Article Text

Download PDFPDF
Original research
A carving method to determine an optimal working projection using three-dimensional volume rendering digital subtraction angiography in coil embolization of cerebral aneurysms
  1. Keiko Kitazawa,
  2. Takatoshi Sorimachi,
  3. Yasushi Ito,
  4. Yukihiko Fujii
  1. Department of Neurosurgery, Brain research Institute, University of Niigata, Niigata, Japan
  1. Correspondence to Dr K Kitazawa, Department of Neurosurgery, Brain Research Institute, University of Niigata, 1-757, Asahimachidori, Chuo-ku, Niigata 951-8585, Japan; keiko413{at}mte.biglobe.ne.jp

Abstract

Background A working projection in coil embolization of a cerebral aneurysm is usually determined using three-dimensional volume rendering digital subtraction angiography (3D VRDSA), in which the boundary between an aneurysm and its parent artery cannot be identified simultaneously on the front and back surfaces of the vessel. A new method was devised to obtain a more optimal working projection.

Methods 83 patients with aneurysms being treated by embolization were retrospectively evaluated. An aneurysm was accurately distinguished from the parent artery by observation of front, back, proximal and distal sides of the vessel on 3D VRDSA (carving method). An optimal working projection with simultaneous identification of the front and back boundary lines was determined using a translucent vessel complex combined with the carved aneurysm.

Results In 32 aneurysms (38.6%), the optimal working projection was consistent with the working projection that had been used during the procedure. In terminal type aneurysms, the angle difference between the optimal and actual working projections was significantly smaller than in the other types (p<0.05). Aneurysms with a maximal diameter <5 mm showed a significantly larger angle difference between the optimal and actual working projections than aneurysms with a maximal diameter ≥5 mm (p<0.05).

Conclusion In more than half of the patients, the actual working projection was inaccurate. The carving method might be useful to determine working projections, especially for aneurysms other than the terminal type and/or those with a maximal diameter <5 mm.

Statistics from Altmetric.com

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.

Introduction

A working projection used during coil embolization of a cerebral aneurysm is an image created by digital subtraction angiography (DSA) which distinctly shows both the aneurysm and the parent artery. During coil embolization procedures, operators manipulate microcatheters and coils based on the working projection image, in which a real time fluoroscopic image is projected on an overlay or roadmap image obtained by two-dimensional DSA (2D DSA). Tight packing of coils into a cerebral aneurysm has been reported to reduce the recanalization rate of the aneurysm in long term follow-up studies.1–5 To obtain the tight packing, an optimal working projection should be created to separate the cerebral aneurysm from the parent artery as clearly as possible.6 ,7

In an image created by three-dimensional volume rendering DSA (3D VRDSA), the boundary between an aneurysm and its parent artery can be delineated on both the front and back surfaces of vessels. In an optimal working projection, boundary lines on the front and back surfaces should be united. The working projection is usually determined by use of a 3D VRDSA image obtained just before coil embolization. The boundary between an aneurysm and its parent artery on both front and back surfaces of the vessel cannot be observed simultaneously even using a translucent mode of the volume rendering image; therefore, the front and back boundaries are often dissociated from each other on a working projection. To obtain an optimal working projection, we devised the carving method, a new method for confirming the boundary lines on both the front and back surfaces. With the carving method, an aneurysm is accurately delineated from the parent artery by observation of the front, back, proximal and distal sides of the vessel on a 3D VRDSA image at the work station. The carved aneurysm is combined with another 3D VRDSA image taken in the translucent mode. Then, an optimal working projection, which depicts the boundary lines on both the front and back surfaces, is obtained.

In the present study, we retrospectively determined the optimal working projection using the carving method in patients who had previously undergone coil embolization of cerebral aneurysms. We then compared the optimal working projections and working projections that were actually used during the embolization to evaluate aneurysm types, and observed a significant difference between the optimal and actual working projections.

Methods

Subjects

We conducted a retrospective analysis of 142 consecutive patients who underwent coil embolization of a cerebral aneurysm between September 2005 and February 2010 at our institution. Of the 142 patients, three-dimensional rotational angiography (3D RA) data were not available for 31 patients, and data on working projections used during the embolization were lost for another 26 patients. Two patients underwent embolization for recurrent aneurysms. A total of 59 patients were thus excluded from the study. Sixty-eight of the remaining 83 patients included in the study had unruptured aneurysms, and the remaining 15 had ruptured aneurysms. Table 1 shows the aneurysm locations. Based on the relationship between the aneurysm locations and the parent arteries, we categorized the aneurysms into three types: a terminal type, including basilar tip and internal carotid artery (ICA) tip aneurysms; a side wall type, including ICA paraclinoid, ICA–anterior choroidal artery, ICA–posterior communicating artery, vertebral artery (VA) trunk, VA–posterior inferior cerebellar artery, basilar artery trunk and posterior cerebral artery aneurysms; and the remaining type, including middle cerebral artery, anterior communicating artery and basilar artery–superior cerebellar artery aneurysms.8–10 Aneurysms were classified as terminal type, side wall type and the remaining type in 18, 37 and 28 patients, respectively.

Table 1

Relationship between the location of the aneurysms and slant ratios

Digital subtraction angiography

In the 83 patients, DSA was performed just before coil embolization using an AXIOM Artis system (Siemens, Munich, Germany). Conventional 2D DSA and 3D DSA using a 6 Fr catheter advanced into the ICA or VA were performed with the Seldinger method via the femoral artery under general anesthesia. For 2D DSA, we used ioxaglic acid (Hexabrix: Guerbet Japan, Tokyo, Japan) as the contrast medium, which was administered automatically with a power injector at a rate of 3 ml/s, with a total amount of 5 ml for the ICA and a rate of 2.5 ml/s, and a total amount of 4.5 ml for the VA. For 3D RA, the contrast medium was administered at a rate of 2.5 ml/s and a total amount of 15 ml DSA series (three frames per second) were obtained in two (posteroanterior and lateral) views, routinely followed by 3D RA. Using the AXIOM Artis system, the C arm was rotated 190° within 5 s. A total of 126 images with a 512×512 matrix were acquired. All images were immediately transferred to the work station (Syngo Workplace: Siemens) via a network and stored on the hard disk to serve as the basis for various reconstruction algorithms.

Actual working projection used during coil embolization

Coil embolization was performed by board certified neurointerventionalists as well as board certified neurosurgeons with more than 15 years of experience in their specialties. An operator determined a working projection based on angiography performed just before the coil embolization. A 3D VRDSA image in the surface mode and/or the translucent mode was rotated on a monitor of the work station to distinguish the aneurysm from the parent artery and then the projection angle was decided. A C arm system was moved to the projection angle based on the 3D VRDSA image, and 2D DSA was performed. After confirmation that both the aneurysm and the parent artery were clearly depicted in the 2D DSA image, the angle of the working projection was finally determined.

Working projection using the carving method

Three-dimensional RA data were stored on the hard disk of a work station as dataset A; the same data were copied and stored as dataset B. Figure 1 shows a schema of the carving method for deciding the optimal working projection. Dataset A was evaluated in a volume rendering mode, and proximal and distal parts of the parent artery were cut out just at the distal and proximal sides of the aneurysm on a monitor of the work station (figure 1A). Boundary lines between the aneurysm and the parent artery were determined on the vessel surface by careful observation of the aneurysm–parent artery complex from the front, back, proximal and distal sides of the vessel in a 3D VRDSA image (figure 1B). The aneurysm was accurately carved out from the parent artery by the boundary lines (figure 1C). The carved aneurysm was combined with another 3D VRDSA image created from dataset B evaluated in a translucent mode (figure 1D). The image was rotated to a projection in which the boundary lines on both the front and back surfaces were united; this projection was defined as an optimal working projection (figure 1E).

Figure 1

Schema of the carving method procedure for an optimal working projection. (A) A schematic image of an aneurysm–parent artery complex on three-dimensional volume rendering digital subtraction angiography displayed on a monitor of a work station. The proximal and distal parts of the parent artery are cut out from the complex. (B) Observation of the complex from its front, back, proximal and distal sides. Boundary lines between the aneurysm and the parent artery on the surface of the vessel complex are delineated. (C) An aneurysm carved out from the parent artery by the boundary lines. (D) Fusion of the carved aneurysm and a vessel complex image on a translucent mode. (E) The image is rotated to an optimal working projection in which the slant area is shown as small as possible. (F) Upper: The carved aneurysm at the actual working projection. Lower: The carved aneurysm observed from the base. The carved aneurysm base (the area delineated by a yellow circle) is measured at each projection.

Comparison between an actual working projection used in coil embolization and a working projection obtained by the carving method

The area of the carved aneurysm base was measured using Image J (National Institutes of Health, Bethesda, Maryland, USA). Then, a combined image of the carved aneurysm and the parent artery in a translucent mode was rotated to the actual working projection used during coil embolization at the work station. The base area observed in this image was also measured as the slant area of the aneurysm base (figure 1F). A slant ratio is defined by the following equation: slant area/base area of the carved aneurysm × 100 (%). When an actual working projection was consistent with an optimal working projection determined by the carving method, the slant ratio was 0%. On the other hand, the wider the angle difference between an actual working projection and an optimal working projection, the larger the slant ratio. Types of aneurysms with large slant ratios were also investigated.

Statistical analysis

For analysis of a correlation between a slant ratio and types of aneurysms, a one way factorial ANOVA was performed. If statistically significant differences were observed by ANOVA, then a Tukey HSD was used to compare the groups. For analysis of a correlation between a slant ratio and a maximal diameter of an aneurysm, a Mann–Whitney test was performed. A value of p<0.05 was considered significant. All statistical analysis was performed with SPSS version 17 (IBM).

Results

In all 83 aneurysms of the present series, working projections were determined using the carving method. The average slant ratio for the 83 aneurysms was 20.9% (range 0–88%). In 32 of the 83 aneurysms (38.6%), the slant ratio was 0%, which indicated that the actual working projection used during embolization was consistent with the optimal working projection. In 68 patients with unruptured aneurysms, the average slant ratio was 20.9% (range 0–88%). In 27 of the 68 unruptured aneurysms, the slant ratio was 0%. In 15 patients with ruptured aneurysms, the average slant ratio was 21.1% (range 0–70%). In five ruptured aneurysms, a slant ratio of 0% was found. Table 1 shows the relationship between aneurysm locations and slant ratios.

A slant ratio of 0% was found in 12 of 18 aneurysms of the terminal type, in 10 of 37 aneurysms of the side wall type and in 10 of 28 other aneurysms. Figure 2 shows the relationship between the type of aneurysm and the slant ratio, and an analysis of variance demonstrated a significant positive correlation (p<0.05). The multiple comparison test (Tukey HSD) showed that the slant ratio for terminal type aneurysms was significantly smaller in comparison with side wall type aneurysms (p<0.05). The slant ratio for terminal type aneurysms tended to be smaller than that for other type aneurysms but the difference was not significant (p=0.077).

Figure 2

Relationship between three aneurysm types classified by location of the aneurysm to its parent artery and slant ratios. A significant difference in slant ratio is shown between terminal type aneurysms and side wall type aneurysms (p<0.05). An open circle indicates the mean slant ratio.

Average slant ratios were 24.5% (range 0–88%) and 15.0% (range 0–67%) in 42 aneurysms with a maximum diameter <5 mm and in 41 aneurysms with a maximum diameter ≥5 mm, respectively. Aneurysms with a maximal diameter <5 mm showed a significantly larger slant ratio than aneurysms with a maximal diameter ≥5 mm (p<0.05).

Representative cases

Case No 1: side wall type aneurysm

An incidental cerebral aneurysm was found in a 58-year-old woman by MR angiography which was performed due to vertigo and headache. DSA showed an aneurysm with a maximal diameter of 5 mm in the left ICA. Intra-aneurysmal coil embolization was performed with a balloon remodeling technique, which resulted in a neck remnant of the aneurysm. Figure 3 shows images of 3D VRDSA in a translucent mode of the actual working projection used during the procedure and the optimal working projection obtained by the carving method. The slant ratio of the aneurysm was 15.1%.

Figure 3

Translucent image of three-dimensional volume rendering digital subtraction angiography in case No 1, in which the optimal working projection was inconsistent with the actual working projection. (A) An image of the actual working projection used during the coil embolization procedure. (B) An actual working projection image combined with the carved aneurysm. (C) An actual working projection image displayed from the back side. The bottom of the carved aneurysm is shown in this image (the shaded slant area). This indicates that the actual working projection is inaccurate. (D) An optimal working projection obtained by rotation of the image on a work station to diminish the slant area.

Case No 2: terminal type aneurysm

A cerebral aneurysm was incidentally found in a 77-year-old man by CT scans which were performed following a head injury. On DSA, an aneurysm with a maximum diameter of 14 mm was shown in the bifurcation of the left ICA. Coil embolization was performed with a balloon remodeling technique, and DSA immediately after the embolization showed a slight neck remnant. Figure 4 shows that the actual working projection was consistent with the optimal working projection determined by the carving method. The slant ratio was 0% in this patient.

Figure 4

Translucent image of three-dimensional volume rendering digital subtraction angiography in case No 2, in which the optimal working projection was consistent with the actual working projection. (A) An image of the actual working projection used during the coil embolization procedure. (B) An actual working projection image, in which the carved aneurysm is combined with the vessel complex. (C) An actual working projection image displayed from the back side. A slant bottom area of the aneurysm is not observed. A straight black line indicates the bottom of the aneurysm. This demonstrates the accuracy of the actual working projection.

Discussion

Recanalization of embolized aneurysms after coil embolization is a major inherent problem of coil embolization. Many studies have reported that loose packing of an aneurysm during the procedure is related to the recanalization.1–5 11–13 Therefore, tight packing of an aneurysm is essential to prevent the recanalization.1–5 To achieve the tight packing, an accurate working projection, which shows both the aneurysm and the parent artery distinctly, is important. Currently, the working projection is usually created with a 3D VRDSA image, which has been reported to be useful for endovascular therapy of cerebral aneurysms in several studies.6 ,14–18 In the 3D VRDSA image, boundary lines between an aneurysm and a parent artery on both the front and back surfaces of the vessel cannot be identified simultaneously; therefore, the front and back boundary lines are often dissociated from each other on the working projection.

The carving method presented in this study, in which the boundary lines between an aneurysm and a parent artery on both the front and back surfaces of the vessel were identified simultaneously, provided a more accurate working projection than the ordinary method. To the best of our knowledge, this is the first report of this method of achieving an optimal working projection. In 32 of the 83 aneurysms (38.6%) in the present series, working projections obtained with the carving method were consistent with the working projections used during the procedures. This study demonstrated that a working projection determined by experienced neurointerventionalists using the ordinary method was less accurate than the working projection obtained by using the carving method in more than half of the procedures.

In the present study, the angle difference between the optimal and actual working projections was significantly smaller in terminal type aneurysms, including basilar top aneurysms and ICA bifurcation aneurysms, in comparison with other types of aneurysms. A terminal type aneurysm tends to develop on an extension of its parent artery; therefore, the boundary between the aneurysm and the parent artery is often perpendicular to the long axis of the parent artery. This anatomical feature of terminal type aneurysms is considered to result in a more accurate actual working projection compared with the other type of aneurysms. This result might suggest that the carving method is more useful in the other type of aneurysms than in terminal type aneurysms.

This study indicated that aneurysm size could affect the accuracy of a working projection. Aneurysms with a maximal diameter <5 mm showed a significantly larger angle difference between the optimal working projection obtained using the carving method and the actual working projection used during procedures than aneurysms with a maximal diameter ≥5 mm (p<0.05). The relationship between a large aneurysm and the parent artery can be observed in greater detail compared with that between a small aneurysm and its parent artery. The close observation of a large aneurysm complex might explain the more accurate actual working projection obtained for the large aneurysms. The carving method is considered valuable, especially in the embolization of small aneurysms.

This study has some limitations. In some aneurysms, determination of the working projection is multifactorial. In cases of an adjacent artery branching from the orifice of the aneurysms, visualization of the arterial branch must be considered to prevent the occlusion of the branch. For those aneurysms with a wide neck and located on the curve of the parent artery, the virtual plane separating the aneurysm from the parent artery may not be a single flat plane. In these cases, the ideal working angle might be sacrificed to prioritize other angiographic information. To reveal the effects of the carving method on the angiographic volume embolization ratio and the consequent recanalization rate after coil embolization, a future prospective randomized study is necessary between the groups of aneurysmal embolization with and without the use of the curving method.

Conclusion

We have devised a carving method for confirming the boundary lines between a cerebral aneurysm and its parent artery on both the front and back surfaces to obtain an optimal working projection using 3D VRDSA images. In 83 patients with aneurysms undergoing coil embolization, we retrospectively compared optimal working projections created using the carving method and the actual working projections used in the procedures. In only 38.6% of the 83 aneurysms was the actual working projection consistent with the optimal working projection. The carving method was considered useful to determine a working projection, particularly for the embolization of aneurysms other than terminal type and/or aneurysms with a maximal diameter <5 mm.

References

Footnotes

  • Competing interests None.

  • Patient consent Obtained.

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