See the corresponding editorial in this issue, pp 970–972.
J Neurosurg 120:973–981, 2014 ©AANS, 2014
Stereotactic radiosurgery for Spetzler-Martin Grade III arteriovenous malformations Clinical article Hideyuki Kano, M.D., Ph.D.,1 John C. Flickinger, M.D., 2 Huai-che Yang, M.D.,1,4 Thomas J. Flannery, M.D., Ph.D.,1 Daniel Tonetti, M.S., 3 Ajay Niranjan, M.Ch., M.B.A.,1 and L. Dade Lunsford, M.D.1 Departments of 1Neurological Surgery and 2Radiation Oncology, and 3University of Pittsburgh School of Medicine, Pittsburgh, Pennsylvania; and 4Department of Neurosurgery, Taipei Veterans General Hospital, Taipei, Taiwan Object. The purpose of this study was to define the outcomes and risks of stereotactic radiosurgery (SRS) for Spetzler-Martin (SM) Grade III arteriovenous malformations (AVMs). Methods. Between 1987 and 2009, SRS was performed in 474 patients with SM Grade III AVMs. The AVMs were categorized by scoring the size (S), drainage (D), and location (L): IIIa was a small AVM (S1D1L1, N = 282); IIIb was a medium/deep AVM (S2D1L0, N = 44); and IIIc was a medium/eloquent AVM (S2D0L1, N = 148). The median target volume was 3.8 ml (range 0.1–26.3 ml) and the margin dose was 20 Gy (range 13–25 Gy). Eighty-one patients (17%) underwent prior embolization, and 58 (12%) underwent prior resection. Results. At a mean follow-up of 89 months, the total obliteration rates documented by angiography or MRI for all SM Grade III AVMs increased from 48% at 3 years to 69% at 4 years, 72% at 5 years, and 77% at 10 years. The SM Grade IIIa AVMs were more likely to obliterate than other subgroups. The cumulative rate of hemorrhage was 2.3% at 1 year, 4.4% at 2 years, 5.5% at 3 years, 6.4% at 5 years, and 9% at 10 years. The SM Grade IIIb AVMs had a significantly higher cumulative rate of hemorrhage. Symptomatic adverse radiation effects were detected in 6%. Conclusions. Treatment with SRS was an effective and relatively safe management option for SM Grade III AVMs. Although patients with residual AVMs remained at risk for hemorrhage during the latency interval, the cumulative 10-year 9% hemorrhage risk in this series may represent a significant reduction compared with the expected natural history. (http://thejns.org/doi/abs/10.3171/2013.12.JNS131600)
Key Words • arteriovenous malformation • Gamma Knife • hemorrhage • stereotactic radiosurgery • Spetzler-Martin grade • vascular disorders
T
Spetzler-Martin (SM) grading system is a simple, widely accepted, practical tool for assessing the outcomes associated with microsurgical management of brain arteriovenous malformations (AVMs). At experienced vascular centers, the grading system has demonstrated that microsurgery is an effective and relatively safe option for patients with SM Grade I or II AVMs.3,4,6,7,10,11,13,16 In contrast, Grade IV and V AVMs are associated with higher risks and less success regardless of the option selected.13 The SM Grade III AVMs are a heterogeneous group that includes different subtypes of he
Abbreviations used in this paper: ARE = adverse radiation effect; AVM = arteriovenous malformation; HR = hazard ratio; SM = Spetzler-Martin; SRS = stereotactic radiosurgery.
J Neurosurg / Volume 120 / April 2014
AVMs according to their size, location in critical brain regions, and venous drainage. Nonetheless, relatively few surgical series use a subclassification system to report outcomes of treatment.4,12,21 Pandey et al.17 were among the first to report the multimodality management of SM Grade III AVMs based on subgroup analysis. Stereotactic radiosurgery (SRS) has been widely used to manage SM Grade III AVMs. The goal of SRS is complete obliteration of the AVM nidus while avoiding postprocedure adverse radiation effects (AREs). Total obliteration appears to reduce the cumulative residual lifetime risk of hemorrhage, even though patients remain at risk for hemorrhage during the latency interval between SRS and obliteration.15 In this report we describe the outcomes of patients who underwent SRS for such AVMs, 973
H. Kano et al. and we use subgroup analysis to evaluate total obliteration, risk of hemorrhage, and complications of Grade III AVM radiosurgery.
Methods Patient Population
Between 1987 and 2009, single-stage SRS was performed using the Leksell Gamma Knife (Elekta AB) in 474 patients with SM Grade III AVMs. Eighty-one patients (17%) underwent prior embolization and 58 (12%) underwent prior resection. The AVMs were categorized as follows: Grade IIIa denoted a small AVM (size < 3 cm, presence of deep venous drainage, critical location); Grade IIIb was a medium/deep AVM (size 3–6 cm, presence of deep venous drainage, noncritical location); Grade IIIc was a medium/eloquent AVM (size 3–6 cm, absence of deep venous drainage, critical location); and Grade IIId was a large AVM (size > 6 cm, absence of deep venous drainage, noncritical location). Of the 474 patients with SM Grade III AVMs, 282 (59%) had Grade IIIa, 44 (9%) had Grade IIIb, and 148 (31%) had Grade IIIc. No Grade IIId AVMs were treated in this study. A history of prior hemorrhage before SRS was seen in 74% of Grade IIIa AVMs, in 32% of Grade IIIb AVMs, and in 30% of Grade IIIc AVMs. The patient demographics and AVM characteristics are summarized in Table 1.
Radiosurgery Technique
Our radiosurgical technique has been described in detail in previous reports.10 The margin SRS dose was crafted to include the entire AVM nidus volume, defined as the shunt between the afferent arteries and draining veins. Successfully embolized volumes were not included in the SRS target volume. The SRS was performed with either a model U, B, C, or 4-C Leksell Gamma Knife (Elekta AB). The median diameter of the AVM nidus was 19 mm (range 6–29 mm) in Grade IIIa, 32 mm (range 30–47 mm) in Grade IIIb, and 30 mm (range 30–45 mm) in Grade IIIc. The median target volume was 2.2 ml in Grade IIIa, 7.7 ml in Grade IIIb, and 6.6 ml in Grade IIIc. The median margin dose was 20 Gy in Grade IIIa, 17 Gy in Grade IIIb, and 18 Gy in Grade IIIc. Patient Follow-Up
After radiosurgery, patients were instructed to undergo clinical and imaging assessments at 6-, 12-, 24-, and 36-month intervals. At the end of 3 years, if MRI suggested complete obliteration, a follow-up angiogram was requested. Complete AVM obliteration was defined as no detectable residual nidus at the time of follow-up angiography, or no evidence of residual flow voids on contrast-enhanced T1- and T2-weighted MRI. Complete angiographic AVM obliteration was defined as disappearance of the nidus and absence of early venous drainage. At any time when a new neurological symptom or sign developed, the patient underwent CT and/or MRI studies to rule out hemorrhage or AREs. This retrospective study was approved by the University of Pittsburgh Institutional Review Board.
974
Statistical Analysis
Kaplan-Meier survival analysis was carried out to calculate rates of total obliteration and hemorrhage in each subgroup. Patients were censored when lost to follow-up or at the time of an event such as total obliteration and hemorrhage. Consequently, rates of hemorrhage reflect a single hemorrhage experienced by a patient even if he or she had more than one hemorrhage. Annual hemorrhage rates were calculated based on years of followup and total number of hemorrhages. The log-rank test was used to assess differences in survival curves, and Cox regression was used to assess hazard ratios (HRs) in multivariate analysis. A value of p < 0.05 was used for statistical significance.
Case Matching of SM Grade IIIb and IIIc AVMs
Forty-four patients with SM Grade IIIb AVMs were eligible to be case control matches in a comparison with 148 patients with SM Grade IIIc AVMs. Propensity score matching was performed between the SM Grade IIIb case cohort and the SM Grade IIIc control cohort.2,8,9 The propensity score was calculated by fitting a logistic regression model in which the following 7 variables were used: age, number of prior hemorrhages, prior embolization, target volume, presence of coexisting aneurysm, radiosurgery-based AVM score, and follow-up duration. We used a nearest neighbor 1:2 matching algorithm based on the propensity score. After propensity score matching, the Mann-Whitney U-test for continuous data and the Fisher exact test for categorical data were used to compare both groups.
Results
At the time of assessment, 426 patients were alive and 48 had died. Twenty patients died of a brain hemorrhage, 1 patient died of untreated symptoms of AREs, and 27 patients died of causes unrelated to their AVM. The median follow-up after SRS was 89 months (range 2–278 months). All living patients had at least 1 year of imaging and clinical assessment. Response to Radiosurgery
Obliteration of AVM was documented by MRI in 258 patients and by angiography in 198 patients at intervals that ranged from 3 to 10 years after SRS. The total obliteration rates (based on either angiography or MRI criteria) were 48% at 3 years, 69% at 4 years, 72% at 5 years, and 77% at 10 years. The total obliteration rates based on angiography alone were 39% at 3 years, 57% at 4 years, 59% at 5 years, and 62% at 10 years. Obliteration rates calculated by angiography alone in this experience are biased (artificially lowered) by the exclusion of patients with MRI-defined obliteration who declined to undergo angiography. The exact time of obliteration remains unknown because the timing of documentation by imaging varies. The 5-year total obliteration rates after SRS were 74% in Grade IIIa, 72% in Grade IIIb, and 69% in Grade IIIc (Fig. 1). In univariate analysis, factors associated with a J Neurosurg / Volume 120 / April 2014
Stereotactic radiosurgery for Spetzler-Martin Grade III AVMs TABLE 1: Demographic data and AVM characteristics in 474 patients whose lesions were treated with SRS* SM Grade† Characteristic
IIIa, S1D1L1
IIIb, S2D1L0
IIIc, S2D0L1
total no. median age in yrs (range) sex M/F location frontal temporal parietal occipital corpus callosum basal ganglia thalamus cerebellum brainstem intraventricular pineal region prior hemorrhages pre-SRS prior resection alone prior embolization alone prior embolization + surgery presence of coexisting aneurysm median diameter of nidus in mm (range) median target vol in ml (range) median margin dose in Gy (range)
282 33 (3–79)
44 37 (7–65)
148 35 (5–76)
152:130
23:21
77:71
21 25 15 18 11 46 63 21 55 3 4 210 (74%) 36 (13%) 28 (10%) 10 (4%) 15 (5%) 19 (6–29) 2.2 (0.1–16) 20 (13–25)
10 7 6 9 3 3 1 4 1 0 0 14 (32%) 4 (9%) 8 (18%) 2 (5%) 3 (7%) 32 (30–47) 7.7 (3.6–21) 17 (14–22)
42 22 41 34 0 0 0 9 0 0 0 44 (30%) 4 (3%) 31 (21%) 2 (1%) 14 (9%) 30 (30–45) 6.6 (1.2–26) 18 (13–25)
* Values are number of patients (%) unless stated otherwise. S, D, L = size, drainage, location. † Grade IIIa = small; IIIb = medium/deep; IIIc = medium/eloquent.
higher rate of total obliteration on MRI included maximum diameter < 3 cm (Grade IIIa vs IIIb and IIIc) (Fig. 1), eloquent location (Grade IIIc vs IIIa and IIIb), Grade IIIa compared with IIIb, Grade IIIa compared with IIIc, smaller target volume, higher margin dose, compact nidus shape, no prior embolization, prior hemorrhage, and lower radiosurgery-based AVM score (Table 2). In multivariate analysis, factors associated with a higher rate of total obliteration on MRI included higher margin dose (p < 0.0001, HR 1.11, 95% CI 1.07–1.15) and no prior embolization (p = 0.001, HR 0.52, 95% CI 0.36–0.77) (Table 3). In patients with prior hemorrhage, the median target volume was 5.9 ml (range 0.17–26.3 ml), the median maximum diameter of the nidus was 2.8 cm (range 0.9–4.7 cm), and the median margin dose was 18.5 Gy (range 13.8–25.2 Gy). In patients without prior hemorrhage, the median target volume was 2.5 ml (range 0.1–18.6 ml), the median maximum diameter of the nidus was 2 cm (range 0.6–4.2 cm), and the median margin dose was 20 Gy (range 13–25 Gy). In patients with prior hemorrhage, this was significantly associated with smaller nidus volume (p < 0.001), smaller maximum diameter (p < 0.001), and higher margin dose (p < 0.001) than in those without prior hemorrhage. J Neurosurg / Volume 120 / April 2014
Hemorrhage Before and After Radiosurgery
Thirty-four patients had 38 AVM hemorrhages after SRS. Two patients had 2 hemorrhages, the second of which was fatal in both patients. One patient had 3 hemorrhages and underwent repeat SRS at another hospital. Eighteen patients died after a single AVM hemorrhage at a median of 20 months (range 4–152 months) after SRS. The mortality rate due to AVM hemorrhage after SRS was 4.2%. In 1385 patient-years of estimated hemorrhage risk (the interval from the date of SRS to the date of total obliteration on angiography or the date of last followup imaging showing a residual AVM), we confirmed 38 hemorrhages, which corresponds to an annual hemorrhage rate of 2.7%. In the 1st year after SRS (462.2 patient-years with 474 patients) 12 hemorrhages occurred; in the 2nd year (415 patient-years with 466 patients) 8 occurred; in the 3rd year (374.7 patient-years with 393 patients) 4 occurred; in the 4th–5th years (584 patient-years with 347 patients) 5 occurred; and in the 6th–10th years (875.7 patient-years with 252 patients) 5 hemorrhages occurred. The annual hemorrhage rates in patients without obliteration, in the interval of 0–1 year, was 2.6%, at 1–2 years it was 1.9%, at 2–3 years it was 1.1%, at 3–5 years it was 0.9%, and at 975
H. Kano et al.
Fig. 1. Left: Kaplan-Meier curves for total obliteration on MRI or angiography after radiosurgery for AVMs with SM Grade IIIa (small) versus IIIb (medium/deep) versus IIIc (medium/eloquent). Right: Kaplan-Meier curves for total obliteration on MRI or angiography after radiosurgery for AVMs with SM Grade IIIa versus IIIb plus IIIc. The SM Grade IIIa subgroup was significantly associated with a higher rate of total obliteration.
5–10 years it was 0.6%. Based on Kaplan-Meier analysis (excluding the second hemorrhage in the patient who had more than 2 ruptures), the cumulative rate of AVM hemorrhage after SRS was 2.3% at 1 year, 4.4% at 2 years, 5.5% at 3 years, 6.4% at 5 years, and 9% at 10 years. Eighteen patients (6.4%) with SM Grade IIIa AVMs, 7 (15.9%) with SM Grade IIIb, and 9 (6.1%) with SM Grade IIIc had hemorrhages after SRS. The cumulative 5-year hemor-
rhage rates were as follows: Grade IIIa AVMs 4.9%, IIIb 14.9%, and IIIc 7.6%. No patient bled after documentation of AVM obliteration was obtained using either MRI or angiography. The detailed outcomes of univariate and multivariate analyses of hemorrhage rates after SRS are shown in Tables 2 and 3. In univariate analysis, factors associated with an increased hemorrhage rate after SRS included older age,
TABLE 2: Univariate analysis of total obliteration and bleeding after radiosurgery* Factor
TO (angio or MRI)
TO (angio)
Bleeding After SRS
age (younger) sex max diameter
J Neurosurg 120:973–981, 2014 ©AANS, 2014
Stereotactic radiosurgery for Spetzler-Martin Grade III arteriovenous malformations Clinical article Hideyuki Kano, M.D., Ph.D.,1 John C. Flickinger, M.D., 2 Huai-che Yang, M.D.,1,4 Thomas J. Flannery, M.D., Ph.D.,1 Daniel Tonetti, M.S., 3 Ajay Niranjan, M.Ch., M.B.A.,1 and L. Dade Lunsford, M.D.1 Departments of 1Neurological Surgery and 2Radiation Oncology, and 3University of Pittsburgh School of Medicine, Pittsburgh, Pennsylvania; and 4Department of Neurosurgery, Taipei Veterans General Hospital, Taipei, Taiwan Object. The purpose of this study was to define the outcomes and risks of stereotactic radiosurgery (SRS) for Spetzler-Martin (SM) Grade III arteriovenous malformations (AVMs). Methods. Between 1987 and 2009, SRS was performed in 474 patients with SM Grade III AVMs. The AVMs were categorized by scoring the size (S), drainage (D), and location (L): IIIa was a small AVM (S1D1L1, N = 282); IIIb was a medium/deep AVM (S2D1L0, N = 44); and IIIc was a medium/eloquent AVM (S2D0L1, N = 148). The median target volume was 3.8 ml (range 0.1–26.3 ml) and the margin dose was 20 Gy (range 13–25 Gy). Eighty-one patients (17%) underwent prior embolization, and 58 (12%) underwent prior resection. Results. At a mean follow-up of 89 months, the total obliteration rates documented by angiography or MRI for all SM Grade III AVMs increased from 48% at 3 years to 69% at 4 years, 72% at 5 years, and 77% at 10 years. The SM Grade IIIa AVMs were more likely to obliterate than other subgroups. The cumulative rate of hemorrhage was 2.3% at 1 year, 4.4% at 2 years, 5.5% at 3 years, 6.4% at 5 years, and 9% at 10 years. The SM Grade IIIb AVMs had a significantly higher cumulative rate of hemorrhage. Symptomatic adverse radiation effects were detected in 6%. Conclusions. Treatment with SRS was an effective and relatively safe management option for SM Grade III AVMs. Although patients with residual AVMs remained at risk for hemorrhage during the latency interval, the cumulative 10-year 9% hemorrhage risk in this series may represent a significant reduction compared with the expected natural history. (http://thejns.org/doi/abs/10.3171/2013.12.JNS131600)
Key Words • arteriovenous malformation • Gamma Knife • hemorrhage • stereotactic radiosurgery • Spetzler-Martin grade • vascular disorders
T
Spetzler-Martin (SM) grading system is a simple, widely accepted, practical tool for assessing the outcomes associated with microsurgical management of brain arteriovenous malformations (AVMs). At experienced vascular centers, the grading system has demonstrated that microsurgery is an effective and relatively safe option for patients with SM Grade I or II AVMs.3,4,6,7,10,11,13,16 In contrast, Grade IV and V AVMs are associated with higher risks and less success regardless of the option selected.13 The SM Grade III AVMs are a heterogeneous group that includes different subtypes of he
Abbreviations used in this paper: ARE = adverse radiation effect; AVM = arteriovenous malformation; HR = hazard ratio; SM = Spetzler-Martin; SRS = stereotactic radiosurgery.
J Neurosurg / Volume 120 / April 2014
AVMs according to their size, location in critical brain regions, and venous drainage. Nonetheless, relatively few surgical series use a subclassification system to report outcomes of treatment.4,12,21 Pandey et al.17 were among the first to report the multimodality management of SM Grade III AVMs based on subgroup analysis. Stereotactic radiosurgery (SRS) has been widely used to manage SM Grade III AVMs. The goal of SRS is complete obliteration of the AVM nidus while avoiding postprocedure adverse radiation effects (AREs). Total obliteration appears to reduce the cumulative residual lifetime risk of hemorrhage, even though patients remain at risk for hemorrhage during the latency interval between SRS and obliteration.15 In this report we describe the outcomes of patients who underwent SRS for such AVMs, 973
H. Kano et al. and we use subgroup analysis to evaluate total obliteration, risk of hemorrhage, and complications of Grade III AVM radiosurgery.
Methods Patient Population
Between 1987 and 2009, single-stage SRS was performed using the Leksell Gamma Knife (Elekta AB) in 474 patients with SM Grade III AVMs. Eighty-one patients (17%) underwent prior embolization and 58 (12%) underwent prior resection. The AVMs were categorized as follows: Grade IIIa denoted a small AVM (size < 3 cm, presence of deep venous drainage, critical location); Grade IIIb was a medium/deep AVM (size 3–6 cm, presence of deep venous drainage, noncritical location); Grade IIIc was a medium/eloquent AVM (size 3–6 cm, absence of deep venous drainage, critical location); and Grade IIId was a large AVM (size > 6 cm, absence of deep venous drainage, noncritical location). Of the 474 patients with SM Grade III AVMs, 282 (59%) had Grade IIIa, 44 (9%) had Grade IIIb, and 148 (31%) had Grade IIIc. No Grade IIId AVMs were treated in this study. A history of prior hemorrhage before SRS was seen in 74% of Grade IIIa AVMs, in 32% of Grade IIIb AVMs, and in 30% of Grade IIIc AVMs. The patient demographics and AVM characteristics are summarized in Table 1.
Radiosurgery Technique
Our radiosurgical technique has been described in detail in previous reports.10 The margin SRS dose was crafted to include the entire AVM nidus volume, defined as the shunt between the afferent arteries and draining veins. Successfully embolized volumes were not included in the SRS target volume. The SRS was performed with either a model U, B, C, or 4-C Leksell Gamma Knife (Elekta AB). The median diameter of the AVM nidus was 19 mm (range 6–29 mm) in Grade IIIa, 32 mm (range 30–47 mm) in Grade IIIb, and 30 mm (range 30–45 mm) in Grade IIIc. The median target volume was 2.2 ml in Grade IIIa, 7.7 ml in Grade IIIb, and 6.6 ml in Grade IIIc. The median margin dose was 20 Gy in Grade IIIa, 17 Gy in Grade IIIb, and 18 Gy in Grade IIIc. Patient Follow-Up
After radiosurgery, patients were instructed to undergo clinical and imaging assessments at 6-, 12-, 24-, and 36-month intervals. At the end of 3 years, if MRI suggested complete obliteration, a follow-up angiogram was requested. Complete AVM obliteration was defined as no detectable residual nidus at the time of follow-up angiography, or no evidence of residual flow voids on contrast-enhanced T1- and T2-weighted MRI. Complete angiographic AVM obliteration was defined as disappearance of the nidus and absence of early venous drainage. At any time when a new neurological symptom or sign developed, the patient underwent CT and/or MRI studies to rule out hemorrhage or AREs. This retrospective study was approved by the University of Pittsburgh Institutional Review Board.
974
Statistical Analysis
Kaplan-Meier survival analysis was carried out to calculate rates of total obliteration and hemorrhage in each subgroup. Patients were censored when lost to follow-up or at the time of an event such as total obliteration and hemorrhage. Consequently, rates of hemorrhage reflect a single hemorrhage experienced by a patient even if he or she had more than one hemorrhage. Annual hemorrhage rates were calculated based on years of followup and total number of hemorrhages. The log-rank test was used to assess differences in survival curves, and Cox regression was used to assess hazard ratios (HRs) in multivariate analysis. A value of p < 0.05 was used for statistical significance.
Case Matching of SM Grade IIIb and IIIc AVMs
Forty-four patients with SM Grade IIIb AVMs were eligible to be case control matches in a comparison with 148 patients with SM Grade IIIc AVMs. Propensity score matching was performed between the SM Grade IIIb case cohort and the SM Grade IIIc control cohort.2,8,9 The propensity score was calculated by fitting a logistic regression model in which the following 7 variables were used: age, number of prior hemorrhages, prior embolization, target volume, presence of coexisting aneurysm, radiosurgery-based AVM score, and follow-up duration. We used a nearest neighbor 1:2 matching algorithm based on the propensity score. After propensity score matching, the Mann-Whitney U-test for continuous data and the Fisher exact test for categorical data were used to compare both groups.
Results
At the time of assessment, 426 patients were alive and 48 had died. Twenty patients died of a brain hemorrhage, 1 patient died of untreated symptoms of AREs, and 27 patients died of causes unrelated to their AVM. The median follow-up after SRS was 89 months (range 2–278 months). All living patients had at least 1 year of imaging and clinical assessment. Response to Radiosurgery
Obliteration of AVM was documented by MRI in 258 patients and by angiography in 198 patients at intervals that ranged from 3 to 10 years after SRS. The total obliteration rates (based on either angiography or MRI criteria) were 48% at 3 years, 69% at 4 years, 72% at 5 years, and 77% at 10 years. The total obliteration rates based on angiography alone were 39% at 3 years, 57% at 4 years, 59% at 5 years, and 62% at 10 years. Obliteration rates calculated by angiography alone in this experience are biased (artificially lowered) by the exclusion of patients with MRI-defined obliteration who declined to undergo angiography. The exact time of obliteration remains unknown because the timing of documentation by imaging varies. The 5-year total obliteration rates after SRS were 74% in Grade IIIa, 72% in Grade IIIb, and 69% in Grade IIIc (Fig. 1). In univariate analysis, factors associated with a J Neurosurg / Volume 120 / April 2014
Stereotactic radiosurgery for Spetzler-Martin Grade III AVMs TABLE 1: Demographic data and AVM characteristics in 474 patients whose lesions were treated with SRS* SM Grade† Characteristic
IIIa, S1D1L1
IIIb, S2D1L0
IIIc, S2D0L1
total no. median age in yrs (range) sex M/F location frontal temporal parietal occipital corpus callosum basal ganglia thalamus cerebellum brainstem intraventricular pineal region prior hemorrhages pre-SRS prior resection alone prior embolization alone prior embolization + surgery presence of coexisting aneurysm median diameter of nidus in mm (range) median target vol in ml (range) median margin dose in Gy (range)
282 33 (3–79)
44 37 (7–65)
148 35 (5–76)
152:130
23:21
77:71
21 25 15 18 11 46 63 21 55 3 4 210 (74%) 36 (13%) 28 (10%) 10 (4%) 15 (5%) 19 (6–29) 2.2 (0.1–16) 20 (13–25)
10 7 6 9 3 3 1 4 1 0 0 14 (32%) 4 (9%) 8 (18%) 2 (5%) 3 (7%) 32 (30–47) 7.7 (3.6–21) 17 (14–22)
42 22 41 34 0 0 0 9 0 0 0 44 (30%) 4 (3%) 31 (21%) 2 (1%) 14 (9%) 30 (30–45) 6.6 (1.2–26) 18 (13–25)
* Values are number of patients (%) unless stated otherwise. S, D, L = size, drainage, location. † Grade IIIa = small; IIIb = medium/deep; IIIc = medium/eloquent.
higher rate of total obliteration on MRI included maximum diameter < 3 cm (Grade IIIa vs IIIb and IIIc) (Fig. 1), eloquent location (Grade IIIc vs IIIa and IIIb), Grade IIIa compared with IIIb, Grade IIIa compared with IIIc, smaller target volume, higher margin dose, compact nidus shape, no prior embolization, prior hemorrhage, and lower radiosurgery-based AVM score (Table 2). In multivariate analysis, factors associated with a higher rate of total obliteration on MRI included higher margin dose (p < 0.0001, HR 1.11, 95% CI 1.07–1.15) and no prior embolization (p = 0.001, HR 0.52, 95% CI 0.36–0.77) (Table 3). In patients with prior hemorrhage, the median target volume was 5.9 ml (range 0.17–26.3 ml), the median maximum diameter of the nidus was 2.8 cm (range 0.9–4.7 cm), and the median margin dose was 18.5 Gy (range 13.8–25.2 Gy). In patients without prior hemorrhage, the median target volume was 2.5 ml (range 0.1–18.6 ml), the median maximum diameter of the nidus was 2 cm (range 0.6–4.2 cm), and the median margin dose was 20 Gy (range 13–25 Gy). In patients with prior hemorrhage, this was significantly associated with smaller nidus volume (p < 0.001), smaller maximum diameter (p < 0.001), and higher margin dose (p < 0.001) than in those without prior hemorrhage. J Neurosurg / Volume 120 / April 2014
Hemorrhage Before and After Radiosurgery
Thirty-four patients had 38 AVM hemorrhages after SRS. Two patients had 2 hemorrhages, the second of which was fatal in both patients. One patient had 3 hemorrhages and underwent repeat SRS at another hospital. Eighteen patients died after a single AVM hemorrhage at a median of 20 months (range 4–152 months) after SRS. The mortality rate due to AVM hemorrhage after SRS was 4.2%. In 1385 patient-years of estimated hemorrhage risk (the interval from the date of SRS to the date of total obliteration on angiography or the date of last followup imaging showing a residual AVM), we confirmed 38 hemorrhages, which corresponds to an annual hemorrhage rate of 2.7%. In the 1st year after SRS (462.2 patient-years with 474 patients) 12 hemorrhages occurred; in the 2nd year (415 patient-years with 466 patients) 8 occurred; in the 3rd year (374.7 patient-years with 393 patients) 4 occurred; in the 4th–5th years (584 patient-years with 347 patients) 5 occurred; and in the 6th–10th years (875.7 patient-years with 252 patients) 5 hemorrhages occurred. The annual hemorrhage rates in patients without obliteration, in the interval of 0–1 year, was 2.6%, at 1–2 years it was 1.9%, at 2–3 years it was 1.1%, at 3–5 years it was 0.9%, and at 975
H. Kano et al.
Fig. 1. Left: Kaplan-Meier curves for total obliteration on MRI or angiography after radiosurgery for AVMs with SM Grade IIIa (small) versus IIIb (medium/deep) versus IIIc (medium/eloquent). Right: Kaplan-Meier curves for total obliteration on MRI or angiography after radiosurgery for AVMs with SM Grade IIIa versus IIIb plus IIIc. The SM Grade IIIa subgroup was significantly associated with a higher rate of total obliteration.
5–10 years it was 0.6%. Based on Kaplan-Meier analysis (excluding the second hemorrhage in the patient who had more than 2 ruptures), the cumulative rate of AVM hemorrhage after SRS was 2.3% at 1 year, 4.4% at 2 years, 5.5% at 3 years, 6.4% at 5 years, and 9% at 10 years. Eighteen patients (6.4%) with SM Grade IIIa AVMs, 7 (15.9%) with SM Grade IIIb, and 9 (6.1%) with SM Grade IIIc had hemorrhages after SRS. The cumulative 5-year hemor-
rhage rates were as follows: Grade IIIa AVMs 4.9%, IIIb 14.9%, and IIIc 7.6%. No patient bled after documentation of AVM obliteration was obtained using either MRI or angiography. The detailed outcomes of univariate and multivariate analyses of hemorrhage rates after SRS are shown in Tables 2 and 3. In univariate analysis, factors associated with an increased hemorrhage rate after SRS included older age,
TABLE 2: Univariate analysis of total obliteration and bleeding after radiosurgery* Factor
TO (angio or MRI)
TO (angio)
Bleeding After SRS
age (younger) sex max diameter
