- Email: [email protected]
Stereotactic Radiosurgery for High-Grade Intracranial Dural Arteriovenous Fistulas
Original Article
Stereotactic Radiosurgery for High-Grade Intracranial Dural Arteriovenous Fistulas Ching-Jen Chen1, Thomas J. Buell1, Joshua Diamond1, Dale Ding2, Jeyan S. Kumar1, Davis G. Taylor1, Cheng-Chia Lee3, Jason P. Sheehan1
OBJECTIVE: Factors associated with favorable outcome after stereotactic radiosurgery (SRS) for dural arteriovenous fistulas (DAVFs) with cortical venous reflux (CVR) are not completely understood. The aim of this retrospective cohort study was to assess the outcomes after SRS for high-grade DAVFs and identify predictors.
-
METHODS: We performed a retrospective review of consecutive patients with high-grade DAVFs, defined as the presence of CVR, who underwent SRS between 1989 and 2017. The primary outcome was defined as DAVF obliteration without a new permanent neurologic deficit. Predictors of outcomes were determined using multivariate logistic regression.
selected patients with high-grade DAVFs after treatment with SRS. NHND at presentation is a risk factor for new permanent neurologic deficit after SRS.
-
RESULTS: The study cohort was composed of 41 highgrade DAVF patients with a mean age of 52 years. DAVF obliteration without a new permanent neurologic deficit was achieved in 62% of patients (13/21). The rates of complete obliteration and new permanent neurologic deficit were 63% (17/27) and 23% (7/30) of patients, respectively. No independent predictors of the primary outcome or angiographic obliteration were identified in the multivariate model. Presentation with a nonhemorrhagic neurologic deficit (NHND) was found to be an independent predictor of a new permanent neurologic deficit after SRS (odds ratio, 14.176; 95% confidence interval, 1.119e179.540; P [ 0.041).
-
CONCLUSIONS: Obliteration without a new permanent neurologic deficit can be achieved in most appropriately
-
Key words Radiosurgery - Dural arteriovenous fistula - Fistula - Intracranial - Hemorrhage - Obliteration -
Abbreviations and Acronyms CI: Confidence interval CVR: Cortical venous reflux DAVF: Dural arteriovenous fistula MRI: Magnetic resonance imaging NHND: Nonhemorrhagic neurologic deficit
INTRODUCTION
D
ural arteriovenous fistulas (DAVFs) comprise approximately 10%e15% of all intracranial vascular malformations.1 DAVFs are categorized into low- versus high-grade lesions based on the absence or presence of cortical venous reflux (CVR), either because of retrograde flow through the draining dural venous sinus (Borden type II) or direct drainage into cortical veins (Borden type III).2 High-grade DAVFs (Borden types II and III) are often associated with an aggressive clinical course that includes hemorrhage, progressive neurologic deficits, and seizures.3-7 Treatment modalities for DAVFs include microsurgical ligation, endovascular embolization, and stereotactic radiosurgery (SRS). SRS is most frequently used as a salvage therapy for DAVFs that fail to obliterate after initial intervention with surgery or embolization. SRS can also be used, albeit uncommonly, as an upfront therapy for DAVF patients with excessive medical comorbidities and unfavorable extra- or intracranial vascular anatomy. Excellent obliteration rates for low-grade DAVFs can be achieved with SRS. However, obliteration rates for high-grade lesions are often lower, with angiographic cure rates of approximately 50%.5,6,8,9 Given the rarity of high-grade DAVFs and their preferred first-line treatment by surgery or embolization, the literature pertaining to SRS for DAVFs with CVR is limited.10
OR: Odds ratio SRS: Stereotactic radiosurgery From the 1Department of Neurological Surgery, University of Virginia Health System, Charlottesville, Virginia, USA; 2Department of Neurosurgery, Barrow Neurological Institute, Phoenix, Arizona, USA; and 3Department of Neurosurgery, Taipei Veterans General Hospital, Taipei, Taiwan To whom correspondence should be addressed: Ching-Jen Chen, M.D. [E-mail: [email protected]] Citation: World Neurosurg. (2018). https://doi.org/10.1016/j.wneu.2018.05.062 Journal homepage: www.WORLDNEUROSURGERY.org Available online: www.sciencedirect.com 1878-8750/$ - see front matter ª 2018 Elsevier Inc. All rights reserved.
WORLD NEUROSURGERY -: ---, - 2018
www.WORLDNEUROSURGERY.org
e1
ORIGINAL ARTICLE CHING-JEN CHEN ET AL.
SRS FOR HIGH-GRADE DAVFS
Table 1. Comparisons of Baseline, Radiographic, and Clinical Characteristics Between Patients With and Without the Primary Outcome Variables
Primary Outcome (n [ 13*)
No Primary Outcome (n [ 8*)
P Value
Age (years)
47.2 15.6
53.3 17
0.417
Female
3 (23.1)
2 (25)
>0.999
MI
1 (7.7)
0/7 (0)
>0.999
CAD
1 (7.7)
0/7 (0)
>0.999
Afib
0 (0)
0/7 (0)
>0.999
Ischemic stroke
0 (0)
0/7 (0)
>0.999
PVD
0 (0)
0/7 (0)
>0.999
DM
1 (7.7)
1/7 (14.3)
>0.999
Hypertension
5 (38.5)
5 (62.5)
0.387
Antiplatelet use
1 (7.7)
0/6 (0)
>0.999
0 (0)
0/6 (0)
>0.999
Never
5/11 (45.5)
3/4 (75)
Past
1/11 (9.1)
0/4 (0)
Current
Anticoagulation use Smoking
0.692
5/11 (45.5)
1/4 (25)
Prior head trauma
2 (15.4)
0/7 (0)
0.521
Baseline mRS score 0e2
13 (100)
5/5 (100)
>0.999
5 (38.5)
3 (37.5)
>0.999
5/12 (41.7)
2/7 (28.6)
0.656
Global deficit
1/12 (8.3)
1/7 (14.3)
Focal deficit
3/12 (25)
3 (37.5)
Flow-related symptoms
3/12 (25)
2/7 (28.6)
Hemorrhage NHND
Seizures
1/12 (8.3)
0 (0)
>0.999
Symptomatic
10 (76.9)
7 (87.5)
>0.999
Borden type III
12 (92.3)
7/7 (100)
>0.999
Deep venous drainage
10 (76.9)
2/6 (33.3)
0.129
Venous ectasia
7 (53.9)
6 (75)
0.400
1/9 (11.1)
0/6 (0)
>0.999
7 (53.9)
6 (75)
0.400
T2/FLAIR hyperintensity Prior embolization Prior surgery
2 (15.4)
3 (37.5)
0.325
Time from diagnosis to SRS (months)
4.1 4.6
5 4.1
0.753
Margin dose (Gy)
18.9 2.9
21 3.6
0.169
Post-SRS hemorrhage Radiologic follow-up (months) Clinical follow-up (months)
1 (7.7)
2 (25)
0.531
56 42.3
80.8 62.4
0.290
34.4 30.6
80.4 67.7
0.065
Values are mean SD, number of patients (%), or as otherwise indicated. MI, myocardial infarction; CAD, coronary artery disease; Afib, atrial fibrillation; PVD, peripheral vascular disease; DM, diabetes mellitus; mRS, modified Rankin scale; NHND, nonhemorrhagic neurologic deficit; FLAIR, fluid-attenuated inversion recovery; SRS, stereotactic radiosurgery. *Or as otherwise indicated.
e2
www.SCIENCEDIRECT.com
WORLD NEUROSURGERY, https://doi.org/10.1016/j.wneu.2018.05.062
ORIGINAL ARTICLE CHING-JEN CHEN ET AL.
Hence, predictors of outcome after SRS for high-grade DAVFs remain largely undefined. Identification of such predictors may aid in the patient selection and counseling processes for SRS. Therefore, the aims of this retrospective cohort study are to evaluate the outcomes after SRS for high-grade DAVFs and determine predictors of obliteration and new neurologic deficit. MATERIALS AND METHODS Study Design We retrospectively reviewed an institutional review boarde approved, prospectively collected database of consecutive patients with DAVFs who underwent SRS at our institution between 1989 and 2017. Patients with high-grade (Borden types II or III) DAVFs were selected for inclusion in the study cohort. SRS Procedure All SRS procedures were performed in a single session using the Gamma Knife (Elekta AB, Stockholm, Sweden). Our institution’s technique has been previously described in detail.6 Briefly, all patients underwent placement of a stereotactic Leksell frame (Elekta AB, Stockholm, Sweden) under monitored anesthesia. Next, catheter cerebral angiography and stereotactic thin-slice (slice width 1 mm) magnetic resonance imaging (MRI) or computed tomography scan were performed to allow treatment planning. Dose planning was performed using Kula software (Stockholm, Sweden) until June 1994, and then Gamma Plan software (Stockholm, Sweden) was used thereafter. Baseline Data Baseline data included patient demographic, clinical, DAVF, and treatment variables. Patient demographic variables included age and sex. Clinical variables included medical comorbidities (history of myocardial infarction, coronary artery disease, atrial fibrillation, ischemic stroke, peripheral vascular disease, diabetes mellitus, and hypertension), use of antiplatelet or anticoagulation therapy, smoking status, prior head trauma, baseline modified Rankin scale score, seizures, prior hemorrhage, and nonhemorrhagic neurologic deficit (NHND), which includes global neurologic deficits (i.e., dementia, psychiatric symptoms, altered level of consciousness), focal neurologic deficits (i.e., cranial nerve palsies except ophthalmoplegia, motor weakness, sensory disturbances), and flow-related symptoms (i.e., headaches not attributable to hemorrhage, tinnitus, ophthalmoplegia, proptosis, visual disturbances). DAVF variables included Borden type, presence of deep venous drainage, venous ectasia, and T2 or fluid-attenuated inversion recovery hyperintensity associated with DAVF on MRI. Treatment variables included prior embolization or surgical ligation of the DAVF, margin dose, and time from diagnosis to SRS. Follow-Up Patients underwent an MRI every 6 months for the first 2 years after SRS, and then annually thereafter. Confirmatory catheter angiography was recommended to patients who were found to have DAVF obliteration on MRI. Obliteration was defined by the absence of abnormal arteriovenous shunting and early venous drainage on catheter angiography or by a lack of flow voids on MRI. Clinical follow-up was performed at clinical visits to our
WORLD NEUROSURGERY -: ---, - 2018
SRS FOR HIGH-GRADE DAVFS
institution or through contact with the referring physician and patient if the patient did not reside in the surrounding area. Additional neuroimaging and clinical follow-up was obtained for patients who experienced new or worsening neurologic symptoms. The primary outcome was defined as DAVF obliteration without a new permanent neurologic deficit. The secondary outcomes were DAVF obliteration and development of a new permanent neurologic deficit. Statistical Analyses All statistical analyses were performed using Stata (version 14.2 [StataCorp LLC, College Station, Texas, USA]). Baseline, clinical, radiologic, and treatment characteristics were compared between patients with and without the primary outcome, those with and without obliteration, and those with and without a new permanent neurologic deficit. Continuous variables were compared using Student t or Mann-Whitney U tests, as appropriate. Categorical variables were compared using Pearson c2 or Fisher exact tests, as appropriate. Univariate predictors with P < 0.2 were entered into a multivariate logistic regression model to identify independent predictors of the primary outcome, obliteration, and new permanent neurologic deficit. The missing data were assumed to be missing at random. To avoid listwise deletions in the multivariable logistic regression model, multiple imputation by chained equations with m ¼ 20 was used. Logistic and linear regressions were used for binary and continuous variables, respectively. Parameter estimates from analyzing the imputed datasets were pooled according to Rubin’s rules.11 Statistical significance was defined as P < 0.05, and all tests were 2-tailed. RESULTS Primary Outcome From 1989 to 2017, 60 patients with DAVFs underwent SRS at our institution. Of these, 41 patients with high-grade DAVFs were included in the study cohort. The mean age of the study cohort was 52 years, and 29% of patients (12/41) were women. DAVF locations comprised the anterior cranial fossa (n ¼ 3), convexity or superior sagittal sinus (n ¼ 2), cavernous sinus (n ¼ 4), middle cranial fossa (n ¼ 1), transverse/sigmoid sinus (n ¼ 3), torcula (n ¼ 2), tentorial (n ¼ 15), petrosal sinus (n ¼ 2), and unspecified (n ¼ 9). Of the 21 patients with sufficient data to allow evaluation for the primary outcome, the primary outcome was achieved in 13 patients (62%). Table 1 compares the baseline, radiographic, and clinical characteristics between patients with (n ¼ 13) and without (n ¼ 8) the primary outcome. There were no significant
Table 2. Multivariable Logistic Regression Analysis for Predictors of the Primary Outcome* Predictors
Odds Ratio 95% Confidence Interval P Value
Deep venous drainage
7.150
0.728e70.227
0.090
Margin dose
0.916
0.705e1.190
0.503
*Multivariable logistic regression model based on pooled parameter estimates from multiple imputation by chained equations with m ¼ 20.
www.WORLDNEUROSURGERY.org
e3
ORIGINAL ARTICLE CHING-JEN CHEN ET AL.
SRS FOR HIGH-GRADE DAVFS
Table 3. Comparisons of Baseline, Radiographic, and Clinical Characteristics Between Patients with and without Dural Arteriovenous Fistulas Obliteration Variables
Obliteration (n [ 17*)
No Obliteration (n [ 10*)
Age (years)
45.9 15.5
53 15.6
0.266
6 (35.3)
2 (20)
0.666
MI
1/16 (6.3)
0/9 (0)
>0.999
CAD
1/16 (6.3)
0/9 (0)
>0.999
Afib
0/16 (0)
0/9 (0)
>0.999
Ischemic stroke
0/16 (0)
0/9 (0)
>0.999
Female
P Value
PVD
0/16 (0)
0/9 (0)
>0.999
DM
1/16 (6.3)
1/9 (11.1)
>0.999
Hypertension
6/16 (37.5)
5/10 (50)
0.689
Antiplatelet use
1/15 (6.7)
0/8 (0)
>0.999
0/15 (0)
0/8 (0)
>0.999
Never
5/12 (41.7)
3/5 (60)
Past
1/12 (8.3)
1/5 (20)
Current
6/12 (50)
1/5 (20)
2/16 (12.5)
1/9 (11.1)
>0.999
Baseline mRS score 0e2
17 (100)
7/7 (100)
>0.999
Hemorrhage
8 (47.1)
4 (40)
>0.999 0.669
Anticoagulation use Smoking
Prior head trauma
NHND
0.462
5/15 (33.3)
2/9 (22.2)
Global deficit
1/15 (6.7)
1/9 (11.1)
Focal deficit
4/16 (25)
3 (30)
4/15 (26.7)
2/9 (22.2)
2/16 (12.5)
0 (0)
0.508
14 (82.4)
8 (80)
>0.999
Flow-related symptoms Seizures Symptomatic Borden type III
16 (94.1)
9/9 (100)
>0.999
Deep venous drainage
12/16 (75)
3/8 (37.5)
0.099
Venous ectasia
9 (52.9)
7 (70)
0.448
T2/FLAIR hyperintensity
1/11 (9.1)
0/6 (0)
>0.999
Prior embolization
10 (58.8)
7 (70)
0.692
Prior surgery
3 (17.7)
3 (30)
0.638
Time from diagnosis to SRS (months)
10.3 21.4
5.7 3.8
0.645
Margin dose (Gy)
19.7 3.9
21.1 3.6
0.348
1 (5.9)
2/9 (22.2)
0.268
Radiologic follow-up (months)
49.4 39.5
81.6 66.2
0.124
Clinical follow-up (months)
36.4 29.9
80.4 67.7
0.065
Post-SRS hemorrhage
Values are mean SD, number of patients (%), or as otherwise indicated. MI, myocardial infarction; CAD, coronary artery disease; Afib, atrial fibrillation; PVD, peripheral vascular disease; DM, diabetes mellitus; mRS, modified Rankin scale; NHND, nonhemorrhagic neurologic deficit; FLAIR, fluid-attenuated inversion recovery; SRS, stereotactic radiosurgery. *Or as otherwise indicated.
e4
www.SCIENCEDIRECT.com
WORLD NEUROSURGERY, https://doi.org/10.1016/j.wneu.2018.05.062
ORIGINAL ARTICLE CHING-JEN CHEN ET AL.
SRS FOR HIGH-GRADE DAVFS
Table 4. Comparisons of Baseline, Radiographic, and Clinical Characteristics between Patients with and without a New Permanent Neurologic Deficit Variables
New Deficit (n [ 7*)
No New Deficit (n [ 23*)
Age (years)
P Value
50.4 20.1
52.5 14.7
0.789
Female
0/5 (0)
5/22 (22.7)
0.547
MI
1/5 (20)
1/21 (4.8)
0.354
CAD
2/5 (40)
1/21 (4.8)
0.085
Afib
0/5 (0)
0/21 (0)
>0.999
Ischemic stroke
0/5 (0)
0/21 (0)
>0.999
PVD
0/5 (0)
0/21 (0)
>0.999
DM
1/5 (20)
2/21 (9.5)
0.488
Hypertension
3/5 (60)
11/22 (50)
>0.999
Antiplatelet use
0/5 (0)
1/20 (5)
>0.999
Anticoagulation use
0/5 (0)
0/20 (0)
>0.999
1/1 (100)
8/16 (50)
Past
0/0 (0)
2/16 (12.5)
Current
0/0 (0)
6/16 (37.5)
0/5 (0)
2/21 (9.5)
>0.999
>0.999
Smoking Never
Prior head trauma
19/19 (100)
4/4 (100)
>0.999
Hemorrhage
1/5 (20)
8/22 (36.4)
0.636
NHND
0.012y
Baseline mRS score 0e2
4/4 (100)
5/20 (25)
Global deficit
4/5 (80)
1/20 (5)
Focal deficit
2/5 (40)
5/21 (23.8)
Flow-related symptoms
0/2 (0)
8/20 (40)
0/5 (0)
1/21 (4.8)
>0.999
Symptomatic
5/5 (100)
18/22 (81.8)
0.561
Borden grade III
4/5 (80)
20/21 (95.2)
0.354
Deep venous drainage
4/5 (80)
11/19 (57.9)
0.615
Venous ectasia
3/5 (60)
14/22 (63.6)
>0.999
T2/FLAIR hyperintensity
1/4 (25)
1/15 (6.7)
0.386
Prior embolization
3/5 (60)
14/22 (63.6)
>0.999
Seizures
Prior surgery
2/5 (40)
5/22 (22.7)
0.580
Time from diagnosis to SRS (months)
2.6 0.9
4.7 4.1
0.491
Margin dose (Gy)
23.4 2.3
19 3.1
0.008y
1/5 (20)
2/22 (9.1)
0.474
28.8 35
60.2 53.4
0.226
35.5 47.3
49 51.1
0.674
Post-SRS hemorrhage Radiologic follow-up (months) Clinical follow-up (months)
Values are mean SD, number of patients (%), or as otherwise indicated. MI, myocardial infarction; CAD, coronary artery disease; Afib, atrial fibrillation; PVD, peripheral vascular disease; DM, diabetes mellitus; mRS, modified Rankin scale; NHND, nonhemorrhagic neurologic deficit; FLAIR, fluid-attenuated inversion recovery; SRS, stereotactic radiosurgery. *Or as otherwise indicated. yStatistically significant at P < 0.05.
WORLD NEUROSURGERY -: ---, - 2018
www.WORLDNEUROSURGERY.org
e5
ORIGINAL ARTICLE CHING-JEN CHEN ET AL.
SRS FOR HIGH-GRADE DAVFS
of a new neurologic deficit. The presence of a NHND at presentation was an independent predictor of a new permanent neurologic deficit after SRS (OR, 14.176; 95% CI, 1.119e179.540; P ¼ 0.041) in the multivariate model.
Table 5. Multivariable Logistic Regression Analysis for Predictors of a New Permanent Neurologic Deficit After Stereotactic Radiosurgery* Predictors
Odds Ratio
95% Confidence Interval
P Value
CAD
1.913
0.124e29.461
0.642
NHND
14.176
1.119e179.540
0.041y
Margin dose
1.045
0.793e1.377
0.751
CAD, coronary artery disease; NHND, nonhemorrhagic neurologic deficit. *Multivariable logistic regression model based on pooled parameter estimates from multiple imputation by chained equations with m ¼ 20. yStatistically significant at P < 0.05.
differences in these characteristics between the 2 groups. Three patients suffered a hemorrhage after SRS during the cumulative follow-up of 1095 patient-months, yielding an annual post-SRS hemorrhage rate of 3.3%. No significant difference in the incidences of post-SRS hemorrhage was found between those with (8%) and without (25%) the primary outcome (P ¼ 0.531). The durations of radiographic and clinical follow-up were not significantly different between the 2 groups. Univariate factors with P < 0.2 were the presence of deep venous drainage (P ¼ 0.129) and margin dose (P ¼ 0.169). Table 2 details the multivariable logistic regression analysis for predictors of the primary outcome. Neither the presence of deep venous drainage (odds ratio [OR], 7.15; 95% confidence interval [CI], 0.728e70.227; P ¼ 0.090) nor margin dose (OR, 0.916; 95% CI, 0.705e1.190; P ¼ 0.503) were found to be significantly associated with the primary outcome in the multivariate model. Obliteration Of the 27 patients with available neuroimaging follow-up, DAVF obliteration was achieved in 17 patients (63%), including 11 (65%) confirmed by angiography and 6 (35%) determined by MRI alone. Table 3 compares the baseline, radiographic, and clinical characteristics between patients with (n ¼ 17) and without (n ¼ 10) DAVF obliteration. No factors were found to be significantly associated with obliteration in the univariate analysis. Of the 10 patients who did not achieve obliteration after SRS, 1 (10%) underwent microsurgical ligation of the DAVF. Data regarding post-SRS interventions for the remaining 9 patients were not available. Post-SRS Neurologic Deficits Of the 30 patients with available clinical follow-up, 7 (23%) developed a new permanent neurologic deficit. Table 4 compares the baseline, radiographic, and clinical characteristics between patients with (n ¼ 7) and without (n ¼ 23) a new permanent neurologic deficit. NHND was significantly more common (100% vs. 25%, P ¼ 0.012) and the margin dose was significantly higher (mean, 23 vs. 19, P ¼ 0.008) in patients with a new permanent neurologic deficit, respectively. Univariate factors with P < 0.2 were coronary artery disease (P ¼ 0.085), NHND (P ¼ 0.012), and margin dose (P ¼ 0.008). Table 5 details the multivariate logistic regression analysis for predictors
e6
www.SCIENCEDIRECT.com
Illustrative Case A 71-year-old man who presented with right-sided synchronous pulsatile tinnitus was found to have a Borden type III right tentorial supplied by the right meningohypophyseal trunk, right inferolateral trunk, right occipital artery, petrosal branch of the right middle meningeal artery, and right accessory meningeal artery (Figure 1). There was direct pial venous drainage into the pontomesencephalic veins draining into the right basal vein of Rosenthal, and anterograde drainage via the right superior petrosal sinus and transverse sinus. The patient initially underwent endovascular treatment of the DAVF with coil embolization, followed by ethylene vinyl alcohol copolymer (Onyx [Medtronic Neurovascular, Irvine, California, USA]). However, after the embolization procedure, there was evidence of residual arteriovenous shunting. Therefore, the residual DAVF was treated with Gamma Knife SRS using a margin dose of 18 Gy to the 50% isodose line. The volume of the prescription isodose was 2.3 cm3. Follow-up angiography 3 years after SRS showed complete obliteration of the DAVF. DISCUSSION High-grade DAVFs are often associated with a relatively more aggressive clinical course and a higher risk of hemorrhage than low-grade DAVFs.12 Among 85 patients with high-grade DAVFs, Söderman et al.13 reported annual hemorrhage rates of 1.5% and 7.4% for unruptured and ruptured lesions, respectively. Similarly, Strom et al.14 reported annual hemorrhage rates of 1.4% and 7.6% for unruptured and ruptured high-grade DAVFs, respectively. The risk of rehemorrhage after the initial hemorrhage has been reported to be as high as 35% within the first 2 weeks.15 In addition to the hemorrhagic risk associated with high-grade DAVFs, persistence of CVR is associated with annual NHND and mortality rates of 6.9% and 10.4%, respectively.16 Given the considerable morbidity and mortality rates of untreated or partially treated high-grade DAVFs, urgent treatment is typically indicated. Embolization and microsurgical ligation are the preferred first-line interventions for symptomatic DAVFs because they can afford immediate closure of the arteriovenous shunt in most cases.17 In a pooled analysis of 19 studies comprising 743 DAVFs treated with SRS, we reported an obliteration rate of 56% for high-grade DAVFs versus 75% for low-grade DAVFs.10 Despite the relatively lower obliteration rates for high- versus low-grade DAVFs, SRS still confers therapeutic benefit in those who have failed or are unable to undergo surgery or embolization.10 In this single-center retrospective cohort study, we reported the outcomes specifically pertaining to patients with high-grade DAVFs who underwent SRS. Of the patients with neuroimaging records maintained at our institution, DAVF obliteration was achieved in 63%. In contrast with the prior DAVF SRS series from our institution, which included DAVFs of all types, the present study sought to identify predictors of post-SRS outcome specifically for high-grade DAVFs.6 Our analysis did not identify a statistically significant
WORLD NEUROSURGERY, https://doi.org/10.1016/j.wneu.2018.05.062
ORIGINAL ARTICLE CHING-JEN CHEN ET AL.
Figure 1. (A) Initial cerebral angiography, lateral view of a right internal carotid artery injection, shows that the right tentorial dural arteriovenous fistula (DAVF) is supplied by the right meningohypophyseal trunk and right inferolateral trunk with direct pial venous drainage into the pontomesencephalic veins draining into the right basal vein of Rosenthal. Initial cerebral angiography, (B) lateral and (C) anteroposterior views of a selective right occipital artery injection, shows that the right tentorial DAVF is supplied by the petrosal branch of the right middle meningeal artery and the right accessory meningeal artery, with direct pial venous drainage to
predictor of the primary outcome (i.e., DAVF obliteration without a new neurologic deficit) or obliteration after SRS. However, presentation with a NHND was found to be an independent predictor of a new permanent neurologic deficit after SRS. This may reflect the natural course of high-grade DAVFs with aggressive versus benign presentations.7 Strom et al.18 had suggested the classification of high-grade DAVFs into symptomatic versus asymptomatic subgroups by presenting symptoms, such that patients who presented with hemorrhage or NHND were classified as symptomatic, whereas those with incidentally diagnosed DAVFs or who presented with pulsatile tinnitus or orbital phenomena
WORLD NEUROSURGERY -: ---, - 2018
SRS FOR HIGH-GRADE DAVFS
the pontomesencephalic veins draining into the right basal vein of Rosenthal, and the right superior petrosal and transverse sinuses. (D) Postembolization (coils and Onyx) cerebral angiography, lateral view of a right common carotid artery injection, shows residual arteriovenous shunting. (E) The Gamma Knife treatment plan for the residual DAVF, which was treated with a margin dose of 18 Gy to the 50% isodose line (yellow). The volume of the prescription isodose was 2.3 cm3. The 8-Gy isodose line is shown in green.
were classified as asymptomatic. In the study comprising 11 symptomatic and 17 asymptomatic high-grade DAVFs, the authors reported significantly higher rates of hemorrhage or NHND in the symptomatic subgroup (45.5% vs. 5.9%, respectively; P ¼ 0.022).18 Higher rates of hemorrhage or NHND may indicate worsening venous outflow obstruction or impending overload of the drainage pathways, such that these aggressive subtypes of highgrade DAVFs are more vulnerable to subsequent venous hypertension or infarction.4,5,19 In a recent pooled analysis comparing SRS outcomes between aggressive (n ¼ 73) and nonaggressive (n ¼ 124)
www.WORLDNEUROSURGERY.org
e7
ORIGINAL ARTICLE CHING-JEN CHEN ET AL.
SRS FOR HIGH-GRADE DAVFS
high-grade DAVFs, Tonetti et al.20 reported significantly higher rates of post-SRS hemorrhage (7% vs. 0%, P ¼ 0.003) and radiation-induced complications (8% vs. 0%, P ¼ 0.001) for aggressive lesions, respectively. SRS could cause delayed occlusion of venous outflow pathways, thereby increasing the stress to an already tenuous venous drainage system in the aggressive subtype of high-grade DAVFs. Finally, lack of obliteration of the DAVF should be considered as a possible cause of new or worsening neurologic symptoms after SRS.9 Furthermore, high-grade DAVFs with an aggressive clinical presentation may be less likely to obliterate after SRS and therefore are more prone to causing subsequent neurologic deterioration.6,9,20 DAVF treatment should be individualized to each unique case, with consideration of both the DAVF’s angioarchitecture and the patient’s clinical presentation, baseline neurologic status, medical comorbidities, and preferences. Hemorrhage, progressive NHND, or signs of intracranial hypertension mandate urgent intervention. Treatment should also be offered to patients who have intolerable nonaggressive symptoms or high-grade DAVFs. Mounting evidence has suggested that asymptomatic high-grade DAVFs may confer low risks of hemorrhage and NHND with observation.12-14 However, nonaggressive high-grade DAVFs may transform into aggressive high-grade DAVFs over time. The risk and temporal course of this progression from a nonaggressive to aggressive subtype has yet to be fully delineated. Therefore, given the favorable overall risk to benefit profile of SRS for these lesions, intervention for nonaggressive high-grade DAVFs with SRS may be a reasonable management strategy for appropriately selected patients. DAVFs which fail to obliterate after SRS should be considered for salvage treatment with embolization, microsurgical ligation, or repeat SRS, after discussion with an experienced, multidisciplinary cerebrovascular team.
REFERENCES 1. Newton TH, Cronqvist S. Involvement of dural arteries in intracranial arteriovenous malformations. Radiology. 1969;93:1071-1078. 2. Borden JA, Wu JK, Shucart WA. A proposed classification for spinal and cranial dural arteriovenous fistulous malformations and implications for treatment. J Neurosurg. 1995;82:166-179. 3. Davies MA, TerBrugge K, Willinsky R, Coyne T, Saleh J, Wallace MC. The validity of classification for the clinical presentation of intracranial dural arteriovenous fistulas. J Neurosurg. 1996;85: 830-837. 4. Oh JT, Chung SY, Lanzino G, Park KS, Kim SM, Park MS, et al. Intracranial dural arteriovenous fistulas: clinical characteristics and management based on location and hemodynamics. J Cerebrovasc Endovasc Neurosurg. 2012;14:192. 5. Söderman M, Edner G, Ericson K, Karlsson B, Rähn T, Ulfarsson E, et al. Gamma knife surgery for dural arteriovenous shunts: 25 years of experience. J Neurosurg. 2006;104:867-875.
e8
www.SCIENCEDIRECT.com
It is important to recognize the limitations of this study, which is subject to the inherent weaknesses and biases of its retrospective and single-center design. Our analyses were limited by the relatively modest size of the study cohort, which may have caused our statistical tests to be underpowered. Because of institutional policies, data storage by the image management department may expire and become inaccessible, thereby resulting in missing data within our dataset. Although missing data were imputed using multiple imputations to avoid listwise deletions, the process itself is vulnerable to bias if the missing at random assumption does not hold true. We acknowledge that approximately one-third of patients who were found to have DAVF obliteration did not undergo confirmatory angiography. Because MRI has not been found to have equivalent accuracy compared with angiography for assessing obliteration of DAVFs treated with SRS, the obliteration rate of our study cohort may be as low as 41%. Given the rarity of DAVFs and the general utilization of SRS as a salvage therapy, single-center analysis is challenging. Future multicenter efforts, including those through the Consortium for Dural Arteriovenous Fistula Outcomes Research and the International Gamma Knife Research Foundation, may allow for more robust analyses for predictors of success and complications after SRS for high-grade DAVFs. CONCLUSIONS Most appropriately selected patients with high-grade DAVFs who undergo treatment with SRS achieve lesional obliteration without suffering a new permanent neurologic deficit. Presentation with a NHND is a risk factor for new permanent neurologic deficit after SRS. Future investigations of larger, multicenter cohorts may afford more adequately powered analyses and provide additional insights regarding predictors of outcome after SRS for these uncommon vascular lesions.
6. Cifarelli CP, Kaptain G, Yen CP, Schlesinger D, Sheehan JP. Gamma Knife radiosurgery for dural arteriovenous fistulas. Neurosurgery. 2010;67: 1230-1235. 7. Della Pepa GM, Parente P, D’Argento F, Pedicelli A, Sturiale CL, Sabatino G, et al. Angioarchitectural features of high-grade intracranial dural arteriovenous fistulas: correlation with aggressive clinical presentation and hemorrhagic risk. Neurosurgery. 2017;81:315-330. 8. Seo Y, Kim Dong G, Dho YS, Kim JW, Kim YH, Park CK, et al. A retrospective analysis of the outcomes of dural arteriovenous fistulas treated with gamma knife radiosurgery: a singleinstitution experience. Stereotact Funct Neurosurg. 2018;96:46-53. 9. Hanakita S, Koga T, Shin M, Shojima M, Igaki H, Saito N. Role of stereotactic radiosurgery in the treatment of high-grade cerebral arteriovenous malformation. Neurol Med Chir (Tokyo). 2012;52: 845-851. 10. Chen CJ, Lee CC, Ding D, Starke RM, Chivukula S, Yen CP, et al. Stereotactic radiosurgery for intracranial dural arteriovenous
fistulas: a systematic review. J Neurosurg. 2015;122: 353-362. 11. Rubin DB. Multiple Imputation for Nonresponse in Surveys. New York, NY: John Wiley & Sons, Inc.; 1987. 12. Gross BA, Du R. The natural history of cerebral dural arteriovenous fistulae. Neurosurgery. 2012;71: 594-603. 13. Söderman M, Pavic L, Edner G, Holmin S, Andersson T. Natural history of dural arteriovenous shunts. Stroke. 2008;39:1735-1739. 14. Strom R, Botros J, Refai D, Moran CJ, Cross DT 3rd, Chicoine MR, et al. E-004 Cranial dural arteriovenous fistulas: asymptomatic cortical venous drainage portends less aggressive clinical course. J Neurointerv Surg. 2010;2(suppl 1):A27. 15. Duffau H, Lopes M, Janosevic V, Sichez JP, Faillot T, Capelle L, et al. Early rebleeding from intracranial dural arteriovenous fistulas: report of 20 cases and review of the literature. J Neurosurg. 1999;90:78-84. 16. van Dijk JM, terBrugge KG, Willinsky RA, Wallace MC. Clinical course of cranial dural
WORLD NEUROSURGERY, https://doi.org/10.1016/j.wneu.2018.05.062
ORIGINAL ARTICLE CHING-JEN CHEN ET AL.
arteriovenous fistulas with long-term persistent cortical venous reflux. Stroke. 2002;33:1233-1236. 17. Gross BA, Albuquerque FC, Moon K, McDougall CG. Evolution of treatment and a detailed analysis of occlusion, recurrence, and clinical outcomes in an endovascular library of 260 dural arteriovenous fistulas. J Neurosurg. 2017;126:1884-1893. 18. Strom RG, Botros JA, Refai D, Moran CJ, Cross DT 3rd, Chicoine MR, et al. Cranial dural arteriovenous fistulae: asymptomatic cortical venous drainage portends less aggressive clinical course. Neurosurgery. 2009;64:241-247 [discussion: 247-248].
WORLD NEUROSURGERY -: ---, - 2018
SRS FOR HIGH-GRADE DAVFS
19. Kwon BJ, Han MH, Kang HS, Chang KH. MR imaging findings of intracranial dural arteriovenous fistulas: relations with venous drainage patterns. Am J Neuroradiol. 2005;26:2500-2507. 20. Tonetti DA, Gross BA, Jankowitz BT, Kano H, Monaco EA 3rd, Niranjan A, et al. Reconsidering an important subclass of high-risk dural arteriovenous fistulas for stereotactic radiosurgery. J Neurosurg. 2018:1-5.
commercial or financial relationships that could be construed as a potential conflict of interest. Received 24 March 2018; accepted 9 May 2018 Citation: World Neurosurg. (2018). https://doi.org/10.1016/j.wneu.2018.05.062 Journal homepage: www.WORLDNEUROSURGERY.org Available online: www.sciencedirect.com 1878-8750/$ - see front matter ª 2018 Elsevier Inc. All rights reserved.
Conflict of interest statement: The authors declare that the article content was composed in the absence of any
www.WORLDNEUROSURGERY.org
e9
Stereotactic Radiosurgery for High-Grade Intracranial Dural Arteriovenous Fistulas Ching-Jen Chen1, Thomas J. Buell1, Joshua Diamond1, Dale Ding2, Jeyan S. Kumar1, Davis G. Taylor1, Cheng-Chia Lee3, Jason P. Sheehan1
OBJECTIVE: Factors associated with favorable outcome after stereotactic radiosurgery (SRS) for dural arteriovenous fistulas (DAVFs) with cortical venous reflux (CVR) are not completely understood. The aim of this retrospective cohort study was to assess the outcomes after SRS for high-grade DAVFs and identify predictors.
-
METHODS: We performed a retrospective review of consecutive patients with high-grade DAVFs, defined as the presence of CVR, who underwent SRS between 1989 and 2017. The primary outcome was defined as DAVF obliteration without a new permanent neurologic deficit. Predictors of outcomes were determined using multivariate logistic regression.
selected patients with high-grade DAVFs after treatment with SRS. NHND at presentation is a risk factor for new permanent neurologic deficit after SRS.
-
RESULTS: The study cohort was composed of 41 highgrade DAVF patients with a mean age of 52 years. DAVF obliteration without a new permanent neurologic deficit was achieved in 62% of patients (13/21). The rates of complete obliteration and new permanent neurologic deficit were 63% (17/27) and 23% (7/30) of patients, respectively. No independent predictors of the primary outcome or angiographic obliteration were identified in the multivariate model. Presentation with a nonhemorrhagic neurologic deficit (NHND) was found to be an independent predictor of a new permanent neurologic deficit after SRS (odds ratio, 14.176; 95% confidence interval, 1.119e179.540; P [ 0.041).
-
CONCLUSIONS: Obliteration without a new permanent neurologic deficit can be achieved in most appropriately
-
Key words Radiosurgery - Dural arteriovenous fistula - Fistula - Intracranial - Hemorrhage - Obliteration -
Abbreviations and Acronyms CI: Confidence interval CVR: Cortical venous reflux DAVF: Dural arteriovenous fistula MRI: Magnetic resonance imaging NHND: Nonhemorrhagic neurologic deficit
INTRODUCTION
D
ural arteriovenous fistulas (DAVFs) comprise approximately 10%e15% of all intracranial vascular malformations.1 DAVFs are categorized into low- versus high-grade lesions based on the absence or presence of cortical venous reflux (CVR), either because of retrograde flow through the draining dural venous sinus (Borden type II) or direct drainage into cortical veins (Borden type III).2 High-grade DAVFs (Borden types II and III) are often associated with an aggressive clinical course that includes hemorrhage, progressive neurologic deficits, and seizures.3-7 Treatment modalities for DAVFs include microsurgical ligation, endovascular embolization, and stereotactic radiosurgery (SRS). SRS is most frequently used as a salvage therapy for DAVFs that fail to obliterate after initial intervention with surgery or embolization. SRS can also be used, albeit uncommonly, as an upfront therapy for DAVF patients with excessive medical comorbidities and unfavorable extra- or intracranial vascular anatomy. Excellent obliteration rates for low-grade DAVFs can be achieved with SRS. However, obliteration rates for high-grade lesions are often lower, with angiographic cure rates of approximately 50%.5,6,8,9 Given the rarity of high-grade DAVFs and their preferred first-line treatment by surgery or embolization, the literature pertaining to SRS for DAVFs with CVR is limited.10
OR: Odds ratio SRS: Stereotactic radiosurgery From the 1Department of Neurological Surgery, University of Virginia Health System, Charlottesville, Virginia, USA; 2Department of Neurosurgery, Barrow Neurological Institute, Phoenix, Arizona, USA; and 3Department of Neurosurgery, Taipei Veterans General Hospital, Taipei, Taiwan To whom correspondence should be addressed: Ching-Jen Chen, M.D. [E-mail: [email protected]] Citation: World Neurosurg. (2018). https://doi.org/10.1016/j.wneu.2018.05.062 Journal homepage: www.WORLDNEUROSURGERY.org Available online: www.sciencedirect.com 1878-8750/$ - see front matter ª 2018 Elsevier Inc. All rights reserved.
WORLD NEUROSURGERY -: ---, - 2018
www.WORLDNEUROSURGERY.org
e1
ORIGINAL ARTICLE CHING-JEN CHEN ET AL.
SRS FOR HIGH-GRADE DAVFS
Table 1. Comparisons of Baseline, Radiographic, and Clinical Characteristics Between Patients With and Without the Primary Outcome Variables
Primary Outcome (n [ 13*)
No Primary Outcome (n [ 8*)
P Value
Age (years)
47.2 15.6
53.3 17
0.417
Female
3 (23.1)
2 (25)
>0.999
MI
1 (7.7)
0/7 (0)
>0.999
CAD
1 (7.7)
0/7 (0)
>0.999
Afib
0 (0)
0/7 (0)
>0.999
Ischemic stroke
0 (0)
0/7 (0)
>0.999
PVD
0 (0)
0/7 (0)
>0.999
DM
1 (7.7)
1/7 (14.3)
>0.999
Hypertension
5 (38.5)
5 (62.5)
0.387
Antiplatelet use
1 (7.7)
0/6 (0)
>0.999
0 (0)
0/6 (0)
>0.999
Never
5/11 (45.5)
3/4 (75)
Past
1/11 (9.1)
0/4 (0)
Current
Anticoagulation use Smoking
0.692
5/11 (45.5)
1/4 (25)
Prior head trauma
2 (15.4)
0/7 (0)
0.521
Baseline mRS score 0e2
13 (100)
5/5 (100)
>0.999
5 (38.5)
3 (37.5)
>0.999
5/12 (41.7)
2/7 (28.6)
0.656
Global deficit
1/12 (8.3)
1/7 (14.3)
Focal deficit
3/12 (25)
3 (37.5)
Flow-related symptoms
3/12 (25)
2/7 (28.6)
Hemorrhage NHND
Seizures
1/12 (8.3)
0 (0)
>0.999
Symptomatic
10 (76.9)
7 (87.5)
>0.999
Borden type III
12 (92.3)
7/7 (100)
>0.999
Deep venous drainage
10 (76.9)
2/6 (33.3)
0.129
Venous ectasia
7 (53.9)
6 (75)
0.400
1/9 (11.1)
0/6 (0)
>0.999
7 (53.9)
6 (75)
0.400
T2/FLAIR hyperintensity Prior embolization Prior surgery
2 (15.4)
3 (37.5)
0.325
Time from diagnosis to SRS (months)
4.1 4.6
5 4.1
0.753
Margin dose (Gy)
18.9 2.9
21 3.6
0.169
Post-SRS hemorrhage Radiologic follow-up (months) Clinical follow-up (months)
1 (7.7)
2 (25)
0.531
56 42.3
80.8 62.4
0.290
34.4 30.6
80.4 67.7
0.065
Values are mean SD, number of patients (%), or as otherwise indicated. MI, myocardial infarction; CAD, coronary artery disease; Afib, atrial fibrillation; PVD, peripheral vascular disease; DM, diabetes mellitus; mRS, modified Rankin scale; NHND, nonhemorrhagic neurologic deficit; FLAIR, fluid-attenuated inversion recovery; SRS, stereotactic radiosurgery. *Or as otherwise indicated.
e2
www.SCIENCEDIRECT.com
WORLD NEUROSURGERY, https://doi.org/10.1016/j.wneu.2018.05.062
ORIGINAL ARTICLE CHING-JEN CHEN ET AL.
Hence, predictors of outcome after SRS for high-grade DAVFs remain largely undefined. Identification of such predictors may aid in the patient selection and counseling processes for SRS. Therefore, the aims of this retrospective cohort study are to evaluate the outcomes after SRS for high-grade DAVFs and determine predictors of obliteration and new neurologic deficit. MATERIALS AND METHODS Study Design We retrospectively reviewed an institutional review boarde approved, prospectively collected database of consecutive patients with DAVFs who underwent SRS at our institution between 1989 and 2017. Patients with high-grade (Borden types II or III) DAVFs were selected for inclusion in the study cohort. SRS Procedure All SRS procedures were performed in a single session using the Gamma Knife (Elekta AB, Stockholm, Sweden). Our institution’s technique has been previously described in detail.6 Briefly, all patients underwent placement of a stereotactic Leksell frame (Elekta AB, Stockholm, Sweden) under monitored anesthesia. Next, catheter cerebral angiography and stereotactic thin-slice (slice width 1 mm) magnetic resonance imaging (MRI) or computed tomography scan were performed to allow treatment planning. Dose planning was performed using Kula software (Stockholm, Sweden) until June 1994, and then Gamma Plan software (Stockholm, Sweden) was used thereafter. Baseline Data Baseline data included patient demographic, clinical, DAVF, and treatment variables. Patient demographic variables included age and sex. Clinical variables included medical comorbidities (history of myocardial infarction, coronary artery disease, atrial fibrillation, ischemic stroke, peripheral vascular disease, diabetes mellitus, and hypertension), use of antiplatelet or anticoagulation therapy, smoking status, prior head trauma, baseline modified Rankin scale score, seizures, prior hemorrhage, and nonhemorrhagic neurologic deficit (NHND), which includes global neurologic deficits (i.e., dementia, psychiatric symptoms, altered level of consciousness), focal neurologic deficits (i.e., cranial nerve palsies except ophthalmoplegia, motor weakness, sensory disturbances), and flow-related symptoms (i.e., headaches not attributable to hemorrhage, tinnitus, ophthalmoplegia, proptosis, visual disturbances). DAVF variables included Borden type, presence of deep venous drainage, venous ectasia, and T2 or fluid-attenuated inversion recovery hyperintensity associated with DAVF on MRI. Treatment variables included prior embolization or surgical ligation of the DAVF, margin dose, and time from diagnosis to SRS. Follow-Up Patients underwent an MRI every 6 months for the first 2 years after SRS, and then annually thereafter. Confirmatory catheter angiography was recommended to patients who were found to have DAVF obliteration on MRI. Obliteration was defined by the absence of abnormal arteriovenous shunting and early venous drainage on catheter angiography or by a lack of flow voids on MRI. Clinical follow-up was performed at clinical visits to our
WORLD NEUROSURGERY -: ---, - 2018
SRS FOR HIGH-GRADE DAVFS
institution or through contact with the referring physician and patient if the patient did not reside in the surrounding area. Additional neuroimaging and clinical follow-up was obtained for patients who experienced new or worsening neurologic symptoms. The primary outcome was defined as DAVF obliteration without a new permanent neurologic deficit. The secondary outcomes were DAVF obliteration and development of a new permanent neurologic deficit. Statistical Analyses All statistical analyses were performed using Stata (version 14.2 [StataCorp LLC, College Station, Texas, USA]). Baseline, clinical, radiologic, and treatment characteristics were compared between patients with and without the primary outcome, those with and without obliteration, and those with and without a new permanent neurologic deficit. Continuous variables were compared using Student t or Mann-Whitney U tests, as appropriate. Categorical variables were compared using Pearson c2 or Fisher exact tests, as appropriate. Univariate predictors with P < 0.2 were entered into a multivariate logistic regression model to identify independent predictors of the primary outcome, obliteration, and new permanent neurologic deficit. The missing data were assumed to be missing at random. To avoid listwise deletions in the multivariable logistic regression model, multiple imputation by chained equations with m ¼ 20 was used. Logistic and linear regressions were used for binary and continuous variables, respectively. Parameter estimates from analyzing the imputed datasets were pooled according to Rubin’s rules.11 Statistical significance was defined as P < 0.05, and all tests were 2-tailed. RESULTS Primary Outcome From 1989 to 2017, 60 patients with DAVFs underwent SRS at our institution. Of these, 41 patients with high-grade DAVFs were included in the study cohort. The mean age of the study cohort was 52 years, and 29% of patients (12/41) were women. DAVF locations comprised the anterior cranial fossa (n ¼ 3), convexity or superior sagittal sinus (n ¼ 2), cavernous sinus (n ¼ 4), middle cranial fossa (n ¼ 1), transverse/sigmoid sinus (n ¼ 3), torcula (n ¼ 2), tentorial (n ¼ 15), petrosal sinus (n ¼ 2), and unspecified (n ¼ 9). Of the 21 patients with sufficient data to allow evaluation for the primary outcome, the primary outcome was achieved in 13 patients (62%). Table 1 compares the baseline, radiographic, and clinical characteristics between patients with (n ¼ 13) and without (n ¼ 8) the primary outcome. There were no significant
Table 2. Multivariable Logistic Regression Analysis for Predictors of the Primary Outcome* Predictors
Odds Ratio 95% Confidence Interval P Value
Deep venous drainage
7.150
0.728e70.227
0.090
Margin dose
0.916
0.705e1.190
0.503
*Multivariable logistic regression model based on pooled parameter estimates from multiple imputation by chained equations with m ¼ 20.
www.WORLDNEUROSURGERY.org
e3
ORIGINAL ARTICLE CHING-JEN CHEN ET AL.
SRS FOR HIGH-GRADE DAVFS
Table 3. Comparisons of Baseline, Radiographic, and Clinical Characteristics Between Patients with and without Dural Arteriovenous Fistulas Obliteration Variables
Obliteration (n [ 17*)
No Obliteration (n [ 10*)
Age (years)
45.9 15.5
53 15.6
0.266
6 (35.3)
2 (20)
0.666
MI
1/16 (6.3)
0/9 (0)
>0.999
CAD
1/16 (6.3)
0/9 (0)
>0.999
Afib
0/16 (0)
0/9 (0)
>0.999
Ischemic stroke
0/16 (0)
0/9 (0)
>0.999
Female
P Value
PVD
0/16 (0)
0/9 (0)
>0.999
DM
1/16 (6.3)
1/9 (11.1)
>0.999
Hypertension
6/16 (37.5)
5/10 (50)
0.689
Antiplatelet use
1/15 (6.7)
0/8 (0)
>0.999
0/15 (0)
0/8 (0)
>0.999
Never
5/12 (41.7)
3/5 (60)
Past
1/12 (8.3)
1/5 (20)
Current
6/12 (50)
1/5 (20)
2/16 (12.5)
1/9 (11.1)
>0.999
Baseline mRS score 0e2
17 (100)
7/7 (100)
>0.999
Hemorrhage
8 (47.1)
4 (40)
>0.999 0.669
Anticoagulation use Smoking
Prior head trauma
NHND
0.462
5/15 (33.3)
2/9 (22.2)
Global deficit
1/15 (6.7)
1/9 (11.1)
Focal deficit
4/16 (25)
3 (30)
4/15 (26.7)
2/9 (22.2)
2/16 (12.5)
0 (0)
0.508
14 (82.4)
8 (80)
>0.999
Flow-related symptoms Seizures Symptomatic Borden type III
16 (94.1)
9/9 (100)
>0.999
Deep venous drainage
12/16 (75)
3/8 (37.5)
0.099
Venous ectasia
9 (52.9)
7 (70)
0.448
T2/FLAIR hyperintensity
1/11 (9.1)
0/6 (0)
>0.999
Prior embolization
10 (58.8)
7 (70)
0.692
Prior surgery
3 (17.7)
3 (30)
0.638
Time from diagnosis to SRS (months)
10.3 21.4
5.7 3.8
0.645
Margin dose (Gy)
19.7 3.9
21.1 3.6
0.348
1 (5.9)
2/9 (22.2)
0.268
Radiologic follow-up (months)
49.4 39.5
81.6 66.2
0.124
Clinical follow-up (months)
36.4 29.9
80.4 67.7
0.065
Post-SRS hemorrhage
Values are mean SD, number of patients (%), or as otherwise indicated. MI, myocardial infarction; CAD, coronary artery disease; Afib, atrial fibrillation; PVD, peripheral vascular disease; DM, diabetes mellitus; mRS, modified Rankin scale; NHND, nonhemorrhagic neurologic deficit; FLAIR, fluid-attenuated inversion recovery; SRS, stereotactic radiosurgery. *Or as otherwise indicated.
e4
www.SCIENCEDIRECT.com
WORLD NEUROSURGERY, https://doi.org/10.1016/j.wneu.2018.05.062
ORIGINAL ARTICLE CHING-JEN CHEN ET AL.
SRS FOR HIGH-GRADE DAVFS
Table 4. Comparisons of Baseline, Radiographic, and Clinical Characteristics between Patients with and without a New Permanent Neurologic Deficit Variables
New Deficit (n [ 7*)
No New Deficit (n [ 23*)
Age (years)
P Value
50.4 20.1
52.5 14.7
0.789
Female
0/5 (0)
5/22 (22.7)
0.547
MI
1/5 (20)
1/21 (4.8)
0.354
CAD
2/5 (40)
1/21 (4.8)
0.085
Afib
0/5 (0)
0/21 (0)
>0.999
Ischemic stroke
0/5 (0)
0/21 (0)
>0.999
PVD
0/5 (0)
0/21 (0)
>0.999
DM
1/5 (20)
2/21 (9.5)
0.488
Hypertension
3/5 (60)
11/22 (50)
>0.999
Antiplatelet use
0/5 (0)
1/20 (5)
>0.999
Anticoagulation use
0/5 (0)
0/20 (0)
>0.999
1/1 (100)
8/16 (50)
Past
0/0 (0)
2/16 (12.5)
Current
0/0 (0)
6/16 (37.5)
0/5 (0)
2/21 (9.5)
>0.999
>0.999
Smoking Never
Prior head trauma
19/19 (100)
4/4 (100)
>0.999
Hemorrhage
1/5 (20)
8/22 (36.4)
0.636
NHND
0.012y
Baseline mRS score 0e2
4/4 (100)
5/20 (25)
Global deficit
4/5 (80)
1/20 (5)
Focal deficit
2/5 (40)
5/21 (23.8)
Flow-related symptoms
0/2 (0)
8/20 (40)
0/5 (0)
1/21 (4.8)
>0.999
Symptomatic
5/5 (100)
18/22 (81.8)
0.561
Borden grade III
4/5 (80)
20/21 (95.2)
0.354
Deep venous drainage
4/5 (80)
11/19 (57.9)
0.615
Venous ectasia
3/5 (60)
14/22 (63.6)
>0.999
T2/FLAIR hyperintensity
1/4 (25)
1/15 (6.7)
0.386
Prior embolization
3/5 (60)
14/22 (63.6)
>0.999
Seizures
Prior surgery
2/5 (40)
5/22 (22.7)
0.580
Time from diagnosis to SRS (months)
2.6 0.9
4.7 4.1
0.491
Margin dose (Gy)
23.4 2.3
19 3.1
0.008y
1/5 (20)
2/22 (9.1)
0.474
28.8 35
60.2 53.4
0.226
35.5 47.3
49 51.1
0.674
Post-SRS hemorrhage Radiologic follow-up (months) Clinical follow-up (months)
Values are mean SD, number of patients (%), or as otherwise indicated. MI, myocardial infarction; CAD, coronary artery disease; Afib, atrial fibrillation; PVD, peripheral vascular disease; DM, diabetes mellitus; mRS, modified Rankin scale; NHND, nonhemorrhagic neurologic deficit; FLAIR, fluid-attenuated inversion recovery; SRS, stereotactic radiosurgery. *Or as otherwise indicated. yStatistically significant at P < 0.05.
WORLD NEUROSURGERY -: ---, - 2018
www.WORLDNEUROSURGERY.org
e5
ORIGINAL ARTICLE CHING-JEN CHEN ET AL.
SRS FOR HIGH-GRADE DAVFS
of a new neurologic deficit. The presence of a NHND at presentation was an independent predictor of a new permanent neurologic deficit after SRS (OR, 14.176; 95% CI, 1.119e179.540; P ¼ 0.041) in the multivariate model.
Table 5. Multivariable Logistic Regression Analysis for Predictors of a New Permanent Neurologic Deficit After Stereotactic Radiosurgery* Predictors
Odds Ratio
95% Confidence Interval
P Value
CAD
1.913
0.124e29.461
0.642
NHND
14.176
1.119e179.540
0.041y
Margin dose
1.045
0.793e1.377
0.751
CAD, coronary artery disease; NHND, nonhemorrhagic neurologic deficit. *Multivariable logistic regression model based on pooled parameter estimates from multiple imputation by chained equations with m ¼ 20. yStatistically significant at P < 0.05.
differences in these characteristics between the 2 groups. Three patients suffered a hemorrhage after SRS during the cumulative follow-up of 1095 patient-months, yielding an annual post-SRS hemorrhage rate of 3.3%. No significant difference in the incidences of post-SRS hemorrhage was found between those with (8%) and without (25%) the primary outcome (P ¼ 0.531). The durations of radiographic and clinical follow-up were not significantly different between the 2 groups. Univariate factors with P < 0.2 were the presence of deep venous drainage (P ¼ 0.129) and margin dose (P ¼ 0.169). Table 2 details the multivariable logistic regression analysis for predictors of the primary outcome. Neither the presence of deep venous drainage (odds ratio [OR], 7.15; 95% confidence interval [CI], 0.728e70.227; P ¼ 0.090) nor margin dose (OR, 0.916; 95% CI, 0.705e1.190; P ¼ 0.503) were found to be significantly associated with the primary outcome in the multivariate model. Obliteration Of the 27 patients with available neuroimaging follow-up, DAVF obliteration was achieved in 17 patients (63%), including 11 (65%) confirmed by angiography and 6 (35%) determined by MRI alone. Table 3 compares the baseline, radiographic, and clinical characteristics between patients with (n ¼ 17) and without (n ¼ 10) DAVF obliteration. No factors were found to be significantly associated with obliteration in the univariate analysis. Of the 10 patients who did not achieve obliteration after SRS, 1 (10%) underwent microsurgical ligation of the DAVF. Data regarding post-SRS interventions for the remaining 9 patients were not available. Post-SRS Neurologic Deficits Of the 30 patients with available clinical follow-up, 7 (23%) developed a new permanent neurologic deficit. Table 4 compares the baseline, radiographic, and clinical characteristics between patients with (n ¼ 7) and without (n ¼ 23) a new permanent neurologic deficit. NHND was significantly more common (100% vs. 25%, P ¼ 0.012) and the margin dose was significantly higher (mean, 23 vs. 19, P ¼ 0.008) in patients with a new permanent neurologic deficit, respectively. Univariate factors with P < 0.2 were coronary artery disease (P ¼ 0.085), NHND (P ¼ 0.012), and margin dose (P ¼ 0.008). Table 5 details the multivariate logistic regression analysis for predictors
e6
www.SCIENCEDIRECT.com
Illustrative Case A 71-year-old man who presented with right-sided synchronous pulsatile tinnitus was found to have a Borden type III right tentorial supplied by the right meningohypophyseal trunk, right inferolateral trunk, right occipital artery, petrosal branch of the right middle meningeal artery, and right accessory meningeal artery (Figure 1). There was direct pial venous drainage into the pontomesencephalic veins draining into the right basal vein of Rosenthal, and anterograde drainage via the right superior petrosal sinus and transverse sinus. The patient initially underwent endovascular treatment of the DAVF with coil embolization, followed by ethylene vinyl alcohol copolymer (Onyx [Medtronic Neurovascular, Irvine, California, USA]). However, after the embolization procedure, there was evidence of residual arteriovenous shunting. Therefore, the residual DAVF was treated with Gamma Knife SRS using a margin dose of 18 Gy to the 50% isodose line. The volume of the prescription isodose was 2.3 cm3. Follow-up angiography 3 years after SRS showed complete obliteration of the DAVF. DISCUSSION High-grade DAVFs are often associated with a relatively more aggressive clinical course and a higher risk of hemorrhage than low-grade DAVFs.12 Among 85 patients with high-grade DAVFs, Söderman et al.13 reported annual hemorrhage rates of 1.5% and 7.4% for unruptured and ruptured lesions, respectively. Similarly, Strom et al.14 reported annual hemorrhage rates of 1.4% and 7.6% for unruptured and ruptured high-grade DAVFs, respectively. The risk of rehemorrhage after the initial hemorrhage has been reported to be as high as 35% within the first 2 weeks.15 In addition to the hemorrhagic risk associated with high-grade DAVFs, persistence of CVR is associated with annual NHND and mortality rates of 6.9% and 10.4%, respectively.16 Given the considerable morbidity and mortality rates of untreated or partially treated high-grade DAVFs, urgent treatment is typically indicated. Embolization and microsurgical ligation are the preferred first-line interventions for symptomatic DAVFs because they can afford immediate closure of the arteriovenous shunt in most cases.17 In a pooled analysis of 19 studies comprising 743 DAVFs treated with SRS, we reported an obliteration rate of 56% for high-grade DAVFs versus 75% for low-grade DAVFs.10 Despite the relatively lower obliteration rates for high- versus low-grade DAVFs, SRS still confers therapeutic benefit in those who have failed or are unable to undergo surgery or embolization.10 In this single-center retrospective cohort study, we reported the outcomes specifically pertaining to patients with high-grade DAVFs who underwent SRS. Of the patients with neuroimaging records maintained at our institution, DAVF obliteration was achieved in 63%. In contrast with the prior DAVF SRS series from our institution, which included DAVFs of all types, the present study sought to identify predictors of post-SRS outcome specifically for high-grade DAVFs.6 Our analysis did not identify a statistically significant
WORLD NEUROSURGERY, https://doi.org/10.1016/j.wneu.2018.05.062
ORIGINAL ARTICLE CHING-JEN CHEN ET AL.
Figure 1. (A) Initial cerebral angiography, lateral view of a right internal carotid artery injection, shows that the right tentorial dural arteriovenous fistula (DAVF) is supplied by the right meningohypophyseal trunk and right inferolateral trunk with direct pial venous drainage into the pontomesencephalic veins draining into the right basal vein of Rosenthal. Initial cerebral angiography, (B) lateral and (C) anteroposterior views of a selective right occipital artery injection, shows that the right tentorial DAVF is supplied by the petrosal branch of the right middle meningeal artery and the right accessory meningeal artery, with direct pial venous drainage to
predictor of the primary outcome (i.e., DAVF obliteration without a new neurologic deficit) or obliteration after SRS. However, presentation with a NHND was found to be an independent predictor of a new permanent neurologic deficit after SRS. This may reflect the natural course of high-grade DAVFs with aggressive versus benign presentations.7 Strom et al.18 had suggested the classification of high-grade DAVFs into symptomatic versus asymptomatic subgroups by presenting symptoms, such that patients who presented with hemorrhage or NHND were classified as symptomatic, whereas those with incidentally diagnosed DAVFs or who presented with pulsatile tinnitus or orbital phenomena
WORLD NEUROSURGERY -: ---, - 2018
SRS FOR HIGH-GRADE DAVFS
the pontomesencephalic veins draining into the right basal vein of Rosenthal, and the right superior petrosal and transverse sinuses. (D) Postembolization (coils and Onyx) cerebral angiography, lateral view of a right common carotid artery injection, shows residual arteriovenous shunting. (E) The Gamma Knife treatment plan for the residual DAVF, which was treated with a margin dose of 18 Gy to the 50% isodose line (yellow). The volume of the prescription isodose was 2.3 cm3. The 8-Gy isodose line is shown in green.
were classified as asymptomatic. In the study comprising 11 symptomatic and 17 asymptomatic high-grade DAVFs, the authors reported significantly higher rates of hemorrhage or NHND in the symptomatic subgroup (45.5% vs. 5.9%, respectively; P ¼ 0.022).18 Higher rates of hemorrhage or NHND may indicate worsening venous outflow obstruction or impending overload of the drainage pathways, such that these aggressive subtypes of highgrade DAVFs are more vulnerable to subsequent venous hypertension or infarction.4,5,19 In a recent pooled analysis comparing SRS outcomes between aggressive (n ¼ 73) and nonaggressive (n ¼ 124)
www.WORLDNEUROSURGERY.org
e7
ORIGINAL ARTICLE CHING-JEN CHEN ET AL.
SRS FOR HIGH-GRADE DAVFS
high-grade DAVFs, Tonetti et al.20 reported significantly higher rates of post-SRS hemorrhage (7% vs. 0%, P ¼ 0.003) and radiation-induced complications (8% vs. 0%, P ¼ 0.001) for aggressive lesions, respectively. SRS could cause delayed occlusion of venous outflow pathways, thereby increasing the stress to an already tenuous venous drainage system in the aggressive subtype of high-grade DAVFs. Finally, lack of obliteration of the DAVF should be considered as a possible cause of new or worsening neurologic symptoms after SRS.9 Furthermore, high-grade DAVFs with an aggressive clinical presentation may be less likely to obliterate after SRS and therefore are more prone to causing subsequent neurologic deterioration.6,9,20 DAVF treatment should be individualized to each unique case, with consideration of both the DAVF’s angioarchitecture and the patient’s clinical presentation, baseline neurologic status, medical comorbidities, and preferences. Hemorrhage, progressive NHND, or signs of intracranial hypertension mandate urgent intervention. Treatment should also be offered to patients who have intolerable nonaggressive symptoms or high-grade DAVFs. Mounting evidence has suggested that asymptomatic high-grade DAVFs may confer low risks of hemorrhage and NHND with observation.12-14 However, nonaggressive high-grade DAVFs may transform into aggressive high-grade DAVFs over time. The risk and temporal course of this progression from a nonaggressive to aggressive subtype has yet to be fully delineated. Therefore, given the favorable overall risk to benefit profile of SRS for these lesions, intervention for nonaggressive high-grade DAVFs with SRS may be a reasonable management strategy for appropriately selected patients. DAVFs which fail to obliterate after SRS should be considered for salvage treatment with embolization, microsurgical ligation, or repeat SRS, after discussion with an experienced, multidisciplinary cerebrovascular team.
REFERENCES 1. Newton TH, Cronqvist S. Involvement of dural arteries in intracranial arteriovenous malformations. Radiology. 1969;93:1071-1078. 2. Borden JA, Wu JK, Shucart WA. A proposed classification for spinal and cranial dural arteriovenous fistulous malformations and implications for treatment. J Neurosurg. 1995;82:166-179. 3. Davies MA, TerBrugge K, Willinsky R, Coyne T, Saleh J, Wallace MC. The validity of classification for the clinical presentation of intracranial dural arteriovenous fistulas. J Neurosurg. 1996;85: 830-837. 4. Oh JT, Chung SY, Lanzino G, Park KS, Kim SM, Park MS, et al. Intracranial dural arteriovenous fistulas: clinical characteristics and management based on location and hemodynamics. J Cerebrovasc Endovasc Neurosurg. 2012;14:192. 5. Söderman M, Edner G, Ericson K, Karlsson B, Rähn T, Ulfarsson E, et al. Gamma knife surgery for dural arteriovenous shunts: 25 years of experience. J Neurosurg. 2006;104:867-875.
e8
www.SCIENCEDIRECT.com
It is important to recognize the limitations of this study, which is subject to the inherent weaknesses and biases of its retrospective and single-center design. Our analyses were limited by the relatively modest size of the study cohort, which may have caused our statistical tests to be underpowered. Because of institutional policies, data storage by the image management department may expire and become inaccessible, thereby resulting in missing data within our dataset. Although missing data were imputed using multiple imputations to avoid listwise deletions, the process itself is vulnerable to bias if the missing at random assumption does not hold true. We acknowledge that approximately one-third of patients who were found to have DAVF obliteration did not undergo confirmatory angiography. Because MRI has not been found to have equivalent accuracy compared with angiography for assessing obliteration of DAVFs treated with SRS, the obliteration rate of our study cohort may be as low as 41%. Given the rarity of DAVFs and the general utilization of SRS as a salvage therapy, single-center analysis is challenging. Future multicenter efforts, including those through the Consortium for Dural Arteriovenous Fistula Outcomes Research and the International Gamma Knife Research Foundation, may allow for more robust analyses for predictors of success and complications after SRS for high-grade DAVFs. CONCLUSIONS Most appropriately selected patients with high-grade DAVFs who undergo treatment with SRS achieve lesional obliteration without suffering a new permanent neurologic deficit. Presentation with a NHND is a risk factor for new permanent neurologic deficit after SRS. Future investigations of larger, multicenter cohorts may afford more adequately powered analyses and provide additional insights regarding predictors of outcome after SRS for these uncommon vascular lesions.
6. Cifarelli CP, Kaptain G, Yen CP, Schlesinger D, Sheehan JP. Gamma Knife radiosurgery for dural arteriovenous fistulas. Neurosurgery. 2010;67: 1230-1235. 7. Della Pepa GM, Parente P, D’Argento F, Pedicelli A, Sturiale CL, Sabatino G, et al. Angioarchitectural features of high-grade intracranial dural arteriovenous fistulas: correlation with aggressive clinical presentation and hemorrhagic risk. Neurosurgery. 2017;81:315-330. 8. Seo Y, Kim Dong G, Dho YS, Kim JW, Kim YH, Park CK, et al. A retrospective analysis of the outcomes of dural arteriovenous fistulas treated with gamma knife radiosurgery: a singleinstitution experience. Stereotact Funct Neurosurg. 2018;96:46-53. 9. Hanakita S, Koga T, Shin M, Shojima M, Igaki H, Saito N. Role of stereotactic radiosurgery in the treatment of high-grade cerebral arteriovenous malformation. Neurol Med Chir (Tokyo). 2012;52: 845-851. 10. Chen CJ, Lee CC, Ding D, Starke RM, Chivukula S, Yen CP, et al. Stereotactic radiosurgery for intracranial dural arteriovenous
fistulas: a systematic review. J Neurosurg. 2015;122: 353-362. 11. Rubin DB. Multiple Imputation for Nonresponse in Surveys. New York, NY: John Wiley & Sons, Inc.; 1987. 12. Gross BA, Du R. The natural history of cerebral dural arteriovenous fistulae. Neurosurgery. 2012;71: 594-603. 13. Söderman M, Pavic L, Edner G, Holmin S, Andersson T. Natural history of dural arteriovenous shunts. Stroke. 2008;39:1735-1739. 14. Strom R, Botros J, Refai D, Moran CJ, Cross DT 3rd, Chicoine MR, et al. E-004 Cranial dural arteriovenous fistulas: asymptomatic cortical venous drainage portends less aggressive clinical course. J Neurointerv Surg. 2010;2(suppl 1):A27. 15. Duffau H, Lopes M, Janosevic V, Sichez JP, Faillot T, Capelle L, et al. Early rebleeding from intracranial dural arteriovenous fistulas: report of 20 cases and review of the literature. J Neurosurg. 1999;90:78-84. 16. van Dijk JM, terBrugge KG, Willinsky RA, Wallace MC. Clinical course of cranial dural
WORLD NEUROSURGERY, https://doi.org/10.1016/j.wneu.2018.05.062
ORIGINAL ARTICLE CHING-JEN CHEN ET AL.
arteriovenous fistulas with long-term persistent cortical venous reflux. Stroke. 2002;33:1233-1236. 17. Gross BA, Albuquerque FC, Moon K, McDougall CG. Evolution of treatment and a detailed analysis of occlusion, recurrence, and clinical outcomes in an endovascular library of 260 dural arteriovenous fistulas. J Neurosurg. 2017;126:1884-1893. 18. Strom RG, Botros JA, Refai D, Moran CJ, Cross DT 3rd, Chicoine MR, et al. Cranial dural arteriovenous fistulae: asymptomatic cortical venous drainage portends less aggressive clinical course. Neurosurgery. 2009;64:241-247 [discussion: 247-248].
WORLD NEUROSURGERY -: ---, - 2018
SRS FOR HIGH-GRADE DAVFS
19. Kwon BJ, Han MH, Kang HS, Chang KH. MR imaging findings of intracranial dural arteriovenous fistulas: relations with venous drainage patterns. Am J Neuroradiol. 2005;26:2500-2507. 20. Tonetti DA, Gross BA, Jankowitz BT, Kano H, Monaco EA 3rd, Niranjan A, et al. Reconsidering an important subclass of high-risk dural arteriovenous fistulas for stereotactic radiosurgery. J Neurosurg. 2018:1-5.
commercial or financial relationships that could be construed as a potential conflict of interest. Received 24 March 2018; accepted 9 May 2018 Citation: World Neurosurg. (2018). https://doi.org/10.1016/j.wneu.2018.05.062 Journal homepage: www.WORLDNEUROSURGERY.org Available online: www.sciencedirect.com 1878-8750/$ - see front matter ª 2018 Elsevier Inc. All rights reserved.
Conflict of interest statement: The authors declare that the article content was composed in the absence of any
www.WORLDNEUROSURGERY.org
e9