Comparison of dynamic susceptibility contrast (DSC) using gadolinium and iron-based contrast agents in high-grade glioma at high-field MRI

Abstract

Purpose

To compare DSC-MRI using Gadolinium (GBCA) and Ferumoxytol (FBCA) in high-grade glioma at 3T and 7T MRI field strengths. We hypothesized that using FBCA at 7T would enhance the performance of DSC, as measured by contrast-to-noise ratio (CNR).

Methods

Ten patients (13 lesions) were assigned to 3T (6 patients, 6 lesions) or 7T (4 patients, 7 lesions). All lesions received 0.1 mmol/kg of GBCA on day 1. Ten lesions (4 at 3T and 6 at 7T) received a lower dose (0.6 mg/kg) of FBCA, followed by a higher dose (1.0-1.2 mg/kg), while 3 lesions (2 at 3T and 1 at 7T) received only a higher dose on Day 2. CBV maps with leakage correction for GBCA but not for FBCA were generated. The CNR and normalized CBV (nCBV) were analyzed on enhancing and non-enhancing high T2W lesions.

Results

Regardless of FBCA dose, GBCA showed higher CNR than FBCA at 7T, which was significant for high-dose FBCA (p < .05). Comparable CNR between GBCA and high-dose FBCA was observed at 3T. There was a trend toward higher CNR for FBCA at 3T than 7T. GBCA also showed nCBV twice that of FBCA at both MRI field strengths with significance at 7T.

Conclusion

GBCA demonstrated higher image conspicuity, as measured by CNR, than FBCA on 7T. The stronger T2* weighting realized with higher magnetic field strength, combined with FBCA, likely results in more signal loss rather than enhanced performance on DSC. However, at clinical 3T, both GBCA and FBCA, particularly a dosage of 1.0–1.2 mg/kg (optimal for perfusion imaging), yielded comparable CNR.

Introduction

Dynamic susceptibility contrast (DSC) magnetic resonance imaging (MRI) is a T2* technique widely used for assessing physiologic perfusion within the brain. Quantification of cerebral blood volume (CBV) has proven to be clinically valuable in several aspects. It is particularly effective in distinguishing between neoplastic and non-neoplastic lesions.^1,2 Additionally, CBV quantification plays a crucial role in identifying neoplastic etiologies.^3,4 Moreover, it serves as a reliable parameter for assessing tumor grade.^5,6 Beyond diagnosis, CBV measurement is instrumental in evaluating tumor response to therapy; differentiating true tumor progression from the treatment-induced inflammatory response known as pseudoprogression.^7,8

The DSC technique tracks MR signal changes during a bolus injection of a paramagnetic contrast agent, typically a gadolinium-based contrast agent (GBCA). The contrast agent causes local magnetic susceptibility (T2*) effects due to its compartmentalization within the blood vessels which leads to signal loss on the images.

Ferumoxytol (Feraheme®), an iron-based contrast agent (FBCA), is a long-circulating macromolecular carbohydrate-coated iron oxide particle. It is FDA approved for the treatment of anemia in adult patients with renal failure. FBCA has been extensively studied as an off-label alternative to GBCA.^9–11 For the purpose of diagnostic MRI, a small fraction of the therapeutic dose (1–2 mg/kg) bolus injection provides an alternative method for quantifying CBV using a DSC technique. Importantly, the physical properties of FBCA allow it to serve as a blood pool intravascular agent without leakage during the first pass of DSC-MRI after intravenous administration. To this end, FBCA has shown great potential to differentiate true tumor progression from pseudoprogression with the advantage of not requiring leakage correction.^12,13

Increasing magnetic field strength (B₀) can lead to higher image resolution (smaller voxel size) while still potentially maintaining a high signal due to greater sensitivity to susceptibility effects (increased ΔR2*) in the DSC sequence.^14–16 Therefore, performing DSC-MRI at higher field strengths could be advantageous for detecting subtle CBV signal change. However, there remains a paucity of literature regarding the clinical use of FBCA at 7T field strength.

The aim of this study was to compare DSC-MRI using GBCA and FBCA in high-grade glioma at 3T and 7T MRI field strengths. We hypothesized that the performance of the DSC-MRI sequence, as measured by contrast-to-noise ratio (CNR), would be improved using FBCA at 7T, thereby providing better image conspicuity.

Materials and methods

Subjects

This prospective observational case-control study (NCT00659126) received institutional review board approval and included participants who met the following inclusion criteria: (i) histologically proven primary brain tumor (any grade) either before or after treatment, (ii) radiographically evaluable or measurable disease with standard MRI, (iii) Eastern Cooperative Oncology Group performance status =< 3 (Karnofsky performance status >= 30), (iv) ability to understand and willingness to sign a written informed consent document or have a representative able to consent for the subject, (v) sexually active women of child-bearing potential must agree to use adequate contraception (hormonal or barrier method of birth control, or abstinence) prior to study treatment and for the duration of study treatment, and (vi) 18 years or older. Patients with a history of hepatic disease, iron overload, pregnancy or lactation, three or more drug allergies, brain herniation, or contraindication to MRI were excluded from enrollment. Thirty-one patients were screened at our institution between April 2008 and November 2021.

Each patient was assigned to undergo DSC-MRI perfusion at either 3T or 7T MRI field strength. Sequential imaging sessions were undertaken over the course of 2 days. Intravenous GBCA injection on protocol Day 1 and FBCA injection on protocol Day 2 (16 to 24 h after GBCA). Each patient maintained the same MRI field strength for both visit days.

This study was conducted at a time when assessing IDH mutational status was not standard of care. Therefore, “Not Otherwise Specified” (NOS) will be applied, following the 2021 WHO classification of CNS tumors. ¹⁷

Twenty-one out of 31 patients were excluded from the analysis due to screen failure prior to MRI (1 glioblastoma, NOS, WHO grade 4 with MRI-incompatible implantation), presence of other types of primary brain tumors (8 patients; 1 astroblastoma, 1 pineoblastoma, 3 oligodendroglioma, NOS, WHO grade 2, 1 meningioma grade 1, 1 atypical meningioma grade 2, and 1 solitary fibrous tumor grade 3), absence of measurable disease (2 glioblastoma, NOS, WHO grade 4), or non-assessable DSC data including incomplete MRI scan due to patient anxiety/claustrophobia or DSC post-processing errors (10 patients; 4 glioblastoma, NOS, WHO grade 4, 1 CNS lymphoma, 1 astrocytoma, NOS, WHO grade 2, 2 astrocytoma, NOS, WHO grade 3, and 2 oligodendroglioma, NOS, WHO grade 3). Finally, a total of 10 high-grade glioma patients (8 glioblastomas, NOS, WHO grade 4, and 2 astrocytomas, NOS, WHO grade 3) with evaluable data were enrolled in this analysis.

Patient demographics of the final evaluable cohort, along with the number of assessed lesions, and the visit schedule according to MRI field strength, are provided in Table 1. One patient who underwent imaging at 7T, was diagnosed with multifocal glioblastomas (2 lesions; one on the right and one on the left parietal lobes) and had two MRI visits 4 months apart. In total, the cohort consisted of 10 patients with 13 brain tumor lesions (6 patients with 6 lesions at 3T and 4 patients with 7 lesions at 7T).

Table 1.

Demonstrates the demographic data of patients, number of lesions and MRI visit counts for each scanner, tumor types, treatment details, steroid (dexamethasone) usage, disease status, and types of contrast agents administered during the MRI study.

MRI	Number of lesions (visit)	Gender	Age (years)	Tumor types	Treatment	Steroid usage	Disease Status	GBCA	Low dose FBCA	High dose FBCA
3 T	1	Male	55	Glioblastoma	Surgery + CCRT	No	SD	√	-	√
	1	Male	64	Glioblastoma	Surgery + CCRT	No	SD	√	-	√
	1	Male	73	Glioblastoma	Surgery + CCRT	No	PD	√	√	√
	1	Female	62	Astrocytoma grade 3	Surgery	Yes	SD	√	√	√
	1	Female	45	Glioblastoma	Surgery + CCRT	Yes	PR	√	√	√
	1	Female	36	Astrocytoma grade 3	Surgery + CCRT	No	SD	√	√	√
7 T	1	Male	35	Glioblastoma	Surgery + CCRT	No	SD	√	-	√
	1	Female	62	Glioblastoma	Surgery + avastin, carboplatin + IMRT	No	PsP	√	√	√
	1	Male	48	Glioblastoma	Surgery + CCRT	Yes	PD	√	√	√
	2 (1)	Male	66	Multifocal glioblastoma	Surgery + CCRT	Yes	PsP	√	√	√
	2 (2)	Male	66	Multifocal glioblastoma	Surgery + CCRT + avastin	No	SD	√	√	√

Note: CCRT = concurrent chemoradiation; IMRT = intensity-modulated radiation therapy, SD = Stable disease, PD = Progressive disease, PR = Partial response, PsP = Pseudoprogression.

Contrast agents

Gadoteridol (ProHance, Bracco Diagnostic Inc., Princeton, NJ) was administered using a standard dose of 0.1 mmol/kg at an injection rate of 3 mL/s, followed by a saline flush for DSC-MRI on Day 1.

Ferumoxytol (Feraheme®, AMAG Pharmaceuticals Inc. Cambridge, MA) (FBCA) was diluted 1:3 or 1:2 with normal saline to make a 7.5 mg/mL or 10 mg/mL Fe concentration solution, respectively. The FBCA with a 10 mg/mL Fe concentration solution was split between two sequential DSC scans. First, a “lower dose” of FBCA bolus at 0.6 mg/kg Fe was administered during DSC-MRI acquisition. This was immediately followed by the second “higher dose” of FBCA bolus at 1.0–1.2 mg/kg Fe (optimal dose for perfusion imaging ¹¹ ) during the second DSC-MRI acquisition. The FBCA with a 7.5 mg/mL Fe concentration solution was administered with a dose of 1.0–1.2 mg/kg Fe for DSC-MRI acquisition. A total maximum dose of 510 mg Fe, not to be exceeded, was administered. The FBCA was administered at a flow rate of 3 mL/s followed by a saline flush for DSC-MRI on Day 2.

A power injector (Spectris Solaris - MEDRAD INC, PA, USA) was used for both contrast agents through an 18-gauge intravenous line.

Finally, all 10 patients (13 lesions; 6 at 3T and 7 at 7T) received a standard dose of GBCA on Day 1. Among them, 7 patients (10 lesions; 4 at 3T and 6 at 7T) initially received a lower dose of FBCA, immediately followed by a higher dose, while 3 patients (3 lesions; 2 at 3T and 1 at 7T) received only a higher dose of FBCA on Day 2. (Table 1).

MRI imaging

3 Tesla MRI parameters

Six patients were scanned on a 3T MRI scanner (TIM Trio, Siemens, Erlangen, Germany) with a 12-channel phased-array head receive RF coil. For anatomical imaging, T2-weighted 2D turbo spin echo (TR/TE: 9000/93 ms, flip angle 140°, field of view 180 mm × 240 mm, matrix 192 × 256, 44 axial slices, slice thickness 2 mm, no gap, voxel size 0.9 mm × 0.9 mm x 2 mm, scan time 3 min 30 s) and post-Gd T1-weighted 3D magnetization-prepared rapid gradient-echo (MPRAGE) (TR/TE/TI: 2300/3/900 ms, flip angle 12°, field of view 180 mm × 240 mm, matrix 192 × 256, 128 axial slices, slice thickness 1 mm with no gap, voxel size 0.9 mm × 0.9 mm x 1.0 mm, scan time 8 min 20 s) sequences were acquired. DSC-MRI was acquired using a 2D GRE echo planar imaging sequence (TR/TE: 1500/20 ms, flip angle 45°, field of view (192 mm), ² matrix 64 × 64, 27 axial slices, slice thickness 3 mm with 0.9 mm gap, voxel size 3 mm × 3 mm x 3.9 mm, scan time 2 min 25 s).

7 Tesla MRI parameters

Four patients were scanned on a 7T MRI scanner (MAGNETOM, Siemens, Erlangen, Germany) with an 8-channel phased-array transmit/receive RF head coil (Rapid Biomedical, Wurzburg, Germany) which was designated for research purposes only. For anatomical imaging, T2-weighted 2D turbo spin echo (TR/TE: 10,000/91 ms, flip angle 120°, field of view 180 mm × 240 mm, matrix 384 × 512, 9 axial slices, slice thickness 3 mm, no gap, voxel size 0.5 mm × 0.5 mm x 3 mm, scan time 4 min 20 s) and post-Gd T1-weighted 3D MPRAGE (TR/TE/TI: 3000/2.7/1100 ms, flip angle 6°, field of view 181 mm × 223 mm, matrix 260 × 320, 192 axial slices, slice thickness 0.7 mm with no gap, voxel size 0.7 mm × 0.7 mm x 0.7 mm, scan time 7 min) sequences were acquired. DSC-MRI was acquired using a 2D GRE echo planar imaging sequence (TR/TE: 1500/20 ms, flip angle range from 20 to 45°, field of view (192 mm)², matrix 64 × 64, 24 axial slices, slice thickness 2 mm with 0.6 mm gap, voxel size 3 mm × 3 mm x 2.6 mm, scan time 4 min 20 s).

DSC-MRI perfusion analysis

All MRI DICOM files, including DSC-MRI perfusion and anatomical images (T1W-MPRAGE post-Gd and T2W), were transferred to an offline workstation, NordicBrainEx (version 2.3.5, NordicNeuroLab, Bergen, Norway, an FDA (Food and Drug Administration)-approved commercially available image processing software), for post-processing analysis. Each DSC-MRI image set was co-registered with the corresponding anatomical image using an automated process. After motion correction was performed, cerebral blood volume (CBV) maps with leakage correction^18,19 were generated for GBCA, while no leakage correction ¹³ was applied for FBCA (Figure 1). Single-plane freehand ROIs were drawn (by D.W., an attending neurological radiologist with more than 10 years of experience) on enhancing tumor portion, non-enhancing hyperintense T2-weighted tumor portion, and contralateral normal-appearing white matter (NAWM). ROIs excluded large vessels, areas of acute or subacute blood products, necrotic regions, and regions with intrinsic blooming on susceptibility-weighted images (SWI) or T2*-GRE sequences. Signal intensity time curves corresponding with each ROI were generated (Figures 2 and 3) for further automatic regional CBV measurement.

Figure 1.

Two cases of glioblastoma (GBM) after surgery and concurrent chemoradiation (CCRT) are presented. The first case underwent a 3T MRI scan (upper row) with a left parietal enhancing lesion shown on T1 MPRAGE post-gadolinium (a). The other case underwent a 7T MRI scan (lower row) revealing bilateral parietal enhancing lesions (e). Upon visual assessment, on both MRI field strengths, GBCA-derived CBV maps overlaid on post-Gd images (b, f) show increased perfusion of the enhancing mass with some areas displaying hot spots (red arrows). FBCA-derived CBV maps (c, d, g, h) exhibit poorer image conspicuity and low tumor perfusion at a low dose (0.6 mg/kg) FBCA (c, g), while they exhibit better image conspicuity and higher perfusion at a higher dose (1.0–1.2 mg/kg) of FBCA (d, h).

Figure 2.

A 73-year-old male with GBM, post-surgery, and concurrent chemoradiation (CCRT), underwent a 3 T MRI scan. T1 MRPAGE post-gadolinium (a) displays ROIs defined on the enhancing tumor in the left parietal lobe (yellow color) and contralateral normal-appearing white matter (NAWM) in the right frontal lobe (cyan color). DSC images with signal intensity time curves corresponding to each ROI after GBCA, low-dose FBCA, and high-dose FBCA administration are shown (b, c, d).

Figure 3.

A 66-year-old male with GBM, post-surgery, and concurrent chemoradiation (CCRT), underwent a 7 T MRI scan. T1 MRPAGE post-gadolinium (a) displays ROIs defined on enhancing tumor in the left parietal lobe (cyan color) and contralateral normal-appearing white matter (NAWM) in the right frontal lobe (pink color). DSC images with signal intensity time curves corresponding to each ROI after GBCA, low-dose FBCA, and high-dose FBCA administration are shown (b, c, d).

The mean regional CBV value and standard deviation (SD) of each ROI were collected. The CNR was calculated using the formula:

$$\text{CNR}=\frac{\left|\mathrm{\mu }\hspace{0.17em}\text{tumor}-\hspace{0.17em}\mathrm{\mu }\hspace{0.17em}\text{white}\hspace{0.17em}\text{matter}\right|}{\sqrt{{\mathrm{\sigma }\hspace{0.17em}\text{tumor}}^2+{\mathrm{\sigma }\hspace{0.17em}\text{white}\hspace{0.17em}\text{matter}}^2}}$$

where μ indicates the mean regional CBV value and σ is the SD of regional CBV value in the corresponding ROIs on the CBV maps.^20–22

The normalized cerebral blood volume (nCBV) of tumors was calculated from mean CBV of enhancing tumor portion or non-enhancing hyperintense T2-weighted tumor portion divided by mean CBV of contralateral normal-appearing white matter.

Statistical analysis

ROI analysis was undertaken with descriptive statistics. All data were tested for normal distribution before further appropriate statistical analysis. Comparison of CNR and nCBV between GBCA and FBCA either at 3T or 7T field strength was performed with paired sample t test. Comparison of CNR and nCBV between the two MRI field strengths (3T and 7T) with either GBCA or FBCA was performed with independent sample t test. p-value <.05 indicated a significant difference. All statistical analyses were performed using SPSS Statistics version 28.0 (SPSS, Inc., Chicago, IL, USA).

Results

The mean CNR of lesions scanned at 7T MRI was significantly higher when using GBCA compared to higher FBCA dose (Table 2 and Figure 4). At 3 T MRI, there was no significant difference in CNR among lesions when using GBCA compared to either FBCA dose (Tables 2 and 3 and Figure 5). Importantly, among lesions that received FBCA, the highest CNR was produced by higher dose FBCA at 3T. Conversely, the lowest CNR was generated using a lower FBCA dose at 7T (Tables 2 and 3).

Table 2.

Demonstration of CNR, CBV-tumor, CBV-NAWM, and nCBV of GBCA and high dose (1.0–1.2 mg/kg) FBCA at two MRI field strengths. The results are presented as mean ± standard deviation.

MRI field strength	Number of lesions	CNR GBCA	CNR FBCA	CBV-GBCA tumor	CBV-GBCA NAWM	nCBV GBCA	CBV-FBCA tumor	CBV-FBCA NAWM	nCBV FBCA	p-value
3 T	6	1.06 ± 0.7	1.04 ± 0.9	1.74 ± 0 .56	0.46 ± 0 .22	4.88 ± 3.12	1.22 ± 0.23	0.68 ± 0 .25	1.85 ± 0.72	0.98 a , 0.08 b
7 T	7	1.06 ± 0.48	0.64 ± 0.29	1.05 ± 0 .18	0.55 ± 0 .07	1.94 ± 0.42	0.85 ± 0.41	0.73 ± 0.16	1.23 ± 0.67	0.04a,c 0.04b,c
p-value		0.98	0.33			0.06			0.12

^ap-value comparison between CNR-GBCA and CNR-FBCA.

^bp-value comparison between nCBV-GBCA and nCBV-FBCA.

^cSignificant difference (p < .05).

Figure 4.

Boxplot with connecting line comparison mean CNR between GBCA and FBCA at 7T MRI. Note: * Significant difference (p < .05) between GBCA and high dose FBCA.

Table 3.

Demonstration of CNR, CBV-tumor, CBV-NAWM, and nCBV of GBCA and low dose (0.6 mg/kg) FBCA at two MRI field strengths. The results are presented as mean ± standard deviation.

MRI field strength	Number of lesions	CNR GBCA	CNR FBCA	CBV-GBCA tumor	CBV-GBCA NAWM	nCBV GBCA	CBV-FBCA tumor	CBV-FBCA NAWM	nCBV FBCA	p-value
3 T	4	1.05 ± 0.90	0.77 ± 0.42	1.63 ± 0.62	0.52 ± 0.24	4.12 ± 3.03	1.39 ± 0.97	0.73 ± 0.33	2.15 ± 1.62	0.49 a , 0.13 b
7 T	6	1.08 ± 0.52	0.52 ± 0.47	1.05 ± 0.20	0.54 ± 0.06	1.98 ± 0.45	0.8 ± 0.57	0.81 ± 0.28	0.93 ± 0.35	0.11 a , 0.001b,c
p-value		0.95	0.4			0.25			0.23

^ap-value comparison between CNR-GBCA and CNR-FBCA.

^bp-value comparison between nCBV-GBCA and nCBV-FBCA.

^cSignificant difference (p < .05).

Figure 5.

Boxplot with connecting line comparison mean CNR between GBCA and FBCA at 3T MRI.

In comparison between the two MRI field strengths, where all lesions (6 at 3T and 7 at 7T) received both GBCA and higher dose FBCA (1.0–1.2 mg/kg), there was no significant difference in the mean CNR of GBCA or the mean CNR of FBCA. However, there was a trend toward a higher CNR with FBCA on 3T versus 7T field strength (Table 2).

The nCBV of lesions scanned at the 7T was significantly higher when using GBCA compared to using FBCA at either low or higher dose (Tables 2 and 3 and Figure 6). Even though not reaching statistical significance, there was a trend toward higher nCBV-GBCA about 2 times compared to nCBV-FBCA of either low or higher dose at 3T (Tables 2 and 3 and Figure 7).

Figure 6.

Boxplot with connecting line comparison mean nCBV between GBCA and FBCA at 7T MRI. Note: * Significant difference (p < .05) between mean GBCA and both low dose and high dose FBCA.

Figure 7.

Boxplot with connecting line comparison mean nCBV between GBCA and FBCA at 3T MRI.

Discussion

In this prospective observational case-control study, we investigated whether 3T and 7T magnetic field strength affect CNR obtained from T2* weighted DSC perfusion MRI using FBCA and GBCA.

According to recommendations for the standardized brain tumor imaging protocol (BTIP)-compliant DSC-MRI for high-grade gliomas, ²³ a full dose of GBCA preload followed by a full dose of GBCA bolus injection using an intermediate flip angle (60) and TE of 25-35 ms at 3T provides overall best accuracy and precision for CBV estimation.

For our study, we aimed to maintain a balance of DSC parameters to optimize bias reduction when comparing CNR across the two MRI field strengths (3T and 7T). Since preload was not performed in this study, we utilized a low flip angle (45) at 3T to minimize T1 effect. ²⁴ At 7 T, we employed an even lower flip angle (range of 20–45), aiming for a more substantial reduction in the longer T1 value expected at a higher magnetic field strength.

At 7T, it enabled us to employ a thinner slice (2 mm) compared to 3T (3 mm), providing higher image resolution. However, this may lead to a reduction in CNR. Despite the expectation that a higher magnetic field strength would at least maintain a high image signal due to increased sensitivity to the T2* effect (resulting in greater first-pass signal loss). We utilized a short (sub-optimal) TE of 20 ms for both 3T and 7T, aiming to avoid any reduction in CNR, ²³ particularly at 7T. Nevertheless, this may enhance the T1 effect as well, which can be mitigated by using a lower flip angle.

Paradoxically, our results suggest that the use of higher dose 1.0–1.2 mg/kg Fe as a bolus contrast agent at 7T resulted in almost the lowest CNR. In comparison to GBCA, FBCA consistently produced a significantly lower CNR at 7T. At 3T CNR was similar between the two contrast types. Indeed, the highest CNR observed was generated using higher dose FBCA at 3T magnetic field strength. Taken together, these findings suggest that the use of FBCA at higher dose produces a similar CNR to GBCA at clinical field strengths, but, not at ultra-high field strengths. The observations of our study have important clinical implications for performing DSC perfusion MRI in patients with high grade glioma, if validated in a larger cohort. The biophysical properties of FBCA provide for an MRI contrast that has exclusive intravascular properties for the initial 14 h after administration. This may negate the need for leakage correction and pre-load administration thereby simplifying the clinical imaging methodology required for CBV calculations. ¹¹ However, at 7T magnetic field strength FBCA clearly produces a lower dose-dependent CNR compared to GBCA.

DSC-MRI, widely used for brain tumor perfusion evaluation, relies on the transient loss in T2*-signal resulting from the first passage of an intravascular contrast bolus. Previous studies ²⁵ have compared different doses of FBCA to a standard dose of GBCA for DSC perfusion in normal brain regions at 3T MRI, they found that 1 mg/kg Fe provides similar imaging results to GBCA, and 2 mg/kg Fe has a benefit of increased CNR. These findings are comparable to our study of high-grade glioma. We also observed comparable CNR-GBCA and CNR-FBCA at a higher dose (1.0–1.2 mg/kg Fe) and reduced CNR-FBCA at a lower dose (0.6 mg/kg Fe) at 3T MRI. This implies that the DSC imaging results using FBCA are dose-dependent. Higher FBCA dose increases the T2* effect leading to better image conspicuity at 3T (Figure 1).

Conversely, we found that at ultra-high field strength (7T), there was significantly higher CNR when using GBCA compared to those with either low dose or high dose of FBCA. These findings were not what we expected initially. Here, FBCA caused more DSC image blurring than GBCA, even at a low dose of FBCA at 7T. We presume that the observed results at 7T ultra-high magnetic field strength, together with FBCA, induce more image noise that outweighs the benefit of increased susceptibility effect. In addition, at ultra-high field strength (7T), it was theoretically anticipated to have better DSC image contrast than at lower field strength (3T). But this study showed no significant difference in CNR-GBCA or CNR-FBCA between the two MRI field strengths. Instead, there was a trend toward higher CNR-FBCA at 3T than at 7T. However, the differences in current treatment and number of lesions between the two field strength groups might introduce bias to this point.

The same phenomena occurred with nCBV measurements at ultra-high field strength (7T) as well, showing significantly higher nCBV when using GBCA compared to either low or higher dose FBCA. Moreover, the nCBV results in this study demonstrated approximately 2 times higher nCBV using GBCA compared to FBCA at either low or high dose and at either 3T or 7T field strength. These findings may imply that obtaining CBV by using GBCA as a contrast agent on leaky tumors, even with post-processing leakage correction, still remains doubtful in terms of accuracy and reliability. However, to validate whether FBCA might be superior to GBCA in brain tumor perfusion evaluation, particularly for leaky tumors, further studies on FBCA-derived CBV related to prognosis or disease outcome are warranted.

Our study has several important limitations. These include a small sample size and a limited set of DSC-MRI sequence parameters that were tested. We chose to minimize the T1 effect of the DSC-MRI sequence by utilizing a low flip angle technique. The echo time (TE) of the DSC-MRI was fixed for both 3T and 7T, which was sub-optimal for 3T. This certainly affected observed CNR values, which may differ based on alternative DSC parameters. Additionally, we included patients who had received various therapies, including dexamethasone and/or chemotherapies at the time of MRI, which may alter cerebral blood volume.^26–28 Even though 5 out of 6 patients (scanned at 3T) and 3 out of 4 patients (scanned at 7T) were defined as having non-progressive disease status, the absence of supporting evidence, such as tissue pathology, to confirm whether the assessed lesion at the time of the MRI scan was a treated lesion or a mix of tumor and treatment-related change might introduce study bias. Moreover, the direct comparison between the two MRI field strengths could be considered as another weakness of this study. Conducting further research with the same tumor type/grade and treatment within the same patient on both 3T and 7T MRI would help mitigate this bias. Almost all higher doses of FBCA were sequentially administered a few minutes after an initial lower dose scan. While allowing for a separate DSC-MRI acquisition based on the first injection, this could cause background contamination (T2* effect) by the distribution of FBCA from the first dose, which might lead to inaccurate CNR measurement of the second dose. A prior study ²⁹ found that GBCA leakage correction algorithms, either bidirectional or unidirectional, could change CBV estimation in enhancing glioma (BBB disruption) with a lesser degree in non-enhancing glioma (BBB intact). Therefore, future work should examine these lesions separately.

Conclusion

GBCA showed better image conspicuity than FBCA on 7T MRI, as measured by CNR, which could be due to an extreme T2* effect. This might result from the ultra-high magnetic field strength combined with FBCA itself, inducing more image noise that outweighs the increased susceptibility effect. However, at the clinical 3T MRI field strength, both GBCA and FBCA, particularly at a dosage of 1.0–1.2 mg/kg (optimal for perfusion imaging), yielded comparable CNR. This suggests that the biophysical properties of FBCA may improve upon GBCA as an MRI contrast agent for quantitative DSC perfusion techniques. Further investigation with appropriate adjustments of DSC sequence parameters (TE, flip angle) or alternatives to spin-echo DSC acquisition may be logical next steps to explore the potential usefulness of FBCA under ultra-high magnetic field strength.

ORCID iDs

Doonyaporn Wongsawaeng https://orcid.org/0000-0002-0418-5191

Jared Stoller https://orcid.org/0000-0002-3558-8372

Ramon F Barajas https://orcid.org/0000-0001-5093-4168