Changes in contrast transit times on digital subtraction angiography post-Pipeline Embolization Device deployment

Abstract

It is postulated that hemodynamic changes occur in the distal vascular bed post-deployment of Pipeline Embolization Devices (PEDs). In this paper, we evaluate changes in the contrast transit times (TTs) on digital subtraction angiography (DSA) post-PED interventions. DSA films were analyzed using custom-made software for the time-density relationship at baseline and compared to post-PED deployment. All analyses were performed within the middle cerebral artery (MCA) M1 segment. Analyses included TT_10%–100% (time needed for the contrast to change from 10% image intensity to 100%), TT_100%–10%, and TT_25%–25%. Forty-four patients were included. We found a significant decrease in TT_10%–100% (2.79 to 2.24 seconds, p < 0.001) post-PED. There was a significant correlation between the percentage change in TT_100%–10% and aneurysm size (p = 0.02). There was also a significant decrease in TT_25%–25% (7.07 to 6.41 seconds, p = 0.02) post-PED. Moreover, there was a significant correlation between the absolute or percentage changes in TT_25%–25% and aneurysm size (rho = 0.54, p = 0.05 and rho = 0.29, p = 0.05, respectively). Statistically significant distal intracranial hemodynamic changes occur post-PED deployment. These hemodynamic changes appear to be more pronounced with large and giant aneurysms.

Introduction

The Pipeline Embolization Device (PED; ev3 Neurovascular, Irvine, CA) was approved by the Food and Drug Administration in April 2011 for the treatment of complex anterior circulation cerebral aneurysms. Recently, however, the use of PED has been expanded to include posterior circulation aneurysms, small aneurysms less than 1 cm, and distal anterior circulation aneurysms. Flow diversion results in vessel reconstruction and flow disruption away from the aneurysm, with subsequent aneurysm thrombosis.^1,2 With the expanded use of PED, unexpected complications have emerged. These complications include rupture of the aneurysm and delayed ipsilateral intraparenchymal hemorrhage (DIPH),^1,3–5 both of which have morbid outcomes³ and still no consensus agreement on their etiology. The effects of PED on intracranial hemodynamics beyond the aneurysm have not been clearly described. In this study, we evaluate changes in distal intracranial hemodynamics using digital subtraction angiography (DSA) and a contrast transit time (TT)-density analysis pre- and post-PED deployment.

Materials and method

Patient selection

Following institutional review board approval, records of patients with intracranial aneurysms treated at our institution between 2008 and 2015 were retrospectively reviewed. In order to standardize the analysis, patients were included if the intracranial aneurysm was treated with PED and located proximal to the internal carotid artery (ICA) terminus, and if complete DSA films in digital imaging and communications in medicine (DICOM) format were available pre- and post-PED treatment. Patients with posterior circulation aneurysms, ruptured aneurysms, and distal anterior circulation aneurysms were excluded. Of the 78 patients treated with PED during this period, 46 patients met the inclusion criteria.

Image acquisition

All embolization procedures were performed under general anesthesia and with systemic heparinization as well as with dual antiplatelet therapy. A trans-femoral approach was used in all cases. A total of 12 ml of the iodine-based contrast agent iohexol (300 mg/ml, Omnipaque, GE Healthcare, Chicago, IL) was injected into the cervical ICA. Contrast was injected by a power-contrast injector (Medrad, Bayer HealthCare, Whippany, NJ) over two seconds. All settings of the injector were the same for each angiogram. The power-contrast injector was synchronized to a fluoroscopy angiographic machine with a 1.2-second delay in the contrast injection. DSA images at a rate of four frames/s were acquired in anterior-posterior, lateral, and oblique trans-orbital projections using a biplane neuro-angiography suite (Artis zee, Siemens Healthineers, Erlangen, Germany). DSA images were routinely saved to the picture archiving and communication system (PACS), and the entire unedited DSA was archived on a separate digital video disc (DVD) in DICOM format.

DSA TT analysis

Archived DICOM images of the DSA for the PED procedure at baseline as well as post-procedure were downloaded onto a central computer for analysis. Baseline and post-PED imaging were acquired under general anesthesia during the same procedure and under a comparable blood pressure measurement. To interpret the relative changes in blood flow, we developed a custom-made software code (MATLAB, MathWorks, Natick, MA). Individual DSA runs were analyzed for contrast image intensity throughout the angiogram cycle. This technique identified various intensity plots, including the maximum gray intensity in consecutive DICOM images.⁶ Region of interest (ROI) was selected over the middle segment of the middle cerebral artery (MCA) M1 segment at baseline and compared to post intervention for the same ROI. The time-intensity plot was divided into three components as shown in Figure 1, and estimated TTs were defined as follows:

TT_10%–100% (time needed for the contrast to change from 10% image intensity to 100%)

TT_100%–10% (time needed for the contrast to change from 100% image intensity to 10%)

TT_25%–25% (time needed for the contrast to change from 25% image intensity up slope to 25% down slope of the curve)

Figure 1.

(a) TT_10%–100% at baseline for a patient with a paraophthalmic ICA aneurysm treated with PED. (b) TT_10%–100% for the same patient post-PED showing significant shortening of the early phase of time-intensity indicating hemodynamic changes over the selected ROI on the M1 segment. TT: transit time; ICA: internal carotid artery; PED: Pipeline Embolization Device; ROI: region of interest.

Statistical analysis

TT_10%–100%, TT_100%–10%, and TT_25%–25% selected over the M1 segment ROI were compared between pre- and post-PED intervention using a paired Student's t-test or Wilcoxon matched pairs signed rank test, based on results of the skewness and kurtosis normality tests. Pearson and Spearman correlations were used to measure statistical dependence between the absolute and percentage changes in all TTs post-PED deployment versus the number of PEDs deployed or the length of the PED used. P ≤ 0.05 was defined as a significant difference. Data analysis was performed using STATA software (Version 12.0 for Macintosh, StataCorp, College Station, TX).

Results

Patient characteristics

The study cohort consisted of 44 patients with a mean age of 56 years, and 85% were female. Mean aneurysm size was 8 mm (median 7 mm) and all aneurysms were unruptured. Mean length of PED used was 18 mm and mean diameter used was 4.4 mm. In five patients, the aneurysm was treated with partial coiling in addition to the PED. Study cohort demographics and PED characteristics are provided in Table 1.

Table 1.

Study cohort and PED characteristics.

Mean age, years (range)	56.4 (37–82)
Female gender, n (%)	39 (84.8%)
Clinical presentation, n (%)
• Incidental	31 (67.4%)
• Visual symptoms	7 (15.2%)
• Recurrence	8 (17.4%)
Aneurysm site, n (%)
• Cavernous ICA	10 (21.7%)
• Paraophthalmic ICA	32 (69.6%)
• Posterior communicating ICA	2 (4.3%)
• Anterior choroidal ICA	1 (2.2%)
• Superior hypophyseal ICA	1 (2.2%)
Mean aneurysm size, mm (range)	8.0 mm (2–28 mm)
• < 10 mm, n	34
• ≥ 10 mm, n	12
PED dimensions
• Diameter: mean, mm (range)	4.4 mm (3.25–5.0 mm)
• Length: mean, mm (range)	18 mm (12–30 mm)
Number of PEDs (% of cohort)	1 (93.5%) 2 (6.5%)
Associated comorbidities, n (%)
• Hypertension	25 (54.3%)
• Hyperlipidemia	10 (21.7%)
• Smoking	12 (26.1%)

PED: Pipeline Embolization Device; ICA: internal carotid artery.

TT_10%–100%

There was a significant decrease in TT_10%–100% (2.79 to 2.24 seconds, p <  0.001) post-PED (Figure 1, Table 2). There was no correlation between the absolute or percentage change in TT_10%–100% and aneurysm size (p = 0.25 and p = 0.50, respectively), PED diameter (p =  0.48 and p = 0.58 respectively), or PED length (p = 0.90 and p = 0.80, respectively).

Table 2.

Summary of DSA TT results pre- and post-PED deployment.

	Baseline	Post-PED	p value
TT10%–100% (s)	2.79	2.24	<0.001
TT100%–10% (s)	5.94	5.68	0.33
TT25%–25% (s)	7.07	6.41	0.02

DSA: digital subtraction angiography; TT: transit time; PED: Pipeline Embolization Device; s: seconds.

TT_100%–10%

There was a trend toward a decrease in TT_100%–10% (5.94 to 5.68 seconds, p = 0.33) (Figure 2, Table 2). There was a significant correlation between the percentage change in TT_100%–10% and aneurysm size (r =  0.34, p = 0.02), but not between the absolute change in TT_100%–10% and aneurysm size (p = 0.15) (Figure 3). There was no correlation between the absolute or the percentage change in TT_100%–10% and PED diameter (p =  0.68 and p = 0.64, respectively) or PED length (p = 0.14 and p = 0.07, respectively).

Figure 2.

Box plots comparing differences in TTs pre- and post-PED. Asterisk represents a statistically significant difference. Boxes represent the interquartile range, horizontal line the median, and whiskers the minimum and maximum values. TTs: transit times; PED: Pipeline Embolization Device.

Figure 3.

Best fit correlation between the percentage change in TT_100%–10% and aneurysm size post-PED. Percentage change in TT_100%–10% correlates with aneurysm size (r = 0.34, p = 0.02). TT: transit time; PED: Pipeline Embolization Device.

TT_25%–25%

There was a significant decrease in TT_25%–25% (7.07 to 6.41 seconds, p =  0.02) post-PED (Figure 1, Table 2). There was also a significant correlation between the absolute or percentage changes in TT_25%–25% and aneurysm size (rho = 0.54, p = 0.05 and rho = 0.29, p =  0.05, respectively) (Figure 4). TT_25%–25% post-PED deployment decreased among aneurysms smaller than 15 mm and increased among aneurysms larger than 15 mm (Figure 4). There was no correlation between the absolute or the percentage change in TT_25%–25% and PED diameter (p =  0.71 and p = 0.76, respectively) or PED length (p = 0.19 and p = 0.16, respectively).

Figure 4.

Best fit correlation between TT_25%–25% and aneurysm size at baseline and post-PED. Absolute change in TT_25%–25% correlates with aneurysm size (rho = 0.54, p = 0.05). TT: transit time; PED: Pipeline Embolization Device.

Discussion

Flow diverters were introduced recently in the treatment of intracranial aneurysms. They divert flow away from the neck of the aneurysm, slowly inducing thrombosis of the aneurysm sac while preserving flow in the parent vessel and adjacent branches.⁷ However, these devices are associated with perioperative morbidity and mortality rates up to 10%.¹ DIPH is a poorly understood and often fatal complication seen following PED. The majority occur in the ipsilateral hemisphere (82%) and within the first month after PED deployment (86%).⁸ Multiple mechanisms for DIPH have been hypothesized, including intraprocedural microwire perforation, hemorrhagic transformation of periprocedural infarct, and changes in parent vessel and distal intracranial hemodynamics. The fact that in nearly 20% of cases DIPH occurs in a separate vascular territory might suggest a mechanism that is not related to deleterious hemodynamic changes or ischemia related to the procedure, but rather may be related to dual antiplatelet therapy. Yet, this complication is not reported as much in stent-assisted coiling cases⁹ or in patients with intracranial atherosclerosis, where patients are routinely placed on dual antiplatelet therapy.¹⁰

Additionally, PED treatment of a giant aneurysm may result in the sudden loss of a large capacitance chamber (i.e. the giant aneurysm) potentially resulting in cerebral hyperperfusion distal to the aneurysm and reversal of the Windkessel effect, as suggested by Chiu and Wenderoth.¹¹ Specifically, reconstruction of the carotid artery with PED could result in reduced vascular compliance in that region, leading to alteration of the blood pressure waveform that is transmitted distally.¹² Eker et al.¹³ evaluated volumetric flow rate waveforms, using magnetic resonance angiography, in the ICA at baseline and post-PED. In this study, they demonstrated a significant, inverse correlation between the size of the aneurysm and the pulsatility index ratio. This might suggest an existing Windkessel effect at baseline for larger aneurysms, where a large aneurysm acts as a reservoir for retained blood in systole, which flows out in diastole. This results in a decrease in the systolic peak and a relative increase in diastolic flow. In a small subgroup of patients with symptomatic cavernous ICA aneurysms, there was significant change in the pulsatility index ratio from 0.61 at baseline to 0.95 post-PED, corresponding to a 25% increase in diastolic velocity.¹⁴ Similarly, Shakur et al.¹⁵ examined changes in distal intracranial hemodynamics post-PED and found that the Lindegaard ratio in diastole was significantly higher ipsilaterally compared to contralaterally to the PED.

The fact that DIPH occurs in a proportionally larger percentage of patients with giant aneurysms points to possible hyperperfusion when the Windkessel effect is reversed upon PED deployment.¹ Indeed, results from a large, multicenter retrospective study reported that the odds ratio (OR) for DIPH is significantly increased (OR = 4.44; p = 0.04) with the use of three PEDs.³ Our data also revealed a positive relationship between large aneurysm size and the change in TTs. TT_25%–25% post-PED deployment decreased among aneurysms smaller than 15 mm but increased among aneurysms larger than 15 mm (Figure 4). These results demonstrate reversal of the Windkessel effect post-PED. Increased TTs detected among aneurysms larger than 15 mm likely reflects contrast stasis within the aneurysm after PED deployment.

Prior reports on DSA image analysis demonstrated that the time-intensity curve depends on vessel size, amount of contrast injected, and frame rate.¹⁶ DSA TT analysis was performed recently in a subset of patients with aneurysmal subarachnoid hemorrhage. In this study, Ivanov et al. found significantly longer TTs in poor-grade versus good-grade patients.⁶ The prolonged TTs may be attributed to poor cerebral perfusion in the setting of elevated intracranial pressure and subarachnoid hemorrhage. Prolonged intra-procedural cerebral circulation time, measured as the maximal opacification of the terminal portion of the ICA and a cortical vein on DSA, was shown to be associated with higher risk of hyperperfusion syndrome post-carotid artery stenting.¹⁷ In this study, we reveal hemodynamic changes that occur immediately post-flow diversion, which is seen in the change in the contrast time-density curve. The three parameters assessed (TT_10%–100%, TT_100%–10%, TT_25%–25%) represent the main phases seen on DSA, including arterial and venous phases. Additionally, our previous study describing the technique of DSA TT analysis⁶ demonstrated less artifact with these particular parameters compared to other TTs. Changes in TTs post-PED were most significant for TT_10%–100% and TT_25%–25%, and so these parameters are likely to be more clinically useful. Specifically, TT_10%–100% and TT_25%–25% decreased significantly post-PED, indicating faster intracranial flow and reversal of the Windkessel effect where the pulse pressure reaches the intracranial circulation more quickly. These results suggest that tight blood pressure control post-PED may be warranted to avert hyperperfusion-related complications.

Limitations of this study include its retrospective design and small sample size. A few patients in our cohort had large and giant aneurysms, which affected TT. Consequently, the relationship between baseline TTs and aneurysm size should be investigated further with a more normally distributed population and using the contralateral ICA as an internal control to control for systemic factors, such as age, cardiovascular risk factors, and other variables that may affect DSA TT analysis. Additionally, because of the small sample size we were unable to assess the additive effect of multiple PEDs on changes in contrast TTs. Future studies with larger sample sizes may better address this question.

Conclusions

Significant changes in contrast TTs within the distal intracranial circulation occur post-PED. These hemodynamic changes appear to be more pronounced when larger aneurysms are treated with PED.

Declaration of conflicting interests

The authors declared the following potential conflicts of interest with respect to the research: Ali Alaraj, M.D.: research grant: National Institutes of Health, and consultant: Cordis-Codman. Fady T. Charbel, M.D.: ownership interest: VasSol Inc, and consultant: Transonic. The other authors have nothing to declare.

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.