Comparison of non-contrast magnetic resonance perfusion and phase-contrast angiography for the quantitative assessment of cerebral blood flow: a prospective cross-sectional study

Full Text

BACKGROUND

Arterial spin labeling (ASL) is a non-contrast magnetic resonance imaging (MRI) perfusion technique that enables quantitative measurement and visualization of cerebral blood flow (CBF). Its advantages include non-invasiveness and the absence of contrast administration, which is particularly important for longitudinal monitoring [1, 2]. Limitations of the technique include relatively low spatial resolution, prolonged acquisition time (approximately 5 minutes), dependence on magnetic field homogeneity and the patient’s cardiovascular status, as well as susceptibility to motion-related and metallic-implant–related artifacts [3, 4]. Quantitative assessment additionally requires post-processing. Based on contemporary studies [1, 5], we previously proposed an algorithm for calculating cerebral perfusion parameters [6].

Non-contrast perfusion MRI is widely used in the evaluation of cerebral ischemia [7, 8], demyelinating diseases [9], brain neoplasms [10], epilepsy [11], migraine [12], and hydrocephalus [13]. However, modern contrast-enhanced and non-contrast perfusion sequences rely on similar physiological models of tissue perfusion and properties of endogenous and exogenous contrast agents. This similarity may introduce errors related to shared acquisition and interpretation principles [14]. The results of perfusion techniques largely depend on technical and methodological measurement parameters. Post-processing features, as well as the lack of on-scanner quantitative output in systems from certain manufacturers, necessitate verification of derived metrics. One consequence of these limitations is the broad reported range of normal perfusion values in gray matter (40–60 mL/100 g/min) and white matter (15–20 mL/100 g/min) across studies [1, 15].

Among non-invasive approaches to cerebral blood flow assessment, a notable method is macrovascular flow quantification followed by mathematical conversion. Two-dimensional phase-contrast angiography (PCA) provides quantitative blood-flow measurements in the target vessels with appropriate selection of peak velocity parameters. Its advantages, such as rapid data acquisition (<4 min) and straightforward post-processing, have supported its wide adoption for quantitative macrovascular flow assessment [16, 17] and in cardiovascular MRI [18, 19]. Flow-velocity measurements in PCA are obtained through phase-encoding gradients, and the accuracy of this technique has been demonstrated in phantom studies. The PCA sequence is based on a physiological model fundamentally different from ASL, allowing for the verification of quantitative perfusion estimates by measuring major arterial blood flow with consideration of brain volume and mass characteristics [20].

AIM

The study aimed to verify the quantitative assessment of cerebral blood flow obtained by ASL-MRI against 2D PCA in healthy adults.

METHODS

Study Design

This was a single-center, prospective, cross-sectional study.

Study Setting

The study was carried out at the Laboratory of Functional Neuroimaging, International Tomography Center, Siberian Branch of the Russian Academy of Sciences (Novosibirsk, Russia). Volunteers were recruited based on the inclusion criteria from June 2023 to August 2024.

Eligibility Criteria

Inclusion criteria:

• Age 18 to 75 years;
• Absence of focal or space-occupying brain lesions on structural MRI (isolated chronic microvascular lesions were permitted);
• Absence of hemodynamically significant stenoses of the major cervical arteries according to 3D time-of-flight magnetic resonance angiography (TOF-MRA);
• Absence of clinical symptoms and neurological deficits;
• No history of acute cerebrovascular events, including hypertensive crises, ischemic/
• hemorrhagic stroke, intracranial hemorrhage, or traumatic intracranial injury.

Non-inclusion criteria: presence of visualized limited/multiple lesions and/or space-occupying lesions on routine MRI.

Exclusion criteria: technically inadequate imaging or quantitative assessment of cerebral perfusion and macrovascular blood flow (including data post-processing stages).

Non-contrast Perfusion Magnetic Resonance Imaging

Non-contrast perfusion MRI was performed on a 3.0-T Ingenia® scanner (Philips, the Netherlands) using a routine brain imaging protocol (T1-weighted imaging [T1-WI], T2-WI, three-dimensional fluid-attenuated inversion recovery [3D-FLAIR], and time-of-flight magnetic resonance angiography [TOF-MRA]) to assess structural brain morphology. Image interpretation for each patient was performed once by two qualified radiologists (over 10 years of practice in neuroradiology), who had access to all clinical information required by the study protocol (symptoms, medical history, and results of routine MRI).

Cerebral tissue perfusion was assessed using a pseudocontinuous ASL (pCASL) sequence with the following parameters:

• Field of view (FOV): 240 × 240 × 99 mm
• Repetition time (TR): 4550 ms
• Echo time (TE): 16 ms
• Labeling duration (LD): 1800 ms
• Post-labeling delay (PLD): 1800 ms.

The imaging slab was oriented axially along the long axis of the corpus callosum, covering the cerebral hemispheres. The labeling plane was positioned parallel to it and perpendicular to the course of the cervical segments of the major arteries, at a distance of 90 mm from the inferior edge of the FOV, with adjustments based on individual vascular anatomy as assessed on 3D TOF-MRA. Cerebral perfusion maps were generated using FSL® (BASIL, UK) by loading the native images and subtracting control and label acquisitions, incorporating the corresponding scan parameters. This was followed by proton-density calibration (TR and TE of 4550 and 13 ms, respectively) and reslicing of the resulting images with motion-correction algorithms. At the final stage, automated partial-volume correction of gray and white matter boundaries was performed using T1-weighted images for tissue segmentation and perfusion quantification. The output included a cerebral perfusion map and quantitative cerebral blood flow metrics (CBF-ASL) [6]. Mean values were calculated voxel-wise for gray and white matter separately within a single slice for each participant.

2D Phase-Contrast Angiography

The examination was performed on an Ingenia® MRI scanner (Philips, the Netherlands) operating at a magnetic field strength of 3.0 T. To assess macrovascular cerebral blood flow, a PCA sequence was acquired with the following parameters:

• FOV, 150 × 101 mm
• TR, 9.1 ms
• TE, 5.4 ms
• Number of signal acquisitions (NSA), 2
• Velocity encoding (VE), 100 cm/s
• Retrospective cardiac gating with 15 phases.

The imaging slab was oriented perpendicular to the cervical segments of the internal carotid arteries (C1 segments

according to Bouthillier) and the vertebral arteries (V7 segments). Post-processing was performed using the manufacturer’s software by manually delineating the region of interest along the inner vessel boundary to obtain volumetric blood-flow values (mL/s) in the lumen of vertebral and internal carotid arteries.

To evaluate brain volume and mass, a three-dimensional T1-weighted Turbo Field Echo (3D T1-TFE) sequence was acquired with the following parameters:

• FOV, 250 × 250 mm
• TR, 7.6 ms
• TE, 3.7 ms
• Voxel size, 1 × 1 × 2 mm
• NSA, 2.

Brain volume and mass were assessed by segmenting and normalizing the T1-WI (3D T1-TFE) in FSLanat® (BASIL, United Kingdom), using the physiological constant of brain tissue density (1.045 g/cm³) [28, 29]. To calculate total cerebral perfusion from the PCA data (CBF-PCA), we propose the following formula, based on the mathematical transformation of volumetric blood-flow values (mL/s) with adjustment for brain tissue density (g/cm³):

$6000\times \frac{(ICAr+ICAl)+(VAr+VAl)}{1.045\times \left(\frac{V}{1000}\right)} \mathrm{mL}/100\mathrm{g}/\min$, (1)

where 6000 is the conversion factor from mL/g/s to mL/100 g/min; ICAr, ICAl, volumetric blood-flow rates in the right and left internal carotid arteries, mL/s; VAr, VAl, volumetric blood-flow rates in the right and left vertebral arteries, mL/s; 1.045, physiological brain tissue density, g/cm³; V, brain volume obtained from segmentation, mm³.

Primary Study Endpoint

Verification of quantitative cerebral tissue blood flow assessment using ASL was performed by comparing perfusion values obtained with pCASL to those derived from volumetric blood-flow measurements acquired with PCA.

Sensitivity Analysis

A sensitivity analysis was conducted to evaluate how differences between perfusion values measured by ASL and PCA varied as a function of age, sex, and brain volume.

Ethics Approval

The study was approved by the Local Ethics Committee of the International Tomography Center, Siberian Branch of the Russian Academy of Sciences (Minutes No. 37 dated April 22, 2024). Written informed consent was obtained from all participants before enrollment.

Statistical Analysis

Sample size determination: a sample size calculation was not performed at the study planning stage.

Statistical data analysis methods. Data analysis was conducted using STATISTICA®, version 10.0 (StatSoft Inc., USA). The distribution of quantitative variables was assessed using the Kolmogorov–Smirnov and Shapiro–Wilk tests (for both tests: pCASL, p > 0.2; PCA, p > 0.2), as well as by inspecting normal probability plots (values aligned along the theoretical normal line) and histograms. Results are presented as M ± SD, where M is the arithmetic mean and SD is the standard deviation. Paired Student’s t-test was applied to compare quantitative parameters between paired observations. Correlation analysis was performed using Pearson’s correlation coefficient (r). To assess sensitivity and evaluate associations between sex, age, and brain volume (independent variables) and the difference between cerebral perfusion values obtained by CBF-ASL and CBF-PCA (dependent variable), multivariable linear regression analysis was used. Differences and associations were considered significant at p < 0.05.

RESULTS

Study Sample Formation

A total of 86 individuals presented for routine clinical screening; all agreed to participate and signed informed consent for participation. Based on structural MRI and TOF-MRA findings, 83 participants were eligible for inclusion (three participants were excluded due to the detection of chronic ischemic foci of varying size within the cerebral hemispheres). Eighty participants completed the ASL and PCA examinations; post-processing of cerebral and macrovascular perfusion data was considered technically unsatisfactory in three cases. Thus, data analysis according to the study objectives was performed for 80 participants. The mean age of the study cohort was 38.7 ± 16.5 years (50% were women).

Primary Results

Cerebral perfusion values obtained using ASL were as follows:

• White matter: 17.8 ± 2.3 mL/100 g/min;
• Gray matter: 42.1 ± 7.1 mL/100 g/min;
• Total perfusion (CBF-ASL): 59.6 ± 8.6 mL/100 g/min.
• PCA yielded the following volumetric flow rates:
• In the right and left internal carotid arteries: 3.8 ± 0.7 and 3.7 ± 0.8 mL/s, respectively (p = 0.497);
• In the right and left vertebral arteries: 1.5 ± 0.6 and 1.6 ± 0.6 mL/s, respectively (p = 0.487).

Segmentation of the T1-weighted images produced a brain volume estimate of 1070.9 ± 103.0 cm³ and a brain mass of 1119.1 ± 107.7 g. Taking into account brain mass, the volumetric arterial inflow through the internal carotid and vertebral arteries was converted into an adjusted cerebral perfusion value (CBF-PCA), which equaled 59.0 ± 8.2 mL/100 g/min. Correlation analysis demonstrated a strong positive correlation between CBF-ASL and CBF-PCA values (Fig. 1).

 

Fig. 1. Correlation between total cerebral perfusion values obtained using non-contrast perfusion magnetic resonance imaging and phase-contrast angiography (r = 0.839, p < 0.001). Histograms above and to the right illustrate the distribution of observations for CBF-ASL and CBF-PCA values, respectively. ASL, arterial spin labeling; CBF, cerebral blood flow; PCA, phase-contrast angiography.

 

Sensitivity Analysis

The difference between CBF-ASL and CBF-PCA values was 1.2 mL/100 g/min (95% CI: 0.87–1.61) (Fig. 2). A statistically significant negative correlation was found between the CBF-ASL minus CBF-PCA difference and age (with a small effect size, R² = 0.09; Fig. 3), as well as a positive correlation with brain volume (R² = 0.09; Fig. 4). Comparison of the CBF-ASL and CBF-PCA difference between sexes revealed no statistically significant differences (t[158] = –1.20, p = 0.230). Testing for equality of variances (F = 58.60, p < 0.001) confirmed heterogeneity, which was considered in the analysis. Multivariable linear regression showed that a greater discrepancy between CBF-ASL and CBF-PCA was associated with female sex (p = 0.014) and larger brain volume (p = 0.016), whereas the effect of age was not significant (p = 0.347).

 

Fig. 2. Distribution of differences in cerebral perfusion values obtained using non-contrast perfusion magnetic resonance imaging and phase-contrast angiography. ASL, arterial spin labeling; CBF, cerebral blood flow; PCA, phase-contrast angiography.

 

Fig. 3. Correlation between age and the difference in cerebral perfusion values obtained using non-contrast perfusion magnetic resonance imaging and phase-contrast angiography (r = −0.300, p = 0.007). ASL, arterial spin labeling; CBF, cerebral blood flow; PCA, phase-contrast angiography.

 

Fig. 4. Correlation between brain volume and the difference in cerebral perfusion values obtained using non-contrast perfusion magnetic resonance imaging and phase-contrast angiography (r = 0.300, p = 0.007). ASL, arterial spin labeling; CBF, cerebral blood flow; PCA, phase-contrast angiography.

 

DISCUSSION

Summary of Primary Results

A strong positive correlation was established between cerebral tissue blood flow values obtained through non-contrast perfusion MRI and those calculated from direct measurements of volumetric blood flow in the major brachiocephalic arteries using quantitative PCA, with adjustment for individual brain volume and the physiological constant of brain tissue density.

Discussion of Primary Results

Non-contrast magnetic resonance perfusion is a valuable noninvasive method for assessing the state of the microcirculatory bed. Despite the challenges associated with acquiring and interpreting quantitative parameters, ASL can be successfully used to diagnose a wide range of pathological conditions [21], providing accuracy comparable to contrast-enhanced magnetic resonance and computed

tomographic perfusion techniques [22, 23]. Existing kinetic and mathematical models of invasive perfusion methods used to validate ASL rely on similar processing algorithms that involve sequential subtraction of native and post-contrast images followed by the generation of qualitative or quantitative perfusion maps. Therefore, confirming ASL-derived results using an alternative methodological approach based on a different physiological model of tissue perfusion assessment remains an important objective [24, 25]. In addition, calibration of ASL in a control population is necessary to account for technical factors and variability [26]. For these reasons, in the present study we selected and analyzed a perfusion calculation model distinct from those used previously to verify ASL-derived measurements, one based on the in vivo assessment of brain volume and mass characteristics obtained through T1-WI segmentation and on the calculation of volumetric blood flow in the major cervical arteries using PCA. PCA is considered reliable for assessing blood flow and, consequently, cardiovascular function and cerebrospinal fluid state [33], providing robust qualitative and quantitative measurements [34], which have been validated, including through simulated experiments [35, 36].

Only a few studies in the international publications [37, 38] have simultaneously examined quantitative cerebral perfusion measurements obtained using ASL and PCA.

In our study, we demonstrated a strong positive correlation between perfusion values obtained with these methodologically distinct sequences, confirming the reliability of the results and supporting the feasibility of applying the previously developed algorithm for quantitative brain perfusion assessment [6] in both clinical practice and research settings. Additionally, the study demonstrates the feasibility of using quantitative PCA to indirectly assess cerebral blood supply.

Sensitivity analysis revealed that the degree of discrepancy between cerebral tissue perfusion estimates obtained by the different methods depended on brain volume. This may be related to partial-volume effects of ASL in specific brain structures [39], as well as to differences in the sensitivity of the techniques to the vascularization of brain volumes of varying size [40]. Moreover, multivariable analysis showed that participant’s sex was associated with the magnitude of disagreement between the quantitative perfusion estimates: men demonstrated a smaller difference between CBF-ASL and CBF-PCA values, which may reflect sex-related differences in cerebrovascular angioarchitecture [41] or methodological aspects of measurement [42]. After adjusting for brain volume, the association between age and discrepancies in cerebral tissue perfusion measurements was not confirmed, indicating that age exerts an indirect effect through anatomical changes [43], thereby underscoring the need to control for confounding factors.

Study Limitations

Limitations related to the control method

Researchers note that non-contrast perfusion assessment using PCA remains an experimental field because of the challenges inherent in estimating human brain volume, density, and mass, which together determine the absolute amount of blood flow per second [44]. Additional difficulties arise in interpreting regional CBF values, as the distribution of blood flow across specific brain regions is heterogeneous in both temporal and volumetric dimensions, whereas microcirculatory model interactions continue to be investigated.

Limitations related to the study sample

This study also has several limitations regarding the applicability of the quantitative perfusion metrics obtained. The sample is not representative and therefore cannot be generalized to the broader population. Another limitation is the absence of participants aged

30–40 years, which is related to technical constraints associated with recruiting volunteers among institutional staff and faculty. This may influence the interpretation of age-related perfusion characteristics and restrict the extrapolation of findings to a middle-aged population.

Limitations related to measurement methods

The reliability of results obtained from T1-weighted image segmentation for estimating brain volume and mass is determined by the preprocessing algorithms used, the processing methods applied, and the selected value of the brain density constant. The high variability among methods indicates reduced accuracy of quantitative assessments.

Limitations related to interpretation

The interpretation of individual participants’ results was performed by a single specialist. In addition, the interpreter was aware of the findings obtained using both methods.

CONCLUSION

A strong positive correlation was established between cerebral perfusion values obtained using ASL and perfusion estimates derived from volumetric macrovascular blood flow measured with PCA. Given that these methods rely on distinct physiological models, the findings refine cerebral perfusion measurements, bringing them closer to true physiological values, and enable selection of the most appropriate approach depending on the clinical scenario, equipment availability, or investigator preference.

ADDITIONAL INFORMATION

Author contributions: V.V. Popov: magnetic resonance imaging, data interpretation, writing—original draft, writing—review & editing; Yu.A. Stankevich: image reading and report generation, data interpretation, writing— original draft, writing—review & editing; O.B. Bogomyakova: image reading and report generation, data interpretation, writing—review & editing; A.A. Tulupov: data interpretation, writing— review & editing. All the authors approved the version of the manuscript to be published and agreed to be accountable for all aspects of the work, ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.

Ethics approval: The study was approved by the Local Ethics Committee of the International Tomography Center, Siberian Branch of the Russian Academy of Sciences (Protocol No. 37 dated April 22, 2024). All patients provided written informed consent to participate in the study.

Funding sources: The study was supported by the Ministry of Science and Higher Education of the Russian Federation (State Assignment No. 1023110800234-5-3.2.25; 3.1.4; 3.2.12, Investigation of Post-Stroke Structural and Functional Brain Reorganization Using Advanced Neuroimaging Methods) as part of a grant competition for the establishment of a youth laboratory. The responsible institution is the International Tomography Center, Siberian Branch of the Russian Academy of Sciences.

Disclosure of interests: The authors have no relationships, activities, or interests for the last three years related to for-profit or not-for-profit third parties whose interests may be affected by the content of the article.

Statement of originality: This study includes ASL-based cerebral blood flow data from 20 individuals, which were originally obtained and analyzed as control group characteristics in our previously published work (Popov V.V. et al., 2024; https://doi.org/10.18699/SSMJ20240622).

Data availability statement: The editorial policy regarding data sharing does not apply to this work.

Generative AI: No generative artificial intelligence technologies were used to prepare this article.

Provenance and peer review: This paper was submitted unsolicited and reviewed following the fast-track procedure. The peer review process involved two external reviewers and the in-house science editor.

About the authors

Vladimir V. Popov

International Tomography Institute, Siberian Branch of the Russian Academy of Sciences; Novosibirsk State University

Author for correspondence.
Email: popov.v@tomo.nsc.ru
ORCID iD: 0000-0003-3082-2315
SPIN-code: 5473-0707

MD

Russian Federation, 3a Institutskaya st, unit 1, Novosibirsk, 630090; Novosibirsk

Yuliya A. Stankevich

International Tomography Institute, Siberian Branch of the Russian Academy of Sciences; Novosibirsk State University

Email: stankevich@tomo.nsc.ru
ORCID iD: 0000-0002-7959-5160
SPIN-code: 6668-5010

MD, Cand. Sci (Medicine)

Russian Federation, 3a Institutskaya st, unit 1, Novosibirsk, 630090; Novosibirsk

Olga B. Bogomyakova

International Tomography Institute, Siberian Branch of the Russian Academy of Sciences; Novosibirsk State University

Email: bogom_o@tomo.nsc.ru
ORCID iD: 0000-0002-8880-100X
SPIN-code: 9172-6975

MD, Cand. Sci (Medicine)

Russian Federation, 3a Institutskaya st, unit 1, Novosibirsk, 630090; Novosibirsk

Andrey A. Tulupov

International Tomography Institute, Siberian Branch of the Russian Academy of Sciences; Novosibirsk State University

Email: taa@tomo.nsc.ru
ORCID iD: 0000-0002-1277-4113
SPIN-code: 6630-8720

MD, Dr. Sci (Medicine), Professor, Corresponding Member of the Russian Academy of Sciences

Russian Federation, 3a Institutskaya st, unit 1, Novosibirsk, 630090; Novosibirsk

References

Supplementary files