Retrospective analysis of intracranial aneurysms after flow diverter treatment including color-coded imaging (syngo iFlow) as a predictor of aneurysm occlusion

Abstract

Purpose

Flow Diverters (FD) have immensely extended the treatment of cerebral aneurysms in the past years. Complete aneurysm occlusion is a process that often takes a certain amount of time and is usually difficult to predict. Our aim was to investigate different syngo iFlow parameters in order to predict aneurysm occlusion.

Methods

Between 2014 and 2018 patients with unruptured cerebral aneurysms treated with a FD were reviewed. Aneurysm occlusion and complication rates have been assessed.

In addition, various quantitative criteria were assessed using syngo iFlow before, after the intervention, and after short and long-term digital subtraction angiography (DSA).

Results

A total of 66 patients hosting 66 cerebral aneurysms were included in this study. 87.9% (n = 58) aneurysms in the anterior and 12.1% (n = 8) in the posterior circulation were treated. Adequate aneurysm occlusion at long-term follow-up (19.05 ± 15.1 months) was achieved in 90.9% (n = 60). Adequately occluded aneurysm revealed a significantly greater peak intensity delay (PI-D, p = 0.008) and intensity decrease ratio (ID-R, p < 0.001) compared to insufficiently occluded aneurysms. Increased intra-aneurysmal contrast agent intensity (>100%) after FD implantation resulted in an ID-R < 1, which was associated with aneurysm growth during follow-up DSA. Retreatment with another FD due to foreshortening and/or aneurysm growth was performed in 10.6% (n = 7). Overall morbidity and mortality rates were 1.5% (n = 1) and 0%.

Conclusion

The applied syngo iFlow parameters were found to be useful in predicting adequate aneurysm occlusion and foresee aneurysm growth, which might indicate the implantation of another FD.

Introduction

For the treatment of cerebral aneurysms, fine-meshed stent systems, so called flow diverters have become widely applied in the past years. Many clinical studies have shown promising results using such devices with an acceptable risk profile.^1–3 Current studies have even shown that the risk profile of such devices has decreased compared to the early years.^4–10 By diverting blood flow away from the aneurysm, they aim to alter the dynamics of intra-aneurysmal flow to induce thrombosis that leads to the occlusion of the aneurysm over time. This is one of the few methods in the treatment of cerebral aneurysms where complete reconstruction of the original vessel wall and thus complete healing of the parent vessel is possible. The duration of complete aneurysm occlusion after FD treatment is very difficult to estimate and still not fully understood due to reports of insufficient occlusions reaching up to 17% in literature. ¹ Many clinical studies have already investigated intra-aneurysmal flow changes of cerebral aneurysms after treatment with flow diverters and the complications associated with them.^11–15 Color-coded imaging using syngo iFlow has already been shown beneficial in assessing various cerebrovascular diseases.^16,17 Previously described preclinical syngo iFlow parameters have shown promising results in the assessment of hemodynamic changes in the field of flow diversion. ¹⁸ Unfortunately, there are only few studies today that have used color-coded imaging to evaluate the prediction of aneurysm occlusion following flow diverter treatment.^19,20

The purpose of this retrospective study was to apply the previously described syngo iFlow parameters ¹⁸ in a clinical setting and to evaluate their predictive value for aneurysm occlusion after flow diverter treatment.

Material and methods

Patient selection

Ethics committee approval was obtained for this retrospective data analysis. Our database was screened for patients with unruptured cerebral aneurysms treated with flow diverters between January 2014 and December 2018. A minimum follow-up time of 3 months was required. Aneurysms additionally treated with coil embolization were excluded because the number was not representative in our study and the main intention was to assess the hemodynamic changes as purely as possible. Clinical outcome was evaluated by using the modified Rankin scale (mRS) before treatment, at discharge and at follow-up.

Interventional procedure

All patients in this study received a dual antiplatelet medication of 100 mg acetylsalicylic acid and clopidogrel 75 mg daily at least 3–5 days prior to the intervention. Endovascular aneurysm treatment was performed under general anesthesia using a standard tri-axial system. Periprocedural a dose of 3000 IU i.v. heparin was administered. Dual antiplatelet therapy (DAPT) was continued for at least 6 months and then continued as life-long monotherapy. Platelet function testing was not performed since its level of evidence remains controversial. ²¹ For the treatment of the aneurysms four different types of FDs have been used; Pipeline Embolization Device (PED; Medtronic Inc., Minneapolis, Minnesota, USA), Derivo Embolization Device (DED; Acandis, Pforzheim, Germany), Flow Redirection Endoluminal Device (FRED; MicroVention, Tustin, CA, USA) and p64 Flow Modulation Device (p64; Phenox GmbH, Bochum, Germany). The selection of FDs was based on the maximum diameter of the parent vessel within the target region in which the FD was to be implanted. The length of the implanted FD was determined based on the width of the aneurysm neck to be covered and the underlying vascular course. All devices were delivered via a 0.027-inch microcatheter using a standard deployment technique. First, the microcatheter was placed distal to the aneurysm neck and then the FD was inserted into the microcatheter. The placement of the FD was achieved by a combination of withdrawing the microcatheter and gently holding/pushing the delivery wire of the FD in place of the desired target area.

Imaging protocol

DSA was performed with either a monoplane flat panel detector angiographic system (Artis zeego; Siemens AG, Erlangen, Germany) or with a biplane flat panel detector angiographic system (Artis Q; Siemens AG, Erlangen, Germany) in posterior-anterior and lateral projection (2 D-DSA series) at a rate of four frames/s. Manual contrast agent injection (Ultravist 370, Bayer Vital, Germany) was performed using a variety of standard diagnostic catheters (4–5Fr) and periinterventional intermediate catheters (5–6Fr). DSA images of all the patients were collected and postprocessed on an appropriate workstation (Leonardo Workstation; Siemens AG, Erlangen, Germany) using a dynamic flow evaluating software (syngo iFlow; Siemens AG, Erlangen, Germany). The software visualizes a complete 2 D-DSA run into a colored single image, which reveals the time-dependent intensity of contrast agent within the aneurysm and parent vessel quantifying the time between injection of the contrast agent and maximum opacification. First, a projection plane was selected in which the aneurysm did not overlap with the parent vessel. Regions of interest (ROI) with a surface area between 1.60–9.70 mm², depending on the vessel size, were placed within the parent artery for reference (REF). Depending on the shape of the aneurysm, the reference ROI for saccular aneurysms was placed directly at the level of the aneurysm neck (Figure 1(b)); for fusiform aneurysms, the ROI was placed in the proximal parent vessel (Figure 2(c)). ROIs of the same size were then placed inside the aneurysm at a corresponding location in order to obtain comparable baseline curves. In case of saccular aneurysms, the ROI was placed centrally in the aneurysm (Figure 1(b)), in fusiform cases the ROI was placed in the part of the aneurysm that did not overlap with the original parent vessel (Figure 2(c)). At follow-up, the ROIs were placed in a comparable location, unless the aneurysm changed its morphology, in which case the intra-aneurysmal ROI was adjusted accordingly. Time-density curves (TDC) were generated before and after implantation of the flow diverter and at follow-up DSA (Figure 1(a)). From the TDCs different values (peak intensity (PI), time to peak (TTP), intensity decrease (ID)) have been determined. Based on these values different flow parameters have been defined (peak intensity delay (PI-D), peak intensity ratio (PI-R, intensity decrease ratio ID-R)) like previously described. ¹⁸ The prerequisite for the iFlow evaluation was a DSA run of at least 6 seconds in length prior FD implantation and at least 8 seconds post FD implantation, so that the selected parameters could be applied. These parameters have been applied at each DSA and correlated to the aneurysm occlusion rate observed during follow-up. Two experienced neuroradiologists evaluated and compared the DSA series and the iFlow images, respectively.

Figure 1.

Digital subtraction angiography (DSA) of a saccular paraophthalmic aneurysm (4.8 × 4.5 × 4.5 mm) of the left internal carotid artery (a). iFlow image prior FD implantation including the ROI of the aneurysm (2) and parent vessel (REF) as well as the TDC showing comparable flow patterns (b). iFlow image and TDC after FD implantation with an PI-D of 2.3 and ID-R of 6.7 representing an adequate intra-aneurysmal flow reduction (c). DSA after 4 months showed a complete occlusion of the aneurysm and (d).

Figure 2.

DSA and Flat-Panel Computer Tomography (3 D-DYNA-CT) reconstruction of a fusiform infraophthalmic aneurysm (19.7 × 12.3 × 15.5 mm) of the right internal carotid artery (a and b). iFlow image prior FD implantation including the ROI of the aneurysm (2) and parent vessel (REF) as well as the TDC (c). Fluoroscopy of the implanted Derivo Emolization Device (5.5 × 40 mm) (d). iFlow image and TDC after FD implantation with an PI-D of 1.1 and ID-R of 6.0 representing an adequate intra-aneurysmal flow reduction (e). DSA after 16 months showed a complete occlusion of the aneurysm and reconstruction of the parent vessel (f).

Statistics

Continuous variables are expressed as means ± standard deviations. Categorical variables are presented as absolute and relative frequencies, unless stated otherwise. Fisher’s exact tests were performed for the comparison of categorical variables between the groups. Continuous variables were tested for normal distribution. Differences between the groups were compared with Student’s t-tests. Statistical significance was accepted at a two-sided p value of <0.05. All data analyses were performed using SPSS Statistics 25TM (IBM Inc., Chicago, IL, USA).

Results

Demographics, flow diverter, aneurysm morphology and location

We identified 71 patients who met the requirements for an iFlow evaluation. In these patients a total of 71 unruptured cerebral aneurysms have been treated by the sole implantation of a flow diverter. Unfortunately, 5 patients (7.0%) had to be excluded because of missing follow-up date, leading to a total collective of 66 patient (50 female and 16 male; median age 55, range 27–80) with 66 treated aneurysms. The morphology of the treated aneurysm was saccular in 61 cases (92.4%) and fusiform in 5 cases (7.6%). 7 aneurysms were retreated by implanting an additional flow diverter, resulting in a total of 73 interventional procedures. Most aneurysms have been treated using the PED and DED, while only a few have been treated with the FRED and p64. Parent vessels in which the FDs were implanted were 3.73 ± 0.59 mm in diameter. The selected FDs were 4.2 ± 0.52 in diameter and 21.24 ± 5.19 mm in length. Demographic data, aneurysm sizes and the different flow diverters applied are presented in detail in Table 1. Of the treated aneurysms 58 (87.9%) were located the anterior circulation and 8 (12.1%) in the posterior circulation (Table 2).

Table 1.

Demographic data and aneurysm and FD information presented as mean±SD.

	PED	Derivo	FRED	p64
Mean age	53.7 ± 12.2	57.4 ± 12.9	56.8 ± 12.4	51.0 ± 11.4
Female/male sex	18/6	18/6	8/4	5/1
aneurysm width [mm]	4.9 ± 2.4	4.8 ± 2.6	4.5 ± 2.5	4.7 ± 1.4
aneurysm length [mm]	4.3 ± 1.8	4.1 ± 2.4	4.3 ± 2.3	3.8 ± 1.4
aneurysm depth [mm]	4.1 ± 1.5	4.6 ± 2.9	4.7 ± 3.8	4.3 ± 1.7
aneurysm neck [mm]	4.1 ± 2.0	3.9 ± 1.9	3.3 ± 1.7	3.9 ± 1.5
aneurysm volume [mm³]	119.8 ± 148.5	199.0 ± 534.7	209.2 ± 501.7	88.1 ± 61.6
vessel diameter [mm]	3.7 ± 0.5	3.7 ± 0.7	3.9 ± 0.6	3.8 ± 0.5
FD diameter [mm]	4.1 ± 0.4	4.2 ± 0.5	4.4 ± 0.7	4.3 ± 0.6
FD length [mm]	19.3 ± 5.1	23.7 ± 4.9	21.3 ± 5.4	19.0 ± 1.5

Table 2.

Aneurysm Location.

	No. of treated aneurysms (%)	PED (n = 24)	Derivo (n = 24)	FRED (n = 12)	p64 (n = 6)
Anterior circulation	58 (87.9)
Supraophthalmic ICA	18 (27.3)	4	7	5	2
Paraophthalmic ICA	19 (28.8)	9	8	1	1
Infraophthalmic ICA	21 (31.8)	7	7	5	2
Posterior circulation	8 (12.1)
VA/PICA	8 (12.1)	4	2	1	1

ICA: internal carotid artery; VA: vertebral artery; PICA: posterior inferior cerebellar artery.

Follow-up

The angiographic follow-up ranged from 3 to 59 months. Aneurysm occlusion was assessed using a five-point scale previously described in literature ²² as follows: Grade 0, no endo-aneurysmal flow changes; grade I, residual filling > 50%; grade II, residual filling < 50%; grade III, near complete occlusion with residual filling at the aneurysm neck; grade IV, complete occlusion. Aneurysms showing grade III and IV were classified as adequately occluded (Figures 1 and 2). Adequate aneurysm occlusion was observed in 51 cases (77.3%) at the first follow-up, compared with 60 cases (90.9%) at the long-term follow-up. The longest time we recorded for an aneurysm to completely occlude in our study was 42 months. Inadequate aneurysm occlusion was noticed in 15 cases (22.7%) at the first follow-up, compared with 6 cases (9.1%) at the long-term follow-up. Four of the insufficiently occluded aneurysms showed a basal vessel emerging from the aneurysm (ophthalmic artery or posterior communicant artery), that however, led to occlusion rates of grade I and II. The two non-occluded aneurysms (Grade 0) were due to growth of a partially thrombosed aneurysm on the one hand and to an incomplete coverage of the aneurysm neck after foreshortening of the FD on the other.

Retreatment in these two cases was recommended but never carried out because the patients decided to seek a second opinion in another hospital. In one case a marginal placement of the FD led to a grade II occlusion at follow-up.

Table 3 gives an overview on the occlusion rates in detail and in relation to the used FDs, as well as the duration of dual antiplatelet therapy.

Table 3.

Five-point grading scheme of aneurysm occlusion and DAPT.

	First follow-up					Last follow-up
OcclusionRate	Treated aneurysms (%)	PED	Derivo	FRED	p64	Treated aneurysms (%)	PED	Derivo	FRED	p64
Grade 0	5 (7.6)	3	1	1	0	2 (3.0)	1	1	0	0
Grade I	6 (9.1)	1	2	1	1	1 (1.5)	1	0	0	0
Grade II	4 (6.1)	1	2	1	0	3 (4.5)	1	1	1	0
Grade III	8 (12.1)	2	5	2	0	3 (4.5)	0	2	1	0
Grade IV	43 (65.1)	17	14	7	5	57 (86.4)	21	20	10	6
Months	3.98 ± 2.04	3.9 ± 1.5	4.4 ± 2.8	3.0 ± 0.4	4.0 ± 1.1	19.1 ± 15.1	18.3 ± 16.8	14.9 ± 11.3	28.1 ± 16.5	23.8 ± 14.3
Months until Grade IV	6.81 ± 6.92	5.6 ± 5.9	7.4 ± 5.0	8.8 ± 13.1	4.0 ± 1.2
Duration of DAPT	8.28 ± 6.43	9.0 ± 4.1	7.0 ± 3.3	10.1 ± 14.3	7.3 ± 3.5

Complications

The complication rates in detail are shown in Table 4. Ischemic events were observed in a total of 5 cases. Most ischemic events were minor strokes without neurological impairment or change in modified Rankin scale (mRS). In one case, during treatment of a left ICA aneurysm, the patient suffered a peripheral occlusion in the parietal segment of the middle cerebral artery. After the intervention the patient reported a discreet right sided hemihypesthesia, MRI revealed a small cortical infarction in the parietal lobe resulting in a postinterventional mRS score of 1. At follow-up the patient recovered fully and the mRS score declined back to 0. In one case a major stroke occurred due to an in-stent-thrombosis followed by an immediate occlusion of the FD after treatment of a right ICA aneurysm. Revascularization maneuvers were unsuccessful, the patient suffered a territorial infarction of the right middle cerebral artery and a complete media syndrome. Recovery from a postinterventional mRS score of 4 was only possible to a mRS score of 3 at last seen follow-up after 18 months.

Table 4.

Complications.

Complication	PED	Derivo	FRED	p64	Total (%)
Stroke	4	1	0	0	5 (7.6)
Stent thrombosis/occlusion	0	1	1	0	2 (3.0)
Relevant Foreshortening	5	0	1	1	7 (10.6)
Retreatment	4	0	2	1	7 (10.6)
Aneurysm growth	1	1	1	0	3 (4.5)
Deployment problems	0	1	0	0	1 (1.5)

The clinical outcome, measured by the mRS scale of the 66 patients during this study, were observed as follows: preinterventional mRS 0 = 61 (92.4%), 1 = 4 (6.1%) and 2 = 1 (1.5%), postinterventional mRS 0 = 59 (89.4%), 1 = 5 (7.6%), 2 = 1 (1.5%) and 5 = 1 (1.5%), follow-up mRS 0 = 60 (90.9%), 1 = 4 (6.1%), 2 = 1 (1.5%) and 4 = 1 (1.5%). Respectively leading to an overall morbidity rate of 1.5% in this study.

In total we observed an acute in-stent-thrombosis followed by immediate occlusion of the FDs in two cases (FRED and DED). In one case, complete revascularization by aspiration thrombectomy was possible; the patient had no clinical sequelae.

In total we observed a foreshortening of the FDs during follow-up in 15 cases (9 PED, 3 DED, 2 FRED, 1 p64). The vessel diameter in these cases were 3.7 ± 0.6 mm and the selected FDs measured a diameter 4.2 ± 0.5 mm and length of 20.9 ± 5.9 mm, respectively. The length of the implanted FDs was 23.8 ± 6.7 mm immediately after deployment and 18.9 ± 6.6 mm at the first follow-up leading to significant foreshortening with a p-value of < 0.001. In 7 cases (5 PED, 1 FRED, 1 p64) the foreshortening observed led to insufficient or marginal overlap of the aneurysm neck. In these cases, the FDs showed a length of 20.7 mm immediately after deployment compared to 15.9 mm at the first follow-up. In three cases (PED, DED and FRED) the DSA follow-up showed growth of the treated fusiform aneurysms. In one case, using a PED, the aneurysm growth was caused by a foreshortening of the FD into the distal portion of a fusiform aneurysm (Figure 3(b) and (d)). In another case, using a FRED, the aneurysm initially revealed an adequate occlusion (grade III) after 6 months (Figure 4(d)). However, after 18 months, significant regrowth from almost the initial aneurysm size was observed (Figure 4(f)). Morphological changes of the implanted FD or foreshortening were not observed in this case. In the third case a partially thrombosed, fusiform aneurysm of the right vertebral artery (V4 segment) was treated with a DED. Follow-up after 6 months revealed a distinct aneurysm growth despite adequate overlapping of the aneurysm. In a total of 7 cases retreatment with another FD due to foreshortening and/or aneurysm growth was performed. Of the retreated aneurysms, 4 (57.1%) were saccular and 3 (42.9%) fusiform in morphology.

Figure 3.

Flat-Panel Computer Tomography (3 D-DYNA-CT) reconstruction of a fusiform aneurysm (11.2 × 7.0 × 7.4 mm) of the right vertebral artery (V4 segment) (a). DSA after 6 months revealed an incomplete occluded aneurysm due to a foreshortening of the implanted PED (3.75 × 20mm) into the distal portion of the aneurysm (b). iFlow image and the TDC showed an increased intra-aneurysmal contrast agent intensity (>100%) and an ID-R of 0.78 compared to the parent vessel (c). DSA after 12 months revealed a slight aneurysm growth and prompted the implantation of a further FD (DED 3.5 × 20 mm) (d and e). iFlow image and TDC after the second FD implantation with an PI-D of 1.7 and ID-R of 5.5 representing an adequate intra-aneurysmal flow reduction (f). MR-Angiography 3 months after placement of the second FD showed a complete occlusion of the aneurysm (g).

Figure 4.

DSA of a fusiform infraophthalmic aneurysm (10.8 × 15.6 × 9.4 mm) of the left internal carotid artery (a). iFlow image and TDC prior FD implantation showing comparable flow patterns (b). iFlow image and the TDC after FD implantation showed an increased intra-aneurysmal contrast agent intensity (>100%) and an ID-R of 0.89 compared to the parent vessel (c). DSA at 6 months showed a promising aneurysm occlusion (d). iFlow image and the TDC also show an increased intra-aneurysmal contrast agent intensity (>100%) and an ID-R of 0.7 (e). DSA at 18 months showed a distinct regrowth of the aneurysm and prompted the implantation of a further FD (DED 4.5 × 20 mm) (f). iFlow image and TDC after the second FD implantation with an PI-D of 1.7 and ID-R of 5.5 representing an adequate intra-aneurysmal flow reduction (g). DSA at 12 months after the second FD implantation showed a complete aneurysm occlusion (h).

Problems during deployment occurred in one case, when a DED (4.5x30mm) was used in a tortious vessel anatomy to treat an ICA aneurysm. Despite a complete expansion of the distal and proximal ends with adequate vessel wall apposition, the stent showed a high-grade constriction in the central section. The constriction could be remedied by a single dilatation with a balloon-catheter (HyperGlide 5 × 15mm, Medtronic/eV3, Irvine, CA, USA).

We observed no periinterventional bleeding or delayed aneurysm rupture or procedural mortality.

iFlow parameters

Prior to the FD implantation the TDCs were comparable with each other in almost all cases demonstrating similar flow conditions within the aneurysm and the parent vessel (Figure 1(b)). Only in a few cases the TTP as well as the remaining course of the curve did differ slightly from one another (Figures 2(c) and 4(b)). When comparing adequately occluded aneurysms (grade III and IV) with insufficiently occluded aneurysms (grade 0-II), we observed a significantly greater peak intensity delay (PI-D, p = 0.008) and intensity decrease ratio (ID-R, p < 0.001). Details are demonstrated in Table 5. The intensity decrease within the aneurysm (ID-IA), measured 3 seconds after the peak intensity of the contrast agent was reached, was significantly lower in adequately occluded aneurysms (p = 0.009). The intensity decrease within the aneurysm (ID-REF) and the peak intensity ratio (PI-R) was not significantly different between both occlusion groups. The duration of DAPT did also not differ significantly between the two occlusion groups. The subgroup analysis of insufficiently occluded aneurysms (Grade 0 vs. Grade I and II) revealed no significant differences for PI-D (0.10 ± 0.14 vs. 0.36 ± 0.43, p = 0.316), PI-R (0.98 ± 0.03 vs. 0.89 ± 0.13, p = 0.260) and ID-R (0.84 ± 0.23 vs. 1.56 ± 0.84, p = 0.208). The subgroup analysis between saccular and fusiform aneurysms showed no significant differences for PI-D (0.95 ± 0.73 vs. 0.76 ± 0.40, p = 0.465) and PI-R (0.76 ± 0.13 vs. 0.78 ± 0.14, p = 0.821). The ID-R of fusiform aneurysms was significantly lower than that of saccular aneurysms (1.32 ± 0.64 vs. 3.77 ± 3.56), resulting in a p-value of <0.001. In three cases of fusiform aneurysm we observed an increased intensity of the intra-aneurysmal contrast agent of more than 100% (133.33 ± 5.77) compared to the parent vessel after FD implantation (Figures 3(c) and 4(c)). In these cases, intensity decrease ratios (ID-R) of less than 1 (0.73 ± 0.07) were observed. The iFlow parameters in the relevant cases of foreshortening changed significantly when the measurements were compared with those seen immediately after the FD implantation. Almost all intra-aneurysmal TDCs normalized and were comparable to the baseline TDCs leading to significant changes of PI-D (0.86 ± 0.31 vs. 0.10 ± 0.22, p = 0.003), PI-R (0.76 ± 0.08 vs. 0.98 ± 0.04, p = 0.001) and ID-R (2.37 ± 0.28 vs. 1.16 ± 0.36, p = 0.002). The subgroup analysis of the iFlow parameters between the different FDs used showed no significant differences between PI-D (p = 0.602), PI-R (p = 0.694) and ID-R (p = 0.957).

Table 5.

Overview of the measured iFlow parameters and period of DAPT.

Occlusion (n)	PI-D	PI-R	ID REF [3sec]	ID IA [3sec]	ID-R	Months of DAPT
Grade III–IV (60)	1.03 ± 0.73	0.76 ± 0.15	83.30 ± 11.89	30.02 ± 14.86	4.20 ± 3.77	8.51 ± 6.90
Grade 0–II (6)	0.47 ± 0.36	0.86 ± 0.11	84.83 ± 7.81	80.17 ± 30.27	1.16 ± 0.36	6.00 ± 3.32
p-Value	0.008	0.100	0.677	0.009	<0.001	0.193

Discussion

The treatment of cerebral aneurysms with flow diverters has developed into a safe and widespread therapeutic concept. Despite the frequent use of these devices, the exact flow changes caused within the aneurysms are not yet fully understood. In addition, it is difficult to make a statement about whether or when the time of complete occlusion of the treated aneurysm can be expected. The reduction of intra-aneurysmal flow remains a central hypothesis for the success after FD treatment. Intra-aneurysmal flow reduction is influenced by a variety of factors (aneurysm size, morphology, localization; FD size and position) and remains a central aspect of many preclinical and clinical studies.^18,23–25 Investigation of the hemodynamic changes within cerebral aneurysms after FD treatment remains desirable, since associated complications such as delayed aneurysm growth or rupture remain topics in current literature.^15,26–28

This study retrospectively examined the predictive value of various iFlow parameters after flow diverter treatment of cerebral aneurysms. Cattaneo et al. previously described promising preclinical evaluation criteria, that are based on a comparison between intraluminal and intra-aneurysmal TDCs, aimed at the difference of maximum contrast agent intensity, delay between intensity peaks and ratio of intensity decrease. ¹⁸ The results of our study show that these parameters could be implemented easily and successfully in a clinical setting. The peak intensity delay (PI-D) in our study revealed significantly greater in adequately occluded aneurysms, indicating a reduced aneurysmal inflow. Furthermore, the observed intensity decrease ratio (ID-R) was significantly higher in adequately occluded aneurysms meaning that less contrast agent was being washed out of the aneurysm over time compared to the parent vessel. These aspects support the theory that decreased intra-aneurysmal flow is due to both a decrease in inflow and outflow and is associated with successful FD treatment. However, the in vivo study of Gölitz et al. showed that a flow diverter mainly influences the intra-aneurysmal outflow and thus promotes aneurysm occlusion. ¹⁹ Despite the fact that their results were preliminary, we believe that two limitations influenced them. First of all, flow changes were only measured intraaneurysmally before and after the implantation of the FD without being related to the target vessel. Secondly, in order to measure these intra-aneurysmal changes, a point of interest was used instead of an area of ​​interest, that can assess only a very small fraction of the flow changes within the aneurysm. These two aspects were taken into account in our study.

The peak intensity ratio (PI-R) observed in this study did not differ significantly, but merely showed a trend between adequately and inadequately occluded aneurysms. Theoretically, this parameter describes the peak intensity ratio between the parent vessel and the aneurysm and should also serve as a marker for decreased aneurysmal inflow. Since this parameter is only a ratio and it has shown fluctuating results in previous reports, its predictive relevance remains uncertain. ¹⁸ The aneurysm occlusion rates in this study were comparable to those reported in the current literature.^4,8,10,29 At the last follow-up we found a persistent patency of the aneurysm in 6 cases. Grade 0 occlusions (2 cases) were due to an incomplete overlap of the aneurysm neck by the implanted FD on the one hand and the growth of the treated aneurysm on the other. In 4 cases, persistent patency of the aneurysm was caused by a branch artery originating from the base of the aneurysm. This is a common finding that should be considered when using FDs and should be avoided if possible, as adequate occlusion cannot be achieved in these cases. ³⁰ Nevertheless, aneurysm occlusion grade II was achieved in these cases in our study. Unfortunately, it turned out that none of the iFlow parameters had a predictive value with regard to the differentiation of an insufficiently occluded aneurysms (grade 0 versus grade I-II).

Another method for evaluating intra-aneurysmal flow modification after FD treatment can be achieved by optical flow imaging. The mean aneurysm flow amplitude (MAFA) analysis is performed by calculating spatial and temporal changes in contrast density between successive images of a DSA run. An in vitro study by Cancelliere et al. showed that a reduced MAFA correlated with a superior flow diversion effect. ³¹ Since the MAFA analysis requires a DSA run of 60 frames/s before and after FD placement, we believe that the iFlow parameters examined in this study are easier to apply given radiation exposure.

The flow diverting effect on intracranial aneurysms can be influenced by many factors. Mechanical aspects such as pore size and pore density play a role as well as the selected FD diameter and the positioning of the FD. ¹⁸ Sizing an FD is essential and should consider two important aspects. On the one hand, sufficient wall apposition must be achieved in order to avoid an endoleak, and on the other hand, pronounced oversizing should be avoided because this can reduce the flow-diverting effect by increasing the size of the stent cells. ¹² An oversizing of 0.25–0.5 mm in relation to the parent vessel has become a well-established compromise in clinical practice. ³² The FDs that were implanted in our study were also oversized with a mean diameter of 0.5 mm compared to the parent vessel. Unfortunately, the subgroup analysis of the iFlow parameters did not reveal any significant advantages in the direct comparison of the various FDs used.

Foreshortening of FDs during or after implantation is a well-known complication that must be considered when using such devices. ³³ Although the devices in our study were relatively oversized, we observed a total of 15 cases of foreshortening, of which only 7 were found to be relevant as they resulted in incomplete or marginal overlap of the aneurysm neck. In these cases, all of the applied iFlow parameters (PI-D, PI-R and ID-R) changed significantly and resembled the normalization of the intra-aneurysmal flow pattern similar to the baseline. As described in the literature, we share the same experience in our study that foreshortenings are more common when using the PED. ³⁴

In three cases of fusiform aneurysms, DSA follow-up revealed growth or regrowth of the treated aneurysms. In these cases, an increased intra-aneurysmal contrast agent intensity of greater than 100% compared to the parent vessel was observed, indicating an increased aneurysmal inflow. The intensity decrease ratio (ID-R) was measured to be less than 1 in these cases and was significantly lower compared to saccular aneurysms (p < 0.001), which also indicates increased aneurysmal outflow. ID-R thus served as a sensitive parameter for a critical flow constellation and potential aneurysm growth. Cattaneo et al. suspected in a preclinical study that increased intra-aneurysmal flow changes are caused by endoleaks after FD implantation and could possibly lead to aneurysm growth or even rupture. ¹⁸ Only in one of the cases observed in our study could the hemodynamic changes that promoted aneurysm growth be associated with an endoleak. The endoleak itself was caused by the displacement of an FD into the distal part of the aneurysm after foreshortening. In the other cases the overlap and apposition of the FDs was appropriate and no endoleak was found angiographically. In both cases, the growth/regrowth of the aneurysm could be associated with thrombus remodeling processes as one aneurysm was partially thrombosed and the other showed regrowth after almost complete occlusion at 18 months. Partially thrombosed cerebral aneurysms are known to have higher growth/re-growth rates than others.^35,36 The exact cause of the increased changes in intra-aneurysmal flow remains uncertain at this time. Hussein et al. also observed increased intra-aneurysmal flow changes after FD treatment in their small cases study. ²⁰ The iFlow TDCs reported in their publication were comparable to those from our study. In contrast to our study, the observed flow changes were not associated with an aneurysm growth, but with the possibility of complete occlusion of the treated aneurysms. A disadvantage of this study was the relatively short observation period of 6 months, as delayed aneurysm growth in our study was observed after 18 months. Further investigations with larger case numbers are needed to validate our findings. Despite the controversial results we believe that an ID-R < 1 can serve as a predictor for potential aneurysm growth during follow-up. Since aneurysm growth can lead to aneurysm rupture and poor clinical outcome, ¹⁵ we believe that the indication for retreatment with another FD should be given generously in the case of increased intra-aneurysmal flow changes. According to the results of this study, we would have treated at least two patients with an additional FD immediately after the periprocedural iFlow assessment.

Most of the aneurysms treated in our study were saccular; fusiform morphology was observed in only 5 cases. Despite this fact, our results show that the treatment of fusiform aneurysms with FDs remains more difficult compared to saccular aneurysms because fusiform aneurysms have the potential to grow in size and thus have shown a higher rate of retreatment. The retreatment rates of cerebral aneurysms in general after FD treatment due to recurrent, remaining aneurysm filling or aneurysm growth are described in the literature between 3 and 16%.^27,37,38 Currently, there is no standard strategy as to whether and when retreatment should be considered, especially given the fact that complete aneurysm occlusion can sometimes take years. Bescke et al. report that complete aneurysm occlusions can take up to 5 years after FD treatment.^4,37 In our study we experienced the latest aneurysm occlusion after a follow-up of 3 ½ years. Retreatment in our study was performed in a total of 7 cases in which foreshortening with insufficient overlap of the aneurysm neck or aneurysm growth/re-growth was observed. Since the iFlow parameters used in these cases were all noticeably different as described above, we believe they can be useful in identifying indications for future retreatment.

Ischemic complications occurred in 7.6% and led to an overall morbidity of 1.5% since one patient suffered a major stroke, after acute thrombosis and occlusion of the FD, with a residual mRS score of 3 at final follow-up after 18 months. During the deployment technical problems were seen in one case as a constriction of the FD occurred in a tortious vessel anatomy. This could be remedied without complications by a one-time balloon angioplasty. The ischemic and technical complications observed in our study are comparable to those reported in the current literature after FD treatment. ²⁹

Limitations

This study is limited due to its retrospective design and the small number of patients who met the criteria for the applied iFlow parameters. The literature shows that hemodynamic analysis of treated cerebral aneurysms using iFlow remains a heterogeneous and highly user-dependent method.^19,20,39,40 The parameters assessed in this study are just one of the many ways to analyze such data, so that a direct comparison of our results with similar studies seems to be difficult. In addition, DSA acquisition was performed by manual injection of a contrast agent rather than mechanically using an injector as this is our clinical standard. For further investigations, a mechanical injector could be used in order to standardize the application of contrast agent. Finally, aneurysm occlusion is multifactorial and depends on more than just hemodynamic changes after FD implantation.

Conclusion

Our study provides insight to hemodynamic changes of cerebral aneurysms after FD treatment using different iFlow parameters. Two of the applied parameters (PI-D and ID-R) were associated with an adequate occlusion rate as they showed a significantly reduced inflow and a delayed outflow of the treated aneurysms. An ID-R of less than 1 was associated with increased aneurysmal inflow and resulted in growth/re-growth of the treated aneurysms during follow-up. The predictive relevance of PI-R remains uncertain because of fluctuating results. Overall, aneurysm occlusion rate, observed complications and retreatment rate were comparable to those reported in current literature. In our opinion, these iFlow parameters served as promising predictors for aneurysm occlusion and can be used in future clinical practice to assess hemodynamic flow changes after FD treatment of cerebral aneurysms. Further validation is needed to determine whether these parameters can help neurointerventionalists to identify retreatment indications.