Abstract
Keywords
Introduction
Endovascular aneurysm sealing (EVAS) is a new technique to treat patients with abdominal aortic aneurysms (AAA). Differing in principle from open or endovascular aneurysm repair,1,2 EVAS employs dual endobags filled with a biocompatible polymer (polyethylene glycol diacrylate hydrogel) surrounding two 10-mm balloon-expandable covered stents that preserve flow into each common iliac artery. The endobags expand to completely fill the aortic lumen to provide stability to the seal.
Endovascular aneurysm sealing is performed with the Nellix system (Endologix Inc, Irvine, CA, USA). Case selection and procedure planning rely on measurements of the aortic lumen, as the Nellix is limited by the volume of the endobags to treating AAAs with 6-cm-diameter flow lumens. 3 Furthermore, adequate quantities of the polymer must be thawed before the procedure, so the quantity of polymer to be used must be estimated based on the preoperative flow lumen measurements. The utility of these measurements to predict the polymer volume relies on the assumption that the aneurysm and its thrombus do not change between the planning computed tomography (CT) scan and the procedure. It also assumes that endobag inflation does not affect the lumen volume; however, previous research from our group documented potential changes in the volume of aortic thrombus post EVAS. 4 The aim of this study was to quantify and further characterize these changes and their potential association with any variation in the density of aortic thrombus.
Methods
Study Sample
The 34 consecutive EVAS procedures performed in our unit between December 2013 and December 2014 were retrospectively reviewed to identify patients whose aneurysms contained at least 5 mm of thrombus at any point within the aneurysm on cross section and who had a postoperative CT scan available. Two patients undergoing EVAS to reline previously placed endografts were excluded. Of the 32 patients remaining, 6 had no thrombus and another did not have a postoperative CT, leaving 25 patients (mean age 78±7 years; 17 men) available for the study. The cohort had a mean preoperative maximum aortic diameter of 59±3.6 mm.
Endovascular Aneurysm Sealing Procedure
The EVAS technique has been described elsewhere.1,5 In brief, 2 suitable Nellix catheters are positioned in the infrarenal aorta and their respective stents inflated simultaneously to a pressure of 7 atm. With a large syringe, air is aspirated from the endobags, which are then simultaneously filled with a test saline solution up to a pressure of 180 mm Hg. The saline solution is removed from the endobags, its volume precisely measured, and the verified quantity of polymer is injected by means of a dispenser in 3-mL increments until the desired 180 mm Hg pressure is achieved. After the polymer fill, the stent balloons are reinflated to a pressure of 10 atm and left inflated until the polymer cures. The balloons are deflated, the catheters removed, and, if deemed necessary, the stents are postdilated with shorter (typically 10×60 or 10×40 mm) balloons at 12 atm.
Computed Tomography Measurements
All CT measurements were made by 2 experienced radiologists (U.S. and T.Y.C.) using Carestream software (version 11.4.1.1011; Carestream Health Inc, Rochester, NY, USA). Different volume measuring techniques were initially tested on images of a phantom of known volume acquired with a 64-slice Siemens Somatom scanner (Siemens Healthcare, Frimley, UK). The original CT slice thickness was 0.625 mm, but the multiplanar reconstructed images were rebuilt to 2 mm with reconstruction intervals of 2 mm.
After this initial test, a protocol was composed and tested on 10 scans (5 preoperative and 5 postoperative); the threshold for acceptance of the technique was an interobserver variability ±5%. For these measurements, pre- and postoperative cross sections were registered on 2 parallel screens using the Carestream “slice alignment tool” so that each measurement was taken exactly at the same level on the 2 scans. For purposes of uniformity, the most proximal measurement was taken 1 cm below the uppermost level of the stents on the post-EVAS CT scan and the most distal measurement at the aortic bifurcation. A further check was performed by ensuring that the scans were aligned at the level of the renal arteries. The Carestream “lesion livewire segmentation tool” was used to measure, at each level, the cross-sectional areas of the aneurysm (on pre- and postoperative scans), the flow lumen (on preoperative scans), and the volume of the endobags (on postoperative scans). At each cross section, this tool generates a semiautomated area, whose contour was then manually adjusted to correspond to the inner aortic wall or the inner thrombus outline. The software automatically generated aortic, flow lumen, and endobag volumes between the most proximal and most distal cross section. Thrombus volumes were calculated by subtracting the lumen volume or the volume of the endobags from the aortic volume. During measurements, each observer was allowed to individually control the CT window levels as this would represent standard clinical practice. The CT attenuation of the thrombus at 5 randomly selected points along the thrombus on the pre- and postoperative scans was sampled, and the mean value in each scan was used to define thrombus density.
Data Collection and Outcome Measures
In addition to CT measurements, variables collected for this study included the time intervals between EVAS and the 2 CT scans, age and sex of the patients, and maximum aortic diameter on CT. For analysis, the time interval between the preoperative CT and EVAS was used because we assumed that further changes in volume would not occur in the interval between EVAS and our routine 1-month post-EVAS CT.
The primary outcome measures were the changes in thrombus and aortic volume between pre- and postoperative scans. Secondary outcome measures were the relationship between changes in volumes and time intervals between preoperative CT and EVAS (CT-to-EVAS time delay) and the change in thrombus density [expressed in Hounsfield units (HU)] between pre- and postoperative scans.
Data Analysis
All distributions were assessed for normality with the Kolmogorov-Smirnov test. Normally distributed variables (the majority) were described with the mean ± standard deviation; skewed variables were described with the median and interquartile range (IQR). Differences between pre- and postoperative volumes were reported as means and 95% confidence intervals (CIs) according to the underlying distributions. Correlations were expressed with the Pearson r coefficient for normally distributed continuous variables. Statistical significance was set at the conventional 5% level. Statistical analysis was performed with SPSS software (version 22.0; SPSS Inc, IBM Corporation, Somers, NY, USA).
Results
Endovascular aneurysm sealing was performed a mean 118±65 days (range 19–271) after the planning CT. Postoperative CTs were performed a median 31 days (95% CI 30 to 37) after EVAS. There was little interobserver variability on the initial test set of 10 CT scan measurements, with a mean bias between operators of −0.6% (range −3% to 2%) for preoperative aortic volumes, −1% (range −3% to 1%) for preoperative lumen volumes, −2% (range −4% to −1%) for postoperative aortic volumes, and −0.9% (range −4% to 1%) for endobag volumes.
Volume measurements and the differences between post- and preoperative CTs are reported in Table 1. Post EVAS, all but one aneurysm had an increase in volume (mean change 17 mL, 95% CI 10.0 to 23.5, p<0.001), and in all cases the volume of the endobags was greater than the preoperative lumen measurements. Aortic thrombus volume (Figure 1) decreased in 20 patients, remained unchanged in one, and increased in 4 (mean change −11 mL, 95% CI −4.7 to −18.2, p<0.001). Of the 4 patients in whom thrombus volume increased, only 1 patient displayed a substantial change (34 vs <5 mL in the other 3 cases). This patient, whose aneurysm had also increased in volume by 46 mL, had the longest interval between the planning CT and EVAS at 271 days.
Volume Measurements.
Abbreviations: CI, confidence interval, EVAS, endovascular aneurysm sealing.
Median (interquartile range) or mean ± standard deviation.
Difference between the volume of endobags on postoperative CT and the lumen volume on preoperative CT.
Axial computed tomography (CT) scans in a patient undergoing endovascular aneurysm sealing (EVAS) 127 days after the planning CT. (A) The aneurysm is largely filled with thrombus on the preoperative scan; (B) postoperatively, the endobags have partially replaced the thrombus. Total thrombus volume decreased from 145 to 127 mL despite a slight increase in aortic volume from 203 to 209 mL.
The median thrombus density was 41 HU (IQR 34–49). Thrombus density decreased in 10 patients and increased in 15; the median overall change was 3 HU (IQR −3.3 to 8.9, p=0.957).
There were good correlations between changes in aneurysm and thrombus volumes (r=0.864, p<0.001; Figure 2); between the planning CT/EVAS time interval and the change in aneurysm volume (r=0.640, p=0.001; Figure 3A), and between the planning CT/EVAS time interval and the change in thrombus volume (r=0.567, p=0.003; Figure 3B).
Relationship between changes in aneurysm volume and changes in thrombus volume. Although there was a strong linear correlation, aneurysm volume increased and thrombus volume decreased in most patients.
Relationship between the interval from the planning computed tomography (CT) to the endovascular aneurysm sealing (EVAS) procedure and changes in (A) aneurysm volume and (B) thrombus volume. Smaller aneurysm growth and greater thrombus shrinkage were seen in patients with short delays between CT and EVAS.
Discussion
This study demonstrated that there are significant changes in aortic volumes between preoperative and post-EVAS CT scans, which may impact EVAS planning and performance. In our study, most aneurysms increased in size between the preoperative and postoperative CT scans, and in 80% of the patients, the volume of aortic thrombus decreased. The change in thrombus volume appeared to have a linear relationship with the time elapsed between the planning CT and EVAS, with patients undergoing EVAS soon after the planning CT displaying more pronounced thrombus shrinkage.
Our research group had already reported post-EVAS reduction in the thickness of aortic thrombus on cross sections 4 ; however, to our knowledge, no one has performed a careful volumetric assessment of aneurysms treated by EVAS. Our findings have implications for the planning and the performance of EVAS: in particular, they suggest that the volume of polymer necessary for sealing can be estimated (rather than precisely measured) only on a planning CT. The Nellix IFU include a “saline prefill” of the endobags prior to polymer injection in order to measure volume precisely. This step was described as optional in the original instructions for use (IFU), 3 but our findings suggest that it should be considered mandatory. Indeed, the manufacturer has recently amended the IFU to make the prefill mandatory.
The causes of the changes in aortic volumes are not known and may be multiple. For example, aortic diameter changes during the cardiac cycle and inflation of the endobags to above-systolic pressure should frame the aneurysm at its maximum (systolic) volume, while the CT is likely to scan the aneurysm during diastole, which encompasses much of the cardiac cycle. 6 Because of this effect, one would expect larger aortic volumes on postoperative CTs. Interval changes between the planning CT and surgery may also play a role, as aneurysms grow over time and thrombus, a dynamic structure, may form or resorb. Intuitively, however, it is the pressure exerted by the endobags on the thrombus that may be responsible for its shrinkage, either by compressing it or, potentially, by “embolizing” some of its components into side branches. Pressure transmission to the aortic wall may also contribute to aortic enlargement and, in rare circumstances, rupture. 4
Interestingly, changes in thrombus, but not aortic volume, were also related to CT-to-EVAS times, with patients undergoing EVAS shortly after the planning CT generally showing larger decreases in thrombus volume. We expected to detect some increased density of the thrombus of patients with thrombus shrinkage; however, in our study, there was no difference between the pre- and post-EVAS CTs, suggesting either little impact of EVAS on this variable or poor sensitivity of CT to detect subtle changes in the composition of thrombus.
Limitations
There are several limitations to our study, including the small sample size, the retrospective design, the potential error related to our method of measurement,7,8 and the limitation of CT in producing aortic images of sufficient quality to allow accurate measurements. Regarding the latter, CT produces static images of the aorta, which is known to change during the cardiac cycle, reaching maximum diameter in systole. As much of the cardiac cycle is spent in diastole, most CT images are acquired when the aortic diameter is smallest; it is not inconceivable, with modern fast scanners, that the whole aneurysm may be imaged in this phase.
Aortic thrombus is also a dynamic entity and has been shown to change with the cardiac cycle. 6 CT scan definition of the lumen-thrombus interface depends on the timing of image acquisition with the contrast bolus: poor definition of this interface may lead to measurement errors. Because of the retrospective nature of our study, preoperative CT protocols were not standardized: as our unit is a tertiary referral center, the CTs reviewed in this study were performed in 7 different hospitals, although all scans were judged of sufficient quality for planning purposes.
A further limitation of our findings is the long CT-to-EVAS time in some patients, as those with short delays were more likely to display thrombus shrinkage. While in some cases the delay in treatment was justified by patient-related factors, our findings may not be replicated in other settings where diagnosis-to-treatment times are less variable.
Conclusion
Endovascular aneurysm sealing is a potentially revolutionary treatment for AAA patients, and its planning requires measurement of aortic volumes. Prior to EVAS, these measurements were not necessary for any type of aortic intervention, thus their importance has been overlooked by the literature. Hopefully, our findings will encourage other researchers to investigate time-related thrombus remodeling and the effect of EVAS on the aneurysm and its thrombus. In the meantime, we suggest that pre-EVAS volume measurements should be considered only an estimate of intraoperative polymer requirements.
Footnotes
Declaration of Conflicting Interests
The author(s) declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: Francesco Torella and Robert K. Fisher have received remuneration by Endologix for proctorship. Francesco Torella and Tze Y. Chan have received educational support from Endologix. Francesco Torella, Robert K Fisher, Richard G. McWilliams, and Andrew England have received research funding from Endologix.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.