Arterial spin labeling magnetic resonance imaging: toward noninvasive diagnosis and follow-up of pediatric brain arteriovenous malformations

Full access

OBJECT

Arterial spin labeling (ASL)-MRI is becoming a routinely used sequence for ischemic strokes, as it quantifies cerebral blood flow (CBF) without the need for contrast injection. As brain arteriovenous malformations (AVMs) are highflow vascular abnormalities, increased CBF can be identified inside the nidus or draining veins. The authors aimed to analyze the relevance of ASL-MRI in the diagnosis and follow-up of children with brain AVM.

METHODS

The authors performed a retrospective analysis of 21 patients who had undergone digital subtraction angiography (DSA) and pseudo-continuous ASL-MRI for the diagnosis or follow-up of brain AVM after radiosurgery or embolization. They compared the AVM nidus location between ASL-MRI and 3D contrast-enhanced T1 MRI, as well as the CBF values obtained in the nidus (CBFnidus) and the normal cortex (CBFcortex) before and after treatment.

RESULTS

The ASL-MRI correctly demonstrated the nidus location in all cases. Nidal perfusion (mean CBFnidus 137.7 ml/100 mg/min) was significantly higher than perfusion in the contralateral normal cortex (mean CBFcortex 58.6 ml/100 mg/min; p < 0.0001, Mann-Whitney test). Among 3 patients followed up after embolization, a reduction in both AVM size and CBF values was noted. Among 5 patients followed up after radiosurgery, a reduction in the nidus size was observed, whereas CBFnidus remained higher than CBFcortex.

CONCLUSIONS

In this study, ASL-MRI revealed nidus location and patency after treatment thanks to its ability to demonstrate focal increased CBF values. Absolute quantification of CBF values could be relevant in the follow-up of pediatric brain AVM after partial treatment, although this must be confirmed in larger prospective trials.

ABBREVIATIONSASL = arterial spin labeling; AVM = arteriovenous malformation; CBF = cerebral blood flow; DSA = digital subtraction angiography; ROI = region of interest.

Abstract

OBJECT

Arterial spin labeling (ASL)-MRI is becoming a routinely used sequence for ischemic strokes, as it quantifies cerebral blood flow (CBF) without the need for contrast injection. As brain arteriovenous malformations (AVMs) are highflow vascular abnormalities, increased CBF can be identified inside the nidus or draining veins. The authors aimed to analyze the relevance of ASL-MRI in the diagnosis and follow-up of children with brain AVM.

METHODS

The authors performed a retrospective analysis of 21 patients who had undergone digital subtraction angiography (DSA) and pseudo-continuous ASL-MRI for the diagnosis or follow-up of brain AVM after radiosurgery or embolization. They compared the AVM nidus location between ASL-MRI and 3D contrast-enhanced T1 MRI, as well as the CBF values obtained in the nidus (CBFnidus) and the normal cortex (CBFcortex) before and after treatment.

RESULTS

The ASL-MRI correctly demonstrated the nidus location in all cases. Nidal perfusion (mean CBFnidus 137.7 ml/100 mg/min) was significantly higher than perfusion in the contralateral normal cortex (mean CBFcortex 58.6 ml/100 mg/min; p < 0.0001, Mann-Whitney test). Among 3 patients followed up after embolization, a reduction in both AVM size and CBF values was noted. Among 5 patients followed up after radiosurgery, a reduction in the nidus size was observed, whereas CBFnidus remained higher than CBFcortex.

CONCLUSIONS

In this study, ASL-MRI revealed nidus location and patency after treatment thanks to its ability to demonstrate focal increased CBF values. Absolute quantification of CBF values could be relevant in the follow-up of pediatric brain AVM after partial treatment, although this must be confirmed in larger prospective trials.

Brain arteriovenous malformations (AVMs) are the main cause of intracerebral hemorrhage in children and represent 30%–50% of pediatric hemorrhagic strokes.12,14 Although successful treatment of small AVMs can be safely and quickly achieved with microsurgery, more complex lesions often require multimodal therapy consisting of endovascular procedures and radiosurgery.3,4,8 Effects of the latter are delayed, and obliteration rarely occurs before 2 or 3 years afterward. Moreover, obliteration is not always achieved, and long-term imaging follow-up is needed to assess the completeness of treatment. The recurrence of pediatric brain AVM has been reported in up to 5% of cases, even as late as 16 years after initial complete treatment.4,8,19 Thus, long-term imaging follow-up is required even after complete occlusion of a brain AVM. Digital subtraction angiography (DSA) is the reference standard for diagnosis and for tailoring the appropriate treatment. Apart from offering high diagnostic performance in the detection of untreated brain AVM, DSA enables visualization of residual shunt after partial treatment.3,10 However, DSA remains an invasive tool, requiring sedation or general anesthesia in children, with the potential risk, albeit very low, for periprocedural complications, and it exposes both patients and medical staff to repeated ionizing radiation, which in children is associated with an increased risk of developing solid cancers or leukemia.7

Numerous noninvasive imaging techniques, such as arterial spin labeling (ASL)-MRI, have already shown promise in the management of acute and chronic cerebrovascular disease in adults.5,18,19 Arterial spin labeling is an MRI technique that can noninvasively assess cerebral blood flow (CBF) by magnetically labeling the arterial water spins with a radiofrequency pulse. The general principle of ASL-MRI consists of using the arterial blood as an endogenous contrast agent. By continuous inversion of the arterial blood's water proton nuclear spins in the cervical arteries, ASL-MRI has allowed accurate and reliable quantification of CBF. A direct application is acute and chronic cerebrovascular disease because quantification of reduced CBF allows one to detect areas of penumbra and vascular collaterals and to objectively follow cerebral reperfusion after therapeutic interventions.25 Moreover, ASL-MRI gives an overview of cortical perfusion in the brain besides the pathological process. We hypothesized that—as recently shown in small adult cohorts2,5,20—by enabling focal detection of increased CBF in the nidus of a malformation,24 ASL-MRI may have interesting applications in the management of pediatric brain AVMs.

Our aim was to study the ability of ASL-MRI to demonstrate nidus patency and localization in untreated and partially treated pediatric brain AVMs.

Methods

Study Design

The study conformed to the scientific principles and research ethics standards of our institution. The paper was prepared in accordance with Standards for the Reporting of Diagnostic Accuracy Studies (STARD) guidelines.1 Our institutional review board waived the need for written informed consent from participants. We retrospectively analyzed all children (age < 18 years) with a brain AVM, either untreated or partially treated, who had undergone ASL-MRI and DSA during the same week in the last 3 years.

Magnetic Resonance Imaging Procedures

Magnetic resonance imaging was performed during the last 3 years on a 1.5-T GE Signa HDxt system (General Electric Healthcare) using a 12-channel head-neck-spine coil. The brain MRI protocol included at least 3D T1 sequences before and after Gd injection (3D-Gd-T1 MRI), axial FLAIR sequences, axial T2 sequences, and diffusion and noncontrast perfusion sequences with 3D pseudo-continuous ASL-MRI (axial partitions 40, field of view 240 × 240 × 4 mm3, acquisition matrix 8 spiral arms in each 3D partition, TE 10.5 msec, TR 4428 msec, post-labeling delay 1025 msec, flip angle 155°, acquisition time 4 min 17 sec). No anesthesia or sedation was required during MRI acquisition. When consecutive ASL-MRI studies were conducted following radiosurgery or embolization, the last imaging session was performed between 5 days and 2 years after the last procedure.

Catheter Angiography

All endovascular procedures were performed with the patient under general anesthesia, using 3- or 4-Fr catheters introduced via the right femoral route, according to patient age and weight. A 3- or 4-Fr coaxial system (guiding catheter, Glidecath, Terumo) was positioned over a 0.28 or 0.35 guide successively in the proximal internal and external carotid artery and cervical vertebral artery. Face, lateral, and oblique views (IGS Innova 530, GE Healthcare) were obtained after selective injection of 3–5 ml of iodinated contrast (Guerbet).

Imaging Analysis

All imaging data were anonymized and analyzed on a dedicated workstation (Advantage Windows, version 4.2). Three-dimensional T1 sequences and ASL sequences were automatically realigned and fused in a volume viewer. A CBF map was automatically generated, and the color scale was set to rainbow.

Two readers (a neuroradiologist and a neurosurgeon with 10 and 8 years of experience in pediatric imaging, respectively) decided by consensus whether CBF maps and 3D-Gd-T1 sequences showed anatomical colocalization of the AVM nidus. For each patient, we performed, in the axial plane of the ASL-MRI study in which the nidus was the largest, a manual segmentation of the AVM to define the region of interest (ROI) in which the AVM nidus area and nidus CBF (CBFnidus) mean values were automatically calculated by the software. Average CBF values for normal cortex of the contralateral hemisphere (CBFcortex) were calculated using the average values of 2 cortical ROIs of 2 cm2 that were positioned in the frontal and parietal regions. This allowed us to calculate the relative AVM CBF, defined as the ratio of CBFnidus/CBFcortex.

Statistical Analysis

We compared values between CBFnidus and CBFcortex using a nonparametric Mann-Whitney test. Significance level was set at p = 0.05. We did not perform statistical analysis for patients who had CBF quantification before and after a therapeutic procedure (embolization or Gamma Knife) because of the small number of patients. Values are expressed as the mean ± standard deviation, unless indicated otherwise.

Results

Patients and AVM Characteristics

Among 21 included patients (14 males), whose mean age was 10.8 ± 3.9 years (range 3.6–17 years), 14 had imaging for a ruptured AVM associated with an intracerebral hematoma (Table 1). Magnetic resonance imaging was performed for headaches in 3 cases, for late follow-up of a presumably obliterated AVM in 2 cases (4 and 5 years after treatment), for epilepsy in 1 case, and for developmental delay in 1 case.

TABLE 1

Summary of patients and imaging characteristics*

Case No.Patient Age (yrs), Sex at DiagnosisMode of RevelationAVM CharacteristicsAVM Treatment (date)ASL1 Characteristics (date of imaging after diagnosis; CBFnidus [ml/100 mg/min]; nidus diameter [cm2]; CBFcortex [ml/100 mg/min])ASL2 Characteristics (date of imaging; CBFnidus [ml/100 mg/min]; nidus diameter [cm2]; CBFcortex [ml/100 mg/min])
113.5, MICHRt temporal lobe, SM GIEmbol ×2, GK (01/12)06/2012; 148.7; 2.4; 5906/2013 (18 mos after GK); 156; 1.2; 54
28.6, FHeadachesLt basal ganglia, SM GVEmbol ×1, GK (02/12)05/2011; 135; 3.5; 6512/2012 (10 mos after GK); 113; 2.7; 68
313.7, MICHRt basal ganglia, SM GVEmbol ×6, GK (05/11)10/2011; 112; 5; 503/2013 (22 mos after GK); 97; 2.7; 48
44.1, FICHLt temporal lobe, SM GIIIEmbol ×1, GK (3/13)6/2012; 71; 3.4; 51.67/2013 (4 mos after GK); 70; 1.7; 60
59.5, MICHRt basal ganglia, SM GIVEmbol ×2, GK (12/11)5/2012; 244; 3.1; 875/2013 (18 mos after GK); 72; 0.9; 66
611, MICHRt frontal, SM GIEmbol ×312/2012 (pre-embol); 260; 2.15; 5907/2013 (incomplete embol); 128; 1.4; 62
713.4, MEpilepsyLt central, SM GIVEmbol ×4At diagnosis; 242; 11.4; 59
813.3, FICHCerebellar culmen, SM GIVEmbol ×3NA
98.1, MICHLt temporal lobe, SM GIINTAt diagnosis; 124; 0.5; 68
103.5, FDevelopmental delayBrainstem, SM GIVNTAt diagnosis; 139; 0.6; 57.5
1116.9, MICHLt parietal, SM GIEmbol ×1, surgeryAt diagnosis; 115; 0.7; 69
1210, FICHLt occipital, SM GIIIEmbol ×2 GKAt diagnosis; 98; 1.2; 68
138.5, FICH, status epilepticusRt occipital, SM GIIIEmbol ×1At diagnosis; 128; 81; 137
1415, MFU for obliterated AVMLt frontal, SM GIIISurgery (4 yrs earlier)At diagnosis; 92; 0.9; 63Postop; no AVM on ASL MRI
155.7, MICHRt central, SM GIIEmbol ×112/2013 (pre-embol); 92; 1; 5702/2014 (complete embol); No AVM on ASL-MRI
1616.5, FFU for obliterated AVMLt frontal, SM GIIGK (5 yrs earlier)At diagnosis; 117.8; 2; 71
178.4, MICHLt frontal, SM GIVNTAt diagnosis; 105; 1.4; 42
1810, MICHLt central, SM GIIINTAt diagnosis; 123; 3.7; 51
1912.9, MICHLt occipital, SM GIIEmbol ×102/12/213 (pre-embol); 131; 1.3; 5107/12/2013 (incomplete embol); 86; 0.7; 45
2017, MHeadachesBrainstem, SM GIVNTAt diagnosis; 145; 5.1; 53
218.2, MHeadachesLt basal ganglia, SM GIIINTAt diagnosis; 121.6; 1.6; 41

Embol = embolization; FU = follow-up; GK = Gamma Knife; ICH = intracerebral hemorrhage; NA = not available; NT = not treated; SM G(I–V) = Spetzler-Martin grade.

Comparison of 3D-Gd-T1 sequences with ASL-MRI demonstrated colocalization in all patients.

The AVM was located on the left side in 12 cases, and lobar localization was as follows (number of patients): basal ganglia (4), central region (3), frontal lobe (4), occipital lobe (3), temporal lobe (3), posterior fossa (3: medulla oblongata, mesencephalon, and cerebellum vermis), and parietal lobe (1). Spetzler-Martin grades were as follows: Grade I, 3; Grade II, 4; Grade III, 6; Grade IV, 6; and Grade V, 2.

Diagnosis of AVM and ASL-MRI

Visual analysis of the CBF maps revealed the AVM nidus in all 21 cases. Fusion of 3D-Gd-T1 images and ASL-MRI studies showed anatomical colocalization of the nidus in all 21 cases (Figs. 13).

FIG. 1.
FIG. 1.

Conventional DS angiograms (1), 3D-Gd-T1 MR images (2), and ASL-MRI fusion with 3D-Gd-T1 MRI (3) in 3 cases of large AVMs: Cases 3 (A), 1 (B), and 8 (C). Increased CBF values appear in red. There is colocalization of the AVM nidus in both MRI sequences. Figure is available in color online only.

FIG. 2.
FIG. 2.

Conventional DS angiograms (1), 3D-Gd-T1 MR images (2), 3D time of flight (TOF; 3), and ASL-MRI fusion with 3D-Gd-T1 MRI (4) in 2 cases of small AVMs: Cases 12 (A) and 11 (B). There is a colocalization of the AVM nidus (arrows) in the 3D-Gd-T1, 3D-TOF, and ASL sequences. Figure is available in color online only.

FIG. 3.
FIG. 3.

Case 6. Digital subtraction angiogram (A), 3D-Gd-T1 MR image (B), and raw ASL CBF maps (C) in a patient with a large left frontal hematoma associated with an AVM. Increased CBF values appear in red. Fusion of ASL and 3D-Gd-T1 MRI (D). Figure is available in color online only.

The mean AVM nidus area was 2.7 ± 2.5 cm2 (range 0.5–11.4 cm2). The mean CBFnidus (137.7 ± 51 ml/100 mg/min, range 92–260 ml/100 mg/min) was significantly higher than CBFcortex (58.6 ± 9 ml/100 mg/min, range 41–71 ml/100 mg/min; p < 0.0001). The mean relative AVM CBF was 2.25 ± 0.7 (range 1.43–4.10). The mean CBFnidus was higher in larger AVMs, and there was a correlation between mean CBFnidus and mean AVM nidus areas as well (Pearson correlation, r = 0.631, p = 0.003). There was no difference in the mean values for CBFnidus according to venous drainage type (p = 0.739, Mann-Whitney test) or localization (superficial vs deep, p = 0.824, Mann-Whitney test).

Follow-Up of AVM

Among the 21 patients, 9 had undergone ASL-MRI both before and after therapeutic procedures: surgery (1), embolization (3), or radiosurgery (5; Table 1). After surgical treatment, no nidus was left, and therefore CBF was normal. After complete embolization in 1 case, ASL-MRI was normal. In 2 cases of incomplete embolization, we measured a reduction in the AVM nidus surface (2.15 and 1.3 cm2 before vs 1.4 and 0.7 cm2 after treatment) and the mean CBFnidus (260 and 131 ml/100 mg/min before vs 128 and 86 ml/100 mg/min after embolization). As 3D angiography was not performed, we could not correlate the reduction in nidal volume on angiography with the reduction in the surface of the nidus on ASL-MRI. After radiosurgery (Fig. 4), the mean CBFnidus (142.1 vs 102.8 ml/100 mg/min) and AVM nidus areas (3.48 vs 0.82 cm2) decreased before and after treatment, respectively, whereas CBFcortex was stable (62.5 vs 59.2 ml/100 mg/min). However, statistical analysis could not be performed owing to the small number of patients.

FIG. 4.
FIG. 4.

Case 5. Conventional DS angiogram obtained before radiosurgery (A); ASL-MRI obtained 6 months after Gamma Knife (B); and ASL-MRI obtained 18 months after Gamma Knife (C). Graph (D) showing the evolution of nidal CBF in Cases 1–5 after radiosurgery. Graph (E) showing evolution of the nidal area, in cm2, in Cases 1–5 after radiosurgery. There was a reduction in the nidus area but not in the perfusion values inside the nidus. Case 6. Initial angiogram (F) showing a right frontal lobe AVM. Initial ASL-MRI (G). Last angiogram (H) showing partial treatment of the AVM. Last ASL-MRI (I) showing a decrease in the size of the nidus. Figure is available in color online only.

Pseudo-Continuous ASL-MRI in a Case of Seizure

One patient was admitted for a right-sided occipital hematoma (Case 13), intracranial hypertension, and partial status epilepticus with visual hallucinations and metamorphopsia in the left visual field that disappeared after 2 days of antiepileptic drugs but were present during MRI acquisition. Interestingly, we observed hyperperfusion (CBF: 135 ml/100 mg/min in epileptic cortex vs 81 ml/100 mg/min in normal cortex and 128 ml/100 mg/min in AVM) in the parietooccipital junction, where there was no AVM or draining veins (Fig. 5). We postulate that this area was the symptomatogenic epileptic zone. A control MRI study performed 2 months later, when the patient was without seizure, showed the normalization of CBF in the parietooccipital cortex (104 ml/100 mg/min), whereas the nidus remained patent.

FIG. 5.
FIG. 5.

Case 13. Conventional DS angiogram (A) shows a small compact right occipital AVM with 2 feeders (middle cerebral artery and posterior cerebral artery). An axial 3D-Gd-T1 MR image (B) shows the right occipital hematoma and a small compact AVM nidus. An ASL-MR image (C) shows hyperperfusion of the nidus at the posterior part of the hematoma. An ASL-MR image (D) shows hyperperfusion of the parietooccipital area, compatible with ictal hyperperfusion of the epileptogenic zone. Axial FLAIR MR image (E) reveals the absence of cerebral edema in the parietooccipital junction. An ASL-MR image (F) performed 2 months after intracerebral hemorrhage, when the patient was seizure free. The CBF map is back to normal. Arrows indicate the ROI—that is, the AVM or the hyperperfusion related to the seizure. Figure is available in color online only.

Discussion

In this preliminary study, we found that 1.5-T ASL-MRI demonstrated nidus location and patency in a cohort of 21 pediatric patients with brain AVM, in accordance with DSA and contrast-enhanced T1 sequences. To our knowledge, this is the first description of clinical interest in ASL-MRI in pediatric brain AVMs. We quantified the hyperperfusion of the nidus as compared with perfusion in remote normal cerebral cortex, and we objectively measured the modifications of CBF after therapeutic interventions.

Different imaging techniques are available to evaluate brain hemodynamics, such as PET, SPECT, xenonenhanced CT, susceptibility-weighted MRI perfusion, or Doppler ultrasound.2,20 All of these techniques except ultrasound require the injection of a contrast agent.

We demonstrated that visual detection of nidus patency was possible in all cases, and image fusion showed excellent anatomical colocalization between ASL-MRI and 3D-Gd-T1 sequences. These results are in accordance with those in a recent study on 26 adults with 15 cases of DSA-proven vascular malformations (8 dural arteriovenous fistulas and 7 AVMs) and 11 cases of intracerebral hemorrhage without vascular malformations.5 Detection of venous ASL signal hyperintensity enabled the authors to detect vascular lesions with a sensibility and specificity of 78% and 85%, respectively. Moreover, multivariate analysis showed that the detection of ASL-MRI abnormality increased 17 times the chance of detecting an AVM on DSA. Increased CBF was also noted in 7 cases of AVM by using ASL-MRI, and these authors observed a “steal phenomenon,” as decreased CBF was noted in the ipsilateral basal ganglia as compared with that in the contralateral side.21,23

We encountered one false-positive case in a patient with status epilepticus and increased CBF values distant from the AVM nidus. Focally increased signal intensity on ASL-MRI may be produced by true hyperperfusion related to neurovascular coupling or by an artifact such as motion or intravascular spin label. Arterial spin labeling is sensitive in revealing hyperperfusion in a number of conditions related to tumor,17 seizure, or loss of autoregulatory function of blood vessels. The regional hyperperfusion implicated in hypertensive encephalopathy and reversible encephalopathy syndrome can be readily disclosed by ASL-MRI, which on serial follow-up can demonstrate the normalization often observed in these conditions.6,23

Our results showed that ASL-MRI offers reproducible quantitative measurement of CBF. This imaging modality uses radiofrequency-labeled water protons in the neck vessels as an endogenous tracer. Hence, there is no variability in the concentration of the contrast agent. Moreover, in children, the absence of atherosclerotic stenosis of the carotid or vertebral arteries is another advantage in regard to the reproducibility of blood flow analysis. We observed a small standard deviation (6.5 ml/100 mg/min) in the CBF of normal cortical areas, in agreement with published data, and we did not observe significant variations in CBF in the normal cortex on subsequent acquisitions despite an interval of 18 months between MRIs.

We observed mean CBFnidus values that were twice those of CBFcortex. This result may explain why visual analysis could easily detect the nidus and could be relevant in the early diagnosis of brain AVM in children with intracerebral hematoma, although this must be confirmed in dedicated prospective studies.

One can argue that ASL-MRI performed at 1.5 T has limitations. In adults, 1.5-T ASL-MRI suffers from relatively low signal-to-noise ratio. Interestingly, in children, a consistent pattern of increased signal-to-noise ratio as well as globally elevated absolute CBF has been observed, as compared with that in adults. Possible explanations for this globally increased signal intensity include higher baseline CBF, faster mean transit time, increased baseline magnetization values in gray and white matter, and increased T1 values in blood and tissue (that is, a longer tracer half-life). Decreased susceptibility artifact at the base of the skull from immature paranasal sinus development may also play a role in improved image quality and added signal intensity in the frontal and inferior regions. Finally, the ASL-MRI sequence lasts 5 minutes, is feasible with all MRI setups, and does not require intravenous injection. In the pediatric population, these factors are highly relevant for the etiological diagnosis of an intracerebral hemorrhage or the diagnosis of an AVM recurrence. Another major application of ASL-MRI is the follow-up of AVMs treated with radiosurgery even though DSA is still the gold standard for proving AVM obliteration.15,22 Numerous noninvasive imaging modalities, such as time-of-flight or dynamic contrast MRA, have also proven useful in monitoring obliteration dynamics, although they only provide volumetry of the nidus and do not quantify CBF variations.11,16,22 In the present study and as previously reported in an adult cohort with 8 AVMs, 1 dural fistula, and 1 vein of Galen malformation treated with partial embolization, quantification of the nidus diameter and perfusion values allowed us to follow up patients after embolization or radiosurgery. Some authors have noted an average reduction of 36% in the relative amount of signal measured with ASL.16 There has been good correlation between the estimation of nidus obliteration with DSA and the ASL-MRI observations. As obliteration after radiosurgery rarely occurs before 2 or 3 years after treatment, with an obliteration rate of about 70%–80% and a rebleeding rate of 6%, long-term follow-up is mandatory in such patients.9,14 Thus, ASL-MRI could help in the noninvasive monitoring of nidus obliteration before DSA confirms it, as previously suggested.13 Moreover, this imaging modality could decrease the economic cost of the follow-up of brain AVMs. Last but not least, ASL-MRI studies can be used during therapeutic procedures. The fusion of CBF maps with anatomical MRI can be integrated into neuronavigation software during microsurgical or radiosurgical procedures.

Our study suffered several limitations. First, although data were prospectively acquired, analysis was retrospective. Magnetic resonance imaging sequences were acquired at different times after therapeutic procedures, and this could affect their interpretation. Second, the population consisted only of patients referred for brain AVM. Because readers were aware that all patients had been referred for the brain AVMs, sensitivity in detecting the nidus location might have been artificially increased. In addition, we did not analyze the specificity of ASL-MRI. Third, pediatric CBF is prone to large variation from a neonatal age until 8 years old and then gradually decreases to adult levels. This may hamper the possibility of a quantitative long-term follow-up based on CBF values. These limitations might be overcome in the near future with the further refinement of ASL-MRI techniques and the use of 3-T MRI. At our institution, we now use 1.5-T ASL-MRI in all patients harboring a brain AVM. We continue to strongly advocate for the use of DSA in patients with presumed nidus obliteration on ASL-MRI. However, the latter technique might help to reduce the number of DSA examinations performed for brain AVMs, and thus saving children from potential procedural risks. This strategy may change as we gain more experience and with technical improvements in this sequence.

Conclusions

Our preliminary observations support ASL as a sensitive MRI sequence for the diagnosis of nidus patency in brain AVMs. Quantification of nidal CBF may enable objective monitoring of AVM obliteration after radiosurgery or partial embolization. Additional, larger prospective studies are required to confirm our initial results.

Author Contributions

Conception and design: Blauwblomme, Sainte-Rose. Acquisition of data: Blauwblomme, Brunelle, Grévent, Di Rocco, Beccaria, Paternoster, Bourgeois, Kossorotoff. Analysis and interpretation of data: Blauwblomme, Naggara, Brunelle. Drafting the article: Blauwblomme. Critically revising the article: Naggara, Grévent, Puget, Kossorotoff, Sainte-Rose. Reviewed submitted version of manuscript: Naggara, Brunelle, Puget, Sainte-Rose. Approved the final version of the manuscript on behalf of all authors: Blauwblomme. Statistical analysis: Blauwblomme. Administrative/technical/material support: Zerah, Sainte-Rose. Study supervision: Zerah, Sainte-Rose.

References

  • 1

    Bossuyt PMReitsma JBBruns DEGatsonis CAGlasziou PPIrwig LM: Towards complete and accurate reporting of studies of diagnostic accuracy: the STARD Initiative. Radiology 226:24282003

  • 2

    Calamante FThomas DLPell GSWiersma JTurner R: Measuring cerebral blood flow using magnetic resonance imaging techniques. J Cereb Blood Flow Metab 19:7017351999

  • 3

    Darsaut TEGuzman RMarcellus MLEdwards MSTian LDo HM: Management of pediatric intracranial arteriovenous malformations: experience with multimodality therapy. Neurosurgery 69:5405562011

  • 4

    Kader AGoodrich JTSonstein WJStein BMCarmel PWMichelsen WJ: Recurrent cerebral arteriovenous malformations after negative postoperative angiograms. J Neurosurg 85:14181996

  • 5

    Le TTFischbein NJAndré JBWijman CRosenberg JZaharchuk G: Identification of venous signal on arterial spin labeling improves diagnosis of dural arteriovenous fistulas and small arteriovenous malformations. AJNR Am J Neuroradiol 33:61682012

  • 6

    Lüdemann LJedrzejewski GHeidenreich JHan ETBruhn H: Perfusion imaging of cerebral arteriovenous malformations: a study comparing quantitative continuous arterial spin labeling and dynamic contrast-enhanced magnetic resonance imaging at 3 T. Magn Reson Imaging 29:115711642011

  • 7

    Miglioretti DLJohnson EWilliams AGreenlee RTWeinmann SSolberg LI: The use of computed tomography in pediatrics and the associated radiation exposure and estimated cancer risk. JAMA Pediatr 167:7007072013

  • 8

    Morgan MKPatel NJSimons MRitson EAHeller GZ: Influence of the combination of patient age and deep venous drainage on brain arteriovenous malformation recurrence after surgery. J Neurosurg 117:9349412012

  • 9

    Nataf FGhossoub MSchlienger MMoussa RMeder JFRoux FX: Bleeding after radiosurgery for cerebral arteriovenous malformations. Neurosurgery 55:2983062004

  • 10

    Niazi TNKlimo P JrAnderson RCRaffel C: Diagnosis and management of arteriovenous malformations in children. Neurosurg Clin N Am 21:4434562010

  • 11

    Nogueira RGBayrlee AHirsch JAYoo AJCopen WA: Dynamic contrast-enhanced MRA at 1.5 T for detection of arteriovenous shunting before and after Onyx embolization of cerebral arteriovenous malformations. J Neuroimaging 23:5145172013

  • 12

    O'Lynnger TMAl-Holou WNGemmete JJPandey ASThompson BGGarton HJ: The effect of age on arteriovenous malformations in children and young adults undergoing magnetic resonance imaging. Childs Nerv Syst 27:127312792011

  • 13

    Pollock JMWhitlow CTSimonds JStevens EAKraft RABurdette JH: Response of arteriovenous malformations to gamma knife therapy evaluated with pulsed arterial spinlabeling MRI perfusion. AJR Am J Roentgenol 196:15222011

  • 14

    Smith ERButler WEOgilvy CS: Surgical approaches to vascular anomalies of the child's brain. Curr Opin Neurol 15:1651712002

  • 15

    Steinberg GKFabrikant JIMarks MPLevy RPFrankel KAPhillips MH: Stereotactic heavy-charged-particle Bragg-peak radiation for intracranial arteriovenous malformations. N Engl J Med 323:961011990

  • 16

    Suazo LFoerster BFermin RSpeckter HVilchez COviedo J: Measurement of blood flow in arteriovenous malformations before and after embolization using arterial spin labeling. Interv Neuroradiol 18:42482012

  • 17

    Tourdias TRodrigo SOppenheim CNaggara OVarlet PAmoussa S: Pulsed arterial spin labeling applications in brain tumors: practical review. J Neuroradiol 35:79892008

  • 18

    van Laar PJvan der Grond JHendrikse J: Brain perfusion territory imaging: methods and clinical applications of selective arterial spin-labeling MR imaging. Radiology 246:3543642008

  • 19

    Williams DSDetre JALeigh JSKoretsky AP: Magnetic resonance imaging of perfusion using spin inversion of arterial water. Proc Natl Acad Sci U S A 89:2122161992

  • 20

    Wintermark MSesay MBarbier EBorbély KDillon WPEastwood JD: Comparative overview of brain perfusion imaging techniques. Stroke 36:e83e992005

  • 21

    Wolf RLWang JDetre JAZager ELHurst RW: Arteriovenous shunt visualization in arteriovenous malformations with arterial spin-labeling MR imaging. AJNR Am J Neuroradiol 29:6816872008

  • 22

    Wowra BMuacevic ATonn JCSchoenberg SOReiser MHerrmann KA: Obliteration dynamics in cerebral arteriovenous malformations after CyberKnife radiosurgery: quantification with sequential nidus volumetry and 3-tesla 3-dimensional time-of-flight magnetic resonance angiography. Neurosurgery 64:2 SupplA102A1092009

  • 23

    Yamashita KYoshiura THiwatashi ATogao OYoshimoto KSuzuki SO: Arterial spin labeling of hemangioblastoma: differentiation from metastatic brain tumors based on quantitative blood flow measurement. Neuroradiology 54:8098132012

  • 24

    Yu SLWang RWang RWang SYao YQZhang D: Accuracy of vessel-encoded pseudocontinuous arterial spinlabeling in identification of feeding arteries in patients with intracranial arteriovenous malformations. AJNR Am J Neuroradiol 35:65712014

  • 25

    Zaharchuk GDo HMMarks MPRosenberg JMoseley MESteinberg GK: Arterial spin-labeling MRI can identify the presence and intensity of collateral perfusion in patients with moyamoya disease. Stroke 42:248524912011

If the inline PDF is not rendering correctly, you can download the PDF file here.

Article Information

Correspondence Thomas Blauwblomme, Service de Neurochirurgie Pédiatrique, Hôpital Necker, 149 rue de Sèvres, 75015 Paris, France. email: thomas.blauwblomme@nck.aphp.fr.

INCLUDE WHEN CITING Published online January 30, 2015; DOI: 10.3171/2014.9.PEDS14194.

DISCLOSURE The authors report no conflict of interest concerning the materials or methods used in this study or the findings specified in this paper.

© AANS, except where prohibited by US copyright law.

Headings

Figures

  • View in gallery

    Conventional DS angiograms (1), 3D-Gd-T1 MR images (2), and ASL-MRI fusion with 3D-Gd-T1 MRI (3) in 3 cases of large AVMs: Cases 3 (A), 1 (B), and 8 (C). Increased CBF values appear in red. There is colocalization of the AVM nidus in both MRI sequences. Figure is available in color online only.

  • View in gallery

    Conventional DS angiograms (1), 3D-Gd-T1 MR images (2), 3D time of flight (TOF; 3), and ASL-MRI fusion with 3D-Gd-T1 MRI (4) in 2 cases of small AVMs: Cases 12 (A) and 11 (B). There is a colocalization of the AVM nidus (arrows) in the 3D-Gd-T1, 3D-TOF, and ASL sequences. Figure is available in color online only.

  • View in gallery

    Case 6. Digital subtraction angiogram (A), 3D-Gd-T1 MR image (B), and raw ASL CBF maps (C) in a patient with a large left frontal hematoma associated with an AVM. Increased CBF values appear in red. Fusion of ASL and 3D-Gd-T1 MRI (D). Figure is available in color online only.

  • View in gallery

    Case 5. Conventional DS angiogram obtained before radiosurgery (A); ASL-MRI obtained 6 months after Gamma Knife (B); and ASL-MRI obtained 18 months after Gamma Knife (C). Graph (D) showing the evolution of nidal CBF in Cases 1–5 after radiosurgery. Graph (E) showing evolution of the nidal area, in cm2, in Cases 1–5 after radiosurgery. There was a reduction in the nidus area but not in the perfusion values inside the nidus. Case 6. Initial angiogram (F) showing a right frontal lobe AVM. Initial ASL-MRI (G). Last angiogram (H) showing partial treatment of the AVM. Last ASL-MRI (I) showing a decrease in the size of the nidus. Figure is available in color online only.

  • View in gallery

    Case 13. Conventional DS angiogram (A) shows a small compact right occipital AVM with 2 feeders (middle cerebral artery and posterior cerebral artery). An axial 3D-Gd-T1 MR image (B) shows the right occipital hematoma and a small compact AVM nidus. An ASL-MR image (C) shows hyperperfusion of the nidus at the posterior part of the hematoma. An ASL-MR image (D) shows hyperperfusion of the parietooccipital area, compatible with ictal hyperperfusion of the epileptogenic zone. Axial FLAIR MR image (E) reveals the absence of cerebral edema in the parietooccipital junction. An ASL-MR image (F) performed 2 months after intracerebral hemorrhage, when the patient was seizure free. The CBF map is back to normal. Arrows indicate the ROI—that is, the AVM or the hyperperfusion related to the seizure. Figure is available in color online only.

References

1

Bossuyt PMReitsma JBBruns DEGatsonis CAGlasziou PPIrwig LM: Towards complete and accurate reporting of studies of diagnostic accuracy: the STARD Initiative. Radiology 226:24282003

2

Calamante FThomas DLPell GSWiersma JTurner R: Measuring cerebral blood flow using magnetic resonance imaging techniques. J Cereb Blood Flow Metab 19:7017351999

3

Darsaut TEGuzman RMarcellus MLEdwards MSTian LDo HM: Management of pediatric intracranial arteriovenous malformations: experience with multimodality therapy. Neurosurgery 69:5405562011

4

Kader AGoodrich JTSonstein WJStein BMCarmel PWMichelsen WJ: Recurrent cerebral arteriovenous malformations after negative postoperative angiograms. J Neurosurg 85:14181996

5

Le TTFischbein NJAndré JBWijman CRosenberg JZaharchuk G: Identification of venous signal on arterial spin labeling improves diagnosis of dural arteriovenous fistulas and small arteriovenous malformations. AJNR Am J Neuroradiol 33:61682012

6

Lüdemann LJedrzejewski GHeidenreich JHan ETBruhn H: Perfusion imaging of cerebral arteriovenous malformations: a study comparing quantitative continuous arterial spin labeling and dynamic contrast-enhanced magnetic resonance imaging at 3 T. Magn Reson Imaging 29:115711642011

7

Miglioretti DLJohnson EWilliams AGreenlee RTWeinmann SSolberg LI: The use of computed tomography in pediatrics and the associated radiation exposure and estimated cancer risk. JAMA Pediatr 167:7007072013

8

Morgan MKPatel NJSimons MRitson EAHeller GZ: Influence of the combination of patient age and deep venous drainage on brain arteriovenous malformation recurrence after surgery. J Neurosurg 117:9349412012

9

Nataf FGhossoub MSchlienger MMoussa RMeder JFRoux FX: Bleeding after radiosurgery for cerebral arteriovenous malformations. Neurosurgery 55:2983062004

10

Niazi TNKlimo P JrAnderson RCRaffel C: Diagnosis and management of arteriovenous malformations in children. Neurosurg Clin N Am 21:4434562010

11

Nogueira RGBayrlee AHirsch JAYoo AJCopen WA: Dynamic contrast-enhanced MRA at 1.5 T for detection of arteriovenous shunting before and after Onyx embolization of cerebral arteriovenous malformations. J Neuroimaging 23:5145172013

12

O'Lynnger TMAl-Holou WNGemmete JJPandey ASThompson BGGarton HJ: The effect of age on arteriovenous malformations in children and young adults undergoing magnetic resonance imaging. Childs Nerv Syst 27:127312792011

13

Pollock JMWhitlow CTSimonds JStevens EAKraft RABurdette JH: Response of arteriovenous malformations to gamma knife therapy evaluated with pulsed arterial spinlabeling MRI perfusion. AJR Am J Roentgenol 196:15222011

14

Smith ERButler WEOgilvy CS: Surgical approaches to vascular anomalies of the child's brain. Curr Opin Neurol 15:1651712002

15

Steinberg GKFabrikant JIMarks MPLevy RPFrankel KAPhillips MH: Stereotactic heavy-charged-particle Bragg-peak radiation for intracranial arteriovenous malformations. N Engl J Med 323:961011990

16

Suazo LFoerster BFermin RSpeckter HVilchez COviedo J: Measurement of blood flow in arteriovenous malformations before and after embolization using arterial spin labeling. Interv Neuroradiol 18:42482012

17

Tourdias TRodrigo SOppenheim CNaggara OVarlet PAmoussa S: Pulsed arterial spin labeling applications in brain tumors: practical review. J Neuroradiol 35:79892008

18

van Laar PJvan der Grond JHendrikse J: Brain perfusion territory imaging: methods and clinical applications of selective arterial spin-labeling MR imaging. Radiology 246:3543642008

19

Williams DSDetre JALeigh JSKoretsky AP: Magnetic resonance imaging of perfusion using spin inversion of arterial water. Proc Natl Acad Sci U S A 89:2122161992

20

Wintermark MSesay MBarbier EBorbély KDillon WPEastwood JD: Comparative overview of brain perfusion imaging techniques. Stroke 36:e83e992005

21

Wolf RLWang JDetre JAZager ELHurst RW: Arteriovenous shunt visualization in arteriovenous malformations with arterial spin-labeling MR imaging. AJNR Am J Neuroradiol 29:6816872008

22

Wowra BMuacevic ATonn JCSchoenberg SOReiser MHerrmann KA: Obliteration dynamics in cerebral arteriovenous malformations after CyberKnife radiosurgery: quantification with sequential nidus volumetry and 3-tesla 3-dimensional time-of-flight magnetic resonance angiography. Neurosurgery 64:2 SupplA102A1092009

23

Yamashita KYoshiura THiwatashi ATogao OYoshimoto KSuzuki SO: Arterial spin labeling of hemangioblastoma: differentiation from metastatic brain tumors based on quantitative blood flow measurement. Neuroradiology 54:8098132012

24

Yu SLWang RWang RWang SYao YQZhang D: Accuracy of vessel-encoded pseudocontinuous arterial spinlabeling in identification of feeding arteries in patients with intracranial arteriovenous malformations. AJNR Am J Neuroradiol 35:65712014

25

Zaharchuk GDo HMMarks MPRosenberg JMoseley MESteinberg GK: Arterial spin-labeling MRI can identify the presence and intensity of collateral perfusion in patients with moyamoya disease. Stroke 42:248524912011

TrendMD

Metrics

Metrics

All Time Past Year Past 30 Days
Abstract Views 1 1 0
Full Text Views 222 222 62
PDF Downloads 207 207 28
EPUB Downloads 0 0 0

PubMed

Google Scholar