3D rotational angiography with volume rendering: The utility in the detection of intracranial aneurysms
Correspondence Address: Source of Support: The National Natural Scientific Fund of
China, Shanghai Important Subject Fund of Medicine and
Program for Shanghai Outstanding Medical Academic Leader, Conflict of Interest: None DOI: 10.4103/0028-3886.73743
Source of Support: The National Natural Scientific Fund of China, Shanghai Important Subject Fund of Medicine and Program for Shanghai Outstanding Medical Academic Leader, Conflict of Interest: None
Aims: The advent of three-dimensional (3D) rotational angiography (3D DSA) challenged the role of digital subtraction angiography (DSA) as a "gold standard" in the diagnosis of intracranial aneurysms. In this study, we report our experiences in diagnosing intracranial aneurysms by using 3D DSA with volume rendering (VR) technique, particularly focusing on its role in depicting additional aneurysms missed by 2D DSA. Materials and Methods: One hundred and thirty-eight consecutive patients with known or suspected aneurysms (54 men, 84 women; median age, 55 years; age range, 18-83 years) underwent both conventional DSA and 3D DSA with VR examination simultaneously. The images of 2D DSA or 3D DSA with VR were evaluated by two observers independently for the presence of aneurysms. Then additional aneurysms were decided and depicted. Results: 3D DSA with VR showed 146 aneurysms in 123 (89.1%) of 138 patients and no aneurysms in 15 patients. 2D DSA showed 115 aneurysms in 110 of 137 patients, with one aneurysm in 106 patients each, 2 in 3 patients each and 3 in 1 patient. After reaching a consensus, there were 31 additional aneurysms detected by 3D DSA with VR. 30 aneurysms were <3 mm in maximum diameter with 3 aneurysms ruptured. These additional aneurysms were located in internal carotid artery (ICA, n = 28, 90.32%), anterior cerebral artery (ACA, n = 3, 9.68%). No additional aneurysms were found in either middle cerebral artery (MCA) or vertebrobasilar and posterior cerebral artery (PCA) systems. Conclusions: 3D DSA, especially VR images, not only clearly reveals aneurysms and aneurysmal morphology, but also detects additional aneurysms missed by 2D DSA, especially small aneurysms less than 3 mm.
Keywords: Diagnosis, digital subtraction angiography, intracranial aneurysms, volume rendering, three-dimensional rotational angiography
Catheter digital subtraction angiography (DSA) has been the accepted "gold standard" for the diagnosis of intracranial aneurysms due to its excellent capacity to display vasculature without any disturbance of background. , However, additional small angiographic occult aneurysms are commonly found during surgery of symptomatic aneurysms with an incidence of 3.7-12.2%. ,,, Since even small aneurysms of 2 mm or smaller can rupture and that the appropriate choice of either neurosurgical or endovascular treatment is based on the optimal depiction of as many anatomic details of the aneurysm as possible, the role of two-dimensional (2D) rotational angiography (2D DSA) in providing this information is obviously being challenged. Recently, three-dimensional (3D) rotational angiography (3D DSA) has been considered to be superior to 2D DSA in the detection of intracranial aneurysms ,,,, and enabled clinicians to review intracranial aneurysms and other vascular lesions thoroughly in an angiography study or during an interventional procedure. 3D DSA requires substantially fewer projections and thus reduces radiation dose and volume of contrast material to the patient, which may result in a shorter procedure time and fewer risks and complications for the patient. 
The reconstruction techniques for 3D DSA include maximum intensity projection (MIP), shaded surface display (SSD), and volume rendering (VR). Volume rendering seems to be theoretically superior to MIP or SSD because VR technique incorporates entire DSA data into 3D images, whereas MIP and SSD images just use a small fraction of data.  It is widely accepted that VR has been the favored reconstruction procedure used in spiral CT, ,, and we believe that the case is true in DSA as well. To our knowledge, there have been few specialized reports on 3D DSA by VR for detecting and evaluating intracranial aneurysms. ,, So the purpose of this study was to report our experiences of detecting and visualizing intracranial aneurysms with 3D DSA with VR.
The study was conducted in one regional neuroradiological center, Institute of Diagnostic and Interventional Radiology, The Sixth Affiliated People's Hospital, Shanghai Jiao Tong University. In this specialized care center, we conducted a retrospective clinical study to evaluate the utility of 3D DSA with VR in the detection of and characterization of intracranial aneurysms. The institutional review board reviewed and approved the study protocol. Patients or immediate family members (if the patient's clinical status precluded him or her from granting consent) provided informed consent for study participation.
From June 2007 to October 2008, 138 consecutive patients with known or suspected aneurysms (54 men, 84 women; median age, 55 years; age range, 18-83 years) underwent DSA examination during a 16-month period, and 3D DSA with VR, as a reference of golden diagnostic standard, was performed simultaneously in all patients. This study cohort consisted of 89 patients with known or suspected aneurysms who underwent intracranial aneurysmal epidemic investigation, 24 patients with symptoms that might be due to aneurysm or variety of intracranial diseases, and 25 patients with proved subarachnoid hemorrhage (SAH) by computerized tomography (CT), known or suspected aneurysms detected by magnetic resonance imaging (MRI).
An interventional neuroradiologist performed DSA examinations. Conventional 2D DSA was performed on a monoplanar digital angiography unit (Axiom Artis VB22N, Siemens Medical Systems) with a 1024 Χ 1024 matrix and a 17- to 20-cm FOV. The contrast medium was injected at a flow rate of 4-5 mL/s and 2-3 mL/s in two projections, resulting in a total of 10 mL for internal carotid artery (ICA) and 7 mL for vertebral artery. Rotational angiography was performed with an 8-s 200░ rotational run, with acquisition of 200 images and with injection of 3- to 4-mL contrast material per second resulting in a total of 16-20 mL for each artery. The image data were transferred to a workstation (syngoXWP VA70B; Siemens Medical Systems), where reconstruction of 3D images was performed with a software package.
In our practice, in patients with suggestion of intracranial aneurysms in any of the vascular territory, a complete 2D DSA and 3D DSA of the affected artery and the intact ipsilateral artery were performed with 2-4 projections. 2D DSA was only performed for the remaining one artery. So, a complete DSA at least consisted of three-vessel 2D DSA and 2-vessel 3D DSA for each patient. Thus, patients with complete 2-vessel 3D DSA with VR were included in this study. On the contrary, patients who lacked 3D DSA with VR were excluded from this study.
All images of 2D DSA and 3D DSA with VR were independently evaluated by two experienced neurointerventional radiologists with 16-20 years of experience in their specialties. 2D DSA images were analyzed separately from those of 3D DSA with VR; anonymous 3D DSA images were given in random order to the readers 8 weeks after each reader completed the analysis of 2D DSA images. Both readers were blind to all clinical information, including the location of the SAH. For analysis, DSA images were presented on film, and 3D DSA images were presented on the monitor screen of the workstation. All DSA series of the vascular territory covered by 3D DSA with VR (basilar artery or left or right ICA) were presented. To optimize the display, we adjusted the threshold and window levels, selected contiguous objects, and cut overlying structures. For analysis purposes, the presence of aneurysm either on 2D DSA or on 3D DSA images was assessed by using a five-point scale of observer's confidence: 1, definitely not present; 2, probably not present, 3, equivocal; 4, probably present; and 5, definitely present. "Probably present" or "definitely present" were considered positive for aneurysm; all others were considered negative for aneurysm. Interobserver disagreements on either 2D DSA or 3D DSA with VR were resolved by consensus. Interobserver agreement between both readers was determined by calculating κ values (poor agreement, κ=0; slight agreement, κ=0.01-0.20; fair agreement, κ=0.21-0.40; moderate agreement, κ=0.41-0.60; good agreement, κ=0.61-0.80; and excellent agreement, κ= 0.81-1.00).
No procedure-related morbidity or mortality occurred during 3D DSA as well as 2D DSA examination, and the examination was well tolerated by all 138 patients. In 16 (10.8%) patients examined without general anesthesia, complaints centered on the automated administration of contrast medium, which caused a prolonged feeling of heat.
3D DSA with VR
All 3D DSA with VR images in our study were interpretable. Interobserver agreement was considered good and significant (agreement in 94% (130/138) of cases and κ=0.79 ± 0.07). After reaching a consensus, 3D DSA with VR showed 146 aneurysms in 123 (89.1%) of 138 patients and no cerebral aneurysm in 15 patients. The aneurysms were located in ICA (n = 101, 69.2%), middle cerebral artery (MCA) circulation (n = 12, 8.2%), anterior cerebral artery (ACA) circulation (n = 27, 18.5%), vertebrobasilar and posterior cerebral artery (PCA) system (n = 6, 4.1%). Twenty-five aneurysms were ruptured (with evidence of CT-proven SAH) and 121 were unruptured. The mean size of all aneurysms was 3.66 ± 3.06 mm (range 1 to 34 mm) in maximum diameter. Of the 146 aneurysms, 67 were <3 mm, 73 were 3 to 10 mm and 6 were >10 mm in maximum diameter. Four patients had 3 aneurysms each, 15 patients had 2 aneurysms each, and 104 patients had 1 aneurysm each [Table 1].
2D DSA angiograms were interpretable in all but one case. Artifacts were considered major in one case: vessel overlapping was too great for aneurysm assessment. This DSA examination was considered as unreadable. Two patients, diagnosed as harboring 2 aneurysms in C7 segment of ICA by 2D DSA, were proved overlapping and torture of vasculature by 3D DSA with VR, so these patients were considered as negative finally. Interobserver agreement was considered excellent and significant for DSA (agreement in 93% (128/137) of cases and κ = 0.82 ± 0.06). After reaching a consensus, 2D DSA showed 115 aneurysms in 110 of 137 patients, with one aneurysm in 106 patients each, 2 in 3 patients each and 3 in 1 patient. The aneurysms were located in the ICA (n = 73, 63.5%), MCA circulation (n = 12, 10.4%), anterior cerebral artery (ACA) circulation (n = 24, 20.9%), vertebrobasilar and PCA system (n = 6, 5.2%) with a mean size of 4.04 ± 3.33 mm (range 1.73 to 34 mm) in maximum diameter.
Additional aneurysms detected by 3D DSA with VR
Therefore, after reaching a consensus, there were 31 additional aneurysms detected by 3D DSA with VR. [Figure 1] and [Figure 2] The mean size of additional aneurysms was 2.20 ± 0.43 mm (range 1.0 to 3.1 mm). Of these additional aneurysms, 30 aneurysms were <3 mm in maximum diameter, with 3 aneurysms ruptured. These additional aneurysms were located in ICA (n=28, 90.32%) [Figure 2]c and ACA circulation (n = 3, 9.68%)[Figure 1]c. No additional aneurysms were found in either MCA circulation or vertebrobasilar and PCA system.
In this single institute study of patients with suspected or known intracranial aneurysms comparing 3D DSA with VR with 2D DSA, it was found that 3D DSA with VR not only has a reliability for depicting the morphology and volume of aneurysms, but also can detect additional aneurysms missed by 2D DSA, especially small aneurysms less than 3 mm in maximum diameter. Since some of these additional aneurysms can be ruptured at any time point, patients with SAH and negative findings in 2D DSA need further examination by 3D DSA. , In the present study, three additional aneurysms were ruptured and treated with coils. It was also found that 3D DSA with VR can depict more accurately the direction of aneurysms neck than 2D DSA, which is vital for interventional neuroradiologists to determine the appropriate working projection angle for the task of embolization  and to reshape the tip of microcatheter so that the tip of microcatheter can be advanced into aneurysm sac easily and reliably. In our opinion, there are some factors attributed to the excellent ability of 3D DSA with VR to visualize intracranial aneurysms. 3D DSA with VR allows arbitrary angles necessary to check aneurysms, and it allows avoidance of overlapped cerebral vessels. If required, overlapped vessels can be deleted from the images to allow more accurate definition of the aneurysms.
As reported in the literature, CT angiography (CTA) is the most frequently used non-invasive diagnostic tool for the detection of ruptured intracranial aneurysms in the acute setting. Combining the techniques such as SSD, MIP and source images, it makes the diagnosis more accurate than 2D DSA.  However, detection of intracranial aneurysms by CTA is limited because axial source section evaluation is tedious and 3D visualization is hampered by overprojecting bone, especially in the region of the skull base. , Although there are several methods to remove bone, CTA has limited sensitivity in detecting very small aneurysms.  Some reports have also demonstrated that 3D DSA is superior to 2D DSA in detecting intracranial aneurysms, especially those smaller than 2 mm; however, the techniques used for 3D DSA in these studies were typically MIP and/or SSD. ,, MIP and SSD have some potential disadvantages due to inappropriate reconstruction threshold, which would make the evaluation of intracranial aneurysms more or less inaccurate.  In our opinion, 3D DSA with VR theoretically can overcome these disadvantages because the VR technique incorporates the entire DSA data into 3D images, whereas MIP and SSD images just use a small fraction of data.  During the reconstruction process, the volume-rendered 3D DSA is not influenced by the threshold because it is volume determined. We can adjust brightness of screen to obtain the best quality 3D DSA images; however, the process is individually determined. Inappropriate adjustment may filter some tiny vasculatures which can be compensated by repeated adjustment of the brightness of screen. Some authors agreed on the role of 3D DSA in detecting additional aneurysms missed by 2D DSA and in evaluating aneurysmal morphology, its relationship with neighboring arteries and residual aneurysms postoperatively, but they did not describe their 3D reconstruction techniques, with no mention of the role of 3D DSA with VR in an analysis of intracranial aneurysms. ,, To our best knowledge, there are only a few studies concerning 3D DSA with VR. ,, Thines et al., who were from the surgeon's view, showed 3D DSA, especially with VR technique, was useful for analysis of the arterial relationships and projection of intracranial aneurysms and, therefore, might help to predict the pitfalls and risks of surgery.  Kawashima et al. argued that 3D DSA, especially VR images, had greater advantage in depicting ruptured intracranial aneurysms than 2D DSA.  Our results demonstrated that 3D DSA with VR has great advantages in detecting additional aneurysms missed by 2D DSA, and because of its excellent ability of depicting spatial relationship between aneurysms and neighboring arteries, it has been useful for interventionalists to determine the therapeutic protocol and the working projection angle.
Some potential disadvantages of 3D DSA should be mentioned. Firstly, the accuracy of 3D DSA in depicting the volumes of aneurysm remains a controversial issue. Jou et al. in his study reported two basilar aneurysms whose volumes were underestimated by 3D DSA and concluded that pulsation and gravity were two possible causes of aneurysm under estimation.  Conversely, another study conducted by Kawashima. et al. confirmed that the 3D DSA, especially VR images, tend to depict the volumes of aneurysms more accurately than 2D DSA.  Secondly, 3D DSA has not demonstrated superiority to 2D DSA in determining one of the most clinically important geometric factors of an aneurysm, dome-to-neck ratio.  Finally, compared to 2D DSA, 3D DSA is relatively time consuming, for example, the average time for 3D reconstruction in the present study is no less than 7 min from transfer of the data to workstation to availability of 3D images.
Several limitations of the present study should be noticed. Firstly, our study is a retrospective study with a non-randomized trial, which makes it difficult to confirm the real value of 3D DSA with VR in diagnosis. Although the diagnostic performance of 3D DSA with VR looks more promising than 2D DSA and theoretically stronger than 3D DSA with SSD or MIP, further studies with randomized and controlled trials are needed to confirm and extend our observations. Secondly, some additional aneurysms detected by 3D DSA with VR have not been treated with either endovascular coiling or surgical clipping, so their existence needs further confirmation. However, our follow-up for these untreated aneurysms with both MRA and DSA at a repeated 3-6 month interval has confirmed their existence. Finally, because our aim was to confirm that 3D DSA with VR is superior to 2D DSA, we did not compare 3D DSA with VR with either MRA or CTA, which is still under studying in our center.
This study has been supported by the National Natural Scientific Fund of China (Contract number: 30570540), Shanghai Important Subject Fund of Medicine (Contract number: 05 III 023) and Program for Shanghai Outstanding Medical Academic Leader (Contract number: LJ 06016).
[Figure 1], [Figure 2]