Summary

Construction and Application of Cerebral Functional Region-Based Cerebral Blood Flow Atlas Using Magnetic Resonance Imaging-Arterial Spin Labeling

Published: May 31, 2024
doi:

Summary

This study integrated magnetic resonance imaging- arterial spin labeling images to derive cerebral blood flow (CBF) atlas for cerebral functional regions. Comparing typical healthy and chronic cerebral ischemia CBF atlases revealed significant differences in regional CBF distributions, enabling rapid, noninvasive assessments of functional CBF to assist in diagnosis and evaluate therapeutics.

Abstract

Cerebral conditions often require precise diagnosis and monitoring, necessitating advanced imaging techniques. Current modalities may not adequately detect early signs of reversible tissue damage, underlining the need for innovative diagnostic tools that can quantify changes in cerebral blood flow (CBF) with high specificity and sensitivity. This study integrates three-dimensional arterial spin labeling (3D-ASL) with structural MRI to develop comprehensive CBF atlases that cover all main functional regions of the brain. This innovative magnetic resonance imaging- arterial spin labeling (MRI-ASL) methodology provides a rapid and noninvasive means of quantifying region-specific CBF, offering a detailed view of CBF levels across different functional regions.The comparison between chronic cerebral ischemia (CCI) patients and healthy subjects revealed significantly diminished CBF across the cerebral functional regions in the constructed CBF atlases for the former. This approach not only allows for the efficient identification of CCI by analyzing concurrent decreases in CBF across critical areas relative to healthy distributions but also enables the tracking of treatment responses and rehabilitation progress through longitudinal CBF atlases.The CBF atlas developed using the MRI-ASL technique represents a novel advancement in the field of cerebral diagnostics and patient care. By comparing regional CBF levels against normative standards, this method enhances diagnostic capabilities, enabling clinicians to provide personalized care to patients with cerebral conditions.

Introduction

In the realm of neuroimaging, the quest for precise, noninvasive tools to assess cerebral function and pathology remains paramount. Among these, cerebral blood flow (CBF) stands as a vital indicator, reflecting the metabolic demands and health status of brain tissue1. Traditional approaches often entail empirical assessments, relying heavily on the expertise of clinicians to interpret images and discern pathological changes2. However, advancements in magnetic resonance imaging (MRI) techniques, particularly arterial spin labeling (ASL)3, offer a promising avenue for quantifying CBF with greater accuracy and objectivity4,5.

This study presents a pioneering methodology that integrates three-dimensional ASL (3D-ASL) with structural MRI to construct a comprehensive CBF atlas across cerebral functional regions6. By leveraging this novel approach, clinicians can not only obtain a global perspective of CBF but also delve into specific functional areas, allowing for a nuanced understanding of cerebral perfusion patterns7,8. This improvement in resolution is a direct result of technological progress in imaging equipment rather than the use of interpolated voxels. It is worth noting that the majority of mainstream MRI devices available on the market today typically offer imaging precision better than 1.5 mm9. These advancements in imaging technology have paved the way for more detailed and accurate CBF assessments. This represents a paradigm shift from conventional imaging, which often lacks the resolution to detect subtle changes in CBF associated with early-stage pathologies10.

The genesis of this methodology lies in the imperative to address the diagnostic challenges posed by cerebral conditions, including chronic cerebral ischemia (CCI) and other neurological disorders11,12. These conditions necessitate precise and timely assessments to guide therapeutic interventions effectively13,14. By comparing CBF atlases between healthy individuals and patients with CCI, this study unveils significant disparities in regional CBF distributions, offering insights into disease pathology and potential treatment avenues.

The utility of this MRI-ASL approach extends beyond diagnosis, encompassing therapeutic evaluation and monitoring of disease progression15. Longitudinal CBF atlases hold promise in tracking treatment responses and rehabilitation outcomes, providing clinicians with invaluable tools for personalized patient management. Moreover, the ability to discern subtle CBF changes may serve as an early biomarker for impending tissue abnormalities, enabling proactive interventions to mitigate neurological damage before it becomes irreversible16.

While this methodology represents an advanced tool, several avenues for refinement and expansion merit consideration. Standardizing scanning protocols, CBF normalization techniques, and constructing multi-subject healthy CBF atlases are crucial steps toward enhancing diagnostic accuracy and clinical utility. Collaborative efforts across diverse cerebral pathologies are essential to validate and refine this approach for widespread clinical adoption.

This study introduces a novel approach whereby MRI-derived CBF atlases offer clinicians deep insights into cerebral function and pathology. By bridging the gap between imaging group and clinical interpretation, this methodology has the potential to revolutionize the diagnosis and management of a myriad of neurological conditions, ushering in a future of precision medicine tailored to the unique needs of each patient.

Protocol

This study was approved by the Institutional Review Board of Beijing Dongzhimen Hospital, Beijing, China. An MRI scanner was used with pulsed ASL (PASL) sequence based on turbo gradient spin echo (TGSE) for 3D arterial spin labeling (3D-ASL) with the following parameters: TR 4000 ms, TE 25 ms, bolus duration 700 ms, inversion time 1990 ms. The software tools used in this research are listed in the Table of Materials.

1. Data collection and preparation

NOTE: The variance in parameters remains unaffected by the research approach. Both DICOM and NIFTI formats are used for storing digital medical imaging data, with DICOM being the usual output from clinical imaging devices. However, the NIFTI format is often preferred for computational convenience in research activities. Converting from DICOM to NIFTI is a straightforward and common practice17. In this study, authentic DICOM data were acquired and converted into NIFTI format. The data was acquired using a 1.5 Tesla MRI scanner. In the cross-image registration process of this study, the fluid-attenuated inversion recovery (FLAIR) sequence was mainly used and fused with CBF images. The CBF Atlas tool (Table of Materials) used in this study is commercial software.

  1. Copy data to the designated working directory.
    1. Copy all NIFTI data to a customized working directory.
      NOTE: The working directory is the same in the operating system and MATLAB. The research follows the Right Anterior Superior (RAS) orientation standard.
    2. Go to the directory housing the data within MATLAB's current working directory, and use the niftiread function to load FLAIR data into the workspace. Use the size function to check the dimensions of the FLAIR sequence. Call the Flair_Slice command to view the FLAIR sequence (as shown in Figure 1). Use the specific commands as follows:
      FLAIR_XLF = niftiread('FLAIR_XFL.nii');
      size(FLAIR_XLF)
      Flair_Slice(FLAIR_XLF);
    3. Refer to Figure 1 for an image of the interactive graphic user interface (GUI) for the FLAIR sequence. Use the bottom scroll bar to quickly browse the different sequences.
  2. Quickly check images of CBF.
    1. Use the niftiread function to load CBF data into the workspace. Use the size function to check the dimensions of the CBF sequence. Call the CBF_Slice command to view the CBF sequence (as shown in Figure 2). Use the specific commands as follows:
      CBF_XLF = niftiread('CBF_XFL.nii');
      size(CBF_XLF)
      CBF_Slice(CBF_XLF);
    2. Refer to Figure 2 for an interactive GUI screenshot of the CBF sequence. Use the bottom scroll bar to quickly browse the different sequences.
      NOTE: In Figure 2, the CBF value range is usually 0-120 mL/100 g/min. Figure 2 uses a color map to represent different CBF levels in different colors.

2. Segmentation of cerebral functional regions from FLAIR sequence

NOTE: The FLAIR sequence serves as both structural imaging and provides excellent pathological diagnostic capabilities. Therefore, fusing FLAIR with CBF holds important diagnostic value in clinics. This study segments major cerebral functional regions from the FLAIR sequence.

  1. In the workspace, call the FLAIR_Segment function and run the pretrained 3-D U-Net-based image segmentation program to automatically generate triplanar views of cerebral functional regions segmentation, as shown in Figure 3. Each color in Figure 3 represents a distinct functional region.
  2. For real-time inspection of different cerebral functional regions, use the crosshair interaction (Figure 3). Click and drag the center of the crosshair for an arbitrary 3D examination of the reconstructed brain anatomy.
    NOTE: The GUI in Figure 3 also enables adjustment of the grayscale intensity range, contrast, and brightness of the triplanar views.
  3. Press and drag the left mouse button over any region of the images for real-time modification of brightness and contrast levels. Release the mouse button to confirm and finalize the adjustments.

3. Triplanar views of CBF distribution across cerebral functional regions

NOTE: Examining CBF distribution across different cerebral functional regions facilitates precise clinical judgments of patient conditions. Under the functional region framework from Figure 3, incorporating the exact CBF values from the CBF sequence and presenting them in triplanar views allows comprehensive physician inspection.

  1. Call the CBF_triplanar function to generate the triplanar GUI view shown in Figure 4, displaying CBF spatial distribution across functional regions. Move the crosshair to allow examination of CBF distribution in regions of interest.
  2. Click the Data Tips button on the top right corner of the GUI to display the CBF values at any position.
  3. Press and drag the left mouse button over any region of the images for real-time modification of brightness and contrast levels. Release the mouse button to confirm and finalize the adjustments.

4. CBF atlas across major cerebral functional regions

NOTE: Normalizing the CBF probability distributions across different functional regions generates the Cerebral Functional Regions CBF Atlas, expressing the CBF levels across cerebral functional regions of the subject.

  1. Call the CBF_Atlas function to convert the CBF spatial distribution shown in Figure 4 into a CBF Atlas (as shown in Figure 5).
    NOTE: In Figure 5, the x-axis represents different cerebral functional regions, and the y-axis represents different CBF levels; the different colors denote different probability levels (the redder the color, the more voxels present).
  2. Click the Zoom In/Out button on the top right of the GUI shown in Figure 5 to scale partial images.
    NOTE: The curve in Figure 5 connects the average CBF across regions.

5. Significant differences in CBF_Atlas between healthy subjects and CCI patients

NOTE: Using the same process described in sections 1-4, the average CBF values across different cerebral functional regions can be obtained for CCI patients.

  1. Utilize the CBF_Compare function to generate a comparative plot of the CBF curve from a healthy subject versus a CCI patient (Figure 6).
  2. Observe the significant differences between the patient's CBF curve (colored in black) and the healthy curve (colored in red) (Figure 6). Identify the functional regions with more pronounced CBF drops in the patient. Integrating step 3.3, re-examine the regions where the patient performed worse compared to the other regions.
    NOTE: Step 3.3 allows for the viewing of the patient's CBF levels in three-dimensional space, meaning doctors can utilize the triplanar GUI view shown in Figure 4 to examine the CBF at any position of the patient. Here, the degree of CBF decline varies in different locations for CCI patients. Doctors can revisit Figure 4 to focus on brain regions with significant CBF decline.
  3. Use the icons in the GUI's upper right corner (Figure 5) to access functions such as zooming out, zooming in, returning to the global view, and marking the coordinates of the selected pixel.
    NOTE: Using the same principle, pre-treatment and post-treatment results from the same patient can also be compared using the approach in Figure 6 to evaluate clinical efficacy over time.

Representative Results

This investigation utilizes actual patient data acquired using a 1.5 T MRI scanner to validate the cerebral blood flow (CBF) quantification and atlas construction methodology. The preprocessing steps included FLAIR structural images (Figure 1), CBF images (Figure 2), and triplanar fused images (Figure 3 and Figure 4).

Cerebral functional regions were segmented from FLAIR scans using a pre-trained deep learning model (Figure 3). The ASL-CBF images were then registered to align with the FLAIR space. Fusing the region labels with corresponding CBF values generated triplanar views exhibiting CBF distribution across segmented areas (Figure 4).

Further processing constructed a CBF atlas for the healthy subject (Figure 5) and CCI patient (Figure 6). The x-axis denotes regions, the y-axis denotes CBF levels, and color intensity reflects probability. The curve indicates the mean regional CBF. Comparing the two CBF atlases revealed significantly reduced CBF across all cerebral functional regions in the CCI patient versus the healthy subject.

This technique successfully differentiated normal versus abnormal CBF distribution at global and regional resolution levels. The construction of 3D cerebral functional regions CBF atlases based on MRI-ASL imaging holds substantial potential for assisted diagnosis of various cerebral pathophysiologies.

Figure 1
Figure 1: Interactive graphical user interface for FLAIR slice view. The figure shows the GUI for viewing FLAIR slices. Please click here to view a larger version of this figure.

Figure 2
Figure 2: Interactive graphical user interface for CBF slice view. The figure shows the GUI for viewing CBF slices. Please click here to view a larger version of this figure.

Figure 3
Figure 3: Segmentation of cerebral functional regions based on FLAIR imaging sequence. Cerebral functional regions extraction and triplanar views based on FLAIR sequence. Please click here to view a larger version of this figure.

Figure 4
Figure 4: Triplanar views of CBF spatial distribution across cerebral functional regions. Triplanar views showing CBF distribution across cerebral regions. Please click here to view a larger version of this figure.

Figure 5
Figure 5: Cerebral Functional Regions CBF Atlas. CBF atlas constructed by the probability distribution of region-specific CBF levels. This study adopts a commonly used segmentation of 32 regions. These regions include background, left cerebral white matter, left cerebral cortex, left lateral ventricle, left inferior lateral ventricle, left cerebellum white matter, left cerebellum cortex, left thalamus, left caudate, left putamen, left pallidum, third ventricle, fourth ventricle, brain stem, left hippocampus, left amygdala, left accumbens area, left ventral DC, right cerebral white matter, right cerebral cortex, right lateral ventricle, right inferior lateral ventricle, right cerebellum white matter, right cerebellum cortex, right thalamus, right caudate, right putamen, right pallidum, right hippocampus, right amygdala, right accumbens area, and right ventral DC. Please click here to view a larger version of this figure.

Figure 6
Figure 6: Comparison of Cerebral Functional Regions CBF Atlas. Comparison reveals significantly reduced CBF across cerebral functional regions in a CCI patient vs a healthy subject. Please click here to view a larger version of this figure.

Discussion

The key steps (sections 3 and 4) constitute the basis for constructing the CBF Atlas, quantifying CBF distribution across cerebral functional regions. Step 4.2 explicitly delineates CBF levels for each brain area, pioneering a new technique. This not only provides physicians with a global view of patient CBF but also quantitative measurements of individual functional regions. Step 5.1 demonstrates that the CBF Atlas holds substantial clinical diagnostic utility distinguishing CCI from healthy controls.

Several technical refinements could enhance this promising protocol, mainly by utilizing MRI-ASL imaging. Standardization regarding scanning accuracy, CBF normalization, and synchronization of structural and CBF images is needed. Additionally, constructing a multi-subject healthy CBF Atlas as a diagnostic baseline via expanded cohorts warrants further investigations. Accumulating CBF atlas patterns across diverse cerebral pathophysiologies through clinical collaborations also represents important future work.

A current limitation is CBF, though invaluable, is not the sole biomarker for brain conditions18,19. Integrating other vital clinical markers such as phosphorylated tau protein (p-tau) in cerebrospinal fluid, electroencephalography (EEG), and cerebral metabolic rates (e.g., glucose metabolic rate) could augment diagnostics.

Accurately calculating the CBF distribution for each functional brain region and presenting it in the form of an atlas necessitates aligning brain functional structural imaging with CBF sequences in three-dimensional space and calculating the CBF distribution for each region. This effort represents the main innovation of our study. The resulting CBF atlas, with the x-axis representing brain functional regions and the y-axis showing the CBF distribution for each region, allows for comparing pre- and post-treatment outcomes in patients and enables quantitative comparisons between healthy and diseased groups20. This makes the assessment of CBF more precise (enabling accurate comparison of CBF distribution across regions) rather than relying on experiential judgement21.

Furthermore, the potential of longitudinal CBF atlases to monitor treatment efficacy and track the progression of neurological conditions over time offers a promising avenue for personalized medicine. By establishing a baseline of CBF distribution within healthy cerebral functional regions, clinicians can more accurately gauge the impact of therapeutic interventions and make informed decisions regarding patient care.

In conclusion, developing the CBF Atlas utilizing MRI-ASL imaging is a promising step toward improving the diagnostic and therapeutic landscape for cerebral conditions. This technique has the potential to contribute to a more comprehensive understanding of cerebral blood flow dynamics and may support the development of precision medicine approaches that aim to tailor interventions based on individual patient's cerebral perfusion patterns. Future directions for this research include refining the methodology, standardizing protocols, and expanding clinical validations to fully realize the potential of CBF atlases in neurology and beyond.

Declarações

The authors have nothing to disclose.

Acknowledgements

This study received significant support and modeling assistance from the R&D department of Beijing Intelligent Entropy Science & Technology Co Ltd., Beijing, China.

Materials

CBF Atlas Intelligent Entropy CBF Atlas V1.0 Beijing Intelligent Entropy Science & Technology Co Ltd.
Modeling for Thyroid Disease
MATLAB MathWorks 2023B Computing and visualization
MRI Device Siemens Amria 1.5 T MRI scanner

Referências

  1. Rodriguez, G., et al. Regional cerebral blood flow in chronic stroke patients. Stroke. 24 (1), 94-99 (1993).
  2. Kwok, C. H. R., et al. Cognition and cerebral blood flow after extracranial carotid revascularization for carotid atherosclerosis: A systematic review. Clin Ther. 45 (11), 1069-1076 (2023).
  3. Togao, O., et al. Arterial spin labeling-based MR angiography for cerebrovascular diseases: Principles and clinical applications. J Magn Reson Imaging. , (2023).
  4. Haidar, H., Majzoub, R. E., Hajeer, S., Abbas, L. A. Arterial spin labeling (ASL-MRI) versus fluorodeoxyglucose-PET (FDG-PET) in diagnosing dementia: a systematic review and meta-analysis. BMC Neurol. 23 (1), 385 (2023).
  5. Xu, Y., Tan, G., Chen, D., Liu, J., Zhou, Z., Liu, L. Arterial spin labeling perfusion MRI applications in drug-resistant epilepsy and epileptic emergency. Acta Epileptol. 5, 23 (2023).
  6. Xie, H., et al. Abnormalities of cerebral blood flow and the regional brain function in Parkinson’s disease: a systematic review and multimodal neuroimaging meta-analysis. Front Neurol. 14, 1289934 (2023).
  7. Tenberg, A., et al. Dysautonomia and activity in the early stroke recovery period. Neurol Sci. 45 (6), 2505-2521 (2024).
  8. Joshi, D., Prasad, S., Saini, J., Ingalhalikar, M. Role of arterial spin labeling (ASL) images in Parkinson’s disease (PD): A systematic review. Acad Radiol. 30 (8), 1695-1708 (2023).
  9. Mohamed, A. Z., Kwiatek, R., Fante, P. D., Calhoun, V. D., Lagopoulos, J., Shan, Z. Y. Functional MRI of the brainstem for assessing its autonomic functions: From imaging parameters and analysis to functional atlas. J Magn Reson Imaging. , (2024).
  10. Vanherle, L., Matuskova, H., Don-Doncow, N., Uhl, F. E., Meissner, A. Improving cerebrovascular function to increase neuronal recovery in neurodegeneration associated to cardiovascular disease. Front Cell Dev Biol. 8, 53 (2020).
  11. Fujimura, M., Tominaga, T. Significance of cerebral blood flow analysis in the acute stage after revascularization surgery for Moyamoya disease. Neurol Med Chir (Tokyo). 55 (10), 775-781 (2015).
  12. Østergaard, L., et al. The role of the microcirculation in delayed cerebral ischemia and chronic degenerative changes after subarachnoid hemorrhage. J Cereb Blood Flow Metab. 33 (12), 1825-1837 (2013).
  13. Mankoo, A., et al. The role of the autonomic nervous system in cerebral blood flow regulation in stroke: A review. Auton Neurosci. 246, 103082 (2023).
  14. Yang, X., Qiang, Q., Li, N., Feng, P., Wei, W., Hölscher, C. Neuroprotective mechanisms of glucagon-like peptide-1-based therapies in ischemic stroke: An update based on preclinical research. Front Neurol. 13, 844697 (2022).
  15. Ho, M. -. L. Arterial spin labeling: Clinical applications. J Neuroradiol. 45 (5), 276-289 (2018).
  16. Sforza, M., Bianchini, E., Alivernini, D., Salvetti, M., Pontieri, F. E., Sette, G. The impact of cerebral vasomotor reactivity on cerebrovascular diseases and cognitive impairment. J Neural Transm (Vienna). 129 (11), 1321-1330 (2022).
  17. Elhadad, A., Jamjoom, M., Abulkasim, H. Reduction of NIFTI files storage and compression to facilitate telemedicine services based on quantization hiding of downsampling approach. Sci Rep. 14 (1), 5168 (2024).
  18. Gonzalez-Ortiz, F., Kac, P. R., Brum, W. S., Zetterberg, H., Blennow, K., Karikari, T. K. Plasma phospho-tau in Alzheimer’s disease: towards diagnostic and therapeutic trial applications. Mol Neurodegener. 18 (1), 18 (2023).
  19. Magliozzi, R., et al. CSF parvalbumin levels reflect interneuron loss linked with cortical pathology in multiple sclerosis. Ann Clin Transl Neurol. 8 (5), 534-547 (2021).
  20. Kawano, H., et al. Aging and sex differences in brain volume and cerebral blood flow. Aging Dis. , (2023).
  21. Jia, R., Solé-Guardia, G., Kiliaan, A. J. Blood-brain barrier pathology in cerebral small vessel disease. Neural Regen Res. 19 (6), 1233-1240 (2023).

Play Video

Citar este artigo
Tan, Z., Xing, F., Zhang, L. Construction and Application of Cerebral Functional Region-Based Cerebral Blood Flow Atlas Using Magnetic Resonance Imaging-Arterial Spin Labeling. J. Vis. Exp. (207), e66853, doi:10.3791/66853 (2024).

View Video