Elsevier

NeuroImage

Volume 51, Issue 2, June 2010, Pages 765-774
NeuroImage

Whole-brain mapping of venous vessel size in humans using the hypercapnia-induced BOLD effect

https://doi.org/10.1016/j.neuroimage.2010.02.037Get rights and content

Abstract

Measuring the morphology of the cerebral microvasculature by vessel-size imaging (VSI) is a promising approach for clinical applications, such as the characterization of tumor angiogenesis and stroke. Despite the great potential of VSI, this method has not yet found widespread use in practice due to the lack of experience in testing it on healthy humans. Since this limitation derives mainly from the need for an invasive injection of a contrast agent, this work explores the possibility to employ instead the easily accessible blood oxygenation level dependent (BOLD) effect for VSI of the venous microstructure. It is demonstrated that BOLD-VSI in humans can be realized by a hypercapnic challenge using a fast gradient-echo (GE) and spin-echo (SE) sequence at 7 T. Reproducible maps of the mean venous vessel radius, based on the BOLD-induced changes in GE and SE relaxation rates, could be obtained within a scan time of 10 min. Moreover, the method yields maps of venous blood volume and vessel density. Owing to its non-invasive character, BOLD-VSI provides a low-risk method to analyze the venous microstructure, which will not only be useful in clinical applications, but also provide a better understanding of BOLD effect.

Introduction

The approach of vessel-size imaging (VSI) (Prinster et al., 1997, Dennie et al., 1998, Troprès et al., 2001) for characterizing cerebral tumor vascularization has recently drawn much attention (Robinson et al., 2003, Troprès et al., 2004, Kiselev et al., 2005, Quarles and Schmainda, 2007, Lin et al., 2008, Valable et al., 2008). This non-invasive method allows quantitative mapping of the microvasculature, i.e. of the mean vessel radius in a voxel on a spatial scale of several micrometers, which is otherwise not directly accessible by magnetic resonance imaging (MRI). VSI employs the ratio of the change between the transverse gradient-echo (GE) and spin-echo (SE) relaxation rates, ΔR2 and ΔR2, in response to an intravascular contrast agent. This method works because the ratio ΔR2R2 is relatively insensitive to the blood volume fraction and the magnitude of the induced susceptibility shift (compare Fig. 1). Yet, ΔR2R2 is strongly influenced by diffusional dephasing which depends on the vessel size (Boxerman et al., 1995).

Accompanied by an increased metabolic activity, the tumor vasculature is structurally and functionally abnormal (Robinson et al., 2003). Thus, VSI is a strong indicator for angiomas and gliomas and permits evaluation of morphological changes in tumor vasculature in response to therapy (Quarles and Schmainda, 2007). VSI correlates strongly with tumor grade (Donahue et al., 2000) and can also be applied to characterize stroke (Bosomtwi et al., 2008). VSI was validated with stereologic measurements on histology data (Troprès et al., 2004) and micro-CT angiography (Ungersma et al., 2008). In summary, VSI has the potential to complement, if not replace, invasive tumor biopsy.

Despite the great potentials of VSI, the method has not found widespread use in clinical practice. Previous studies have largely relied on animal models. We believe that this mismatch between the theoretical possibilities and the practical application of VSI in humans is due to the lack of experience with healthy subjects. A reason for this deficit is that studies using intravenously administered contrast agents, such as gadopentetate dimeglumine (Gd-DTPA), in humans, solely for the purpose of methodological development, are often ethically difficult to justify. We propose, therefore, to employ the easily accessible blood oxygenation level dependent (BOLD) effect for VSI of the venous (i.e. post-capillary) vasculature (Prinster et al., 1997, Jochimsen and Möller, 2008), for instance, induced by a mild hypercapnic challenge, such as breathing carbogen (5% CO2, 95% O2). The common understanding is that hypercapnia elevates blood flow (Robinson et al., 1995) due to a dilation of pial arterioles mediated by an increase in the extracellular pH-value (Kontos et al., 1977). Since oxidative metabolism is not affected, blood oxygenation is increased globally (Schwarzbauer and Heinke, 1999, Sedlacik et al., 2008b). In addition, the high oxygen concentration in carbogen amplifies the effect by increased (arterial) oxygen saturation. Consequently, the concentration of paramagnetic deoxyhemoglobin is decreased, which leads to a shift in intravascular susceptibility, making deoxyhemoglobin effectively an endogenous contrast agent.

Compared to VSI based on an exogenous contrast agent, BOLD-VSI is more susceptible to (physiological) noise due to the smallness of the BOLD-induced ΔR2 and, in particular, ΔR2. However, in BOLD-VSI, transient changes in the relaxation rates can be induced repeatedly. Hence, the smallness of the effect can be outweighed by the statistical power available in repeating the hypercapnic challenge. Moreover, since BOLD-VSI does not require the injection of a contrast agent, potential toxic effects can be avoided, for instance, in patients with an impaired kidney function, which would be aggravated by accumulation due to multiple examinations (Thomsen, 2008). Even if this risk is negligible, the injection may also provoke adverse reactions (Li et al., 2006) and remains overall an unpleasant procedure for the patient. Furthermore, the use of low-molecular-weight contrast agents, such as Gd-DTPA, can be complicated by extravasation through leaky capillaries in tumors, which leads to errors in the calculated VSI (Kiselev et al., 2005). All these difficulties can be avoided by BOLD-VSI. Thus, it would allow a low-risk screening procedure in the attempt to identify abnormal tissue, and it could be repeated frequently in multiple examinations that aim to track tumor development and/or treatment, e.g. during radiotherapy. The combination of BOLD-VSI with direct imaging of the venous macrovasculature by susceptibility-weighted imaging (SWI) (Reichenbach et al., 1997) would also provide a method to examine the whole venous vascular tree. BOLD-VSI will also provide a better understanding of the BOLD effect, and can be used to increase of the specificity in functional magnetic resonance imaging (fMRI) by identifying the origin of the BOLD contrast (Jochimsen and Möller, 2008).

In summary, BOLD-VSI has great potential to make VSI feasible in clinical practice. It is demonstrated in this work that whole-brain BOLD-VSI is possible in humans by using the administration of carbogen as a hypercapnic challenge. Furthermore, it is shown that maps of venous vessel size can be utilized to calculate the venous blood volume fraction and vessel density.

Section snippets

Imaging experiments

Measurements were performed on a 7 T Magnetom (Siemens Medical Solutions, Erlangen, Germany) using a 24-channel head coil (Nova Medical, Wilmington MA, USA). If not mentioned otherwise, the ODIN framework (Jochimsen and von Mengershausen, 2004) was used for sequence design, image reconstruction and Monte-Carlo simulation as described in the following. A single-shot, multi-slice, combined GE–SE sequence was used to obtain maps of ΔR2 and ΔR2 simultaneously and with high temporal resolution. The

Vascular model

The simulated curves for rv(q) in Fig. 1 are very similar in shape over a wide range of different β and Δχ. In particular, the curves are almost identical in the medium range at ≈ 10 μm, which is very close to the maximum of the rv-distributions (compare Fig. 6). This indicates that the calculated value of rv depends very little on the venous blood volume and on the strength of the hypercapnia-induced oxygenation changes. Thus, it can be expected that mapping rv provides reliable structural

Discussion

The present work demonstrates the feasibility of VSI based on the hypercapnia-induced BOLD effect. It was possible to assign a reproducible rv to the majority of voxels within a scan duration of approximately 10 min at 7 T. The relatively high similarity of values across regions and subjects suggests that BOLD-VSI has the potential to identify abnormal tissue where rv would differ significantly.

The high field strength used in the present study leads to large changes in R2 and R2 which allowed a

References (51)

  • P.A. Bandettini et al.

    A hypercapnia-based normalization method for improved spatial localization of human brain activation with fMRI

    NMR Biomed.

    (1997)
  • A. Bosomtwi et al.

    Quantitative evaluation of microvascular density after stroke in rats using MRI

    J. Cereb. Blood Flow Metab.

    (2008)
  • J.L. Boxerman et al.

    MR contrast due to intravascular magnetic susceptibility perturbations

    Magn. Reson. Med.

    (1995)
  • D. Bulte et al.

    Measurement of cerebral blood volume in humans using hyperoxic MRI contrast

    J. Magn. Reson. Imaging

    (2007)
  • J. Dennie et al.

    NMR imaging of changes in vascular morphology due to tumor angiogenesis

    Magn. Reson. Med.

    (1998)
  • K.M. Donahue et al.

    Utility of simultaneously acquired gradient-echo and spin-echo cerebral blood volume and morphology maps in brain tumor patients

    Magn. Reson. Med.

    (2000)
  • C.R. Fisel et al.

    MR contrast due to microscopically heterogenous magnetic susceptibility: numerical simulations and applications to cerebral physiology

    Magn. Reson. Med.

    (1991)
  • G.H. Glover et al.

    Image-based method for retrospective correction of physiological motion effects in fMRI: RETROICOR

    Magn. Reson. Med.

    (2000)
  • X. He et al.

    Quantitative BOLD: mapping of human cerebral deoxygenated blood volume and oxygen extraction fraction: default state

    Magn. Reson. Med.

    (2007)
  • F.G.C. Hoogenraad et al.

    In vivo measurement of changes in venous blood-oxygenation with high resolution functional MRI at 0.95 Tesla by measuring changes in susceptibility and velocity

    Magn. Reson. Med.

    (1998)
  • D. Huo et al.

    Robust GRAPPA reconstruction and its evaluation with the perceptual difference model

    J. Magn. Reson. Imaging

    (2008)
  • J.H. Jensen et al.

    MR imaging of microvasculature

    Magn. Reson. Med.

    (2000)
  • J.H. Jensen et al.

    Microvessel density estimation in the human brain by means of dynamic contrast-enhanced echo-planar imaging

    Magn. Reson. Med.

    (2006)
  • T.H. Jochimsen et al.

    Identifying systematic errors in quantitative dynamic-susceptibility contrast perfusion imaging by high resolution multi-echo parallel EPI

    NMR Biomed.

    (2007)
  • A. Kastrup et al.

    Functional magnetic resonance imaging of regional cerebral blood oxygenation changes during breath holding

    Stroke

    (1998)
  • Cited by (35)

    View all citing articles on Scopus
    View full text