Article Text

Download PDFPDF

Original research
Increased contrast enhancement of the parent vessel of unruptured intracranial aneurysms in 7T MR imaging
  1. Edgar A Samaniego1,
  2. Jorge A Roa2,
  3. Honghai Zhang3,
  4. Timothy R Koscik4,
  5. Santiago Ortega-Gutierrez1,
  6. Girish Bathla5,
  7. Milan Sonka3,
  8. Colin Derdeyn6,
  9. Vincent A Magnotta5,
  10. David Hasan7
  1. 1 Interventional Neuroradiology/Endovascular Neurosurgery Division Department of Neurology, Neurosurgery and Radiology, The University of Iowa Hospitals and Clinics, Iowa City, Iowa, USA
  2. 2 Department of Neurology and Neurosurgery, University of Iowa Hospitals and Clinics, Iowa City, Iowa, USA
  3. 3 Department of Electrical and Computer Engineering, Iowa Institute of Biomedical Imaging, University of Iowa Hospitals and Clinics, Iowa City, Iowa, USA
  4. 4 Department of Psychiatry, University of Iowa Hospitals and Clinics, Iowa City, Iowa, USA
  5. 5 Radiology, University of Iowa Hospitals and Clinics, Iowa City, Iowa, USA
  6. 6 Radiology and Interventional Radiology, University of Iowa Hospitals and Clinics, Iowa City, Iowa, USA
  7. 7 Neurological Surgery, The University of Iowa Hospitals and Clinics, Iowa City, Iowa, USA
  1. Correspondence to Dr Edgar A Samaniego, Interventional Neuroradiology/Endovascular Neurosurgery Division Department of Neurology, Neurosurgery and Radiology, The University of Iowa Hospitals and Clinics, Iowa City 52246, Iowa, USA; edgarsama{at}gmail.com

Abstract

Background Inflammation of the arterial wall may lead to aneurysm formation. The presence of aneurysm enhancement on high-resolution vessel wall imaging (HR-VWI) is a marker of wall inflammation and instability. We aim to determine if there is any association between increased contrast enhancement in the aneurysmal wall and its parent artery.

Methods Patients with unruptured intracranial aneurysms (UIAs) prospectively underwent 7T HR-VWI. Regions of interest were selected manually and with a semi-automated protocol based on gradient algorithms of intensity patterns. Mean signal intensities in pre- and post-contrast T1-weighted sequences were adjusted to the enhancement of the pituitary stalk and then subtracted to objectively determine: circumferential aneurysmal wall enhancement (CAWE); parent vessel enhancement (PVE); and reference vessel enhancement (RVE). PVE was assessed over regions located 3- and 5 mm from the aneurysm’s neck. RVE was assessed in arteries located in a different vascular territory.

Results Twenty-five UIAs were analyzed. There was a significant moderate correlation between CAWE and 5 mm PVE (Pearson R=0.52, P=0.008), whereas no correlation was found between CAWE and RVE (Pearson R=0.20, P=0.33). A stronger correlation was found between CAWE and 3 mm PVE (Pearson R=0.78, P<0.001). Intra-class correlation analysis demonstrated good reliability between measurements obtained using semi-automated and manual segmentation (ICC coefficient=0.790, 95% CI 0.58 to 0.90).

Conclusion Parent arteries exhibit higher contrast enhancement in regions closer to the aneurysm’s neck, especially in aneurysms≥7 mm. A localized inflammatory/vasculopathic process in the wall of the parent artery may lead to aneurysm formation and growth.

  • aneurysm
  • artery
  • inflammation
  • MRI
  • vessel wall

Statistics from Altmetric.com

Introduction

Unruptured intracranial aneurysms (UIAs) are found in 3%–5% of the adult population worldwide.1 It is unclear what prompts the formation, growth, and rupture of cerebral aneurysms. Inflammation of the arterial wall may have a pivotal role in this process.2–5 Computational fluid dynamic (CFD) studies have proposed that hemodynamic insults lead to endothelial dysfunction and trigger complex inflammatory endothelial processes. Intravascular hemodynamic stress is composed of two separate entities: fluid force vectors and blood pressure, which fundamentally translate into perpendicular transmural pressure, circumferential cyclic stretch, and wall shear stress (WSS).6 Both transmural pressure and cyclic stretch are derived from blood pressure, whereas WSS is a tangential force exerted by flow vectors.7 The tortuous angioarchitecture of the circle of Willis exposes endothelial cells and the parent vessel wall to different flow profiles, which lead to a wide range of variation in the WSS exerted along the system. Many studies have shown that deviations from normal WSS contribute to various aspects of endothelial cell dysfunction in cerebral aneurysm pathogenesis,8 9 and may trigger an inflammatory cascade within the vessel wall that ultimately results in the formation of an aneurysm.

It is intuitive to expect changes in the parent vessel of brain aneurysms, since the neck of the aneurysm is a defect of the parent artery. CFD studies and histological analysis of ruptured aneurysms have demonstrated focal changes in the aneurysm wall that ultimately cause its rupture.10 11 High-resolution vessel wall imaging (HR-VWI) has been used to study the structure, thickness, and enhancement of the wall of brain aneurysms. It has suggested a correlation between increased enhancement and inflammation and, therefore, aneurysm instability.12 13 In this study, we aim to determine if similar changes can be observed in the parent arteries of UIAs, which may suggest that aneurysm formation underlies on inflammatory changes of the parent artery.

Methods

Study population and imaging acquisition

After Institutional Review Board approval, patients with UIAs prospectively underwent HR-VWI between June 2018 and October 2019. All UIAs were incidentally found using magnetic resonance angiography or CT angiography. At our institution, most patients with UIAs undergo a 3T HR-VWI scan. Patients with UIAs were screened to participate in our prospective 7T HR-VWI study (GE MR950 7T scan, GE Healthcare, Waukesha, WI). The imaging protocol included: 3D T1-weighted fast-spin-echo (CUBE); T2-weighted CUBE; 3D time-of-flight (TOF); and 3D susceptibility-weighted angiogram (SWAN). Technical parameters for acquisition of each sequence are described in the online supplementary table e–1. Gadobutrol (Gadavist, Bayer Pharmaceuticals, Whippany, NJ) was administered intravenously (0.1 mmol/kg), and a post-contrast 3D TI CUBE sequence was obtained after 5 min. Informed consent was required from each subject before enrollment in the study.

HR-VWI analysis: semi-automated segmentation protocol

Highly accurate 3D segmentation of UIAs was achieved following semi-automated and manual techniques. Semi-automated method: the center of the aneurysm sac and a location not-too-far outside of the aneurysm surface was manually identified (figure 1A,B). Then, the LOGISMOS graph search algorithm14 15 and minimally-interactive Just-Enough-Interaction (JEI) strategy16 were used to delineate the aneurysm’s lumen and wall surfaces, and to define a spherical region of interest (ROI) around it (figure 1C). LOGISMOS uses the spherical ROI mesh surface by considering intensity patterns sampled along the rays cast from the center of the sphere to mesh vertices and searching for desired positions of these mesh vertices to reside on the aneurysm surfaces with respect to given optimality criteria (cost functions). Using gradient-based cost functions, the LOGISMOS algorithm guarantees the resulting surfaces to be globally optimal and consistent within and between images. In addition, prior knowledge of the target shape such as surface smoothness and separation constraints is incorporated.

Figure 1

Semi-automated segmentation technique using the LOGISMOS-JEI algorithm of a complex enhancing basilar tip aneurysm. (A) step 1: co-registration of images in all planes. (B) step 2: manual identification of the center of the aneurysm sac (inner sphere) and a location not-too-far outside of the aneurysm surface (outer sphere). (C) step 3: gradient-based cost functions are applied to ‘expand’ the inner sphere and ‘shrink’ the outer sphere in order to define the lumen (red) and surface (green) of the aneurysm, respectively. JEI corrects localized segmentation inaccuracies as needed (arrowhead). (D) 3D volumetric segmentation of both mesh spheres. The ‘virtual space’ between the spheres is interpreted by the algorithm as the aneurysmal wall.

After the automated LOGISMOS graph segmentation, an optional JEI step allows the expert observer to efficiently fix small localized segmentation inaccuracies. Instead of redrawing surfaces on all affected 2D slices, only a few contours indicating correct surface locations are typically needed to modify both the inner and outer 3D surfaces in their entirety (figure 1C, arrowhead). This segmentation protocol was performed in T1 pre- and post-contrast images, and mean signal intensity (SI) measurements of the aneurysmal wall were obtained in all projections.

HR-VWI analysis: manual segmentation protocol

The semi-automated LOGISMOS-JEI segmentation method was compared with our validated manual segmentation protocol using Picture Archiving Communication System software (PACS Carestream, Rochester, NY) previously described in the literature.17 Briefly, T1 pre- and post-contrast images were manually co-registered with 3D TOF to determine the aneurysm size and correctly delineate the aneurysm’s wall. T2-weighted images and SWAN sequences were also co-registered to better delineate the outer surface of the aneurysm and to exclude surrounding artifacts that may appear hyperintense such as parenchymal brain tissue, dura mater, and veins. In this way, ROIs were generated (figure 2) through manual segmentation in three planes (axial, coronal, and sagittal) at the level of maximal aneurysm diameter.

Figure 2

Manual segmentation technique using PACS Carestream Vue of an ACOM aneurysm. The aneurysm is identified in all the TOF sequences, and T1 pre- and post-contrast images are co-registered. ROIs of the aneurysm’s wall and parent vessel are then drawn on axial (A), coronal (B), and sagittal (C) views of pre- and post-contrast T1-weighted images. Note the two different ROIs (3 mm in red stripes, 5 mm in white) drawn over the parent vessel (schematic representation).

HR-VWI analysis: measurements of enhancement

Roa et al demonstrated that normalization of SI of the aneurysmal wall with the SI of the pituitary stalk improves reproducibility and allows standardization of enhancement measurements in HR-VWI.17 Consequently, the pituitary stalk was also segmented with both semi-automated and manual segmentation techniques. The mean SI of the aneurysmal wall in each ROI plane was adjusted for the mean SI of the pituitary stalk. Then, we calculated the circumferential aneurysmal wall enhancement (CAWE) as follows:

Embedded Image

The high spatial resolution of 7T HR-VWI allows visualization of the arterial walls of the circle of Willis. Parent vessels were analyzed in a similar fashion as the aneurysm sac. ROIs for parent vessel enhancement were sampled to include two different length segments of the parent artery most proximal to the aneurysm’s neck: 3 mm and 5 mm (figure 2). Parent vessel enhancement (PVE) was calculated in a similar way to CAWE:

Embedded Image

A reference vessel wall was sampled in a different vascular territory than the one where the aneurysm was located: for aneurysms in the anterior circulation, the M1 segment of the contralateral middle cerebral artery (MCA) or supraclinoid segment of the contralateral internal carotid artery (ICA) in the axial plane were used as the reference vessel, whereas for aneurysms in the posterior circulation, the reference vessel was the M1 segment of the right MCA in the axial plane. A 5 mm section of the selected reference vessel was sampled. Arteries with atherosclerotic changes in the vessel wall were excluded. In patients with multiple aneurysms in the anterior and posterior circulations, the contralateral supraclinoid segment of the ICA or M1 segment of the MCA was assigned as the reference vessel. The reference vessel enhancement (RVE) was calculated as follows:

Embedded Image

Statistical analysis

Continuous variables are presented as mean±SD, and categorical variables are presented as frequency and percent. Aneurysm size was initially measured as a continuous variable, and then dichotomized in two separate categories:<7 mm and ≥7 mm. Size ≥7 mm was used as a surrogate of aneurysm instability as determined by previous epidemiological studies.18 Distributions of values for CAWE, 3- and 5 mm PVE, and RVE were tested for normality using the Shapiro–Wilk method. For normally-distributed variables, Student’s t tests were used to compare means. For non-parametric variables, Mann–Whitney U tests were used to compare the medians between the size groups. Pearson correlations were used to examine relationships between continuous variables. Intra-class correlation (ICC) coefficients were calculated to assess reliability between the semi-automated and the manual segmentation methodologies. A 2-sided P-value<0.05 was considered significant. All statistical analyses were performed with SPSS Statistics 25.0 (IBM, Armonk, New York).

Results

A total of 42 patients were screened, 15 were excluded: seven patients had metal meshes placed in soft tissues from previous surgeries, five patients had an eGFR <60 mL/min, two overweight patients did not fit in the scanner, and one patient suffered from claustrophobia and did not complete the scan. Four patients with fusiform aneurysms were also excluded from the analysis since the neck of the aneurysm could not be defined precisely. Our final sample consisted of 23 patients with 25 aneurysms. Patient and aneurysm characteristics are described in table 1. Mean age was 67.5±9.2 years' old, and 18 (78.3%) were women. Mean aneurysm diameter was 8.8 mm (range 3–32 mm). Seventeen aneurysms were <7 mm. Twelve aneurysms were in the ICA (eight ophthalmic C6, three terminus, one cavernous C4), four anterior communicating artery, four basilar artery, two posterior inferior cerebellar artery, one anterior cerebral artery, one middle cerebral artery, and one posterior communicating artery. Nine aneurysms presented with symptoms and were treated accordingly. All parent vessels demonstrated a uniform pattern of enhancement, comparable to the circumferential pattern of enhancement in the aneurysmal wall.

Table 1

Baseline characteristics of the sample

Using the SI measurements derived from the semi-automated segmentation method, Shapiro–Wilk tests showed that values for CAWE (W=0.67; P<0.001), 3 mm PVE (W=0.82; P=0.001), and 5 mm PVE (W=0.91; P=0.036) were non-normally distributed in our sample. Therefore, non-parametric statistical methods were used. Adjusted SI measurements in the aneurysmal wall, 3- and 5 mm parent vessel, and reference vessel according to aneurysm size are presented in table 2. Although non-significant, Mann–Whitney U test analyses demonstrated that CAWE (U=43.5; P=0.15) and 3 mm PVE (U=52.0; P=0.37) tend to be higher in aneurysms≥7 mm compared with aneurysms<7 mm. This trend was not observed when using 5 mm PVE (U=58.0; P=0.59) or RVE (U=64.0; P=0.84).

Table 2

Adjusted signal intensity ratios according to aneurysm size

Pearson correlations demonstrated a significant moderate positive correlation between CAWE and 5 mm PVE (r=0.52, Ptwo-tailed=0.008). Moreover, a strong positive correlation coefficient was found between CAWE and PVE using 3 mm (Pearson R=0.78, Ptwo-tailed <0.001), suggesting that the parent artery exhibits higher contrast enhancement in regions closer to the aneurysm neck (figure 3A,B). On the other hand, the correlation between CAWE and RVE was weak and non-significant (Pearson R=0.20, Ptwo-tailed=0.33) (figure 3C).

Figure 3

Scatter plots depicting correlations between (A) CAWE and 5 mm PVE, (B) CAWE and 3 mm PVE, and (C) CAWE and RVE. Continuous line shows overall trend and dashed lines represent 95% confidence intervals.

The comparison between the semi-automated and the manual segmentation methods demonstrated good reliability based on ICC coefficients. For single measures, ICC=0.790 (95% CI 0.58 to 0.90). For average measures (ie, assuming the interaction effect is absent), ICC=0.882 (95% CI 0.73 to 0.95). This suggests that the semi-automated segmentation technique using the LOGISMOS-JEI algorithm attains accurate SI measurements when compared with the manual segmentation protocol.

Discussion

Our study applied two different segmentation methodologies to objectively compare CAWE with PVE and RVE on 7T HR-VWI. There is a strong correlation between aneurysmal wall enhancement and parent artery enhancement, which was not observed in arteries of different vascular territories. This correlation was particularly important for aneurysms≥7 mm in size and stronger when a shorter segment of the parent artery closer to the neck of the aneurysm was sampled. We hypothesize that the increased SI of the parent artery is a result of inflammatory and vasculopathic processes primarily affecting the arterial wall, and that these changes ultimately lead to aneurysm formation.

Aneurysmal wall inflammation and rupture risk

Histological analyses have determined the presence of focal deterioration of the arterial wall structure in the pathogenesis of brain aneurysms. Common findings in the aneurysmal wall that correlate with increased contrast enhancement in HR-VWI include the presence of active macrophages, neovascularization, and decreased elastin. After analyzing 10 post-clipped wall specimens from UIAs with pre-surgical HR-VWI, Hudson et al demonstrated that aneurysms with avid enhancement had increased macrophage infiltration and cellularity in comparison to aneurysms with mild- or no-enhancing wall.11 Similarly, Shimonaga et al found a significant association between aneurysmal wall enhancement and increased vasa vasorum and macrophage infiltration on the histological analysis of nine UIAs.19

The ultra-high spatial resolution of 7T MR imaging is an excellent tool for in vivo visualization of aneurysm microstructures.20 21 Imaging of cerebral arteries may be challenging because of their morphology and thin walls (0.06–0.25 mm).22 We used the high spatial resolution of 7T imaging to determine the presence of increased contrast enhancement in the parent artery close to the neck of the aneurysm, when compared with a reference artery located in another vascular territory of the intracranial circulation. Moreover, we found stronger correlations when the 3 mm ROI of the parent vessel closest to the aneurysm’s neck was sampled, compared with the 5 mm ROI. This finding supports the hypothesis that localized arteriopathy affects the parent vessel and leads to aneurysm formation.

CFD studies of parent vessels have shown that increased WSS accompanied by spatial WSS gradient leads to aneurysm formation.23 24 Kulcsár et al described two cases of peaked WSS and spatial WSS gradient at the site of the parent artery next to the neck of the aneurysm that developed later. WSS values at these points were at least five times higher than the temporal average values of the parent vessel.23 We also demonstrated increased contrast enhancement closer to the aneurysmal neck, which suggests a higher hemodynamic stress in areas of the parent vessel closer to the neck.

Recently, Khan et al studied the association between flow hemodynamics and wall enhancement of 25 UIAs using a 3D vessel wall MRI.25 Enhancing UIAs (n=9) had lower WSS compared with non-enhancing UIAs (n=16). Moreover, sac-averaged normalized SI correlated with WSS and was significantly different in regions of low WSS compared with regions of intermediate and high WSS.25 Low WSS in areas of increased enhancement suggests active growth/remodeling of the aneurysmal wall and higher risk of rupture. Likewise, increased SI in the parent artery may reflect an underlying hemodynamic insult to the endothelium that ultimately promotes inflammation and histological changes.

Aneurysm enhancement on HR-VWI: wall inflammation or flow artifact?

Common confounders of increased aneurysmal wall enhancement include low blood flow velocity (stagnation of flow in the aneurysmal sac) and the presence of contiguous strong-enhancing structures, such as dura mater folds and veins.12 In a study by Cornelissen et al, six of 10 enhancing aneurysms showed diameters notably smaller (≥0.8 mm) on post-contrast imaging compared with their appearance on pre-contrast and TOF sequences.26 The authors suggested that the visualized enhancement was partially located in the aneurysm lumen and could be justified by slow intra-aneurysmal flow. Similar flow-related SI artifacts have been reported by the same group in a phantom model after contrast injection was performed at different flow pressures,26 suggesting insufficient slow-flow suppression on HR-VWI.

It is harder to explain increased SI as a flow artifact in parent arteries, as the flow in this scenario is mainly laminar. In the aneurysms analyzed in this study, flow within the parent artery was not impaired by any flow-limiting stenosis, therefore it is less likely that the increased SI documented along the parent arteries corresponds to pseudo-enhancement. The presence of increased SI along the parent artery close to the neck of the aneurysm suggests, such as in studies of aneurysmal wall enhancement, that an inflammatory process affects this segment of the wall and remains active even after aneurysm formation. The high spatial resolution of 7T imaging also decreases the presence of artifact at the time of selecting the ROIs of the parent and reference arteries.

Edjlali et al recently reported a case of decreased CAWE and symptom resolution in a patient with a large para-ophthalmic aneurysm that was treated with steroids.27 Aneurysm pathogenesis involves a complex interplay of flow dynamics, inflammation, and atherosclerotic changes that ultimately lead to aneurysm rupture. HR-VWI is a valuable tool in the study of aneurysm formation, growth, and ruptured. HR-VWI detects the presence of microhemorrhages within the aneurysm wall,28 quantifies aneurysm wall thickness,20 characterizes morphological aneurysm indexes,29 30 defines SI,12 and aids in the generation of CFD models.25

Limitations

This is a small single-center study. All the symptomatic aneurysms were treated and we have limited follow-up data. The “gold-standard” validation of pro - inflammatory and vasculopathic changes in the parent artery of UIAs would be histological analysis. This would be possible only on post-mortem specimens. Since we measured wall enhancement immediately proximal to the aneurysm lumen, the potential confounding effect of altered hemodynamics in this region was not assessed or correlated with CFDs. Finally, validation of the SI-based quantitative techniques used to measure CAWE, PVE, and RVE require larger follow-up studies to evaluate their clinical relevance for prediction of aneurysm growth or rupture. Further efforts directed toward standardization and automatization of SI measurements in the vessel wall are key to bring HR-VWI of UIAs to prime time.

Conclusion

This 7T HR-VWI study demonstrates an increased contrast enhancement of the parent vessel of UIAs when compared with the arteries of other vascular territories. Parent arteries exhibit higher contrast enhancement in regions closer to the aneurysm’s neck, especially in aneurysms≥7 mm. This observation supports the theory that localized arteriopathy of the parent artery leads to the pathogenesis of cerebral aneurysms.

Acknowledgments

We thank Heena Olalde for her continuous support with this research.

References

  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 16.
  29. 29.
  30. 30.

Footnotes

  • Twitter @jorge_roa93

  • Contributors Conception and study design: EAS; Acquisition of data: JAR and HZ; Analysis and interpretation of results: EAS and JAR; Drafting of the manuscript: EAS and JAR; Critical revision of the study: TRK, SOG, GB, MS, CPD, VM, DMH; Final approval of the version to be published: EAS.

  • Funding This work was supported by the 2019 Brain Aneurysm Research Grant from The Bee Foundation and by a Pilot Research Grant from the Society of Vascular and Interventional Neurology (SVIN), both granted to Edgar Samaniego. This work was conducted on an MRI instrument funded by 1S10RR028821-01.

  • Competing interests None declared.

  • Patient consent for publication Not required.

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

  • Data availability statement Data are available upon reasonable request. Additional unpublished data will be made available by the corresponding author upon appropriate request.

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.