|Home | About | Journals | Submit | Contact Us | Français|
Estimation of the stability of dysmorphic fusiform aneurysms of the intra-cranial internal carotid artery requires precise monitoring of their volumes. In this report we apply a method using MRI and 3D post-processing to study the evolution of these aneurysms on a prospective cohort of patients not immediately suitable for surgery or endovascular treatment.
Ten patients with fusiform aneurysms of the intra-cranial internal carotid artery underwent serial MRI studies. Five patients were studied at two time points and the remainder at multiple time points (mean delay between studies: 12.6 +/− 3.8 months). For each patient, studies from all time points were co-registered. Volumes of each vessel component were calculated.
Mean aneurysm volume was 833 +/− 878 mm3. Mean annual rate of volume progression was 1.37 +/− 2.09 % per year. All the aneurysms were thrombus-free.
This study indicates that, given the relatively low rate of progression of these dysplastic fusiform aneurysms and the complexity of their shape, 3D quantitative volumetric methods can be helpful in monitoring whether any growth has occurred.
Determining the most suitable management for patients with enlargement of segments of the distal internal carotid artery is challenging, particularly when that enlargement does not take the form of a saccular aneurysm. Indeed, treatment options that work well for saccular aneurysms, such as coiling or clipping, are generally not appropriate for these dysplastic vessels. Other options, such as bypassing the involved segment, are often challenging, and have variable success (1). Thus it is very important to evaluate the risk of complication (mainly rupture) of these aneurysms. Information on the natural history of these lesions is limited and several criteria have been proposed as predictors of rupture risk including: size, history of subarachnoid hemorrhage, location (larger risk for posterior circulation aneurysms), hemodynamic parameters and etiology (2–7). In cases where the benefit to risk ratio is so low that surgery or endovascular treatment is not planned, it is important to monitor aneurysm size evolution as aneurysm diameter is a major predictor of rupture risk that can change across time (6). Determination of whether an individual has a stable condition or one that is rapidly progressing can be a major factor in deciding whether to perform an intervention or not. In general clinical routine, monitoring the progression of vascular disease over time is a relatively imprecise art. Progression is typically determined by measuring linear dimensions on images of the flow lumen (preferably along three orthogonal axes). If projection catheter-angiography is used, it is virtually impossible to assure that the views acquired at follow-up are in the same orientation as those acquired at baseline. Imaging modalities that acquire 3 dimensional data (3D rotational angiography, MDCTA, or MRA) provide the possibility to extract multiple views of the vessels of interest in post processing (typically as maximum intensity projections or multi-planar reformations). In all cases, it is also impossible to assure that three linear measurements accurately reflect the global changes in lumen shape and size that might have taken place over time. We have previously reported on a quantitative method for measuring morphology change by determining lumenal volume changes from co-registered 3D volume sets acquired at different time points (8, 9). That method obviates the need for specifying optimal view planes or for achieving equivalent view angles at different time points.
The aim of the present study is to report on the serial monitoring of patients with fusiform, dysmorphic disease of the distal internal carotid artery, and on the use of this volumetric approach to accurately determine whether the vascular morphology remains stable over time.
After IRB approval and informed consent, ten patients (aged from 12 to 81, mean 46 +/− 23 years) were included in this prospective study. All patients presented with known aneurysms of the intracranial internal carotid system. Surgical treatments, such as clipping or coiling, could not be performed because of the unsuitable anatomy. Because other surgical options such as bypass were considered of high risk and unknown durability, a decision was made to follow these patients with careful clinical and radiologic follow-up with the intent to reevaluate options if there were a substantial evolution of the aneurysm size or worsening of symptoms. The radiological monitoring of all patients consisted of repeated MRI studies scheduled at approximately 12 months intervals. Patients were recruited between January 2004 and August 2008.
MR studies were performed on a 1.5-Tesla scanner (Intera®, Philips Medical Systems, Best, The Netherlands). A two-step imaging protocol was used. First, an “anatomical” sequence was performed to assess whether any intra-aneurysmal thrombus was present: 3D balanced Fast Field Echo (bFFE, TR=6 ms, TE=2 ms, flip angle=60°, number of averages=3, field of view=220 mm, matrix=256×256, slice thickness=1mm, image resolution=0.86×0.86×1 mm3, number of partitions=50, and acquisition time of about 4 minutes). A contrast-enhanced MR angiography (CE-MRA) was then performed using a 3D slab covering the vessels of interest with an injection of 18 ml GdDTPA followed by 15ml of saline all delivered at 2 ml/s. (The delay between injection and scan initiation was determined from injection of a 2 ml GdDTPA test bolus.) The CE-MRA sequence used elliptic-centric phase reordering, and data were acquired using parallel imaging with an acceleration factor of 2. Imaging parameters included: TR/TE/flip angle=5/2/30°. Images were acquired from a 54 mm para-coronal slab, with an FOV of 240 mm and an acquisition matrix of 400×380×45 zero-filled to 512×512×90. The resultant images had a resolution of 0.6×0.6×1.2 mm3 and were interpolated to 0.47×0.47×0.6 mm3. Total acquisition time was of the order of 35 s.
The analysis of the aneurysm lumen pursued here was similar to that reported previously(9). In brief, CE-MRA images were directly imported into a commercially available 3D data processing software package RapidForm® (INUS Technology, Seoul, South Korea) which allows creation of isosurfaces directly from a DICOM data set after the user chooses an intensity threshold. In order to ensure consistent volume quantitation, subsequent thresholding is constrained to produce identical measurements on reference healthy vessel segments that are considered to remain unchanged over time. In particular, once the first data set is processed, the threshold of subsequent studies is determined by requiring that the diameter of an undiseased part of a carotid artery remains constant over all subsequent studies of the same patient.
The analysis performed in this study compared changes in volume of the lumen of the dysmorphic segment over time. In order to ensure that volume measurements were made of the same vessel segments at all time points, the regions of interest were first manually co-registered using internal fiducial landmarks. A volume cut tool was then used to define the proximal and distal levels of the volumes of interest, ensuring an analysis that was sensitive to changes resulting from the disease process alone by confining the analysis to the involved segments. Figure 1 provides an example of an aneurysm before the segmentation. Figure 2 shows the registration and the cutting process.
Volumes of the aneurysm were reported for each time point. The rate of volume change was calculated for each patient.
We have previously noted that the deposition of intra-lumenal thrombus can be a concurrent process that might confound the analysis of lumenal volume changes. Two experienced observers (LB, DS) reviewed all MRI anatomical sequence images to confirm that there was no intra-aneurysmal thrombus in any of the segments under evaluation.
All statistical analysis was made with Intercooled Stata 9.1 (StataCorp LP, College Station, Texas, USA).
A total of 10 patients and 29 time points were studied. Each patient underwent a least two studies, while two patients had three studies, two patients four and one five. The mean delay between MRI examinations was 12.6 +/− 3.8 months (range 7–22 months). For each patient, co-registration of the 3D CE-MRA image data of the aneurysmal arterial segments was performed between the baseline data and data from all subsequent time points.
At baseline, mean volume of the aneurysms was 833 +/− 878 mm3 (range 47–2877 mm3). Mean annual rate of volume progression was 1.37 +/− 2.09 % per year (range −1.53 – 4.98). Graphical representation and numerical analysis of the evolution of the volumes of the aneurysms are shown in Figure 3 and Table 1.
All the aneurysms were thrombus-free.
Monitoring the evolution of intra-cranial aneurysms that are initially deemed unsuited for treatment can provide important information that can be used in deciding the time point when endovascular or surgical intervention might be necessary. In this study we report how 3D modeling of intra-cranial aneurysms can accurately be used to monitor changes in aneurysm volume on a prospective cohort of aneurysms of the intracranial internal carotid system.
Our study indicates that the rate of aneurysm progression is individual specific but remains relatively low for this cohort of fusiform, dysmorphic aneurysms of the distal internal carotid artery. Indeed, the highest rate of growth we observed is 4.98 % which is a significant change relative to the accuracy of our method (2 +/− 1% (7)) but still relatively low. This is in agreement with previous knowledge about the natural history of aneurysm growth (6) and supports the fact that precise 3D tools are required to monitor these aneurysms.
Indeed 3D volume monitoring presents several advantages in comparison to traditional maximal diameter measurements. First, using a 3D approach accounts for changes in patient position and image acquisition direction at different time points (10). Further, 3D measurements are often more accurate in the detection of structural growth, particularly if that growth is asymmetric (11, 12). Finally, co-registration of datasets acquired at different time points allows a good overall appreciation of global and local evolution of the aneurysms (Figure 4). This feature is particularly important in the case of long dysmorphic segments with several bulging areas such as in Figure 5. In this case, choosing one diameter as a reference for the follow-up can be very misleading and lead to misinterpretation if the growth occurs elsewhere on the pathologic vessel.
A limitation to our study is the small size of our population. Nevertheless, this study does not intend to provide definitive guidelines for the frequency with which patients should be monitored or to assign change in diameter as the only factor governing rupture risk (as many other cofactors may contribute as decribed in the introduction section) but to illustrate the potential advantages in using a 3D volumetric modeling approach for intra-cranial aneurysm rate of progression.
In conclusion, given the relatively low rate of progression, this study would indicate that patients with dysplastic, fusiform aneurysmal vessel segments of the internal carotid artery that are not saccular in form, should be followed with serial imaging studies, and that quantitative volumetric methods can be helpful in monitoring whether any growth has occurred.
Loic Boussel acknowledges support from the Hospices Civils de Lyon, Hopital Louis Pradel, 69500 Bron, France.
This work has been supported in part by grant NS059944 (DS) and K25NS059891 (VR) from the NIH, and a VA MERIT Review Grant (DS).
No conflict of interest or financial disclosure.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.