BOLD fMRI of cerebrovascular reactivity in the middle cerebral artery territory: A 100 volunteers’ study

+Model NEURAD-516; No. of Pages 7

ARTICLE IN PRESS

Journal of Neuroradiology (2015) xxx, xxx—xxx

Available online at

ScienceDirect www.sciencedirect.com

ORIGINAL ARTICLE

BOLD fMRI of cerebrovascular reactivity in the middle cerebral artery territory: A 100 volunteers’ study夽 Naïla Boudiaf a,b,∗, Arnaud Attyé c,e, Jan M. Warnking c, Irène Troprès c,d, Laurent Lamalle d, Johan Pietras d, Alexandre Krainik c,d,e a

Université Grenoble Alpes 3bis, CNRS, LPNC, 38000 Grenoble, France Université Savoie 3, LPNC, 73000 Chambéry, France c Inserm, université Grenoble Alpes, GIN, CHU de Grenoble, 38000 Grenoble, France d Inserm, université Grenoble Alpes, CNRS, IRMaGe, CHU de Grenoble, 38000 Grenoble, France e Department of Neuroradiology and MRI, University Hospital of Grenoble—IFR1, Grenoble, France b

KEYWORDS Middle cerebral artery stenosis; Cerebral vascular reactivity; Hypercapnia; BOLD-fMRI

Summary Background and purpose: The assessment of cerebrovascular reactivity (CVR) has shown promising results for its use in medical diagnosis and prognosis, especially in patients suffering from severe intracranial arterial stenosis. However, its quantification remains uncertain because of a large variability inherent in brain anatomy and in methodological settings. To overcome this variability, we provide lateralization index (LI) values for CVR within the middle cerebral artery territory to detect CVR impairment. Materials and methods: We assessed CVR in 100 volunteers (41 females; 47.52 ± 21.58 years) without cervico-encephalic arterial stenosis using BOLD-fMRI contrast with a block-design hypercapnic challenge. Averaged end-tidal CO2 was used as a physiological regressor for statistical analyses with a general linear model. We measured %BOLD signal-change in segmented gray matter regions of interest in the middle cerebral artery territory (MCA). We calculated a laterality index according to the following formula: LI = (CVRleft −CVRright )/(CVRleft + CVRright ). We

Abbreviations: CVR, cerebrovascular reactivity; CO2 , carbon dioxide; fMRI, functional magnetic resonance imaging; BOLD, blood oxygen level dependent; IE, ischemic event; TIA, transient ischemic attack; PTAS, percutaneous transluminal angioplasty and stenting; MCA, middle cerebral artery; LI, laterality index; TOF, time of flight; GE, gradient echo; SSH-EPI, single shot echo-planar images; TR, repetition time; TE, echo time; FOV, field of view; ROI, region of interest; ANOVA, analysis of variance; PetCO2 , end-tidal CO2 pressure; OEF, oxygen extraction fraction; ASL, arterial spin labelling; CBF, cerebral blood flow. 夽 This work was presented in oral communication at the 61st edition of the French Congress of Radiology (61es Journées franc ¸aises de radiologie 2013), October 2013, Paris—France. ∗ Corresponding author at: Department of Neuroradiology and MRI, University Hospital of Grenoble, CS 10217, 38043 Grenoble cedex 9, France. Tel.: +33 4 76 76 54 87/+33 4 76 76 54 85. E-mail address: [email protected] (N. Boudiaf). http://dx.doi.org/10.1016/j.neurad.2015.04.004 0150-9861/© 2015 Elsevier Masson SAS. All rights reserved.

Please cite this article in press as: Boudiaf N, et al. BOLD fMRI of cerebrovascular reactivity in the middle cerebral artery territory: A 100 volunteers’ study. J Neuroradiol (2015), http://dx.doi.org/10.1016/j.neurad.2015.04.004

+Model NEURAD-516; No. of Pages 7

ARTICLE IN PRESS

2

N. Boudiaf et al. tested the effects of methodological settings (i.e. hypercapnic gas, gas administration means, MR acquisition and sex) on %BOLD signal change and LI values with analysis of variance. Results: No adverse effects of the hypercapnic challenge were reported. LI values were independent of experimental conditions. Mean LI calculated in MCA territories was 0.016 ± 0.031, giving the lower and upper limits of 95% (m ± 2SD) of this population distribution at]−0.05; 0.08[. Conclusion: LI can effectively help us to overcome measurement variabilities. Therefore, it can be used to detect abnormal asymmetries in CVR and identify patients at higher risk of ischemic stroke. © 2015 Elsevier Masson SAS. All rights reserved.

Introduction The brain has different systems to control blood distribution. Cerebral vasoreactivity (CVR) is one of them. This system responds to the concentration of circulating gas in blood vessels, such as carbon dioxide (CO2 ). Studying CVR using vasoactive stimuli and functional magnetic resonance imaging (fMRI) with blood oxygen level dependent (BOLD) contrast is an emerging technique [1,2]. It has provided interesting data in several brain diseases, such as Alzheimer’s disease or tumors [3—8]. In clinical practice, CVR mapping has provided reliable data to estimate cerebrovascular reserve and to manage patients referred for arterial steno-occlusive disease and stroke [9—12]. Indeed, Mazighi et al. reported that 27.4% of patients with severe intracranial arterial stenosis suffer from hemodynamic ischemic events (IE) such as stroke or transient ischemic accidents (TIA). Among those patients, 60.7% had recurrent IE within two years after the initial event, and 38.2% did not respond to optimal medical treatment [13]. In another study, the SAMMPRIS trial, Derdeyn et al. compared percutaneous transluminal angioplasty and stenting (PTAS) to aggressive medical treatment alone, in patients with intracranial arterial stenosis. Despite advances in PTAS, they showed a higher risk of IE and adverse events in the PTAS group when compared to the aggressive medical treatment group. Nevertheless, in this latter group the risk of IE reached 19% at 3 years follow-up, and it remained unclear whether those patients experienced a hemodynamic or an embolic IE [14]. As patients at risk of hemodynamic IE may have a greater benefit of PTAS, it is critical to better identify this population. CVR impairment has been shown to be associated with hemodynamical IE downstream of the stenotic artery [15—17] and with low-grade ischemic injury [18,19]. Therefore, CVR imaging could be helpful to better select these patients and to justify the increased risk of PTAS adverse events, as well as to monitor post-therapeutic changes [20—22]. Assessing CVR could also be complementary to morphological MRI to better characterize TIA prognosis due to hemodynamic disorders [23—25]. Although CVR BOLD fMRI using hypercapnic challenges is a promising, safe, and reliable approach [26], the quantification of the cerebrovascular reserve remains difficult because

BOLD contrast might be affected by MRI acquisitions and the hypercapnic stimulus [8]. Differences across age, sex, brain regions and vascular territories were also identified [8,27,28]. For this reason, we conducted a retrospective study on data collected in previous studies among 100 volunteers free from extra or intracranial arterial stenosis [4,7,29]. We had two objectives. First, we aimed at finding a way to overcome the variability in absolute CVR BOLD assessment. Second, we aimed at providing physiological CVR range values as a reference to detect CVR impairment downstream from unilateral arterial stenosis. We focused on the middle cerebral artery (MCA) territories because this is where most severe intracranial arterial stenosis occur [14,15]. First, we tested the effect of methodological settings on CVR BOLD amplitudes. Then, we tested the effect of methodological settings on CVR expressed as laterality index (LI). As this index takes into account the difference of CVR between the two hemispheres of the same subject, we hypothesized that it would be helpful in overcoming the inter-subject variability. Furthermore, we aimed at providing LI range values for normal CVR within the MCA territory.

Methods Subjects Data from one hundred non-smoking right-handed volunteers (41 females, 47.52 ± 21.58 years) were gathered from previous studies [4,7,29] performed in Grenoble Hospital, which were approved by the local University Hospital Ethics Committee. All participants gave their informed consent according to the Declaration of Helsinki. Volunteers were free from both extra and intracranial arterial stenosis assessed by means of either Doppler sonography or angio-MRI (TOF). They were asked to avoid drinking coffee at least 2 hours before the examination, and to stay awake during MRI acquisitions.

Hypercapnic stimulus Hypercapnic challenges were conducted according to the following three block-designed paradigm [air (1 ) — hypercapnia (2 ) — air (1 )] × 3, for a total duration of 12 minutes.

Please cite this article in press as: Boudiaf N, et al. BOLD fMRI of cerebrovascular reactivity in the middle cerebral artery territory: A 100 volunteers’ study. J Neuroradiol (2015), http://dx.doi.org/10.1016/j.neurad.2015.04.004

+Model NEURAD-516; No. of Pages 7

ARTICLE IN PRESS

MCA normal CVR using BOLD fMRI GAS factor: each participant was exposed either to a gas mixture of CO2 (7%) and O2 (93%) (n = 77), or a gas mixture of CO2 (7%), O2 (21%), and N2 (72%) (n = 23). ADMINISTRATION factor: each participant was exposed to the gas mixture through a non-rebreathing face mask (n = 47), or nasal cannula (n = 53).

Magnetic resonance imaging All participants had both anatomical and CVR imaging. CVR was estimated using the BOLD contrast. MR factor: each participant was scanned using either 1.5 T MR (n = 43), or 3 T MR scanner (n = 57), at Grenoble Hospital and IRMaGe MRI facility (Grenoble, France). At 1.5 T, exams were performed on a whole-body Achieva MR scanner (Philips Healthcare® ) with a 8-channel receiveonly head coil. CVR imaging was obtained using 240 volumes of T2*-weighted image gradient echo (GE) single shot echoplanar images (SSH-EPI) (TR/TE/␣ = 3000/50 ms/90◦ , 32 axial slices covering the brain from the vertex to lower parts of the cerebellum, no gap, FOV = 256 × 256 mm2 , voxel-size 4 × 4 × 4 mm3 ). An anatomical 3D T1-weighted volume covering the whole brain (TR/TE/TI/␣ = 2100/4.1/1100 ms/15◦ , FOV = 256 × 256 × 130 mm3 , voxel-size 1 × 1 × 1.3 mm3 ) was obtained for each participant. At 3 T, exams were performed on a whole-body MedSpec S300 (Bruker® ) MR scanner with a quadrature transmit-receive head coil. CVR imaging was obtained using 240 volumes of T2*-weighted GE SSHEPI (TR/TE/␣ = 3000/30 ms/77◦ , FOV = 216 × 216 mm2 , 40 axial slices covering whole-brain, no gap, voxel-size 3 × 3 × 3.2 mm3 ). An anatomical 3D T1-weighted volume covering the whole brain (TR/TE/TI/␣ = 2500/3.89/900 ms/ 8◦ , FOV = 256 × 224 × 176 mm3 , voxel-size 1.33 × 1.75 × 1.75 mm3 ) was obtained for each participant.

Data processing Image processing was conducted using in-house Matlab® programs and SPM8 (Wellcome Department of Cognitive Neurology, UK). For each participant, the gray matter was segmented from anatomical images, and further normalized to the Montreal Neurologic Institute template. Regions of interest (ROI) of the MCA territories were obtained using a set of canonical ROIs drawn using the MRIcron software (www.mricron.com), according to the cerebral arterial vascular distribution [30]. Each ROI was downscaled to the spatial resolution of the post-processed BOLD images and smoothed.fMRI data preprocessing consisted in co-registration, motion correction (realignment), normalization to the Montreal Neurologic Institute template, resampling with 2 × 2 × 2 mm resolution, and spatial smoothing with a 6-mm Gaussian kernel. Statistical parametric maps were calculated using a dedicated regressor of the response to hypercapnia previously modeled over 20 healthy subjects [7]. CVR was extracted from effect amplitude (beta) maps, estimating the %BOLD signal change. For each participant, a laterality index was calculated on MCA ROIs according to the following formula: LI = (CVRleft — CVRright )/(CVRleft + CVRright ).

3

Statistical analyses Statistical analyses were conducted using the SPSS software (SPSS® , version 20). We performed factorial analyses of variance (ANOVA) to test the main effects and interactions of all experimental factors (SEX, GAS, ADMINISTRATION, MR) on CVR BOLD and LI values. The age was included as a covariable as it is independent from the other factors. We also conducted mixed ANOVA including SIDE as within-subjects factor, to test the effect of the experimental factors on both brain hemispheres. Where main effects or interactions across factors were significant, we performed post-hoc paired Student’s t-tests, using Bonferroni correction, to assess significant differences across factors’ levels. Statistical significance was set at p < 0.05.

Results The participants did not report any adverse effects of carbon dioxide inhalation such as headache, anxiety, or panic attacks.

CVR BOLD signal change In the MCA territories, mean CVR BOLD signal changes were 2.68 ± 0.77% and 2.58 ± 0.71% in the left and right MCA, respectively. The overall ANOVA conducted on all factors identified a significant effect of the ADMINISTRATION means only (p < 0.001). Higher values were detected using mask (3.15 ± 0.81% and 3.00 ± 0.72% in the left and right MCA, respectively) than nasal cannula (2.25 ± 0.40% and 2.21 ± 0.43% in the left and right MCA, respectively). No interaction between ADMINISTRATION and other factors was detected. We also identified a significant effect of age as covariable (p < 0.01), confirming CVR decrease with age. Furthermore, we conducted mixed ANOVA for each between-subject factor, SIDE as within-subject factor, and age as co-factor. We identified an effect of MRI acquisition (p < 0.001). CVR values were higher when using 1.5 T acquisition (3.01 ± 0.77% and 2.91 ± 0.65% in the left and right MCA, respectively) than when using 3 T acquisition (2.42 ± 0.68% and 2.33 ± 0.65% in the left and right MCA, respectively). No significant effects of GAS (p = 0.52) or SEX (p = 0.40) were found. No interaction with SIDE was detected for these factors. In all these ANOVAs, a SIDE effect was significant. A paired T-test on SIDE revealed a left predominance (2.68 ± 0.77% vs 2.58 ± 0.71%) (p < 0.001). Fig. 1 shows an illustrative CVR map obtained within different experimental conditions. The hemispheric asymmetry was hardly visible on CVR maps. The amplitudes of %BOLD signal change were different within the two participants. For participant (A), BOLD signal change was 3.64% and 3.47% in the left and right MCA territories, respectively. And for participant (B), it was 4.52% and 4.28% in the left and right MCA territories, respectively. However, BOLD CVR differences disappeared when using the LI approach as we

Please cite this article in press as: Boudiaf N, et al. BOLD fMRI of cerebrovascular reactivity in the middle cerebral artery territory: A 100 volunteers’ study. J Neuroradiol (2015), http://dx.doi.org/10.1016/j.neurad.2015.04.004

+Model NEURAD-516; No. of Pages 7

ARTICLE IN PRESS

4

N. Boudiaf et al.

Figure 1 CVR maps obtained using BOLD fMRI during hypercapnic challenge. A. CVR map of a 58-year-old male using a hypercapnic gas mixture of 7% CO2 /93%O2 , a high-concentration mask, and the 1.5-T acquisition protocol. BOLD signal changes were 3.64% and 3.47% in the left and right MCA territories, respectively, giving a LI of 0.02. B. CVR map of a 21-year-old male using a hypercapnic gas mixture of 7% CO2 /21%O2 /72%N2 , a high-concentration mask, and the 3-T acquisition protocol. BOLD signal changes were 4.52% and 4.28% in the left and right MCA territories, giving a LI of 0.03.

Table 1 Summary of mean cerebrovascular reactivity (CVR) laterality indexes and standard deviation in middle cerebral artery (MCA) territories for each experimental condition. Experimental conditions

Mean LI ± SDa

Mask Cannula 1.5 T imager 3 T imager Gas CO2 (7%) + O2 (21%) + N2 (72%) Gas CO2 (7%) + O2 (93%) Female Male

0.02 0.01 0.02 0.02 0.02 0.01 0.02 0.01

± ± ± ± ± ± ± ±

0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03

a

Figure 2 Distribution of the laterality index (LI) values computed from left and right MCA territories. The lines on the histogram shows LI = 0; mean LI= 0.016 ± 0.031 giving the lower and upper limits of 95% of this population distribution at −0.05 and 0.08, respectively. The mean LI value is slightly lateralized to the left hemisphere in these right-handed volunteers.

Mean LI is the mean of laterality index values obtained for all participants within each experimental condition. SD is the standard deviation calculated for each mean laterality index.

values and standard deviations related to each experimental condition are reported in Table 1.

Discussion found similar values for both participants (0.02 and 0.03 for participants A and B, respectively).

CVR laterality indices The mean CVR LI was 0.016 ± 0.031, confirming a slight left predominance in these right-handed volunteers. CVR ranged from −0.07 to 0.10, with 95% of our sample falling in an interval of]−0.05; +0.08[(Fig. 2). No significant effect of age, SEX, MR, GAS, nor ADMINISTRATION on LI was detected by the ANOVA (p > 0.05). LI

Our study aimed at providing measurements of CVR to a hypercapnic stimulus using BOLD fMRI in controls without arterial stenosis. We obtained mean CVR BOLD values similar to those reported in the literature [8,31]. We also aimed at providing a semi-quantitative CVR measurement using laterality index in the MCA territories. We investigated the effects of different settings to estimate CVR. Here, we showed that LI values were independent of the MR acquisition, the gas mixture and the administration means used to elicit CVR to hypercapnia. Mean CVR LI reflected a slight left hemispheric predominance in these

Please cite this article in press as: Boudiaf N, et al. BOLD fMRI of cerebrovascular reactivity in the middle cerebral artery territory: A 100 volunteers’ study. J Neuroradiol (2015), http://dx.doi.org/10.1016/j.neurad.2015.04.004

+Model NEURAD-516; No. of Pages 7

ARTICLE IN PRESS

MCA normal CVR using BOLD fMRI right-handed participants, 95% of LI values were included in an interval of]−0.05; +0.08[. No adverse effects of the hypercapnic challenge were detected, confirming that BOLD fMRI of CVR to hypercapnia is a safe method, which is feasible in clinical practice [4,7,8,20,26,29,32—34]. These results are expected to facilitate the detection of patients at risk of hemodynamical stroke events downstream a severe MCA stenosis. We observed that CVR BOLD values were dependent on age and experimental conditions, contrary to CVR LI values. Thus, CVR measurement using the LI approach is suitable to overcome individual and methodological settings variability. BOLD amplitude was about 20% higher at 1.5 T than at 3 T. This result may be surprising when considering the positive relationship between the amplitude of BOLD signal changes and the MR strength [35]. However, this gain may partially be due to the voxel size [36], that was more than twice larger at 1.5 T (64 mm3 ) than at 3 T (28.8 mm3 ), which might explain an increase of the signal to noise ratio. The effects of MR strength and voxel size on BOLD signal illustrated the limitation of expressing CVR as %BOLD signal change when data acquisition is performed using different MR settings. From a clinical point of view, the independence of LI from the experimental settings is interesting to characterize unilateral disease and to compare CVR all over MR centers. CVR BOLD values were higher when using a mask than when using nasal cannula, because the higher airflow and sealing of the mask lead to a higher fraction of inspired CO2 . Consequently, high-concentration masks allow a higher increase of end-tidal CO2 pressure (PetCO2 ) leading to a more significant vasodilation and higher increase in cerebral blood flow [4,8,32,33]. Thus, high-concentration masks should be recommended to elicit CVR to hypercapnia rather than nasal cannula. Masks also allow recording PetCO2 simultaneously, which can be used as a physiological regressor in order to express CVR in %BOLD/mmHg [1,4,12,16]. Furthermore, it helps correcting any variability in the stimulus administration and the subject’s respiratory response. However, masks might be difficult to be used with some subjects because of the limited space within the head coil, and that might lead to unbearable discomfort. In such cases, nasal cannula could be an alternative choice. The independence of LI values of the administration method is an additional advantage to face variability through subjects and settings. Different gas mixtures have previously been used to elicit hypercapnic vasodilation. CO2 fractions usually range from 5 to 10%. Normoxic hypercapnia or hyperoxic hypercapnia can be obtained when CO2 is mixed with air (21% O2 ) or oxygen (93% O2 ). Normoxic hypercapnia provides BOLD signal changes that are related to the vasodilation because changes in oxygen extraction fraction are negligible [37]. Hyperoxic hypercapnia elicits more complex effects on the BOLD signal. First, the vasodilation due to the hypercapnic CVR is slightly counterbalanced by the vasoconstriction due to the hyperoxic CVR. Moreover, it has been suggested that BOLD signal changes during hypercapnia would reveal changes in the venous components whereas hyperoxia would rather reveal changes in the arterial components of the vascular tree [37]. Second, hyperoxia provides freely diffusible oxygen that reduces oxygen extraction fraction (OEF)

5 and thus the deoxyhemoglobin concentration [38,39]. Taken together, both effects would lead to BOLD signal increase. However, the absence of GAS factor effect on BOLD signal changes suggests two possibilities: either the BOLD signal increase related to the decreased OEF might have been compensated by the hyperxoxic CVR, or the effects of hyperoxia were negligible when compared to the hypercapnic CVR. Because of the more complex and equivocal nature of BOLD signal changes during hyperoxic hypercapnia, it is advocated to perform BOLD CVR mapping using normoxic hypercapnia [40]. However, some ethical committees could be reluctant to allow normoxic hypercapnic challenge rather than hyperoxic hypercapnic challenge, especially in patients referred for stroke and steno-occlusive disease. To avoid further difficulties in the interpretation of BOLD signal change related to gas tension, advances in arterial spin labeling (ASL) imaging are promising because signal changes allow the quantification of cerebral blood flow (CBF). However, ASL CVR mapping remains challenging because this technique is better performed using a 3-T MR scanner. Changes in vessels diameters, CBF, and arterial transit times should be taken into account to adjust inversion time. And last but not least, in patients referred for unilateral arterial stenosis who need vascular reserve assessment, regional changes of arterial transit time downstream of the stenosis may bias CBF quantification. Thus, it may require the use of specific ASL methods sampling the signal at multiple post-label delays [41]. As expected, when taking into account all the experimental factors, the age of participants had a significant effect on the BOLD response. Indeed, as age decreases the vascular dilation capacities, we observed higher BOLD variations in younger adults when compared to older adults in line with previous work. However, this does not mean that older adults have systematically altered CVR [8,27,28,42]. The existence of a physiological asymmetry in CVR has already been shown within 37 controls free from stenosis, using hypercapnic challenge and transcranial Doppler sonography monitoring [43]. However, we did not find literature reporting a left CVR predominance in right-handed subjects. We could make the assumption that this finding would be related to anatomo-functional hemispheric specialisation [44,45]. This hypothesis needs to be tested in both rightand left-handed subjects by combining CVR, motor and language mapping, while paying more attention to variability of the circle of Willis among subjects. In our study, we used a dedicated model regressor [7] because it was impossible for us to measure the PetCO2 for all participants, especially for those who received gas via nasal cannula. The weakness of this approach is to assume individuals’ responses to hypercapnic challenge as identical regardless of the experimental conditions. Furthermore, gas delivery was manually performed. Despite the attention paid for stimuli errors, this may have induced some variations among subjects. In all cases, automated means of gas administration remain preferable. Another point, which may be criticized in our study, is the constitution of our groups. As this study is retrospective we could not have equal numbers of participants in each group through the different factor levels. Nonetheless, independence, normality and homoscedasticity assumptions were verified before conducting the ANOVA.

Please cite this article in press as: Boudiaf N, et al. BOLD fMRI of cerebrovascular reactivity in the middle cerebral artery territory: A 100 volunteers’ study. J Neuroradiol (2015), http://dx.doi.org/10.1016/j.neurad.2015.04.004

+Model NEURAD-516; No. of Pages 7

ARTICLE IN PRESS

6

N. Boudiaf et al.

However, despite some effects of the experimental conditions on BOLD signal changes, the LI approach is robust when considering interhemispheric comparisons. Although BOLD fMRI to hypercapnia is a safe and reliable approach to map CVR, the absolute quantification of CVR remains challenging because of the complex nature of BOLD signal change, which is dependent on individual and technical variability.

Conclusion In conclusion, we provide a range of LI values in volunteers without steno-occlusive disease independent of different settings. This semi-quantitative approach could be used further in clinical practice, to better identify patients with impaired vascular reserve among those referred for symptomatic unilateral carotid or MCA stenosis. Because such patients are expected to be at higher risk of ischemic events, CVR mapping could be helpful to justify the risk of morbidity and mortality related to invasive treatments such as percutaneous transluminal angioplasty and stenting.

Disclosure of interest The authors declare that they have no conflicts of interest concerning this article.

Acknowledgments We acknowledge the support of the IRMaGe MRI facility for data acquisition, and the precious help of Patrice Jousse for artwork. IRMaGe MRI/Neurophysiology facility was partly funded by the French program ‘‘Investissement d’Avenir’’ run by the ‘‘Agence Nationale pour la Recherche’’; grant ‘‘Infrastructure d’avenir en Biologie Santé’’ - ANR-11-INBS0006.

References [1] Yezhuvath US, Lewis-Amezcua K, Varghese R, Xiao G, Lu H. On the assessment of cerebrovascular reactivity using hypercapnia BOLD MRI. NMR Biomed 2009;22(7):779—86. [2] Fierstra J, Sobczyk O, Battisti-Charbonney A, et al. Measuring cerebrovascular reactivity: what stimulus to use? J Physiol 2013;591(23):5809—21. [3] Ben Bashat D, Artzi M, Ben Ami H, et al. Hemodynamic response imaging: a potential tool for the assessment of angiogenesis in brain tumors. PLoS One 2012;7(11):e49416. [4] Cantin S, Villien M, Moreaud O, et al. Impaired cerebral vasoreactivity to CO2 in Alzheimer’s disease using BOLD fMRI. Neuroimage 2011;58(2):579—87. [5] Zacà D, Jovicich J, Nadar SR, Voyvodic JT, Pillai JJ. Cerebrovascular reactivity mapping in patients with low grade gliomas undergoing presurgical sensorimotor mapping with BOLD fMRI. J Magn Reson Imaging 2013;40(2):383—90. [6] Fromm A, Lundervold AJ, Moen G, Skulstad S, Thomassen L. A vascular approach to mild amnestic cognitive impairment: a pilot study. Acta Neurol Scand 2013;127:73—6. [7] Jiang Z, Krainik A, David O, et al. Impaired fMRI activation in patients with primary brain tumors. Neuroimage 2010;52(2):538—48.

[8] Krainik A, Villien M, Troprès I, et al. Functional imaging of cerebral perfusion. Diagn Interv Imaging 2013;94(12): 1259—78. [9] Han JS, Mikulis DJ, Mardimae A, et al. Measurement of cerebrovascular reactivity in pediatric patients with cerebral vasculopathy using blood oxygen level-dependent MRI. Stroke 2011;42(5):1261—9. [10] Heyn C, Poublanc J, Crawley A, et al. Quantification of cerebrovascular reactivity by blood oxygen level—dependent MR imaging and correlation with conventional angiography in patients with Moyamoya disease. Am J Neuroradiol 2010;31(5):862—7. [11] Krainik A, Hund-Georgiadis M, Zysset S, Cramon von DY. Regional impairment of cerebrovascular reactivity and BOLD signal in adults after stroke. Stroke 2005;36(6): 1146—52. [12] Mandell DM, Han JS, Poublanc J, et al. Quantitative measurement of cerebrovascular reactivity by blood oxygen level-dependent MR imaging in patients with intracranial stenosis: preoperative cerebrovascular reactivity predicts the effect of extracranial-intracranial bypass surgery. Am J Neuroradiol 2011;32(4):721—7. [13] Mazighi M, Tanasescu R, Ducrocq X, et al. Prospective study of symptomatic atherothrombotic intracranial stenoses the GESICA study. Neurology 2006;66(8):1187—91. [14] Derdeyn CP, Chimowitz MI, Lynn MJ, et al. Aggressive medical treatment with or without stenting in high-risk patients with intracranial artery stenosis (SAMMPRIS): the final results of a randomised trial. Lancet 2013;383:333—41. [15] Chang T-Y, Kuan W-C, Huang K-L, et al. Heterogeneous cerebral vasoreactivity dynamics in patients with carotid stenosis. PLoS One 2013;8(9):e76072. [16] Mandell DM, Han JS, Poublanc J, et al. Mapping cerebrovascular reactivity using blood oxygen level-dependent MRI in patients with arterial steno-occlusive disease comparison with arterial spin labeling MRI. Stroke 2008;39(7):2021—8. [17] Webster MW, Makaroun MS, Steed DL, Smith HA, Johnson DW, Yonas H. Compromised cerebral blood flow reactivity is a predictor of stroke in patients with symptomatic carotid artery occlusive disease. J Vasc Surg 1995;21(2):338—45. [18] Bouvier J, Detante O, Tahon F, et al. Reduced CMRO2 and cerebrovascular reserve in patients with severe intracranial arterial stenosis: A combined multiparametric qBOLD oxygenation and BOLD fMRI study. Hum Brain Mapp 2015;36(2): 695—706. [19] Conklin J, Fierstra J, Crawley AP, et al. Mapping white matter diffusion and cerebrovascular reactivity in carotid occlusive disease. Neurology 2011;77(5):431—8. [20] Attyé A, Villien M, Tahon F, Warnking J, Detante O, Krainik A. Normalization of cerebral vasoreactivity using BOLD MRI after intravascular stenting. Hum Brain Mapp 2014;35(4): 1320—4. [21] Haller S, Bonati LH, Rick J, et al. Reduced cerebrovascular reserve at CO2 BOLD MR imaging is associated with increased risk of periinterventional ischemic lesions during carotid endarterectomy or stent placement: preliminary results1. Radiology 2008;249(1):251—8. [22] Mutch WAC, Mandell DM, Fisher JA, et al. Approaches to brain stress testing: BOLD magnetic resonance imaging with computer-controlled delivery of carbon dioxide. PLoS One 2012;7(11):e47443. [23] Attyé A, Boncoeur-Martel M-P, Maubon A, et al. Volumétrie de l’ischémie cérébrale en IRM de diffusion et déficit fonctionnel secondaire après thrombolyse intraveineuse. J Neuroradiol 2012;39(2):97—103. [24] Lövblad K-O, Altrichter S, Mendes Pereira V, et al. Imaging of acute stroke: CT and/or MRI. J Neuroradiol 2015;42(1): 55—64.

Please cite this article in press as: Boudiaf N, et al. BOLD fMRI of cerebrovascular reactivity in the middle cerebral artery territory: A 100 volunteers’ study. J Neuroradiol (2015), http://dx.doi.org/10.1016/j.neurad.2015.04.004

+Model NEURAD-516; No. of Pages 7

ARTICLE IN PRESS

MCA normal CVR using BOLD fMRI [25] Souillard-Scemama R, Tisserand M, Calvet D, et al. An update on brain imaging in transient ischemic attack. J Neuroradiol 2015;42(1):3—11. [26] Spano VR, Mandell DM, Poublanc J, et al. CO2 blood oxygen level-dependent MR mapping of cerebrovascular reserve in a clinical population: safety, tolerability, and technical feasibility. Radiology 2013;266(2):592—8. [27] D’Esposito M, Deouell LY, Gazzaley A. Alterations in the BOLD fMRI signal with ageing and disease: a challenge for neuroimaging. Nat Rev Neurosci 2003;4(11):863—72. [28] Riecker A, Grodd W, Klose U, et al. Relation between regional functional MRI activation and vascular reactivity to carbon dioxide during normal aging. J Cereb Blood Flow Metab 2003;23(5):565—73. [29] Krainik A, Maillet A, Fleury V, et al. Levodopa does not change cerebral vasoreactivity in Parkinson’s disease. Mov Disord 2013;28(4):469—75. [30] Tatu L, Moulin T, Vuillier F, Bogousslavsky J. Arterial territories of the human brain. In: Paciaroni M, Agnelli G, Caso V, Bogousslavsky J, editors. Frontiers of neurology and neuroscience, Vol 30. Basel: KARGER; 2012. p. 99—110. [31] Leoni RF, Mazzetto-Betti KC, Silva AC, et al. Assessing cerebrovascular reactivity in carotid steno-occlusive disease using MRI BOLD and ASL techniques. Radiol Res Pract 2012; 2012. [32] Goode SD, Krishan S, Alexakis C, Mahajan R, Auer DP. Precision of cerebrovascular reactivity assessment with use of different quantification methods for hypercapnia functional MR imaging. Am J Neuroradiol 2009;30(5):972—7. [33] Kassner A, Winter JD, Poublanc J, Mikulis DJ, Crawley AP. Blood-oxygen level dependent MRI measures of cerebrovascular reactivity using a controlled respiratory challenge: reproducibility and gender differences. J Magn Reson Imaging 2010;31(2):298—304. [34] Leontiev O, Buxton RB. Reproducibility of BOLD, perfusion, and CMRO2 measurements with calibrated-BOLD fMRI. Neuroimage 2007;35(1):175—84.

7 [35] Van der Zwaag W, Francis S, Head K, et al. fMRI at 1.5, 3 and 7 T: characterising BOLD signal changes. Neuroimage 2009;47(4):1425—34. [36] Gati JS, Menon RS, Uurbil K, Rutt BK. Experimental determination of the BOLD field strength dependence in vessels and tissue. Magn Reson Med 1997;38(2):296—302. [37] Schwarzbauer C, Deichmann R. Vascular component analysis of hyperoxic and hypercapnic BOLD contrast. Neuroimage 2012;59(3):2401—12. [38] Bulte DP, Chiarelli PA, Wise RG, Jezzard P. Cerebral perfusion response to hyperoxia. J Cereb Blood Flow Metab 2006;27(1):69—75. [39] Chiarelli PA, Bulte DP, Wise R, Gallichan D, Jezzard P. A calibration method for quantitative BOLD fMRI based on hyperoxia. Neuroimage 2007;37(3):808—20. [40] Hare HV, Germuska M, Kelly ME, Bulte DP. Comparison of CO2 in air versus carbogen for the measurement of cerebrovascular reactivity with magnetic resonance imaging. J Cereb Blood Flow Metab 2013;33(11):1799—805. [41] Alsop DC, Detre JA, Golay X, et al. Recommended implementation of arterial spin-labeled perfusion MRI for clinical applications: a consensus of the ISMRM perfusion study group and the European consortium for ASL in dementia. Magn Reson Med 2015;73(1):102—16. [42] Thomas BP, Liu P, Park DC, van Osch MJP, Lu H. Cerebrovascular reactivity in the brain white matter: magnitude, temporal characteristics, and age effects. J Cereb Blood Flow Metab 2014;34(2):242—7. [43] Fürst H, Hartl WH, Janssen I. Patterns of cerebrovascular reactivity in patients with unilateral asymptomatic carotid artery stenosis. Stroke 1994;25(6):1193—200. [44] Dassonville P, Zhu X-H, Ugurbil K, Kim S-G, Ashe J. Functional activation in motor cortex reflects the direction and the degree of handedness. Proc Natl Acad Sci 1997;94(25):14015—8. [45] Van Essen DC. A population-average, landmark- and surfacebased (PALS) atlas of human cerebral cortex. Neuroimage 2005;28(3):635—62.

Please cite this article in press as: Boudiaf N, et al. BOLD fMRI of cerebrovascular reactivity in the middle cerebral artery territory: A 100 volunteers’ study. J Neuroradiol (2015), http://dx.doi.org/10.1016/j.neurad.2015.04.004