Ferumoxytol-enhanced MRI for surveillance of pediatric cerebral arteriovenous malformations

Yuhao Huang Division of Pediatric Neurosurgery and

Search for other papers by Yuhao Huang in
jns
Google Scholar
PubMed
Close
 BS
,
Timothy G. Singer Division of Pediatric Neurosurgery and

Search for other papers by Timothy G. Singer in
jns
Google Scholar
PubMed
Close
 MD
,
Michael Iv Department of Radiology, Lucile Packard Children’s Hospital, Stanford University School of Medicine;

Search for other papers by Michael Iv in
jns
Google Scholar
PubMed
Close
 MD
,
Bryan Lanzman Department of Radiology, Lucile Packard Children’s Hospital, Stanford University School of Medicine;

Search for other papers by Bryan Lanzman in
jns
Google Scholar
PubMed
Close
 MD
,
Siddharth Nair Division of Pediatric Neurosurgery and

Search for other papers by Siddharth Nair in
jns
Google Scholar
PubMed
Close
 BS
,
James A. Stadler III Department of Neurosurgery, University of Wisconsin School of Medicine and Public Health, Madison, Wisconsin

Search for other papers by James A. Stadler III in
jns
Google Scholar
PubMed
Close
 MD
,
Jia Wang Environmental Health and Safety, Stanford University, Stanford, California;

Search for other papers by Jia Wang in
jns
Google Scholar
PubMed
Close
 PhD
,
Michael S. B. Edwards Division of Pediatric Neurosurgery and

Search for other papers by Michael S. B. Edwards in
jns
Google Scholar
PubMed
Close
 MD
,
Gerald A. Grant Division of Pediatric Neurosurgery and

Search for other papers by Gerald A. Grant in
jns
Google Scholar
PubMed
Close
 MD
,
Samuel H. Cheshier Department of Neurosurgery, University of Utah School of Medicine, Salt Lake City, Utah; and

Search for other papers by Samuel H. Cheshier in
jns
Google Scholar
PubMed
Close
 MD, PhD
, and
Kristen W. Yeom Department of Radiology, Lucile Packard Children’s Hospital, Stanford University School of Medicine;

Search for other papers by Kristen W. Yeom in
jns
Google Scholar
PubMed
Close
 MD
Full access

OBJECTIVE

Children with intracranial arteriovenous malformations (AVMs) undergo digital DSA for lesion surveillance following their initial diagnosis. However, DSA carries risks of radiation exposure, particularly for the growing pediatric brain and over lifetime. The authors evaluated whether MRI enhanced with a blood pool ferumoxytol (Fe) contrast agent (Fe-MRI) can be used for surveillance of residual or recurrent AVMs.

METHODS

A retrospective cohort was assembled of children with an established AVM diagnosis who underwent surveillance by both DSA and 3-T Fe-MRI from 2014 to 2016. Two neuroradiologists blinded to the DSA results independently assessed Fe-enhanced T1-weighted spoiled gradient recalled acquisition in steady state (Fe-SPGR) scans and, if available, arterial spin labeling (ASL) perfusion scans for residual or recurrent AVMs. Diagnostic confidence was examined using a Likert scale. Sensitivity, specificity, and intermodality reliability were determined using DSA studies as the gold standard. Radiation exposure related to DSA was calculated as total dose area product (TDAP) and effective dose.

RESULTS

Fifteen patients were included in this study (mean age 10 years, range 3–15 years). The mean time between the first surveillance DSA and Fe-MRI studies was 17 days (SD 47). Intermodality agreement was excellent between Fe-SPGR and DSA (κ = 1.00) but poor between ASL and DSA (κ = 0.53; 95% CI 0.18–0.89). The sensitivity and specificity for detecting residual AVMs using Fe-SPGR were 100% and 100%, and using ASL they were 72% and 100%, respectively. Radiologists reported overall high diagnostic confidence using Fe-SPGR. On average, patients received two surveillance DSA studies over the study period, which on average equated to a TDAP of 117.2 Gy×cm2 (95% CI 77.2–157.4 Gy×cm2) and an effective dose of 7.8 mSv (95% CI 4.4–8.8 mSv).

CONCLUSIONS

Fe-MRI performed similarly to DSA for the surveillance of residual AVMs. Future multicenter studies could further investigate the efficacy of Fe-MRI as a noninvasive alternative to DSA for monitoring AVMs in children.

ABBREVIATIONS

ASL = arterial spin labeling; AVM = arteriovenous malformation; Fe-MRI = ferumoxytol-enhanced MRI; Fe-SPGR = ferumoxytol-enhanced T1-weighted spoiled gradient recalled acquisition in steady state; SNR = signal-to-noise ratio; TDAP = total dose area product.

OBJECTIVE

Children with intracranial arteriovenous malformations (AVMs) undergo digital DSA for lesion surveillance following their initial diagnosis. However, DSA carries risks of radiation exposure, particularly for the growing pediatric brain and over lifetime. The authors evaluated whether MRI enhanced with a blood pool ferumoxytol (Fe) contrast agent (Fe-MRI) can be used for surveillance of residual or recurrent AVMs.

METHODS

A retrospective cohort was assembled of children with an established AVM diagnosis who underwent surveillance by both DSA and 3-T Fe-MRI from 2014 to 2016. Two neuroradiologists blinded to the DSA results independently assessed Fe-enhanced T1-weighted spoiled gradient recalled acquisition in steady state (Fe-SPGR) scans and, if available, arterial spin labeling (ASL) perfusion scans for residual or recurrent AVMs. Diagnostic confidence was examined using a Likert scale. Sensitivity, specificity, and intermodality reliability were determined using DSA studies as the gold standard. Radiation exposure related to DSA was calculated as total dose area product (TDAP) and effective dose.

RESULTS

Fifteen patients were included in this study (mean age 10 years, range 3–15 years). The mean time between the first surveillance DSA and Fe-MRI studies was 17 days (SD 47). Intermodality agreement was excellent between Fe-SPGR and DSA (κ = 1.00) but poor between ASL and DSA (κ = 0.53; 95% CI 0.18–0.89). The sensitivity and specificity for detecting residual AVMs using Fe-SPGR were 100% and 100%, and using ASL they were 72% and 100%, respectively. Radiologists reported overall high diagnostic confidence using Fe-SPGR. On average, patients received two surveillance DSA studies over the study period, which on average equated to a TDAP of 117.2 Gy×cm2 (95% CI 77.2–157.4 Gy×cm2) and an effective dose of 7.8 mSv (95% CI 4.4–8.8 mSv).

CONCLUSIONS

Fe-MRI performed similarly to DSA for the surveillance of residual AVMs. Future multicenter studies could further investigate the efficacy of Fe-MRI as a noninvasive alternative to DSA for monitoring AVMs in children.

In Brief

The authors used a blood pool agent called ferumoxytol in conjunction with MRI to detect residual or recurrent brain arteriovenous malformations (AVMs) in children. They found that this imaging modality performed equally well to the gold standard, digital subtraction angiography (DSA), which exposes patients to large doses of radiation. This is important for children, as surveillance of AVMs typically requires multiple DSA studies. Hence, this approach of using ferumoxytol-enhanced MRI is considered to be a safe and comparable alternative for monitoring AVMs, without exposing patients to unwanted effects of radiation.

Despite occurring in only 0.06%–0.11% of the general population, cerebral arteriovenous malformations (AVMs) account for approximately 50% of nontraumatic intracranial hemorrhages in children.2,10,14 Given that children are more likely to present with AVM rupture than adults, AVM management in children merits unique consideration.4 Currently, catheter biplane high-resolution DSA serves as the gold standard for initial AVM characterization, as well as preoperative or radiosurgery planning. However, whether DSA is an ideal diagnostic mechanism for AVM surveillance following initial diagnosis remains an unanswered question. Adverse events from DSA are rare, but they include stroke, intraarterial dissection, iatrogenic vessel damage, hematoma formation, and hypersensitivity reactions to contrast agents.6,7,21 More critically, radiation exposure from a single DSA study may carry risk for both acute and long-term radiation effects.18 Particularly in children, cumulative radiation exposure over lifetime is an important consideration. As such, noninvasive studies for cerebral AVM surveillance are critically needed.

Gadolinium-enhanced MRI/MRA is frequently used in conjunction with DSA for the surveillance of cerebral AVMs, especially after the first surveillance DSA study is negative for residual AVM. However, the sensitivity and specificity for evaluating cerebral AVMs using MRI/MRA have ranged from 76.7% to 84.9% and 88.9% to 95.2%, respectively.11 Due to its imperfect sensitivity, gadolinium-enhanced MRI is not recommended as the sole modality for AVM follow-up.15 A recent study has shown that ferumoxytol (Feraheme, AMI-7228, AMAG Pharmaceuticals, Inc.), a blood pool agent, can serve as an intravascular agent to accurately delineate cerebral AVMs on MRI with performance comparable to that of CTA and DSA.9

Ferumoxytol is an ultra-small, superparamagnetic iron oxide that was originally developed for treating iron-deficiency anemia in patients with chronic kidney disease.8 Subsequently, ferumoxytol has been shown to be clinically safe as an alternative contrast agent for renal failure patients, in whom gadolinium is contraindicated. Furthermore, with its high signal-to-noise ratio (SNR) and long blood half-life, ferumoxytol has shown a superior performance to gadolinium in the evaluation of cerebral AVMs.3

For these reasons and with increasing concerns of permanent gadolinium deposition in the brain,5 our institution has used ferumoxytol as an intravascular agent for MRI (Fe-MRI) when evaluating pediatric AVMs since 20148 in conjunction with DSA for monitoring AVMs as a part of standard clinical care. Despite promising results of recent Fe-MRI studies, given the altered angio-architecture, reduced or shifted flow dynamics, foci of thromboses, perilesional fibrosis, and often smaller size of the treated and/or resolving AVM, the performance of Fe-MRI for the follow-up of previously diagnosed AVMs is unexplored. In this study, we investigated the performance of Fe-MRI for the surveillance of diagnosed AVMs as compared to the gold-standard DSA.

Methods

Cohort Composition

The cases of all patients in whom a DSA study was obtained for the evaluation of a known or highly suspected cerebral AVM at our pediatric institution (2014–2016) were retrospectively reviewed after IRB approval. Our institution has been using ferumoxytol as an intravascular agent for the evaluation of cerebral AVMs in conjunction with DSA since 2014. Whenever possible, Fe-MRI was used, in addition to DSA, to minimize the total of number of DSA studies required for surveillance. Inclusion criteria were as follows: 1) patients who underwent 3-T MRI for follow-up AVM surveillance after the initial diagnosis or intervention; and 2) patients who underwent “paired” high-resolution ferumoxytol-enhanced T1-weighted spoiled gradient recalled acquisition in steady state (Fe-SPGR) and DSA within a 6-month interval, with no intervention between the two paired exams.

Imaging Methods

All MRI studies were performed on a 3-T scanner (GE Discovery 750) using a standard 8-channel head coil.

Ferumoxytol Administration

Prior to T1-weighted SPGR image acquisition, a single dose of ferumoxytol (3 mg/kg bodyweight diluted in normal saline) was administered intravenously over a 10- to 15-minute period, followed immediately by 20 ml of saline flush. Patients were closely monitored for allergic and anaphylactic reactions to the drug, with subjective assessment and vital sign and neurological checks.

T1-Weighted SPGR

Imaging parameters for T1-weighted SPGR were as follows: TR 4 msec, TE 1 msec, slice thickness 1 mm, slice spacing 0.5 mm, 416 × 416 matrix, field of view 24–26 cm, inversion time (performed for fat suppression) 12.6 msec, flip angle 15°, receiver bandwidth 62.5 kHz, and number of excitations 1, with a scan time of approximately 5.5 minutes.

Arterial Spin Labeling Perfusion

Arterial spin labeling (ASL) was performed using a pseudocontinuous labeling period of 1500 msec, followed by a 1500-msec post-labeling delay, requiring a total 5-minute scan time. Whole-brain images were obtained with a 3D background-suppressed fast spin echo stack-of-spirals method, with a TR of approximately 5 seconds. Multiarm spiral imaging was used, with 8 arms and 512 points acquired on each arm (bandwidth 62.5 kHz), yielding an in-plane and through-plane spatial resolution of 3 and 4 mm, respectively. A high level of background suppression was achieved by use of 4 separate inversion pulses spaced around the pseudocontinuous labeling pulse.

DSA

Angiograms were obtained on a Siemens Artis zee biplane system.

Blinded Rater Analysis

Two board-certified neuroradiologists (M.I., B.L.) blinded to all patient data, including the DSA results, independently evaluated randomized MRI data sets comprising Fe-SPGR images and, if available, ASL perfusion scans. They assessed for AVM resolution or residual/recurrence. Both Fe-SPGR images and ASL scans were randomized for all patients, and the reviewers were blinded to their own scores of Fe-SPGR and ASL images for each patient. They also recorded the confidence level of their clinical assessment using a Likert rating scale with scores of 1–5. Their assessments of Fe-SPGR images and ASL perfusion scans were compared to gold-standard DSA results.

Radiation Exposure

For each DSA study, radiation dose information, including total dose area product (TDAP, Gy×cm2), was recorded using commercial dose-tracking software (DoseWatch, GE Healthcare). The effective dose was obtained by multiplying the TDAP by the dose conversion factor (mSv/Gy×cm2).12 For cerebral DSA, the value used is 0.056 mSv/Gy×cm2. Safety thresholds for acute radiation exposure were drawn from international standards.13

Statistical Analysis

The cohort demographic characteristics and imaging modality findings were summarized (Table 1). The R Statistical Language and Environment for Statistical Computing was used for analysis of study data. Cohen’s Kappa statistics for agreement among imaging modalities were calculated utilizing the R package “irr” for interrater reliability. The sensitivity and specificity of ASL and Fe-SPGR series were compared to those of DSA detection of residual AVMs. The sensitivity and specificity of each imaging modality were calculated by summing the total findings of the blinded raters for agreement with the DSA studies to construct two-by-two tables.

TABLE 1.

Patients undergoing comparison Fe-MRI and DSA studies

Case No.Age (yrs)SexLocationInitial AVM GradeResidual Size (mm)Intervention Prior to Fe-MRIFU (yrs)TDAP Exposure per DSA (mGy×cm2)Effective Dose Exposure (mSv)
115FLt mesial temporalIII7Embolization, CyberKnife3.774.54.2
29FRt frontalI8Medical2.821.31.2
311FRt periventricularII8Medical2.360.63.4
47MRt posterior mesial temporalI18Medical3.4103.95.8
55MRt posterior frontal lobeI13Resection2.350.52.8
614FLt temporoparietalIII10CyberKnife, embolization11.624.31.4
712FLt frontalIII11CyberKnife10.878.94.4
812MRt posterior fossaII8Medical3.097.55.5
913FRt posterior cingulateIII5–7Resection1.7104.85.9
1015FCerebellar pontineII11CyberKnife, embolization9.242.22.4
1112FLt periventricularINoneMedical3.3No dataNo data
1210FLt temporalIIINoneResection, embolization5.923.51.3
137FRt basal ganglia & thalamusV30CyberKnife5.3140.27.9
149MLt temporalINoneResection3.113.50.8
153FPial: rt parietooccipital branch of PCA to superior sagittal sinusAVF IIa6Medical2.829.61.7

AVF = arteriovenous fistula; FU = follow-up; PCA = posterior cerebral artery.

Results

Study Cohort

From January 2014 to January 2016, a total of 44 children presented to our institution with a suspected AVM. Of these, 15 children had the requisite paired imaging surveillance of Fe-SPGR and DSA (Table 1) for AVM follow-up. The remaining 29 children were lost to follow-up and did not undergo the paired studies. The mean age of the 15 patients was 10 years (range 3–15 years) at the time of initial AVM diagnosis, and there were 4 boys (27%). The mean time between DSA and Fe-MRI was 17 days (range 2–189 days, SD 47 days). Six children underwent Fe-MRI prior to the DSA, and 9 children underwent Fe-MRI after DSA. At the time of the paired study, 9 of 15 children received invasive interventions, including embolization (n = 4), CyberKnife radiosurgery (n = 5), and resection (n = 4). The remaining 6 patients had an established prior diagnosis of AVM, were medically treated for acute stabilization, and were monitored with paired DSA and Fe-MRI for AVM evolution. All 6 patients subsequently received interventions to treat the residual AVM. The size of residual AVM on follow-up imaging ranged from 7 to 30 mm. Figures 1 and 2 illustrate examples of Fe-SPGR, ASL, and the corresponding DSA images.

FIG. 1.
FIG. 1.

Exemplar of a residual AVM on Fe-MRI. A 9-year-old girl presented for routine follow-up. A: Fe-MRI shows asymmetrical enlarged cortical draining vein (red arrows) traversing anterolateral to the hematoma cavity. B: Focal Fe-enhanced vascularity was seen superomedial to the old hematoma cavity (red arrow) and raised suspicion for residual AVM. C: Corresponding ASL image shows high signal (arrow) that represents shunt signal within the draining cortical vein. D: DSA study shows a focal residual AVM nidus (black arrowhead) with an early draining cortical vein (white arrows) that coursed superiorly and eventually drained into superior sagittal sinus. Figure is available in color online only.

FIG. 2.
FIG. 2.

Exemplar of a residual AVM on Fe-MRI. A 13-year-old girl presented for MRI follow-up after treatment of right posterior cingulate AVM. A: Residual Fe-enhanced abnormal vascular tangle is seen, suspicious for residual lesion (arrows). B: Corresponding region shows asymmetrical low perfusion (arrow), relating to prior hemorrhage/injury. No high ASL signal suggestive of a shunt is seen. C: DSA shows small vascular tangle (arrow). No early venous drainage was seen, but this was considered a very small residual AVM nidus. Figure is available in color online only.

Blinded Rater Analysis

The κ statistic for intermodality agreement between Fe-SPGR study and the most recent DSA study was 1.00, which is considered excellent intermodality agreement (Table 2). Twelve of the 15 children had ASL perfusion scans at the time of Fe-SPGR image acquisition. The κ statistic for ASL scanning was 0.53, which is deemed poor intermodality agreement.

TABLE 2.

Intermodality agreement of SPGR and ASL versus DSA for residual AVMs

Modality & FindingRater 1Rater 2DSAAgreement
Fe-SPGR vs DSA
 Residual121212κ = 1.00
 No residual333
 No. of patients15
ASL vs DSA
 Residual859κ = 0.53
 No residual473
 No. of patients12

Sensitivity and Specificity Analysis

The sensitivity and specificity of Fe-SPGR and ASL scans were calculated by summing the total findings of the blinded rater for agreement with the DSA studies to construct two-by-two tables. Fe-SPGR was highly sensitive (100%) and specific (100%) for the detection of residual AVMs compared with DSA. The sensitivity and specificity of ASL compared to DSA were 72% and 100%, respectively (Table 3).

TABLE 3.

Sensitivity and specificity of Fe-SPGR and ASL for diagnosing residual AVMs

DSA
Study Modality Compared to DSAResidualNo Residual
SPGR
 Residual240
 No residual06
 Sensitivity100%
 Specificity100%
ASL
 Residual130
 No residual56
 Sensitivity72%
 Specificity100%

Rater Confidence Scores

The reviewers scored their own confidence level when evaluating cerebral AVMs using combined Fe-SPGR/ASL compared to standard conventional MRI; they used a Likert scale of 1–5, with a score of 1 representing the lowest diagnostic confidence, score 3 representing similar diagnostic confidence, and score 5 representing highest diagnostic confidence in their task of assessing Fe-SPGR/ASL features as compared to conventional MRI. The mean scores for raters 1 and 2 were 3.73 ± 0.15 and 4.40 ± 0.19, respectively.

Estimation of Radiation Dosage

The mean follow-up duration since the initial AVM diagnosis was 4.7 years (median 3.3 years, range 2.3–11.6 years). On average, each patient received a total of 6 MRI studies (all modalities) and 4 DSA studies during the course of AVM evaluation and management. The mean TDAP and effective dose per DSA study for each patient is outlined in Table 1. Averaging across all patients, the mean TDAP per DSA study was 58.6 mGy×cm2 (95% CI 38.6–78.7 mGy×cm2), with an effective dose of 3.8 mSv (95% CI 2.2–4.43.8 mSv). During the surveillance period, children received an average of two DSA studies. Given this, the amount of exposure received during the surveillance period was twice that of the DSA exposure amount (TDAP of 117.2 Gy×cm2 [95% CI 77.2–157.4 Gy×cm2] and an effective dose of 7.8 mSv [95% CI 4.4–8.8 mSv]).

Illustrative Cases

Case 1

A 9-year-old girl presented with altered mental status and was found to have a right frontoparietal lobe bleed, requiring intubation and placement of an external ventricular drain. Subsequent DSA showed a Spetzler-Martin grade I AVM with superficial drainage. She was discharged to home following inpatient rehabilitation for left hemiplegia. DSA and Fe-MRI performed at her 5-month follow-up demonstrated a stable micro-AVM along the anterior margin of the hematoma cavity (Fig. 1). The AVM was resected, and Fe-MRI completed at 2 months, 1 year, and 2 years showed no residual AVM.

Case 2

A 13-year-old girl presented with sudden-onset dizziness, headache, nausea, and vomiting and was found to have a large occipital hemorrhage. She was intubated and underwent external ventricular drain placement for unresponsiveness. DSA demonstrated a Spetzler-Martin grade III right paramedian parietal AVM, which was supplied by splenial and pericallosal feeders and drained into the vein of Galen complex. A craniotomy was performed and the AVM was resected. At her 1-month follow-up, DSA and Fe-MRI showed a residual nidus in the superior and lateral aspects of the resection cavity (Fig. 2); subsequent CyberKnife radiosurgery was performed to treat the residual AVM.

Discussion

Our results demonstrate that Fe-MRI performs with accuracy comparable to DSA for evaluating residual or recurrent AVMs during the surveillance period. Neuroradiologists also felt overall high diagnostic confidence when using Fe-MRI for the surveillance of treated AVMs.

Ferumoxytol, an ultra-small, superparamagnetic iron oxide particle, remains in the intravascular compartment for more than 12 hours, compared to gadolinium with its 1.5 hours of blood half-life. Ferumoxytol’s long half-life can be advantageous when evaluating multifocal vascular malformations that involve the CNS and other organs that require long scan times or when repeat scans are needed8 (Fig. 3).

FIG. 3.
FIG. 3.

Effects of ferumoxytol’s long blood half-life. A 13-year-old boy presented for treatment of a residual AVM after evolution of prior hemorrhage and initial treatment. A: This presurgical Fe-MRI shows residual AVM (arrow) in the left temporal lobe. B: Thirty-six hours after resection of the lesion, the patient was brought back for postoperative MRI without Fe. Due to longer blood half-life of the ferumoxytol compared to gadolinium, intravascular contrast can still be seen (arrows). The bright T1 signal at the operative site represents postoperative blood products (arrowheads). Figure is available in color online only.

Prior studies of Fe-MRI for AVM evaluation have shown its superior performance over the gadolinium agent, gadoteridol. This is primarily due to the higher SNR of ferumoxytol in the blood pool that improves delineation of the vascular architecture.8 We have since found that Fe-MRI performs similarly to CTA and DSA for initial diagnosis and characterization of cerebral AVMs.9 Our institutional experience has shown that it may also have a clinical role for low-flow vascular lesions that might otherwise be difficult to identify with gadolinium-enhanced MRI (Figs. 4 and 5).

FIG. 4.
FIG. 4.

Cavernous malformation. A: A 13-year-old girl presented acutely with brainstem hemorrhage (white arrow, left). No obvious enhancing lesion was seen on gadolinium administration (right). B: The patient was brought back a few days later for Fe-MRI. Fe-enhanced conventional T1-weighted SPGR (red arrow, left) used in the study showed enhancing puddle-like areas of enhancement. Background fat suppression and short TR/TE showing an increase in contrast between Fe-enhancing vessels and adjacent soft tissues (red arrowhead, right), enabling detailed vascular delineation. Partial resection of the lesion later showed cavernous malformation. Figure is available in color online only.

FIG. 5.
FIG. 5.

Low-flow transitional AVM. A: A 16-year-old boy presented for evaluation of a left parotid tumor, shown on T2-weighted fat-suppressed MRI (T2 FS, arrowhead). Incidentally, a subtle signal abnormality with possible punctate flow voids was seen in the left periventricular region on T2-weighted MRI (T2, arrow). Gadolinium-enhanced T1-weighted MRI (T1 Gad) showed possible faint enhancement (arrow) but did not clearly characterize the lesion. B: The patient was brought back weeks later for Fe-MRI to better assess the lesion. Fe-SPGR (arrow) showed curvilinear foci of enhancement suspicious for an AVM. Fe-enhanced susceptibility weighted imaging (Fe-SWI) showed prominent tangle of vessels (asterisk). Note detailed normal microvenous architecture (two arrows) that can be made visible with Fe-SWI. DSA confirmed presence of an AVM (single arrow) with subtle early venous drainage that did not represent a high-flow AVM but was consistent with a transitional-type vascular malformation. Figure is available in color online only.

Furthermore, our results showed that Fe-SPGR outperformed ASL perfusion studies (Figs. 1 and 2), despite a prior study that showed the utility of ASL for preoperative characterization of AVMs.20 The higher spatial resolution of T1-weighted SPGR combined with the high SNR of ferumoxytol-enhanced residual AVM nidus likely contributed to this finding. Additionally, the ASL signal comprises less than 1% of the inflowing blood. For treated AVMs, reduced blood flow and, therefore, a lower shunt signal could lead to poor signal on ASL studies.

Radiation exposure from DSA in children is not well established. We previously reported an average of 119.4 Gy×cm2 (TDAP) and 6.68 mSv (mean effective dose) for either a diagnostic or interventional DSA.9 In the current study, the mean radiation exposure from a single follow-up DSA was lower at 58.6 mGy×cm2 and 3.8 mSv, as DSA was primarily aimed at lesion surveillance rather than initial diagnosis, lesion characterization, or for an interventional purpose. With an average of two DSA studies per patient, the cumulative radiation exposure on average was twice this value over the duration of this study, but we would expect it to increase over future follow-ups. The principal risks of DSA relate to radiation exposure rather than procedure-related adverse events.1 Cumulative radiation doses from CT scans have been shown to increase risks of leukemia and brain tumors.16 Our work highlights the potential role of Fe-MRI for noninvasively monitoring cerebral AVMs, which could reduce potential radiation risks associated with multiple DSA exams over the course of a lifetime.

While allergic reaction can occur with all contrast agents, the safety profile of ferumoxytol is similar to iodinated contrast agents used for CT.19 In one study of 217 unique patients ages 3 days to 94 years, zero adverse events were reported following ferumoxytol administration.17 Given increasing awareness of permanent gadolinium deposition in the brain and its unknown associated health risks,5 ferumoxytol may be a useful alternative contrast agent for vascular imaging, particularly with its high SNR and longer half-life in the intravascular compartment.

The main limitation of this study is its small sample size, and thus it should be considered a pilot study. The lack of large sample size for comparison within Spetzler-Martin grades, locations of AVMs, and treatment modality limits the applicability of these findings to specific patient experiences beyond the case-level analysis provided here. While direct comparison between gadolinium-enhanced MRI and Fe-MRI might be desirable, intravenous administration of two different contrast agents was not feasible and would carry ethical implications in children. Future multiinstitutional and prospective studies with a larger sample size could further probe the clinical utility of Fe-MRI as an alternative surveillance tool for monitoring treated AVMs.

Conclusions

Fe-MRI performed comparably to DSA for surveillance of pediatric cerebral AVMs. Future multicenter studies could further investigate the efficacy of Fe-MRI as a noninvasive alternative to DSA for monitoring AVMs in children.

Disclosures

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

Author Contributions

Conception and design: Yeom. Acquisition of data: Singer, Iv, Lanzman, Nair, Stadler, Wang. Analysis and interpretation of data: Yeom, Huang, Singer. Drafting the article: Yeom, Huang, Singer. Critically revising the article: Yeom, Huang, Stadler, Edwards, Grant, Cheshier. Reviewed submitted version of manuscript: Yeom, Huang, Singer. Approved the final version of the manuscript on behalf of all authors: Yeom. Statistical analysis: Huang. Study supervision: Yeom.

References

  • 1

    Burger IM, Murphy KJ, Jordan LC, Tamargo RJ, Gailloud P: Safety of cerebral digital subtraction angiography in children: complication rate analysis in 241 consecutive diagnostic angiograms. Stroke 37:25352539, 2006

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 2

    Ding D, Starke RM, Kano H, Mathieu D, Huang PP, Feliciano C, et al.: International multicenter cohort study of pediatric brain arteriovenous malformations. Part 1: Predictors of hemorrhagic presentation. J Neurosurg Pediatr 19:127135, 2017

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 3

    Dósa E, Tuladhar S, Muldoon LL, Hamilton BE, Rooney WD, Neuwelt EA: MRI using ferumoxytol improves the visualization of central nervous system vascular malformations. Stroke 42:15811588, 2011

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 4

    Fullerton HJ, Achrol AS, Johnston SC, McCulloch CE, Higashida RT, Lawton MT, et al.: Long-term hemorrhage risk in children versus adults with brain arteriovenous malformations. Stroke 36:20992104, 2005

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 5

    Gulani V, Calamante F, Shellock FG, Kanal E, Reeder SB: Gadolinium deposition in the brain: summary of evidence and recommendations. Lancet Neurol 16:564570, 2017

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 6

    Haridass A, Maclean J, Chakraborty S, Sinclair J, Szanto J, Iancu D, et al.: Dynamic CT angiography for Cyberknife radiosurgery planning of intracranial arteriovenous malformations: a technical/feasibility report. Radiol Oncol 49:192199, 2015

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 7

    Hoffman CE, Santillan A, Rotman L, Gobin YP, Souweidane MM: Complications of cerebral angiography in children younger than 3 years of age. J Neurosurg Pediatr 13:414419, 2014

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 8

    Hope MD, Hope TA, Zhu C, Faraji F, Haraldsson H, Ordovas KG, et al.: Vascular imaging with ferumoxytol as a contrast agent. AJR Am J Roentgenol 205:W366W373, 2015

  • 9

    Iv M, Choudhri O, Dodd RL, Vasanawala SS, Alley MT, Moseley M, et al.: High-resolution 3D volumetric contrast-enhanced MR angiography with a blood pool agent (ferumoxytol) for diagnostic evaluation of pediatric brain arteriovenous malformations. J Neurosurg Pediatr 22:251260, 2018

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 10

    Karhunen PJ, Penttilä A, Erkinjuntti T: Arteriovenous malformation of the brain: imaging by postmortem angiography. Forensic Sci Int 48:919, 1990

  • 11

    Lee CC, Reardon MA, Ball BZ, Chen CJ, Yen CP, Xu Z, et al.: The predictive value of magnetic resonance imaging in evaluating intracranial arteriovenous malformation obliteration after stereotactic radiosurgery. J Neurosurg 123:136144, 2015

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 12

    Manninen AL, Isokangas JM, Karttunen A, Siniluoto T, Nieminen MT: A comparison of radiation exposure between diagnostic CTA and DSA examinations of cerebral and cervicocerebral vessels. AJNR Am J Neuroradiol 33:20382042, 2012

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 13

    Miller DL, Balter S, Schueler BA, Wagner LK, Strauss KJ, Vañó E: Clinical radiation management for fluoroscopically guided interventional procedures. Radiology 257:321332, 2010

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 14

    Mohr JP, Kejda-Scharler J, Pile-Spellman J: Diagnosis and treatment of arteriovenous malformations. Curr Neurol Neurosci Rep 13:324, 2013

  • 15

    Morgenstern PF, Hoffman CE, Kocharian G, Singh R, Stieg PE, Souweidane MM: Postoperative imaging for detection of recurrent arteriovenous malformations in children. J Neurosurg Pediatr 17:134140, 2016

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 16

    National Council on Radiation Protection and Measurements: Limitation of exposure to ionizing radiation: recommendations of the National Council on Radiation Protection and Measurements. Bethesda, MD: National Council on Radiation Protection and Measurements, 1993

    • PubMed
    • Search Google Scholar
    • Export Citation
  • 17

    Nguyen KL, Yoshida T, Han F, Ayad I, Reemtsen BL, Salusky IB, et al.: MRI with ferumoxytol: a single center experience of safety across the age spectrum. J Magn Reson Imaging 45:804812, 2017

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 18

    Pearce MS, Salotti JA, Little MP, McHugh K, Lee C, Kim KP, et al.: Radiation exposure from CT scans in childhood and subsequent risk of leukaemia and brain tumours: a retrospective cohort study. Lancet 380:499505, 2012

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 19

    Vasanawala SS, Nguyen KL, Hope MD, Bridges MD, Hope TA, Reeder SB, et al.: Safety and technique of ferumoxytol administration for MRI. Magn Reson Med 75:21072111, 2016

  • 20

    Yu SL, Wang R, Wang R, Wang S, Yao YQ, Zhang D, et al.: Accuracy of vessel-encoded pseudocontinuous arterial spin-labeling in identification of feeding arteries in patients with intracranial arteriovenous malformations. AJNR Am J Neuroradiol 35:6571, 2014

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 21

    Zhang B, Dong Y, Liang L, Lian Z, Liu J, Luo X, et al.: The incidence, classification, and management of acute adverse reactions to the low-osmolar iodinated contrast media Isovue and Ultravist in contrast-enhanced computed tomography scanning. Medicine (Baltimore) 95:e3170, 2016

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • Collapse
  • Expand
  • Exemplar of a residual AVM on Fe-MRI. A 9-year-old girl presented for routine follow-up. A: Fe-MRI shows asymmetrical enlarged cortical draining vein (red arrows) traversing anterolateral to the hematoma cavity. B: Focal Fe-enhanced vascularity was seen superomedial to the old hematoma cavity (red arrow) and raised suspicion for residual AVM. C: Corresponding ASL image shows high signal (arrow) that represents shunt signal within the draining cortical vein. D: DSA study shows a focal residual AVM nidus (black arrowhead) with an early draining cortical vein (white arrows) that coursed superiorly and eventually drained into superior sagittal sinus. Figure is available in color online only.

  • Exemplar of a residual AVM on Fe-MRI. A 13-year-old girl presented for MRI follow-up after treatment of right posterior cingulate AVM. A: Residual Fe-enhanced abnormal vascular tangle is seen, suspicious for residual lesion (arrows). B: Corresponding region shows asymmetrical low perfusion (arrow), relating to prior hemorrhage/injury. No high ASL signal suggestive of a shunt is seen. C: DSA shows small vascular tangle (arrow). No early venous drainage was seen, but this was considered a very small residual AVM nidus. Figure is available in color online only.

  • Effects of ferumoxytol’s long blood half-life. A 13-year-old boy presented for treatment of a residual AVM after evolution of prior hemorrhage and initial treatment. A: This presurgical Fe-MRI shows residual AVM (arrow) in the left temporal lobe. B: Thirty-six hours after resection of the lesion, the patient was brought back for postoperative MRI without Fe. Due to longer blood half-life of the ferumoxytol compared to gadolinium, intravascular contrast can still be seen (arrows). The bright T1 signal at the operative site represents postoperative blood products (arrowheads). Figure is available in color online only.

  • Cavernous malformation. A: A 13-year-old girl presented acutely with brainstem hemorrhage (white arrow, left). No obvious enhancing lesion was seen on gadolinium administration (right). B: The patient was brought back a few days later for Fe-MRI. Fe-enhanced conventional T1-weighted SPGR (red arrow, left) used in the study showed enhancing puddle-like areas of enhancement. Background fat suppression and short TR/TE showing an increase in contrast between Fe-enhancing vessels and adjacent soft tissues (red arrowhead, right), enabling detailed vascular delineation. Partial resection of the lesion later showed cavernous malformation. Figure is available in color online only.

  • Low-flow transitional AVM. A: A 16-year-old boy presented for evaluation of a left parotid tumor, shown on T2-weighted fat-suppressed MRI (T2 FS, arrowhead). Incidentally, a subtle signal abnormality with possible punctate flow voids was seen in the left periventricular region on T2-weighted MRI (T2, arrow). Gadolinium-enhanced T1-weighted MRI (T1 Gad) showed possible faint enhancement (arrow) but did not clearly characterize the lesion. B: The patient was brought back weeks later for Fe-MRI to better assess the lesion. Fe-SPGR (arrow) showed curvilinear foci of enhancement suspicious for an AVM. Fe-enhanced susceptibility weighted imaging (Fe-SWI) showed prominent tangle of vessels (asterisk). Note detailed normal microvenous architecture (two arrows) that can be made visible with Fe-SWI. DSA confirmed presence of an AVM (single arrow) with subtle early venous drainage that did not represent a high-flow AVM but was consistent with a transitional-type vascular malformation. Figure is available in color online only.

  • 1

    Burger IM, Murphy KJ, Jordan LC, Tamargo RJ, Gailloud P: Safety of cerebral digital subtraction angiography in children: complication rate analysis in 241 consecutive diagnostic angiograms. Stroke 37:25352539, 2006

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 2

    Ding D, Starke RM, Kano H, Mathieu D, Huang PP, Feliciano C, et al.: International multicenter cohort study of pediatric brain arteriovenous malformations. Part 1: Predictors of hemorrhagic presentation. J Neurosurg Pediatr 19:127135, 2017

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 3

    Dósa E, Tuladhar S, Muldoon LL, Hamilton BE, Rooney WD, Neuwelt EA: MRI using ferumoxytol improves the visualization of central nervous system vascular malformations. Stroke 42:15811588, 2011

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 4

    Fullerton HJ, Achrol AS, Johnston SC, McCulloch CE, Higashida RT, Lawton MT, et al.: Long-term hemorrhage risk in children versus adults with brain arteriovenous malformations. Stroke 36:20992104, 2005

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 5

    Gulani V, Calamante F, Shellock FG, Kanal E, Reeder SB: Gadolinium deposition in the brain: summary of evidence and recommendations. Lancet Neurol 16:564570, 2017

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 6

    Haridass A, Maclean J, Chakraborty S, Sinclair J, Szanto J, Iancu D, et al.: Dynamic CT angiography for Cyberknife radiosurgery planning of intracranial arteriovenous malformations: a technical/feasibility report. Radiol Oncol 49:192199, 2015

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 7

    Hoffman CE, Santillan A, Rotman L, Gobin YP, Souweidane MM: Complications of cerebral angiography in children younger than 3 years of age. J Neurosurg Pediatr 13:414419, 2014

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 8

    Hope MD, Hope TA, Zhu C, Faraji F, Haraldsson H, Ordovas KG, et al.: Vascular imaging with ferumoxytol as a contrast agent. AJR Am J Roentgenol 205:W366W373, 2015

  • 9

    Iv M, Choudhri O, Dodd RL, Vasanawala SS, Alley MT, Moseley M, et al.: High-resolution 3D volumetric contrast-enhanced MR angiography with a blood pool agent (ferumoxytol) for diagnostic evaluation of pediatric brain arteriovenous malformations. J Neurosurg Pediatr 22:251260, 2018

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 10

    Karhunen PJ, Penttilä A, Erkinjuntti T: Arteriovenous malformation of the brain: imaging by postmortem angiography. Forensic Sci Int 48:919, 1990

  • 11

    Lee CC, Reardon MA, Ball BZ, Chen CJ, Yen CP, Xu Z, et al.: The predictive value of magnetic resonance imaging in evaluating intracranial arteriovenous malformation obliteration after stereotactic radiosurgery. J Neurosurg 123:136144, 2015

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 12

    Manninen AL, Isokangas JM, Karttunen A, Siniluoto T, Nieminen MT: A comparison of radiation exposure between diagnostic CTA and DSA examinations of cerebral and cervicocerebral vessels. AJNR Am J Neuroradiol 33:20382042, 2012

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 13

    Miller DL, Balter S, Schueler BA, Wagner LK, Strauss KJ, Vañó E: Clinical radiation management for fluoroscopically guided interventional procedures. Radiology 257:321332, 2010

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 14

    Mohr JP, Kejda-Scharler J, Pile-Spellman J: Diagnosis and treatment of arteriovenous malformations. Curr Neurol Neurosci Rep 13:324, 2013

  • 15

    Morgenstern PF, Hoffman CE, Kocharian G, Singh R, Stieg PE, Souweidane MM: Postoperative imaging for detection of recurrent arteriovenous malformations in children. J Neurosurg Pediatr 17:134140, 2016

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 16

    National Council on Radiation Protection and Measurements: Limitation of exposure to ionizing radiation: recommendations of the National Council on Radiation Protection and Measurements. Bethesda, MD: National Council on Radiation Protection and Measurements, 1993

    • PubMed
    • Search Google Scholar
    • Export Citation
  • 17

    Nguyen KL, Yoshida T, Han F, Ayad I, Reemtsen BL, Salusky IB, et al.: MRI with ferumoxytol: a single center experience of safety across the age spectrum. J Magn Reson Imaging 45:804812, 2017

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 18

    Pearce MS, Salotti JA, Little MP, McHugh K, Lee C, Kim KP, et al.: Radiation exposure from CT scans in childhood and subsequent risk of leukaemia and brain tumours: a retrospective cohort study. Lancet 380:499505, 2012

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 19

    Vasanawala SS, Nguyen KL, Hope MD, Bridges MD, Hope TA, Reeder SB, et al.: Safety and technique of ferumoxytol administration for MRI. Magn Reson Med 75:21072111, 2016

  • 20

    Yu SL, Wang R, Wang R, Wang S, Yao YQ, Zhang D, et al.: Accuracy of vessel-encoded pseudocontinuous arterial spin-labeling in identification of feeding arteries in patients with intracranial arteriovenous malformations. AJNR Am J Neuroradiol 35:6571, 2014

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 21

    Zhang B, Dong Y, Liang L, Lian Z, Liu J, Luo X, et al.: The incidence, classification, and management of acute adverse reactions to the low-osmolar iodinated contrast media Isovue and Ultravist in contrast-enhanced computed tomography scanning. Medicine (Baltimore) 95:e3170, 2016

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation

Metrics

All Time Past Year Past 30 Days
Abstract Views 1903 211 0
Full Text Views 223 74 23
PDF Downloads 237 54 15
EPUB Downloads 0 0 0