus, and calcium. With this system, the physician can select the most suitable endograft by simulating the endograft in the three-dimensional aneurysm image on the computer screen.
Vitrea (Vital Images), a relatively new PC-based program, can be used as well. This program can process raw data or Dicom III images (Figure 7). One of the major advantages of the Vitrea system is that fully automated, three-dimensional, shaded-surface reconstructions are incorporated into the program. Although the program is user-friendly, the range of measurements that can be performed is limited.
As discussed, imaging remains a crucial part of the preprocedural planning for EVAR. We believe that CTA is the current standard of care, especially when combined with postimaging processing.
Maarten J. Van der Laan, MD, is the Research Imaging Resident in the Department of Vascular Surgery at the University Medical Center in Utrecht, The Netherlands. Dr. Van der Laan may be reached at +31 (0) 30 2506965; firstname.lastname@example.org.
Ross Milner, MD, is the Marco Polo Fellow for the Society of Vascular Surgery and a postdoctoral research fellow at the University Medical Center in Utrecht, The Netherlands. Dr. Milner may be reached at +31 (0) 30 2506965; email@example.com.
Jan D. Blankensteijn, MD, is Associate Professor in Vascular Surgery at the University Medical Center in Utrecht, The Netherlands. Dr. Blankensteijn may be reached at +31 (0) 30 2506965; firstname.lastname@example.org.
1. Blankensteijn JD. Preoperative imaging techniques and criteria for endovascular aneurysm repair. In: Branchereau A, Jacobs M, eds. Surgical and Endovascular Treatment of Aortic Aneurysms. Armonk, NY: Futura Publishing; 2000:69-79.
2. Broeders IA, Blankensteijn JD, Olree M, et al. Preoperative sizing of grafts for transfemoral endovascular aneurysm management: A prospective comparative study of spiral CT angiography, arteriography, and conventional CT imaging. J Endovasc Surg. 1997;4(3):252-261.
3. Parry DJ, Kessel DO, Robertson I, et al. Type II endoleaks: Predictable, preventable, and sometimes treatable? J Vasc Surg. 2002;36(1):105-110.
4. Armon MP, Yusuf SW, Latief K, et al. Anatomical suitability of abdominal aortic aneurysms for endovascular repair. Br J Surg. 1997;84(2):178-180.
5. Beebe HG, Kritpracha B, Serres S, et al. Endograft planning without preoperative arteriography: A clinical feasibility study. J Endovasc Ther. 2000;7(1):8-15.
6. Fillinger MF. Imaging of the thoracic and thoracoabdominal aorta. Semin Vasc Surg. 2000;13(4):247-263.
7. Wever JJ, Blankensteijn JD, Mali WP, Eikelboom BC. Maximal aneurysm diameter follow-up is inadequate after endovascular abdominal aortic aneurysm repair. Eur J Vasc Endovasc Surg. 2000;20(2):177-182.
8. Rydberg J, Kopecky KK, Lalka SG, et al. Stent grafting of abdominal aortic aneurysms: Pre- and postoperative evaluation with multislice helical CT. J Comput Assist Tomogr. 2001;25(4):580-586.
9. Broeders IA, Blankensteijn JD. A simple technique to improve the accuracy of proximal AAA endograft deployment. J Endovasc Ther. 2000;7(5):389-393.
10. Bush RL, Lin PH, Bianco CC, et al. Endovascular aortic aneurysm repair in patients with renal dysfunction or severe contrast allergy: Utility of imaging modalities without iodinated contrast. Ann Vasc Surg. 2002;19.
11. White RA, Donayre C, Kopchok G, et al. Intravascular ultrasound: The ultimate tool for abdominal aortic aneurysm assessment and endovascular graft delivery. J Endovasc Surg. 1997;4(1):45-55.