High-Resolution Steady-State Cerebral Blood Volume Maps in Patients with Central Nervous System Neoplasms Using Ferumoxytol,a Superparamagnetic Iron Oxide Nanoparticle

Abstract

Cerebral blood volume (CBV) measurement complements conventional magnetic resonance imaging (MRI) to indicate pathologies in the central nervous system (CNS). Dynamic susceptibility contrast (DSC) perfusion imaging is limited by low resolution and distortion. Steady-state (SS) imaging may provide higher resolution CBV maps but was not previously possible in patients. We tested the feasibility of clinical SS-CBV measurement using ferumoxytol, a nanoparticle blood pool contrast agent. SS-CBV measurement was analyzed at various ferumoxytol doses and compared with DSC-CBV using gadoteridol. Ninety nine two-day MRI studies were acquired in 65 patients with CNS pathologies. The SS-CBV maps showed improved contrast to noise ratios, decreased motion artifacts at increasing ferumoxytol doses. Relative CBV (rCBV) values obtained in the thalamus and tumor regions indicated good consistency between the DSC and SS techniques when the higher dose (510 mg) ferumoxytol was used. The SS-CBV maps are feasible using ferumoxytol in a clinical dose of 510 mg, providing higher resolution images with comparable rCBV values to the DSC technique. Physiologic imaging using nanoparticles will be beneficial in visualizing CNS pathologies with high vascularity that may or may not correspond with blood–brain barrier abnormalities.

Keywords

dynamic MR ferumoxytol relative cerebral blood volume steady state

INTRODUCTION

The use of gadolinium-based contrast agent (GBCA)-enhanced magnetic resonance imaging (MRI) to visualize blood–brain barrier (BBB) dysfunction has improved the diagnosis and treatment of human brain tumors and other central nervous system (CNS) lesions. Perfusion imaging with GBCA can be used to generate cerebral blood volume (CBV) maps that correlate with tissue vascularization, an important physiologic marker of neoplasms.¹ Cerebral blood volume is useful for tumor grading,^1–4 assessment of glioma prognosis,⁵ differentiation of active tumor and necrotic tissue,^6,7 and stereotactic biopsy targeting.⁸

Dynamic susceptibility-weighted contrast-enhanced (DSC) MRI, using rapid serial T2*-weighted images to track the first pass of a contrast bolus through the vasculature, is the most common perfusion imaging technique for relative CBV (rCBV) calculation. However, CBV maps generated via DSC experience low spatial resolution, and spatial distortion tends to corrupt CBV maps obtained with echo planar imaging (EPI) sequences.^9,10 Additionally, CBV calculation using GBCA is affected by extravascular contrast leakage in patients with a compromised BBB,¹¹ making leakage-correction techniques necessary to isolate the effect of extravascular contrast.¹²

Ultrasmall superparamagnetic iron oxide nanoparticles have the potential to overcome some of these inherent challenges in perfusion imaging.^13,14 Ferumoxytol, an ultrasmall superparamagnetic iron oxide that is approved for iron replacement therapy, has shown promise as an magnetic resonance (MR) contrast agent in the CNS.^14,15 Because of its large particle size, ferumoxytol acts as a blood pool agent on post injection MR sequences, with no significant leakage in areas of compromised BBB.¹⁶

Steady-state (SS) imaging does not rely on the first pass of contrast through the brain, therefore, this technique can generate high-resolution CBV maps.¹⁷ The SS-CBV has been applied in animal studies using ultrasmall superparamagnetic iron oxides to detect CBV changes caused by neuronal activation¹⁸ and to test pharmacological CBV changes,¹⁹ and to assess the microvascularity of various tumor models.¹⁷ The high spatial resolution and distortion-free SS-CBV maps could help differentiate active tumor from necrotic tissue and better localize most malignant tumor regions, therefore increase accuracy of targeted biopsy, separate from BBB abnormalities. The SS-CBV mapping has not previously been evaluated in the clinical setting.

The aim of this study was to test the feasibility of SS-CBV measurement using a long circulating iron oxide nanoparticle, ferumoxytol at clinically applicable doses. For this purpose, SS-CBV measurement was analyzed at various ferumoxytol doses and compared with DSC-CBV using gadoteridol.

MATERIALS AND METHODS

PATIENTS

Sixty-five patients with brain neoplasms (n = 49) or other CNS pathologies (multiple sclerosis, stroke, n = 16) (Table 1) were included in the study. Patients were from two imaging protocols sponsored by the NIH and approved by the institutional review board (IRB # 2753 and 1562). Both protocols allowed tumor patients to be scanned at different stages of disease up to 4 times, thus resulting in total 99 cases to analyze. All patients provided written informed consent before enrollment.

Table 1

Patient characteristics

Pathologies	# Of patients
Neoplasms	49
CNS metastatic lesions	6
Glioblastoma	21
Grade III glioma	4
Grade I-II glioma	7
Primary CNS lymphoma/lymphoma	4
Primitive neuroectodermal tumors	3
Meningioma	3
Pituitary adenoma	1

Non neoplastic lesions	16
Vascular lesions/stroke	11
Inflammatory lesions/MS	5

Abbreviations: CNS, central nervous system; MS, multiple sclerosis.

Magnetic Resonance Imaging

The imaging protocol consisted of cranial MRI scans on 2 consecutive days using a 3-tesla clinical scanner (TIM Trio, Siemens, Erlangen, Germany) with a 12 channel matrix receive-only head coil. Axial images were acquired in all sequences. For anatomic imaging, T2-weighted (T2w) turbo spin echo and T1-weighted (T1w) 3D magnetization prepared rapid gradient-echo (MPRAGE) sequences were acquired (Table 2). On day 1 DSC imaging, a gradient-echo, EPI sequence was used with a temporal resolution of 1.5 seconds using a short IV bolus of gadoteridol (ProHance, Bracco Diagnostic Inc., Princeton, NJ, USA) in a standard dose of 0.1 mmol/kg. Post gadoteridol MPRAGE T1w images were acquired to depict areas with BBB defects. For the purpose of SS imaging, the magnitude images of susceptibility weighted imaging (SWI_MAG) were used, which is a highresolution, fully flow-compensated, spoiled gradient-echo sequence with substantial T2* weighting. These were acquired on the second day, when gadoteridol was normally completely eliminated from the system. The SWI_MAG scans were acquired pre and post IV ferumoxytol (Feraheme; AMAG Pharmaceuticals Inc., Cambridge, MA, USA) injection in doses from 2 to 10.7 mg/kg: in protocol 2753 a constant of 2 mg/kg was used, whereas in protocol 1562 a total of 510 mg was given. Therefore, patients in the latter protocol received variable doses per body weight. Nine patients received somewhat lower doses (225 to 460 mg) in this protocol.

Table 2

MR sequence parameters

Short name	Sequence type	TR (ms)	TE (ms)	TI (ms)	FA (deg)	Field of view (FOV) (mm²)	Acquisition matrix	Slice thickness (mm)	Gap (mm)	Voxel dimensions (mm)	# Of slices	Scan time (minutes:seconds)
SWI_MAG	3D gradient-echo T2∗-weighted sequence, with flow compensation	28	20		15	187 × 239	297 × 448	1.2	0	0.64 × 0.53 × 1.2	72	6:20
DSC	3D gradient echo, rapid T2∗-weighted EPI, 90 measurements	1,500	20		45	192 × 192	64 × 64	3	0.9	3 × 3 × 3.9	27	2:15
T2w TSE	2D turbo spin echo, T2 weighted	9,000	97		140	180 × 240	192 × 256	2	0	0.94 × 0.94 × 2	44	2:15
T1w MPRAGE	3D gradient-echo T1-weighted sequence	2,300	3	900	12	180 × 240	192 × 256	1	0	0.94 × 0.94 × 1	128	7:22

Abbreviations: DSC, dynamic susceptibility contrast; EPI, echo planar imaging; FA, flip angle; MR, magnetic resonance; SWI_MAG, magnitude images of susceptibility weighted imaging; T2w TSE, T2-weighted turbo spin echo; T1w MPRAGE, T1-weighted 3D magnetization prepared rapid gradient echo; TE, echo time; TI, inversion time; TR, repetition time.

Image Processing

Steady-state cerebral blood volume map calculation. Pre and post ferumoxytol SWI_MAG images were coregistered and SS-CBV maps were calculated using MIPAV (Medical Image Processing, Advanced Visualization, National Institute of Health; mipav.cit.nih.gov/) by assuming that changes in transverse relaxation rate (ΔR2*) relates linearly to contrast agent concentration and therefore with CBV, since the tracer remains intravascular in the observed time period. Pixel-by-pixel ΔR2* maps were obtained from the formula: ΔR2* = ln(SI_pre/SI_post)/TE, in which SI_pre and SI_post indicate the signal intensities pre and post ferumoxytol, and TE is the echo time.²⁰ Figure 1 shows the process of creating SS-CBV maps.

Figure 1.

The process of creating SS-CBV maps. Using pre (A) and post (B) ferumoxytol T2*-weighted (T2*w) images, ΔR2* maps (C) can be calculated. ΔR2* is assumed to be linearly proportional to contrast agent concentration. Since there is no substantial extravasation of this blood pool agent, ΔR2* is derived from the intravascular compartment only and ΔR2* maps can be used as SS-CBV maps. To eliminate the noisy background caused by the logarithmic calculation, images were masked (D). The CBV maps are typically displayed with color coding (E). CBV, cerebral blood volume; MR, magnetic resonance; SS, steady state.

Dynamic susceptibility contrast cerebral blood volume map calculation. The DSC-CBV maps were calculated from DSC scans on an offline workstation using NordicICE (NordicNeuroLab, Bergen, Norway), an FDA (Food and Drug Administration)-approved, commercially available image processing software. Mathematical correction for contrast agent leakage was applied.¹²

Coregistration. The T1w MPRAGE and post ferumoxytol SWI_MAG scans were coregistered to pre ferumoxytol SWI_MAG scans with the ‘correlation ratio’ cost function, rigid body algorithm (6 degrees of freedom) and trilinear interpolation using MIPAV. Because of poor structural delineation of the DSC-CBV map, the initial volume of the original DSC (T2w EPI) scan was coregistered to the high-resolution pre ferumoxytol SWI_MAG data set with the ‘mutual information’ cost function using FSL (FMRIB Software Library, Oxford, UK, www.fmrib.ox.ac.uk/fsl/). The same transformation was then applied to the derived parametric maps. Coregistration was visually inspected in each individual MR set. Poorly registered images were corrected by changing coregistration cost functions or restricting search ranges for rotation.

Image Analysis

Visual assessment of steady-state cerebral blood volume maps. The SS-CBV maps were visually analyzed for patient motion-related artifacts. Nonmasked images were rated semiquantitatively using a scale of 1 to 5; 1: no visible motion artifacts, 2: visible motion artifacts outside of the brain, 3: mild linear motion artifact pattern affecting the brain region, 4: pronounced linear motion artifacts in the brain region, 5: severe motion artifacts, substantially compromising image evaluation.

Volume of interest-based analysis. Volumes of interest (VOIs) in each individual MR image set were described manually via a semiautomated method in MIPAV. On T1w MPRAGE images normal appearing white matter was described in multiple areas: two VOIs in the frontal, two in the occipital paraventricular region, and one in the semioval center on the opposite side of the lesion. Furthermore, thalamus was described using two VOIs on each side.²¹ Abnormal appearing regions in the thalamus or white matter were not used. In patients with brain neoplasms, VOIs in the gadoteridol enhancing solid tumor areas were described using T1w MPRAGE images acquired pre and post gadoteridol. As a second method, VOIs in tumor regions with visually high CBV on DSC-CBV maps were described.⁴ All saved VOIs were applied on the corresponding SS-CBV and DSC-CBV maps. Mean values as well as standard deviations (s.d.) were recorded. Relative CBV values (relative to normal appearing white matter) were obtained in the thalamus, in the enhancing tumor, and in tumor hot spots.

Contrast-to-noise ratio (CNR) was calculated as ‘conspicuity of the thalamus’ using the formula:

CNR = \frac{| μ_{Τ H} - μ_{WM} |}{\sqrt{σ_{Τ H}^{2} + σ_{WM}^{2}}}

where μ indicates the mean values and σ the standard deviations in the corresponding VOIs (TH: thalamus, WM: white matter) on SS-CBV maps.^22,23

Statistical Analysis

Since there were patients contributing more than one MRI study in the evaluation, the correlation among multiple MRI studies within the same patient was taken into account in all analyses. A linear mixed model was to used to evaluate the change of motion-related artifacts of SS-CBV maps with ferumoxytol doses, with the patients grouped into five dose groups (ranging from 2 to >8 mg/kg in 2 mg/kg increments). A linear mixed model was also used to evaluate the dose dependency of CNR in SS-CBV maps with ferumoxytol doses as a continuous variable. The identification variable for each patient entered the linear mixed model as a random effect to account for the correlation among multiple MRI studies. The difference between SS-rCBV and DSC-rCBV in the thalamus at each dose group was similarly determined using a linear mixed effects model with identification variables for both patients and cases as random effects. Interclass correlation coefficient (ICC) was calculated to assess the consistency between SS-CBV and DSC-CBV techniques, using the rCBV values obtained in the tumoral VOIs. As a rule of thumb, ICC above 0.8 is considered as excellent, those in the range of 0.7 to 0.8 are considered as good, and those in the 0.5 to 0.7 range are considered as fair.²⁴ Data are displayed as least square mean (s.e.), if not indicated otherwise. P value of <0.05 was considered as statistically significant. All analyses were conducted using SAS 9.3 (SAS Institute Inc., Cary, NC, USA).

RESULTS

Visual Assessment of Steady-State Cerebral Blood Volume Maps

High-resolution SS-CBV maps could be obtained in all 99 MRI studies. Visually, the overall quality of SS-CBV maps increased with increasing ferumoxytol dose (Figure 2A) Artifacts related to patient motion were significantly less pronounced at higher ferumoxytol doses (Figure 2B).

Figure 2.

Impact of dose on SS-CBV map quality. Axial SS-CBV maps with five different doses (A) visualize that higher ferumoxytol dose increases the image quality. Semiquantitative assessment (1 to 5: non to severe motion artifacts, see text) (B) shows that increasing ferumoxytol dose decreases motion-related artifacts on SS-CBV maps (P < 0.0001). Error bars indicate the 95% confidence intervals (CIs). Scatter plots (C) show the linear relationship between contrast-to-noise ratio (CNR) and the applied ferumoxytol dose. The solid line indicates the fitted line with dashed lines representing the 95% CI. The fitted equation is given by CNR = 0.04 + 0.12 × dose in (mg/kg). CBV, cerebral blood volume; SS, steady state.

Contrast-to-Noise Ratio Assessment of Steady-State Cerebral Blood Volume Maps

Quantitative assessment using CNR was performed in 91 studies, since thalamus was not imaged in 1 patient and in 7 patients (all in the 2-mg/kg group) pronounced imaging artifacts precluded analysis. The CNR of the SS-CBV maps showed a linear relationship with ferumoxytol dose, with a 0.12 (s.e. 0.01) increase for each mg/kg ferumoxytol dose increase (P < 0.001) (Figure 2C).

Visual Assessment of Steady-State Cerebral Blood Volume and Dynamic Susceptibility Contrast Cerebral Blood Volume

Visually assessed, CBV parametric maps were comparable between the SS and DSC techniques (Figure 3). The CBV maps could be overlaid on anatomic 3D MPRAGE data sets with high accuracy. The high resolution of SS-CBV maps appeared beneficial in differentiating blood vessels from tissue with high CBV, and the ‘blooming’ effect was less pronounced around vascular structures than with the DSC technique. The DSC-CBV maps often suffered from distortions especially in the frontal region, coregistration of DSC-CBV maps to anatomic images had to be corrected in 11 cases. No remarkable distortion was found in the SS-CBV maps. Susceptibility artifacts were present with both techniques especially near air-filled bone compartments, as well as in areas near surgery-related bone defects. These artifacts were more pronounced on DSC-CBV maps.

Figure 3.

Comparison of SS-CBV and DSC-CBV maps. (A–D) Multi plane (axial and coronal) DSC-CBV versus SS-CBV maps of a patient with glioblastoma. The low-resolution DSC-CBV maps (A) accentuate vascular structures, whereas the high-resolution SS-CBV maps (B) can better differentiate between highly vascular tissue and larger vessels. These maps can be accurately overlaid on anatomic T1-weighted (T1w) 3D magnetization prepared rapid gradient-echo (MPRAGE) images (D), which could be useful in stereotactic planning to accurately determine the most malignant tumor areas. Note the frontal distortion on the DSC-CBV map, which can also be noticed on the overlaid image (C). (E–H) Patient with diffuse grade II glioma. The coregistered anatomic T2-weighted (T2w) turbo spin echo (TSE) (E) and T1w MPRAGE post gadoteridol (F) scans describe the multifocal signal abnormalities. In corresponding slices, the SS-CBV (G) and DSC-CBV (H) maps show increased areas of CBV referring to highly vascular tumor areas. Note the mismatch between the most enhancing region (arrow) and the highest CBV values. CBV, cerebral blood volume; DSC, dynamic susceptibility contrast; SS, steady state.

Comparison of Steady-State Relative Cerebral Blood Volume Versus Dynamic Susceptibility Contrast Relative Cerebral Blood Volume in the Thalamus

Figure 4 shows the least square mean of rCBV in the thalamus in various dose groups. The mean (s.e.) SS-rCBV was 1.42 (0.08) at the 2-mg/kg dose group, significantly lower (P < 0.0001) than the mean DSC-rCBV of 1.83 (0.08). The mean (s.e.) of SS-rCBV was 1.77 (0.23), 1.71 (0.09), 1.99 (0.10), and 1.85 (0.23) at the following ferumoxytol dose groups 2.1 to 4 mg/kg, 4.1 to 6 mg/kg, 6.1 to 8 mg/kg, or above 8 mg/kg, respectively, without statistical difference compared with DSC-rCBV.

Figure 4.

Comparison of relative cerebral blood volume (rCBV) values in the thalamus between steady-state (SS) (using various doses of ferumoxytol) and dynamic susceptibility contrast (DSC) (using the standard gadoteridol dose with leakage correction) techniques. In the lowest dose group (2 mg/kg), SS-rCBV shows a significantly lower value (*) compared with DSC-rCBV. In higher doses, no significant difference was found between the two techniques (error bars indicate the 95% confidence interval (CI)).

Comparison of Steady-State Relative Cerebral Blood Volume and Dynamic Susceptibility Contrast Relative Cerebral Blood Volume in the Tumor Regions

Gadoteridol-enhancing solid tumors were found in 57 of the 83 tumor cases. The overall ICC was 0.41, which indicates a poor consistency between the two methods. However when analyzing by subgroups (Figure 5), in the study (protocol 1562) in which higher doses of ferumoxytol were administered (510 mg in total), an excellent consistency was found between the two methods with ICC of 0.84. The ICC at the 2 mg/kg was only 0.29.

Figure 5.

Comparison of SS-rCBV and DSC-rCBV values in the tumor volumes of interest (VOIs). Scatter plots show the rCBV values obtained using ferumoxytol-SS measurements compared with Gadoteridol-DSC. Both in the enhancing tumor and in the hot spots there was an excellent consistency between SS-rCBV and DSC-rCBV values in the studies in which higher doses were used (B, D), and poor correlation in patients receiving 2 mg/kg ferumoxytol only (A, C). (On graph C, there are two data points outside the X-axis limits.) CBV, cerebral blood volume; DSC, dynamic susceptibility contrast; rCBV, relative CBV; SS, steady state.

In 36 studies, tumors showed visually higher CBV values on the parametric maps compared with the surrounding areas. In these cases, hotspot analysis was also performed. Within this subset of tumors, an overall poor consistency was found (ICC = 0.42) between the two techniques. At the 2-mg/kg ferumoxytol dose, the ICC was only 0.30 but in the higher dose group the two techniques showed an excellent consistency, with ICC of 0.88.

DISCUSSION

In this clinical study, we analyzed the feasibility of creating highresolution SS-CBV maps using ferumoxytol in a clinically applicable dose in patients with CNS pathologies.

The SS-CBV maps using iron oxide nanoparticles have been used in preclinical studies;^17,20 however, no clinical studies have been reported, most likely due to the lack of a suitable blood pool imaging agent. Ferumoxytol is a long circulating iron oxide nanoparticle, which has been approved by the FDA for iron replacement therapy in a dose of 2 × 510 mg. Ferumoxytol can be used off label for MRI, including anatomic and dynamic imaging. We evaluated MR scans from two imaging protocols, 2753 and 1562, thus SS-CBV from ferumoxytol doses ranging from 2 to 10.7 mg/kg could be analyzed and compared. The quality of SS-CBV maps was evaluated in normal appearing brain regions regardless of pathology.

Generally, 2 mg/kg ferumoxytol was not satisfactory for SS imaging with poor SS-CBV map quality and low consistency with the DSC technique. Higher doses per body weight improved the image quality, which was detected visually and further supported by CNR values that increase linearly with dose. Overall, in patients with a 510-mg total dose the image quality and the consistency with DSC was good, and the motion-related artifacts were less pronounced.

Although there is no ‘gold standard’ for CBV estimation using MRI, the most common technique is DSC using a short bolus of 0.1 mmol/kg GBCA, and most of the clinical publications use this method. Therefore, this study compared SS-CBV maps obtained with ferumoxytol with leakage-corrected DSC-CBV maps obtained with gadoteridol. Previous publications have used ferumoxytol in DSC imaging,^14,16,25 which has the advantage of strong T2* shortening, no substantial T1 effect during the first pass, and absence of contrast leakage. Nevertheless, the DSC technique with ferumoxytol has to be verified first, by comparing it with DSC with GBCA in a larger patient population.

For comparison of rCBV values, we described VOIs in the white matter, the tumor area, and in the thalamus. Thalamus was chosen to compare values in normal appearing gray matter, and its rCBV has been estimated with various techniques.²¹ Technically, the thalamus was not affected by distortions or susceptibility artifacts of bone or air-filled structures. The rCBV in the thalamus with both SS and DSC techniques was somewhat lower than in the literature,²¹ probably because we excluded larger vessels from the VOIs.

Measuring rCBV values in the enhancing tumor areas versus hot spots can potentially give different results between various techniques;⁴ therefore, we analyzed two different VOI selection methods. The first technique used gadoteridol-enhanced T1w MPRAGE scans to describe contrast-enhancing regions. Intense enhancing, obviously necrotic regions, or linear enhancement along the resection margin were often found, these were not included in the VOI. Still, in many cases enhancing areas did not represent high CBV values. The second method was the so-called ‘hotspot’ method. Similarly to clinical practice, we performed hotspot analysis in cases in which there were visually higher CBV areas on the parametric maps (potentially representing the most malignant tumor regions) compared with the surrounding areas. This also allowed us to test the consistency of SS-CBV and DSC-CBV at higher values. Both in the enhancing area and in the tumor hot spots SS-rCBV values showed an excellent consistency with DSC-rCBV in patients receiving 510 mg ferumoxytol, and poor consistency in patients given 2 mg/kg.

Although both SS and DSC techniques are based on changes in transverse relaxation rate constants and the assumption of a linear relationship with contrast agent concentration, differences in individual rCBV values between DSC and SS methods could arise from several factors: (1) The DSC technique, which generally uses gradient-echo EPI sequences, is known to accentuate large vessels caused by the so-called ‘blooming artifact’ .²⁶ This is less pronounced on SS-CBV maps, because of the high spatial resolution of the source SWI_MAG images.²⁷ Using spin-echo EPI sequences for DSC perfusion imaging would also decrease the susceptibility thus the accentuation of large vessels,⁴ but this is not widely used due to its lower overall sensitivity. (2) Coregistration of DSC-CBV maps to the anatomic images cannot be performed with high accuracy due to the low spatial resolution and the inherent distortion of EPI sequences. B0 field map acquisition along with a mathematical correction algorithm could compensate for this distortion,²⁸ this is however rarely used in clinical practice. Determining the quality of co-registration of individual images was crucial to this study. We applied a method by overlaying co-registered images in red-cyan to effectively screen poorly coregistered images. (3) Motion artifacts affect the two scan types differently. Patient motion during DSC acquisition can usually be compensated for offline by a ‘motion correction’ algorithm while this kind of postacquisition correction is not possible in the SS technique. Thus, the longer MR acquisition for SS-CBV must be virtually free of patient motion, particularly at lower ferumoxytol doses. (4) Areas with contrast agent leakage could be compromised in DSC technique using standard GBCAs.

The CBV is usually underestimated due to the T1 relaxation time shortening effect of interstitial contrast media. Nevertheless, overestimation could also occur in highly vascular tumor regions with strong T2* effects, where extravasation further decreases the signal intensity. There are leakage correction methods for both underestimation and overestimation, which may be used successfully in most cases. Steady-state imaging can only be performed with a contrast agent that is virtually free of vascular leakage during the observed period of time.

The high spatial resolution of SS-CBV maps could make this technique clinically relevant. Albeit DSC imaging can additionally provide various dynamic perfusion measures besides CBV, such as mean transit time, blood flow, time to peak, these are rarely used in tumor perfusion assessment. The good consistency between the two techniques indicates that SS-CBV may serve the same purpose as DSC-CBV, including differential diagnosis, tumor grading, and evaluation of early treatment responses in longitudinal studies. Furthermore, potential uses of SS-CBV include stereotactic biopsy planning using fusion images with MPRAGE, which may or may not correlate with BBB abnormality.

DISCLOSURE/CONFLICT OF INTEREST

Dr Neuwelt's studies involving ferumoxytol were entirely funded by Veterans Administration and NIH research grants, with the ferumoxytol USPIO nanoparticles partially donated by AMAG pharmaceuticals. None of the authors has financial interest in this agent or in its developer AMAG.

Footnotes

ACKNOWLEDGEMENTS

The authors thank Aliana Culp for helping with the manuscript preparation.

References

Law

Yang

Babb

Knopp

Golfinos

Zagzag

. Comparison of cerebral blood volume and vascular permeability from dynamic susceptibility contrast-enhanced perfusion MR imaging with glioma grade. AJNR Am J Neuroradiol 2004; 25: 746–755.

Law

Yang

Wang

Babb

Johnson

Cha

. Glioma grading: sensitivity, specificity, and predictive values of perfusion MR imaging and proton MR spectroscopic imaging compared with conventional MR imaging. AJNR Am J Neuroradiol 2003; 24: 1989–1998.

Sugahara

Korogi

Kochi

Ushio

Takahashi

. Perfusion-sensitive MR imaging of gliomas: comparison between gradient-echo and spin-echo echoplanar imaging techniques. AJNR Am J Neuroradiol 2001; 22: 1306–1315.

Schmainda

Rand

Joseph

Lund

Ward

Pathak

. Characterization of a first-pass gradient-echo spin-echo method to predict brain tumor grade and angiogenesis. AJNR Am J Neuroradiol 2004; 25: 1524–1532.

Law

Young

Babb

Peccerelli

Chheang

Gruber

. Gliomas: predicting time to progression or survival with cerebral blood volume measurements at dynamic susceptibility-weighted contrast-enhanced perfusion MR imaging. Radiology 2008; 247: 490–498.

Hazle

Jackson

Schomer

Leeds

. Dynamic imaging of intracranial lesions using fast spin-echo imaging: differentiation of brain tumors and treatment effects. J Magn Reson Imaging 1997; 7: 1084–1093.

Sugahara

Korogi

Tomiguchi

Shigematsu

Ikushima

Kira

. Post-therapeutic intraaxial brain tumor: the value of perfusion-sensitive contrast-enhanced MR imaging for differentiating tumor recurrence from nonneoplastic contrast-enhancing tissue. AJNR Am J Neuroradiol 2000; 21: 901–909.

Knopp

Cha

Johnson

Mazumdar

Golfinos

Zagzag

. Glial neoplasms: dynamic contrast-enhanced T2*-weighted MR imaging. Radiology 1999; 211: 791–798.

Petrella

Provenzale

. MR perfusion imaging of the brain: techniques and applications. AJR Am J Roentgenol 2000; 175: 207–219.

10.

Siegal

Rubinstein

Tzuk-Shina

Gomori

. Utility of relative cerebral blood volume mapping derived from perfusion magnetic resonance imaging in the routine follow up of brain tumors. J Neurosurg 1997; 86: 22–27.

11.

Jackson

. Analysis of dynamic contrast enhanced MRI. Br J Radiol 2004, 77 Spec No 2: S154–S166.

12.

Boxerman

Schmainda

Weisskoff

. Relative cerebral blood volume maps corrected for contrast agent extravasation significantly correlate with glioma tumor grade, whereas uncorrected maps do not. AJNR Am J Neuroradiol 2006; 27: 859–867.

13.

Weinstein

Varallyay

Dosa

Gahramanov

Hamilton

Rooney

. Superparamagnetic iron oxide nanoparticles: diagnostic magnetic resonance imaging and potential therapeutic applications in neurooncology and central nervous system inflammatory pathologies, a review. J Cereb Blood Flow Metab 2010; 30: 15–35.

14.

Gahramanov

Raslan

Muldoon

Hamilton

Rooney

Varallyay

. Potential for differentiation of pseudoprogression from true tumor progression with dynamic susceptibility-weighted contrast-enhanced magnetic resonance imaging using ferumoxytol versus. gadoteridol: a pilot study. Int J Radiat Oncol Biol Phys 2011; 79: 514–523.

15.

Hamilton

Nesbit

Dosa

Gahramanov

Rooney

Nesbit

. Comparative analysis of ferumoxytol and gadoteridol enhancement using T1- and T2-weighted MRI in neuroimaging. AJR Am J Roentgenol 2011; 197: 981–988.

16.

Neuwelt

Varallyay

Manninger

Solymosi

Haluska

Hunt

. The potential of ferumoxytol nanoparticle magnetic resonance imaging, perfusion, and angiography in central nervous system malignancy: a pilot study. Neurosurgery 2007; 60: 601–611, discussion 611-2.

17.

Bremer

Mustafa

Bogdanov

Jr Ntziachristos

Petrovsky

Weissleder

. Steady-state blood volume measurements in experimental tumors with different angiogenic burdens a study in mice. Radiology 2003; 226: 214–220.

18.

van Bruggen

Busch

Palmer

Williams

de Crespigny

. High-resolution functional magnetic resonance imaging of the rat brain: mapping changes in cerebral blood volume using iron oxide contrast media. J Cereb Blood Flow Metab 1998; 18: 1178–1183.

19.

Perles-Barbacaru

Procissi

Demyanenko

Hall

Uhl

Jacobs

. Quantitative pharmacologic MRI: mapping the cerebral blood volume response to cocaine in dopamine transporter knockout mice. Neuroimage 2011; 55: 622–628.

20.

Wong

Andrassy

Tang

. High-resolution in vivo CBV mapping with MRI in wild-type mice. Magn Reson Med 2003; 49: 765–770.

21.

Smith

Grandin

Duprez

Mataigne

Cosnard

. Whole brain quantitative CBF, CBV, and MTT measurements using MRI bolus tracking: implementation and application to data acquired from hyperacute stroke patients. J Magn Reson Imaging 2000; 12: 400–410.

22.

Haroon

Patankar

Zhu

Thacker

Scott

. Comparison of cerebral blood volume maps generated from T2* and T1 weighted MRI data in intra-axial cerebral tumours. Br J Radiol 2007; 80: 161–168.

23.

Mullins

Lev

Bove

O'Reilly

Saini

Rhea

. Comparison of image quality between conventional and low-dose nonenhanced head CT. AJNR Am J Neuroradiol 2004; 25: 533–538.

24.

Blacker

. Psychiatric rating scales. In: Sadock

Sadock

(eds) Comprehensive Textbook of Psychiatry. 8th ed (Lippincott Williams & Wilkins: Philadelphia, 2005, pp 929–955.

25.

Varallyay

Muldoon

Gahramanov

Goodman

. Dynamic MRI using iron oxide nanoparticles to assess early vascular effects of antiangiogenic versus corticosteroid treatment in a glioma model. J Cereb Blood Flow Metab 2009; 29: 853–860.

26.

Pintaske

Muller-Bierl

Schick

. Geometry and extension of signal voids in MR images induced by aggregations of magnetically labelled cells. Phys Med Biol 2006; 51: 4707–4718.

27.

Haacke

Mittal

Neelavalli

Cheng

. Susceptibility-weighted imaging: technical aspects and clinical applications, part 1. AJNR Am J Neuroradiol 2009; 30: 19–30.

28.

Poot

Pintjens

Verhoye

Van Der Linden

Sijbers

. Improved B(0) field map estimation for high field EPI. Magn Reson Imaging 2010; 28: 441–450.