Abstract
We examine the remodeling of arterial feeders and draining veins following Stereotactic Radiosurgery (SRS) and explore their relationship with radiation-induced edema using retrospective data from 50 patients with cerebral AVMs treated with CyberKnife between 2010 and 2023 at a single center. Univariate analyses were performed. 46% of patients developed post-SRS edema. Patients with edema had larger AVM volumes (4.5 vs. 2.1 cm³; p < 0.01) and showed greater reduction in the diameter of their main draining vein (33% vs. 13%; p < 0.01) and accessory draining vein (24.5% vs. 6%; p < 0.01). Those without edema had a larger reduction in the diameter of the main feeder artery (15% vs. 8%; p = 0.03). Patients with edema showed higher change in resistance to outflow in the main draining vein (406% vs. 71%; p < 0.01) and second largest vein (192% vs. 27%; p < 0.01), while those without edema showed higher resistance to inflow in the arterial feeder (95% vs. 38%; p = 0.03). There were no differences in radiation dosing (p = 0.97), obliteration rates (p = 0.35), or functional outcomes (p = 0.61) at follow-up. Post-SRS edema in AVMs is associated with higher resistance to outflow seen in a disproportionated greater reduction in the size of draining veins compared to arterial feeders.
Keywords: Arteriovenous malformations, draining veins, post-radiation edema, remodeling, stereotactic radiosurgery
Introduction
Cerebral arteriovenous malformations (AVMs) are abnormal tangles of vessels characterized by direct, high-flow connections between arteries and veins within the intracranial circulation that affect 0.01% to 0.05% of the population, with a preference for middle-aged adults. 1 Due to their elevated flow characteristics, untreated cerebral AVMs carry a significantly increased risk of rupture, which has led to a heightened focus on the need for timely intervention 2
Stereotactic radiosurgery (SRS) is a treatment option for AVMs located in deep-structures or eloquent cortices, where microsurgical resection carries higher risks and potential complications. 3 SRS facilitates the gradual obliteration of the AVM by inducing progressive fibrosis and narrowing of the vessels, altering hemodynamics. 4 Previous studies using brain MRI/MRA have commonly reported T2-weighted and FLAIR hyperintensities, representing varying degrees of cerebral edema in the treated regions.5–7
While certain studies have quantified flow alterations in arterial feeders and draining veins using 4D-flow magnetic resonance imaging, 8 and others have investigated the characteristics of draining veins in relation to hemorrhagic risk,9,10 limited attention has been given to the remodeling of draining veins and its potential role in the development of post-radiation edema. We aim to explore the relationship of arterial and venous remodeling post SRS to the development of post SRS edema.
Methods
Study design
Following approval by the Institutional Review Board of Beth Israel Deaconess Medical Center, a retrospective study of patients with brain AVMs treated with SRS between 2010 and 2023 was performed at a single academic institution. Patients with AVMs were reviewed at a multidisciplinary conference and options of radiosurgery, surgical resection, embolization or combined modality therapy were considered. The decision to treat the current cohort of patients with radiosurgery was reached by our multidisciplinary team. Patients without imaging available pre and post-SRS were excluded. Additionally, patients whose scans were of insufficient quality to visualize the AVM vessels and those with significant artifact were excluded. From the original 72 treated with SRS 22 patients were ultimately excluded. All radiology reports pertaining to pre and post-treatment were reviewed. A patient was included in the edema group only if the presence of edema was documented in an official radiology report post-SRS. The median time between SRS and the development of edema was used to choose the time point of the post-SRS scan used for comparison in the non-edema group. When measuring the diameter of vessels anatomical landmarks were used to ensure that measurements were taken as close to the same location as possible. After identifying the diameter, changes in vessel resistance were calculated, taking into consideration that resistance has an inverse relationship to the fourth power of the radius (R = 8ηL/π r4) Likewise, the presence of venous stenosis was determined through the official radiology reports. The study complied with the Declaration of Helsinki. The need for patient consent was waived due to the study’s retrospective nature.
Data source and variables
The variables collected included patient demographics (age, gender, race, smoking status, comorbidities), details of clinical presentation (modified Rankin scale (mRS), symptoms at presentation), characteristics of the AVM (side, size, location, Spetzler-martin grade, presence of associated aneurysms, prior treatments, characteristics of the feeding arteries and draining veins), procedural specifics (treatment dates, type, and detailed procedure descriptions), radiographic changes along treatment course (presence or absence of edema, change in size of draining veins and/or feeding arteries), radiographic outcomes (duration of imaging follow up, obliteration rates), and functional outcomes (duration of clinical follow-up, mRS scores, mortality). The obliteration rate for this study was assessed from the follow-up imaging modality available for each patient and was characterized into the following categories: 100%, 90–99%, 50–89%, and <50%.
Statistical analysis
Observations were defined as patients, AVMs, vessels, or procedures, as applicable. Categorical variables were compared using the [chi] 2 test and continuous variables with the Mann-Whitney U test. Univariate analysis was completed to compare demographic, clinical, AVM, and procedural characteristics between those who developed edema and those who did not. A p-value <0.05 was set for statistical significance. All statistical analyses were performed using Stata 14 software (StataCorp).
Results
Baseline patient characteristics
A total of 50 patients with brain AVMs treated with SRS were included in this study. The median age was 52.5 (IQR = 36–62), and 56.0% were female. The cohort was predominantly White (72.3%), followed by Black and Hispanic (12.8% each), and Asian (2.1%). 16.0% were current smokers, and 36.0% were former smokers. 38.0% had a history of hypertension, 16.0% cardiovascular disease, and 10.0% diabetes (Table 1).
Table 1.
Sample and presentation AVM characteristics.
| Sample and presentation AVM characteristics | n = 50 |
|---|---|
| Female | 28 (56.0%) |
| Age, median (IQR) | 52.5 (36–62) |
| Race (a) | |
| White | 34 (72.3%) |
| Black | 6 (12.8%) |
| Hispanic | 6 (12.8%) |
| Asian | 1 (2.1%) |
| Smoking | |
| Current | 8 (16.0%) |
| Former | 18 (36.0%) |
| Never | 24 (48.0%) |
| Cardiovascular Disease | 8 (16.0%) |
| Hypertension | 19 (38.0%) |
| Diabetes | 5 (10.0%) |
| Rupture prior to SRS | 20 (40.0%) |
| Seizures at presentation | 10 (20.0%) |
| Location | |
| Frontal | 9 (18.0%) |
| Parietal | 10 (20.0%) |
| Temporal | 6 (12.0%) |
| Occipital | 2 (4.0%) |
| Cerebellar | 9 (18.0%) |
| Thalamic | 3 (6.0%) |
| Cingulate gyrus | 1 (2.0%) |
| Choroid Plexus | 1 (2.0%) |
| Combination lobar | 9 (18.0%) |
| Side | |
| Right | 23 (46.0%) |
| Left | 23 (46.0%) |
| Midline | 4 (8.0%) |
| Deep | 4 (8.0%) |
| Grade | |
| 1 | 11 (22.0%) |
| 2 | 20 (40.0%) |
| 3 | 13 (26.0%) |
| 4 | 6 (12.0%) |
| Volume of Nidus (cubic cm), median (IQR) (b) | 3.3 (1.4–4.8) |
| Nidal aneurysm | 6 (12.0%) |
| Drainage | |
| Superficial | 26 (52.0%) |
| Deep | 16 (32.0%) |
| Both | 8 (16.0%) |
| Draining Vein Stenosis | 3 (6.0%) |
| Missing | |
| (a) 3 | |
| (b) 1 |
Initial presentation and AVM description
40% of the AVMs were found to be ruptured prior to SRS treatment. 20% manifested with seizures. The distribution of AVMs was similar on the right or left sides (46.0% each), and midline AVMs were present in 8.0% of the cases. 18.0% of patients had Infratentorial AVMs. The most common location was parietal (20.0%), followed by frontal and cerebellar (18.0% each), temporal (12.0%), thalamic (6.0%), occipital (4.0%), and lastly, choroid plexus and cingulate gyrus (2.0% each). 8.0% of the AVMs were deep-seated. Most of the AVMs were Spetzler Martin grade 2 (40.0%), followed by grade 3 (26.0%), grade 1 (22.0%), and grade 4 (12.0%). The median nidal volume measured prior to the initiation of SRS was 3.3 cm3 (IQR = 1.4–4.8). The venous drainage pattern was predominantly superficial (52.0%). 32.0% of AVMs had deep venous drainage, and 16.0% had both superficial and deep draining veins. Venous stenosis was only present in 6% of the AVMs. A peri-nidal aneurysm was seen in 12.0% of the AVMs (Table 1).
Blood supply and venous drainage
Main arterial feeder (feeding artery #1)
The median diameter of the largest feeding artery of the AVMs pre-SRS was measured to be 1.55 mm (IQR = 1.3–1.8). MCA and PCA branches were the most common primary feeding arteries (26.0% each). ACA branches were the second most common (18.0%), followed by SCA (16.0%). Pericallosal, and PCOM branches were equally prevalent (4.0%). 1% were fed from PICA, and 4.0% from other arteries.
Accessory arterial feeder (feeding artery #2)
The median diameter of the second largest feeding artery of the AVMs pre-SRS was measured to be 1.2 mm (IQR = 1.1–1.5). MCA branches were the most common secondary feeding arteries (33.3%), followed by SCA and “other” (20.0% each), PCA (13.3%), and lastly, ACA or PICA (6.7% each).
Main draining vein (draining vein #1)
The median diameter of the largest draining vein of the AVMs pre-SRS was measured to be 2.4 mm (1.8–3.1). The most common primary pathway of venous drainage was seen to be either through the superior sagittal sinus (14.0%), internal cerebral vein (14.0%), cortical vein (14.0%), or “other” (14.0%). This was followed by the transverse sinus (12.0%), vein of Galen (8.0%), vein of Labbé (8.0%), basal vein of Rosenthal (6.0%), straight sinus (4.0%), sylvian vein (2.0%), vein of Trolard (2.0%), and lastly through the sigmoid sinus (2.0%).
Accessory draining vein (draining vein #2)
The median diameter of the accessory draining vein of the AVMs pre-SRS was measured to be 1.9 mm (IQR = 1.5–2.7). The most common secondary pathway of venous drainage was seen to be through a cortical vein (31.6%). Next was the transverse sinus (15.8%), “other” (15.8%), superior sagittal sinus (10.5%), internal cerebral vein (10.5%), vein of Labbé (10.5%), and lastly through the vein of Trolard (5.3%) (Table 2).
Table 2.
Treatment and outcomes.
| Treatment and outcomes | n = 50 |
|---|---|
| Prior microsurgical resection | 2 (4.0%) |
| Preoperative Embolization | 5 (10.0%) |
| Cyberknife radiation | 50 (100.0%) |
| Dose (Gy) median (IQR) | 20 (20–22) |
| Edema | 23 (46.0%) |
| Percent decrease in diameter Feeding artery #1 median (IQR) | 13 (22–7) |
| Percent decrease in diameter Feeding artery #2 median (IQR) | 14 (22–0) |
| Percent decrease in diameter Draining Vein #1 median (IQR) | 18 (33–11) |
| Percent decrease in diameter Draining Vein #2 median (IQR) | 12 (24–6) |
| Percent increase in resistance Feeding artery #1 median (IQR) | 68 (32–173) |
| Percent increase in resistance Feeding artery #2 median (IQR) | 0 (0–162) |
| Percent increase in resistance Draining Vein #1 median (IQR) | 114 (62–406) |
| Percent increase in resistance Draining Vein #2 median (IQR) | 65 (27–197) |
| Days between first and second timepoint median (IQR) | 263 (217–390) |
| Post SRS rupture | 0 |
| Months until final imaging median (IQR) | 24 (12–40) |
| Obliteration status at last imaging | |
| 100% | 17 (34.0%) |
| 90–99% | 10 (20.0%) |
| 50–89% | 17 (34.0%) |
| <50% | 6 (12.0%) |
| Months until last clinical follow up median (IQR) | 24 (12–38) |
| mRS last follow up | |
| 0 | 27 (54.0%) |
| 1 | 12 (24.0%) |
| 2 | 8 (16.0%) |
| 3 | 1 (2.0%) |
| 4 | 1 (2.0%) |
| 5 | 1 (2.0%) |
| 6 | 1 (2.0%) |
Detailed descriptions of the blood supply and venous drainage of the AVMs can be found in supplementary table 1.
Treatment course and outcomes
4% of AVMs had been previously treated with microsurgery, and 10% underwent pre-operative embolization prior to SRS. All patients (100.0%) received CyberKnife radiation with a median dose of 20 Gy (IQR = 20–22). 46.0% developed post-radiation edema during the latency period. The median time between pre-SRS imaging and the scan used to measure the vascular changes was 263 days (IQR = 217–390). The median percent decrease in the diameter of the largest feeding artery (FA #1) was 13% (IQR = 22%–7%), resulting in a median resistance increase of 68% (IQR = 32%–173%). The median percent decrease in the diameter of the second largest feeding artery (FA #2) was 14% (IQR = 22%–0%), resulting in a median resistance increase of 0% (IQR = 0%–162%). The median percent decrease in the diameter of the largest draining vein (DV #1) was 18% (IQR = 33%–11%), resulting in a median resistance increase of 114% (IQR = 62%–406%). The median percent decrease in the diameter of the second largest draining vein (DV #2) was 12% (24%–6%), resulting in a median resistance increase of 65% (IQR = 27%–197%). Zero patients experience post-SRS hemorrhages throughout the latency period. The median time between treatment and the last imaging follow-up was 24 months (IQR = 12–40). At the last imaging follow-up, 34.0% showed complete obliteration of their AVM, 20.0% showed 90–99% obliteration, 34.0% showed 50–89% obliteration, and 12.0% showed less than 50% decrease in AVM volume. The median time until the last clinical follow-up was 24 months (12–38). At this time 54.0% of patients had an mRS of 0, 24.0% had an mRS of 1, 16.0% had an mRS of 2, 2.0% had an mRS of 3, 2.0% had an mRS of 4, 2.0% had an mRS of 5, 2.0% had an mRS of 6. The patient with an mRS of 6 expired due to reasons unrelated to their AVM and their treatment for it (Table 3).
Table 3.
Comparison of baseline & AVM characteristics.
| Comparison of baseline & AVM characteristics | |||
|---|---|---|---|
| Variable | No edema n = 27 (54.0%) | Edema n = 23 (46.0%) | p value |
| Female | 16 (59.3%) | 12 (52.2%) | 0.62 |
| Age, median (IQR) | 57 (40–63) | 44 (30–62) | 0.22 |
| Race (a) | 0.67 | ||
| White | 18 (72.0%) | 16 (72.7%) | |
| Black | 4 (16.0%) | 2 (9.1%) | |
| Hispanic | 3 (12.0%) | 3 (13.6%) | |
| Asian | 0 | 1 (4.6%) | |
| Smoking | 0.18 | ||
| Current | 2 (7.4%) | 6 (26.1%) | |
| Former | 10 (37.0%) | 8 (34.8%) | |
| Never | 15 (55.6%) | 9 (39.1%) | |
| Cardiovascular disease | 5 (18.2%) | 3 (13.0%) | 0.6 |
| Hypertension | 13 (48.2%) | 6 (26.1%) | 0.11 |
| Diabetes | 4 (14.8%) | 1 (4.4%) | 0.22 |
| Rupture prior to SRS | 14 (51.9%) | 6 (26.1%) | 0.06 |
| Seizures at presentation | 5 (18.5%) | 5 (21.7%) | 0.78 |
| Location AVM | 0.31 | ||
| Deep | 2 (7.4%) | 2 (8.7%) | 0.87 |
| Grade | 0.23 | ||
| 1 | 6 (22.2%) | 5 (21.7%) | |
| 2 | 13 (48.2%) | 7 (30.4%) | |
| 3 | 7 (25.9%) | 6 (26.1%) | |
| 4 | 1 (3.7%) | 5 (21.7%) | |
| Volume of Nidus (cubic cm), median (IQR) | 2.1 (1.3–3.6) | 4.5 (1.6–7.2) | <0.01 |
| Nidal aneurysm | 2 (7.4%) | 4 (17.4%) | 0.23 |
| Drainage | 0.06 | ||
| Superficial | 13 (48.2%) | 13 (56.5%) | |
| Deep | 12 (44.4%) | 4 (17.4%) | |
| Both | 2 (7.4%) | 6 (26.1%) | |
| Draining vein stenosis | 2 (7.4%) | 1 (4.4%) | 0.65 |
Comparison of baseline characteristics between edema and no-edema groups
23 individuals (46.0%) developed edema during the latency period. There was no difference in gender (p = .62), median age (p = .22), race (p = .67), smoking history (p = .18), cardiovascular disease (p = .6), hypertension (p = .11), or diabetes (p = .22) (Table 3).
Comparison of AVM characteristics between those that did and did not develop edema
There was no significant difference between the two groups in pre-SRS rupture (p = .06), seizure at presentation (p = .78), AVM location (p = .31), laterality of AVM (p = .94), presence of deep-seated AVMs (p = .87), and Spetzler martin grade (p = .23). Those that developed edema tended to have AVMs with significantly median larger volumes measured (4.5 cm3 vs. 2.1 cm3; p < .01). There was no difference in the presence of peri-nidal aneurysms (p = .23), deep vs superficial drainage (p = .06), or presence of draining vein stenosis (p = .65) (Table 4).
Table 4.
Treatment course.
| Treatment course | |||
|---|---|---|---|
| variable | No edema n = 27 (54.0%) | Edema n = 23 (46.0%) | p value |
| Prior microsurgical resection | 1 (3.7%) | 1 (4.4%) | 0.91 |
| Preoperative Embolization | 3 (11.1%) | 2 (8.7%) | 0.78 |
| Cyberknife radiation | 27 (100.0%) | 23 (100.0%) | |
| Dose (Gy) median (IQR) | 20 (20–22) | 20 (20–24) | 0.97 |
| Single stage | 27 (100.0%) | 23 (100.0%) | |
| % Decrease in Diameter of Feeding Artery #1 between timepoints | 15 (25–9) | 8 (19–0) | 0.03 |
| % Decrease in Diameter of Feeding Artery #2 between timepoints | 21 (22–0) | 8 (22–0) | 0.7 |
| % Decrease in Diameter of Draining Vein #1 between timepoints | 13 (14–8) | 33 (41–23) | <0.01 |
| % Decrease in Diameter of Draining Vein #2 between timepoints | 6 (9–0) | 24.5 (36–2) | <0.01 |
| % increase in resistance Feeding artery #1 between timepoints | 95 (46–216) | 38 (0–129) | 0.03 |
| % increase in resistance Feeding artery #2 between timepoints | 81 (0–167) | 0 (0–144) | 0.45 |
| % increase in resistance Draining Vein #1 between timepoints | 71 (38–85) | 406 (178–735) | <0.01 |
| % increase in resistance Draining Vein #2 between timepoints | 27 (0–46) | 192 (144–486) | <0.01 |
| Days between first and second timepoint median (IQR) | 258 (224–407) | 268 (180–367) | 0.31 |
Comparison of treatment course and outcomes between edema and no-edema groups
There was no difference observed between the two groups in the rates of prior microsurgical resection (p = .91), pre-operative embolization (p = .78), and dose of radiation (p = .97). However, the diameter of the largest feeding artery decreased significantly more between the two timepoints in patients that did not develop edema (15% vs. 8%; p = .03). This reflected an increase in resistance of 38% in patients who developed edema, compared to a 95% increase in those who did not (p = .03). There was no difference in the change in diameter or resistance of the second largest feeding artery between the two groups (p = .7; p = .45). The diameter of the largest draining vein was seen to decreased significantly more between the two timepoints in patients that developed edema (33% vs. 13%; p < .01) (Figure 1). The median resistance to flow in the largest draining vein was seen to increase significantly more between the two timepoints in patients that did develop edema (406% vs 71%; p < .01). The second largest draining vein had a median decrease of 24.5% (36%–2%) in the post-radiation edema group, compared to 6% (9%–0%) in those that did not develop edema (p < .01). There was also a significant increase in resistance to flow in the second largest draining vein in patients that developed edema compared to those that did not (192% vs 27%; p < .01). The median time of radiographic follow-up to evaluate the change in vessel diameter was 268 days (IQR = 180–367) in patients with edema, compared to 258 days (IQR = 224–407) in those without edema (p = .31). The two groups did not differ in the time between treatment and final follow-up imaging (p = .98), obliteration status at last follow up (p = .35), timing of last clinical follow up (p = .61). Likewise, there was no significant difference between the two groups in their functional outcome measured by mRS at last follow up (p = .61) (Table 5). Only 1 patient without edema expired during the follow up period due to factors unrelated to their treatment of their AVM. Patients that developed edema were treated with a course of steroids, which was managed at the discretion of the radiation oncologist. A comparison of change in DV1 and FA1 shown in Figure 2.
Figure 1.
Adult patient with PMHx significant for EtOH cirrhosis and seizures presented as a transfer after having multiple witnessed seizures. The workup included CTA (left) and cerebral angiogram, which showed a right parietal AVM. The patient was treated with cyberknife radiosurgery on 12/10/2018. He presented to the ED on 4/20/19 with 5 days of progressive left-sided weakness. CT outside the hospital showed significant right-sided vasogenic edema, and he was transferred to BIDMC for neurosurgery evaluation. Prior to transfer patient was given 10 mg decadron. CT/CTA taken on 4/23/19 (above and right) demonstrates the edema as well as significant reduction in the vein draining into the superior sagittal sinus. The patient was placed on dexamethasone 4 mg q6h. His symptoms improved, and he was discharged on 4/26/19 with a month-long steroid taper.
Table 5.
Outcomes.
| Outcomes | |||
|---|---|---|---|
| variable | No edema n = 27 (54.0%) | Edema n = 23 (46.0%) | p value |
| Months until final imaging median (IQR) | 24 (12–45) | 24 (12–36) | 0.98 |
| Obliteration status at last imaging | 0.35 | ||
| 100% | 11 (40.7%) | 6 (26.1%) | |
| 90–99% | 3 (11.1%) | 7 (30.4%) | |
| 50–89% | 10 (37.0%) | 7 (30.4%) | |
| <50% | 3 (11.1%) | 3 (13.0%) | |
| Months until last clinical follow up median (IQR) | 24 (12–45) | 24 (12–36) | 0.61 |
| mRS last follow up | 0.61 | ||
| 0 | 14 (51.9%) | 13 (56.5%) | |
| 1 | 5 (18.5%) | 7 (30.4%) | |
| 2 | 5 (18.5%) | 3 (13.0%) | |
| 3 | 1 (3.7%) | 0 | |
| 4 | 1 (3.7%) | 0 | |
| 5 | 0 | 0 | |
| 6 | 1 (3.7%) | 0 | |
Figure 2.
Slope chart showing the imbalance in the percent change of diameter between the arterial feeders and venous drainage post SRS in all patients that developed edema vs those that did not.
Discussion
In this study, we show that the development of post SRS edema appears to be related to a disproportionate decrease in the diameter of the draining veins to the diameter of the arterial feeders resulting in a significantly greater resistance to blood flow in the veins draining the AVM compared to within its arterial feeders. We hypothesize that these changes set the stage for imbalance between the inflow and outflow and resultant edema. In those patients that developed edema, their largest two draining veins decreased significantly more between pre-SRS and the development of edema compared to the patients that did not experience edema (p < .01). However, in contrast to the phenomenon observed in veins, cases of edema exhibited a significantly smaller reduction in the size of arterial feeders (p = .03). Furthermore edema was more likely to occur in larger AVMs. These findings occurred irrespective of gender, age, comorbidities, or location of the AVM. Despite the development of post SRS edema there was no difference in functional outcomes measured by mRS, and obliteration rates at last follow up.
SRS induces damage to the endothelial lining of the blood vessels that make up the AVM. Intimal fibrosis and hyalinization of vessel walls lead to the narrowing of the vessel lumen. In larger vessels, it is postulated that occlusion may ultimately be the result of smooth muscle cell proliferation in response to this endothelial injury. 4 Following SRS, patients remain in the latent phase lasting from two to five years post-intervention during which the AVM gradually shrinks.11–13 SRS-induced injury is known to potentially lead to vasogenic edema, with varying degrees of presentation ranging from headache to neurological deficits as well as seizures.6,14 The marginal SRS dosage, the nidal volume, and the number of arterial feeders have all been associated with radiation-induced edema.5,7,15 Nonetheless, the degree of vessel remodeling has been hypothesized to play a role in this process.
Edema could be explained by the changes in hydrostatic forces. The Bernoulli equation for fluid dynamics states that the sum of pressure energy, the kinetic energy per unit volume, and the potential energy per unit volume at one point in the fluid flow equals the same sum at a second point. Therefore, as the diameter of the venous drainage shrinks disproportionately to the arterial feeders, one would expect pressure to increase and flow velocity to decrease in the arterial feeders. In the veins, velocity must increase, and pressure must decrease. Following Ohm's law of hemodynamics and shown in our analysis, as the vessels in the AVM align like resistors in series, a decrease in the venous diameter increases outflow resistance, leading to a further buildup of flow and pressure in the nidus. The starling equation shows that the net flow of fluid across a semipermeable membrane can be explained by the vessel lumen-interstitial hydrostatic pressure gradient subtracted by the net difference in osmotic pressures on either side of the membrane. As hydrostatic pressure increases in the arterial feeders of the AVM the starling equation favors fluid extravasation into the intersitium, manifesting as edema. While this process is present in all AVMs it occurs on a larger scale in AVMs with larger volumes due to their increased perfusion. This could explain the increased likelihood of edema in larger AVMs that was observed in our cohort. It may be that the decrease in arterial size serves as a preventive factor against radiation-induced edema.
Our findings regarding the description of the changes in the vascular supply of AVMs post-SRS align with previous work done by Srinivas et al., in which they used 4 D flow MRI to characterize hemodynamic changes in AVMs before and after SRS. In this study, the authors showed that decreased flow within the feeding artery and draining vein precede structural changes in arterial circumference. However, unlike the arterial feeders the venous circumference was shown to decrease at early time points post SRS. 8 Furthermore, this study also found that the pulsatility index (PI) in the primary feeding artery increased significantly from pre-SRS to post-SRS measurements. The PI reflects variations in blood flow between systole and diastole and serves as an indicator of downstream arterial resistance as well as the compliance of large cerebral arteries.16–19 An increase in PI can result in endothelial dysfunction, blood-brain barrier disruption, decreased perfusion during diastole, and increased endothelial sheer stress. 20 While the overall blood flow to the AVM decreases post-SRS, the increase in pulsatility index indicates a change in vessel wall properties, such as increased stiffness or decreased compliance. These structural changes enhance pulsatile flow transmission even as the volume of flow diminishes, highlighting that hemodynamic responses involve both flow reduction and vessel wall remodeling.
The hemodynamic changes from decreased venous outflow may lead to edema through the impairment of cerebrospinal fluid (CSF) flow. The cerebral vasculature is bathed in CSF, and two mechanisms for the role of cerebral arteries in its drainage have been postulated. Perivascular pumping, proposed by Hadaczek et al., refers to spontaneous dilations and constrictions that occur along artery walls in synchrony with the cardiac cycle. 21 Supporting this theory, Iliff et al. showed that macromolecules injected into the CNS in rat models were transported further in those with beating hearts compared to those whose hearts has been recently stopped. 22 However, this theory has been faced with a degree of opposition, as some studies claim that the dilations and contractions seen in perivascular pumping are too weak to affect CSF flow. These critics believe that CSF flow across the cerebral arteries occurs due to spontaneous low-frequency contractions and dilations of cerebral arteries caused by vascular smooth muscle cells independent of the cardiac cycle.23,24 Veluw et al. even showed that increasing the amplitude of spontaneous vasomotions leads to increased clearance. 24 Similarly, post-SRS, respiration-induced pressure changes, imbalances in CSF production and absorption (secondary to microglial inflammatory and pro-angiogenic endothelial cell dysfunction), and elevated CSF protein levels may contribute to the development of post-radiation edema.14,25–27 The true mechanism may be a combination of those described here.
To sum up, we postulate that the relationship between the degree of venous changes compared to the arterial feeders that occurred in the patients that experienced edema results from an interplay of local impairment in the hydrostatic pressure, PI, and CSF flow across the arteries. First, the decrease in venous diameter leads to an increase in the hydrostatic pressure in the nidus. Already this increase in hydrostatic pressure would favor the movement of solutes and fluid outside of the arterial lumen. Next, as previously described, the overall increase in PI that is observed in the feeding arteries post-SRS correlates with increased arterial endothelial damage and reduced compliance of the vessel. Reduced arterial compliance is thought to reduce the artery wall motion caused by the cardiac cycle thus impairing perivascular pumping. Additionally, the damage to the feeding vessel caused by the decrease in venous drainage, in addition to the original radiation, leads to further smooth muscle cell proliferation and dysfunction in the arteries. This can potentially result in impairment of cerebral vessel autoregulation as well as the vessel’s ability to perform spontaneous vasomotion, impairing CSF drainage. These various factors work concurrently to create a local environment where CSF reabsorption is in disequilibrium, contributing to the development of edema. This builds upon the concept of occlusive hyperemia first described post microsurgical resection of AVMs and later applied to SRS by Chapman et al. in which venous occlusion was deemed responsible for increased edema.28,29
Post SRS edema is usually managed medically with steroids or in refractory cases, bevacizumab, a monoclonal antibody targeting vascular endothelial growth factor that has been shown to be efficacious.30,31 In our cohort, there were no statistical differences in the outcome between those who developed edema and those who did not in terms of both AVM obliteration and mRS at the last follow-up. Still, other studies have reported that symptoms of edema can be irreversible, causing significant impairment in the functionality of patients and their relatives. 6 There are no current guidelines on the optimal imaging follow-up timings for these patients, but radiation-induced imaging changes appear to peak at 12 months post-radiation. 7 In the absence of symptoms our institutional practice is to scan at 6 months and then again 12 months for 2 years. Our work suggests that around 9 months post-SRS patients are most likely to develop edema.
Lastly, the findings of our study may offer valuable insights for optimizing radiation planning in the treatment of AVMs with SRS. Given that a reduction in arterial size appears to serve as a protective factor against radiation-induced edema in AVMs, it may be advantageous to emphasize the inclusion of the arterial end over the venous drainage in the radiation treatment plan. Further investigation is needed to assess the impact of marginal dosing, the extent of feeder inclusion, and the exclusion of veins on the overall efficacy and benefits of preventing radiation-induced imaging changes, including edema. These modifications require further exploration in larger, prospective cohorts and, ideally, in clinical trials.
Limitations
This study is limited by its retrospective nature and all its inherent design biases. The sample size is relatively small, which limits the ability to detect statistical significance and prevented us from performing a multivariate analysis to control for other influencing variables. Furthermore, a significant limitation is the lack of standardization in the type of scans used to measure changes in the vessel diameter. The original purpose of the scans was not to detect differences in vessel diameter, and thus, there was no standardization in time between scans or the specific imaging modalities used. Vessel diameter changes were evaluated through side-by-side comparisons of the different time points. Anatomical landmarks were used to ensure that the vessels were measured in the same location in both scans. However, despite these efforts, it is not possible to fully account for the slight variations that differences in the quality of scans, slice gantry angles, or imaging modalities may cause. Lastly, all patients in this study underwent treatment with Cyberknife, and thus, the results may not be applicable to other radiation modalities.
Conclusions
In this single-center retrospective study, we found that the development of post-SRS edema in AVMs treated with Cyberknife was associated with a temporal disproportionately greater reduction in the size of the draining veins in relation to the reduction in size of the arterial feeders. Furthermore the AVMs that developed edema displayed a significantly larger increase in the resistance to blood flow in both their largest and secondary draining vein. This information reinforces the idea that during SRS planning clinicians should preferentially target arterial feeders and avoid draining veins. Additionally these findings may allow clinicians to inform patients of their risk for developing edema during interval follow ups by evaluating the change in diameter of draining veins relative to arterial feeders, which impact overall vessel resistance. Patients could then be more vigilant to the potential development of symptoms and could modify their lifestyle.
Supplemental Material
Supplemental material, sj-pdf-1-jcb-10.1177_0271678X251358986 for Reversible edema after radiosurgery for arteriovenous malformations (AVMs): Inflow and outflow imbalance by Daniel Sconzo, Felipe Ramirez-Velandia, Alejandro Enriquez-Marulanda, Coleman P Riordan, Sandeep Muram, Nima Aghdam, Philipp Taussky and Christopher S Ogilvy in Journal of Cerebral Blood Flow & Metabolism
Footnotes
Funding: The author(s) received no financial support for the research, authorship, and/or publication of this article.
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Authors’ contributions: Daniel Sconzo – Concept, data collection, analysis, drafting manuscript, reviewing final manuscript
Felipe Ramirez-Velandia – Concept, data collection, analysis, drafting manuscript, reviewing final manuscript
Alejandro Enriquez-Marulanda – Concept, analysis, reviewing final manuscript
Coleman P. Riordan – Data collection, reviewing final manuscript
Sandeep Muram–Data collection, reviewing final manuscript
Nima Aghdam – Concept, reviewing final manuscript
Philipp Taussky – Concept, reviewing final manuscript
Christopher S. Ogilvy – Concept, reviewing final manuscript
Supplementary material: Supplemental material for this article is available online.
ORCID iD: Daniel Sconzo https://orcid.org/0000-0001-8933-9083
Ethical considerations
The study was approved by the IRB and given exemption status due to its retrospective nature.
References
- 1.Fleetwood IG, Steinberg GK. Arteriovenous malformations. Lancet 2002; 359: 863–873. [DOI] [PubMed] [Google Scholar]
- 2. Natural history of cerebral arteriovenous malformations: a meta-analysis in. Journal of Neurosurgery 2013; 118: 437–443. [DOI] [PubMed] [Google Scholar]
- 3.Derdeyn CP, Zipfel GJ, Albuquerque FC, et al. Management of brain arteriovenous malformations: a scientific statement for healthcare professionals from the American Heart Association/American Stroke Association. Stroke 2017; 48: e200–e224. [DOI] [PubMed] [Google Scholar]
- 4.Schneider BF, Eberhard DA, Steiner LE. Histopathology of arteriovenous malformations after gamma knife radiosurgery. J Neurosurg 1997; 87: 352–357. [DOI] [PubMed] [Google Scholar]
- 5.Daou BJ, Palmateer G, Wilkinson DA, et al. Radiation-induced imaging changes and cerebral edema following stereotactic radiosurgery for brain AVMs. AJNR Am J Neuroradiol 2021; 42: 82–87. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Yen CP, Matsumoto JA, Wintermark M, et al. Radiation-induced imaging changes following gamma knife surgery for cerebral arteriovenous malformations. J Neurosurg 2013; 118: 63–73. [DOI] [PubMed] [Google Scholar]
- 7.Cohen-Inbar O, Lee CC, Xu Z, et al. A quantitative analysis of adverse radiation effects following gamma knife radiosurgery for arteriovenous malformations. J Neurosurg 2015; 123: 945–953. [DOI] [PubMed] [Google Scholar]
- 8.Srinivas S, Retson T, Simon A, et al. Quantification of hemodynamics of cerebral arteriovenous malformations after stereotactic radiosurgery using 4D flow magnetic resonance imaging. J Magn Reson Imaging 2021; 53: 1841–1850. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Alexander MD, Cooke DL, Nelson J, et al. Association between venous angioarchitectural features of sporadic brain arteriovenous malformations and intracranial hemorrhage. AJNR Am J Neuroradiol 2015; 36: 949–952. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Nico E, Hossa J, McGuire LS, et al. Rupture-Risk stratifying patients with cerebral arteriovenous malformations using quantitative hemodynamic flow measurements. World Neurosurg 2023; 179: 68–76. [DOI] [PubMed] [Google Scholar]
- 11.Moorthy R, Rajshekhar V. Stereotactic radiosurgery for intracranial arteriovenous malformations: a review. Neurol India 2015; 63: 841–851. [DOI] [PubMed] [Google Scholar]
- 12.Maruyama K, Kawahara N, Shin M, et al. The risk of hemorrhage after radiosurgery for cerebral arteriovenous malformations. N Engl J Med 2005; 352: 146–153. [DOI] [PubMed] [Google Scholar]
- 13.Shin M, Kawamoto S, Kurita H, et al. Retrospective analysis of a 10-year experience of stereotactic radio surgery for arteriovenous malformations in children and adolescents. J Neurosurg 2002; 97: 779–784. [DOI] [PubMed] [Google Scholar]
- 14.Kano H, Flickinger JC, Tonetti D, et al. Estimating the risks of adverse radiation effects after gamma knife radiosurgery for arteriovenous malformations. Stroke 2017; 48: 84–90. [DOI] [PubMed] [Google Scholar]
- 15.Yan D, Chen Y, Li Z, et al. Stereotactic radiosurgery with vs. Without prior embolization for brain arteriovenous malformations: a propensity score matching analysis. Front Neurol 2021; 12: 752164. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Wohlfahrt P, Krajcoviechova A, Jozifova M, et al. Large artery stiffness and carotid flow pulsatility in stroke survivors. J Hypertens 2014; 32: 1097–1103; discussion 1103. [DOI] [PubMed] [Google Scholar]
- 17.Henskens LHG, Kroon AA, van Oostenbrugge RJ, et al. Increased aortic pulse wave velocity is associated with silent cerebral small-vessel disease in hypertensive patients. Hypertension 2008; 52: 1120–1126. [DOI] [PubMed] [Google Scholar]
- 18.Kwan J, Hand P. Early neurological deterioration in acute stroke: clinical characteristics and impact on outcome. QJM 2006; 99: 625–633. [DOI] [PubMed] [Google Scholar]
- 19.Calviello LA, de Riva N, Donnelly J, et al. Relationship between brain pulsatility and cerebral perfusion pressure: replicated validation using different drivers of CPP change. Neurocrit Care 2017; 27: 392–400. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Yoo IH, Kim JM, Han SH, et al. Increased pulsatility index of the basilar artery is a risk factor for neurological deterioration after stroke: a case control study. Clin Hypertens 2022; 28: 27. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Hadaczek P, Yamashita Y, Mirek H, et al. The “perivascular pump” driven by arterial pulsation is a powerful mechanism for the distribution of therapeutic molecules within the brain. Mol Ther 2006; 14: 69–78. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Iliff JJ, Wang M, Zeppenfeld DM, et al. Cerebral arterial pulsation drives paravascular CSF-interstitial fluid exchange in the murine brain. J Neurosci 2013; 33: 18190–18199. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Aldea R, Weller RO, Wilcock DM, et al. Cerebrovascular smooth muscle cells as the drivers of intramural periarterial drainage of the brain. Front Aging Neurosci 2019; 11: 1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.van Veluw SJ, Hou SS, Calvo-Rodriguez M, et al. Vasomotion as a driving force for paravascular clearance in the awake mouse brain. Neuron 2020; 105: 549–561.e5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Vinje V, Ringstad G, Lindstrøm EK, et al. Respiratory influence on cerebrospinal fluid flow – a computational study based on long-term intracranial pressure measurements. Sci Rep 2019; 9: 9732. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Ung TH, Belanger K, Hashmi A, et al. Microenvironment changes in arteriovenous malformations after stereotactic radiation. Front Hum Neurosci 2022; 16: 982190. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Roche PH, Khalil M, Soumare O, et al. Hydrocephalus and vestibular schwannomas: considerations about the impact of gamma knife radiosurgery. Prog Neurol Surg 2008; 21: 200–206. [DOI] [PubMed] [Google Scholar]
- 28.Chapman PH, Ogilvy CS, Loeffler JS. The relationship between occlusive hyperemia and complications associated with the radiosurgical treatment of arteriovenous malformations: report of two cases. Neurosurgery 2004; 55: 228–233; discussion 233–234. [DOI] [PubMed] [Google Scholar]
- 29.Al-Rodhan NR, Sundt TM, Piepgras DG, et al. Occlusive hyperemia: a theory for the hemodynamic complications following resection of intracerebral arteriovenous malformations. J Neurosurg 1993; 78: 167–175. [DOI] [PubMed] [Google Scholar]
- 30.Uysal E, Başkurt O, Avcı İ, et al. Late recovery of stereotactic radiosurgery induced perilesional edema of an arteriovenous malformation after bevacizumab treatment. Br J Neurosurg 2021; 35: 22–26. [DOI] [PubMed] [Google Scholar]
- 31.Kwong F, Scarpelli DB, Barajas RF, et al. Resolution of radiation-induced necrosis in arteriovenous malformation with bevacizumab: a case report and review of current literature. Case Rep Neurol 2021; 13: 297–304. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Supplemental material, sj-pdf-1-jcb-10.1177_0271678X251358986 for Reversible edema after radiosurgery for arteriovenous malformations (AVMs): Inflow and outflow imbalance by Daniel Sconzo, Felipe Ramirez-Velandia, Alejandro Enriquez-Marulanda, Coleman P Riordan, Sandeep Muram, Nima Aghdam, Philipp Taussky and Christopher S Ogilvy in Journal of Cerebral Blood Flow & Metabolism