MR perfusion studies with T1-weighted echo planar imaging

Implementation of quantitative FAIR perfusion imaging with a short repetition time in time-course studies

Flow-sensitive alternating inversion recovery (FAIR) is a pulsed arterial spin labeling magnetic resonance imaging method for perfusion quantification. In its standard implementation for quantification with full longitudinal relaxation between acquisitions, its use in time-course investigations of rapidly changing flow values is limited. The time efficiency can be improved by decreasing the repetition time but quantification becomes problematic. This situation is further complicated if a whole-body radiofrequency transmit coil is not used since fresh blood spins will flow in from outside the coil. To alleviate these problems, the use of global pre-saturation is proposed. The resulting expression for the flow signal depends on the relationship between the imaging parameters and the coil inflow time and can be significantly simplified under certain combinations of these parameters. With this implementation of FAIR, quantitative flow maps of gerbil brains were obtained with a 3 minute time resolution in a study of the effects of reperfusion. The pre-occlusion flow measurements were in good agreement with values obtained by the standard FAIR implementation and by other techniques, but the low values following occlusion were underestimated due to the increased transit times.

Parameter estimation from Rician-distributed data sets using a maximum likelihood estimator: application to T1 and perfusion measurements

General expressions are presented to calculate the maximum likelihood (ML) estimator and corresponding Fisher matrix for Rician-distributed data sets. This estimator results in the most precise, unbiased estimations of T1 from magnitude data sets, even when low signal-to-noise ratios (<6) are present. By optimizing the sample point distributions for inversion-recovery experiments, a 32% increase in precision of the estimated T1 is obtained, compared with a linear sampling scheme. Perfusion rates are estimated from combined data sets of the slice- and nonslice-selective inversion-recovery experiments, as obtained with the flow-sensitive alternating inversion recovery (FAIR) technique. The ML estimator for the combined data set results in the most precise, unbiased estimations of the perfusion rate. Error analysis shows that very high signal-to-noise ratios are required for precise estimation of perfusion rates from FAIR experiments.

Transfer insensitive labeling technique (TILT): application to multislice functional perfusion imaging

Cerebral blood flow can be studied in a multislice mode with a recently proposed perfusion sequence using inversion of water spins as an endogenous tracer without magnetization transfer artifacts. The magnetization transfer insensitive labeling technique (TILT) has been used for mapping blood flow changes at a microvascular level under motor activation in a multislice mode. In TILT, perfusion mapping is achieved by subtraction of a perfusion-sensitized image from a control image. Perfusion weighting is accomplished by proximal blood labeling using two 90 degrees radiofrequency excitation pulses. For control preparation the labeling pulses are modified such that they have no net effect on blood water magnetization. The percentage of blood flow change, as well as its spatial extent, has been studied in single and multislice modes with varying delays between labeling and imaging. The average perfusion signal change due to activation was 36.9 +/- 9.1% in the single-slice experiments and 38.1 +/- 7.9% in the multislice experiments. The volume of activated brain areas amounted to 1.51 +/- 0.95 cm3 in the contralateral primary motor (M1) area, 0.90 +/- 0.72 cc in the ipsilateral M1 area, 1.27 +/- 0.39 cm3 in the contralateral and 1.42 +/- 0.75 cm3 in the ipsilateral premotor areas, and 0.71 +/- 0.19 cm3 in the supplementary motor area.

Velocity selective radiofrequency pulse trains

The incorporation of velocity-encoding gradient pulses in RF-pulse trains is proposed and examined. Velocity selective perturbation is shown to be analogous in many respects to the well established use of trains of short RF-pulses for chemical shift selective perturbation. Velocity selective perturbation is viable in a biomedical setting only if additional RF refocusing pulses are inserted between the individual RF-pulse elements. Aspects of velocity selective excitation saturation and inversion are examined, and new inversion pulse trains proposed. The selective perturbation of both flowing and stationary spins is demonstrated in phantoms and possible biomedical applications of these pulse trains are discussed.

Measurement of human myocardial perfusion by double-gated flow alternating inversion recovery EPI

This paper presents a flow-sensitive alternating inversion recovery (FAIR) method for measuring human myocardial perfusion at 1.5 T. Slice-selective/non-selective IR images were collected using a double-gated IR echoplanar imaging sequence. Myocardial perfusion was calculated after T1 fitting and extrapolation of the mean signal difference SI(Sel - SI(NSel). The accuracy of the method was tested in a porcine model using graded intravenous adenosine dose challenge. Comparison with radiolabeled microsphere measurements showed a good correlation (r = 0.84; mean error = 20%, n = 6) over the range of flows tested (0.9-7 ml/g/min). Applied in humans, this method allowed for the measurement of resting myocardial flow (1.04+/-0.37 ml/g/min, n = 11). The noise in our human measurements (SE(flow) = 0.2 ml/g/min) appears to come primarily from residual respiratory motion. Although the current signal-to-noise ratio limits our ability to measure small fluctuations in resting flow accurately, the results indicate that this noninvasive method has great promise for the quantitative assessment of myocardial flow reserve in humans.

Effect of a gadodiamide contrast agent on the reliability of brain tissue T1 measurements

To determine whether brain spin-lattice relaxation time (T1) can routinely be measured after contrast-agent injection, we measured T1 by a precise and accurate inversion-recovery (PAIR) method in five brain tumor patients, before and again after contrast-agent injection. The T1 in at least 20 regions of interest (ROIs) was measured in each patient, avoiding areas of contrast enhancement visible by conventional MR imaging. Contrast-agent injection reduced T1 in 51 regions of interest in white matter by less than 1% (not significant), and in 50 regions of interest in gray matter by less than 2% (p = 0.001). Pixel-by-pixel plots demonstrate that T1 is reduced substantially in extra-parenchymal tissues, but not in brain tissues. Therefore, T1 mapping with the precise and accurate inversion-recovery method can routinely be done after contrast injection. Our results suggest that the precise and accurate inversion-recovery method is not sensitive to the T1 of blood in the presence of an intact blood-brain barrier, although a substantial T1 reduction does occur in the absence of a blood-brain barrier.

Perfusion analysis using dynamic arterial spin labeling (DASL)

A variety of magnetic resonance (MR) techniques have proved useful to quantify perfusion using endogenous water as a blood flow tracer. Assuming that water is a freely diffusable tracer, the model used for these techniques predicts that the quantitation of perfusion is based on three parameters, all of which can depend on blood flow. These are the longitudinal tissue relaxation time, the transit time from point of labeling to tissue, and the difference in tissue MR signal between an appropriate control and the labeled state. To measure these three parameters in parallel, a dynamic arterial spin labeling (DASL) technique is introduced based on the analysis of the tissue response to a periodic time varying degree of arterial spin labeling, called here the labeling function (LF). The LF frequency can be modulated to overdetermine parameters necessary to define the system. MR schemes are proposed to measure the tissue response to different LF frequencies efficiently. Sprague-Dawley rats were studied by DASL, using various frequencies for the LF and various arterial pCO2 levels. During data processing, the periodic behavior of the tissue response to the LF allowed for frequency filtering of periodic changes in signal intensity unrelated to perfusion and arterial spin labeling. Measures of transit time, tissue longitudinal relaxation time, and perfusion agreed well over a range of LF frequencies and with previous results. DASL shows potential for more accurately quantifying perfusion as well as measuring transit times associated with arterial spin labeling techniques.

Activation detection in functional MRI using subspace modeling and maximum likelihood estimation

A statistical method for detecting activated pixels in functional MRI (fMIRI) data is presented. In this method, the fMRI time series measured at each pixel is modeled as the sum of a response signal which arises due to the experimentally controlled activation-baseline pattern, a nuisance component representing effects of no interest, and Gaussian white noise. For periodic activation-baseline patterns, the response signal is modeled by a truncated Fourier series with a known fundamental frequency but unknown Fourier coefficients. The nuisance subspace is assumed to be unknown. A maximum likelihood estimate is derived for the component of the nuisance subspace which is orthogonal to the response signal subspace. An estimate for the order of the nuisance subspace is obtained from an information theoretic criterion. A statistical test is derived and shown to be the uniformly most powerful (UMP) test invariant to a group of transformations which are natural to the hypothesis testing problem. The maximal invariant statistic used in this test has an F distribution. The theoretical F distribution under the null hypothesis strongly concurred with the experimental frequency distribution obtained by performing null experiments in which the subjects did not perform any activation task. Application of the theory to motor activation and visual stimulation fMRI studies is presented.

Rapid and continuous monitoring of cerebral perfusion by magnetic resonance line scan assessment with arterial spin tagging

A new approach is presented for rapid and continuous monitoring of cerebral perfusion which is based upon line-scan MR column imaging with arterial spin tagging (AST) of endogenous water. Spin tagging of arterial water protons is accomplished using adiabatic fast passage inversion, followed by acquisition of the perfusion sensitive MR signal from a column placed at the desired level through the brain using line scan localization techniques. A perfusion sensitive line scan is followed by a non-perfusion sensitive line scan, and perfusion is calculated pixel-by-pixel from the intensity difference of the two lines. Continuous perfusion measurements are reported with temporal resolution of 10 s in pixels of volume 0.027 cm3 or less. Examples of the methodology are given during hypercapnic challenge induced with carbon dioxide, and during an ischemic event induced by reversible middle cerebral artery occlusion. The method is also used to characterize the signal response as a function of arterial inversion time and post inversion acquisition delay. These methods permit rapid and continuous monitoring of cerebral perfusion with high spatial resolution, and can be interleaved with MR measurements of diffusion and T1 to follow the progression of cerebral events during physiological or pharmacological intervention.

Correction for eddy current induced Bo shifts in diffusion-weighted echo-planar imaging

The importance of diffusion-weighted MRI in the assessment of acute stroke is well-recognized, and quantitative maps of the apparent diffusion coefficient (ADC) are now widely used. Echoplanar imaging provides a robust method of acquiring diffusion-weighted images free of motion artifact. However, initial experience with clinical MRI systems indicates that calculation of artifact-free ADC maps from a series of echo-planar diffusion-weighted images is not necessarily straight-forward. One of the problems is that frequency shifts resulting from eddy currents can cause misregistration of base diffusion-weighted images. In this study, an on-line correction method that overcomes this problem is described, and phantom and human images that demonstrate the validity of the technique are presented. The method uses a non-phase-encoded reference scan to correct the phase of each echo in the echo train, and can provide ADC maps that are free of misregistration artifacts, without the need for off-line postprocessing.

Extraslice spin tagging (EST) magnetic resonance imaging for the determination of perfusion

A new magnetic resonance technique to measure perfusion is described in detail. The means by which this is done is to invert all the spins in the radiofrequency RF coil with a non-spatially selective pulse and immediately re-invert the spins in the imaging plane. The net effect is that the spins in the imaging plane experience minimal perturbation of their magnetization while the spins outside the plane (extraslice) are inverted, or tagged. Tagged spins that flow into the imaging plane before image data are acquired decrease the signal intensity in the imaging plane when compared with an image in which the inflowing spins are not tagged. This decrease in signal can be used to calculate the number of spins that have flowed into the imaging plane, i.e., can be used to calculate the perfusion in mL x 100 g(tissue)(- 1)x min(-1). The extraslice spin tagging (EST) magnetization preparation period was coupled with a fast imaging sequence to obtain perfusion maps for normal volunteers.

Using flow relaxography to elucidate flow relaxivity

We have investigated the theoretical and experimental linear dependence of the reciprocal of the apparent longitudinal relaxation time [(T*1)-1] of the NMR signal from spins in a flowing fluid on the volume flow rate, Fv, the so-called inflow effect. We refer to the coefficient of this dependence as the longitudinal flow relaxivity, r1F. A very simple model predicts that, under a range of conditions pertinent to modern flow studies and perfusion imaging experiments, r1F is controlled by the volume of the fluid in which the magnetization is perturbed by pulsed RF inversion or saturation, not the detection volume, and that it can be approximated as the reciprocal of half of the inversion volume. Phantom sample experiments, using a new, quantitative approach that we call flow relaxography, confirm the general predictions of this simple model. There are two intriguing implications of these findings for general NMR flow studies as well as for medical applications. It should be possible to vary the value of r1F by simply (noninvasively) adjusting the inversion slice thickness, and thus measure the value of (blood 1H2O, for example) Fv in a vessel without changing Fv, from the resultant varying T*1 values. Also, it should be possible to extrapolate to the intrinsic T1 value of the fluid signal (as if it were stationary), without altering or stopping the flow. Again, these are quite successful in phantom sample studies. Imaging versions of the flow relaxographic experiments are also possible. The twin goals of flow studies in medical MRI are the quantitative discrimination of the signals from flowing and nonflowing spins, and the accurate measurement of the flow rate of the former.

Perfusion imaging using FOCI RF pulses

Pulsed arterial spin-tagging techniques for perfusion measurements (e.g., echo planar MR imaging and signal targeting with alternating radiofrequency (EPISTAR), flow-sensitive alternating inversion recovery (FAIR), quantitative imaging of perfusion using a single subtraction (QUIPPS), uninverted FAIR (UNFAIR)) generally use hyperbolic secant (HS) pulses for spin inversion. The performance of these techniques depends on the inversion efficiency, as well as the sharpness of the slice profiles. Frequency offset corrected inversion (FOCI) pulses, a recently proposed HS variant, can provide slice profiles with edges that can be up to 10 times sharper than those obtained with conventional HS pulses. In this communication, the implementation and application of the C-shape FOCI pulse for perfusion imaging in rat brain with the FAIR technique is summarized. Despite providing a more rectangular slice profile than a conventional HS pulse, it is demonstrated both theoretically and experimentally that the FAIR perfusion signal is not increased by using a FOCI tagging pulse. However, the use of a FOCI inversion pulse is shown to significantly minimize static signal subtraction errors that are common with conventional HS pulses. Finally, the suitability of the pulse for perfusion studies is demonstrated, in vivo, on rat brain.

Simultaneous oxygenation and perfusion imaging study of functional activity in primary visual cortex at different visual stimulation frequency: quantitative correlation between BOLD and CBF changes

A relationship between regional cerebral blood flow (CBF) and blood oxygenation level dependent (BOLD) changes in the primary visual cortex (V1) at varied visual stimulation frequency has been examined quantitatively using the multislice FAIR technique. A linear correlation in the common activation areas between functional BOLD and CBF maps was observed. This supports the hypothesis that the task-stimulated BOLD changes in microvasculature are correlated with the CBF changes that presumably reflect the degree of neuronal activity. The linear correlation coefficients for intrasubject comparisons are more significant than those for intersubject comparisons. This suggests that using intrasubject comparisons for quantitative studies of neuronal activity related to different task stimuli and task performances should be more reliable than using intersubject comparisons.

FAIR excluding radiation damping (FAIRER)

It is shown that the flow-sensitive alternating inversion recovery (FAIR) technique is complicated by the effect of radiation damping, leading to problems in calibrating this method on phantoms and to inaccuracies in measured flows. A modified scheme called FAIRER (FAIR excluding radiation damping) is proposed, which suppresses the damping effects by employing very weak magnetic field gradients (0.06 G/cm) during the inversion recovery, spin-echo, and predelay periods. Results on phantoms and in vivo on cat brain are presented that demonstrate that FAIRER effectively solves these problems.

Continuous assessment of perfusion by tagging including volume and water extraction (CAPTIVE): a steady-state contrast agent technique for measuring blood flow, relative blood volume fraction, and the water extraction fraction

A new technique, CAPTIVE, that is a synthesis of arterial spin labeling (ASL) blood flow and steady-state susceptibility contrast relative blood volume imaging is described. Using a single injection of a novel, long half-life intravascular magnetopharmaceutical with a high tissue:blood susceptibility difference (deltachi) to deltaR1 ratio, changes in tissue transverse relaxivity (deltaR2 or deltaR2*) that arise from changes in blood volume were measured, while preserving the ability to measure blood flow using traditional T1-based ASL techniques. This modification permits the continuous measurement of both blood flow and blood volume. Also, because the contrast agent can be used to remove the signal from intravascular spins, it is possible to measure the first-pass water extraction fraction. Contrast-to-noise is easily traded off with repetition rate, allowing the use of non-EPI scanners and more flexible imaging paradigms. The basic theory of these measurements, several experimental scenarios, and validating results are presented. Specifically, the PaCO2-reactivity of microvascular and total relative cerebral blood volume (rCBV), cerebral blood flow (CBF), and the water extraction-flow product (EF) in rats with the new contrast agent MPEG-PL-DyDTPA is measured, and the values are concordant with those of previous literature. As an example of one possible application, continuous flow and volume measurements during transient focal ischemia are presented. It is believed that CAPTIVE imaging will yield a more complete picture of the hemodynamic state of an organ, and has further application for understanding the origins of the BOLD effect.

In vivo quantitative mapping of cardiac perfusion in rats using a noninvasive MR spin-labeling method

Measurement of myocardial perfusion is important for the functional assessment of heart in vivo. Our approach is based on the modification of the longitudinal relaxation time T1 induced by magnetic spin labeling of endogenous water protons. Labeling is performed by selectively inverting the magnetization within the detection slice, and longitudinal relaxation is measured using a fast gradient echo MRI technique. As a result of blood flow, nonexcited spins enter the detection slice, which leads to an acceleration of the relaxation rate. Incorporating this phenomenon in a mathematical model that describes tissue as two compartments yields a simple expression that allows the quantification of perfusion from a slice-selective and a global inversion recovery experiment. This model takes into account the difference between T1 in blood and T1 in tissue. Our purpose was to evaluate the feasibility and reproducibility of this technique to map quantitatively myocardial perfusion in vivo in rats. Quantitative maps of myocardial blood flow were obtained from nine rats, and the reproducibility of the technique was evaluated by repeating the whole perfusion experiment four times. Evaluation of regions of interest within the myocardium yielded a mean perfusion value of 3.6 +/- .5 ml x min(-1) x g(-1) over all animals, which is in good agreement with previously reported literature values.

Visual activation in infants and young children studied by functional magnetic resonance imaging

The purpose of this study was to determine whether visual stimulation in sleeping infants and young children can be examined by functional magnetic resonance imaging. We studied 17 children, aged 3 d to 48 mo, and three healthy adults. Visual stimulation was performed with 8-Hz flickering light through the sleeping childs' closed eyelids. Functional magnetic resonance imaging was performed with a gradient echoplanar sequence in a l.5-T magnetic resonance scanner. Six subjects were excluded because of movement artifacts; the youngest infant showed no response. In 10 children, we could demonstrate areas of signal decrease during visual stimulation in the occipital cortex (mean decrease 2.21%), contrary to the signal increase observed in the adult controls (mean increase 2.82%). This decrease may be due to a higher proportional increase in oxygen extraction compared with increase in cerebral blood flow during activation. The different response patterns in young children and adults can reflect developmental or behavioral differences. Localization of the activation seemed to be age-dependent. In the older children and the adults, it encompassed the whole length of the calcarine sulcus, whereas it was restricted to the anterior and medial part of the calcarine sulcus in the younger infants. This may reflect a different functional organization of the young child's visual cortex or the on-going retinal development.

A general kinetic model for quantitative perfusion imaging with arterial spin labeling

Recently, several implementations of arterial spin labeling (ASL) techniques have been developed for producing MRI images sensitive to local tissue perfusion. For quantitation of perfusion, both pulsed and continuous labeling methods potentially suffer from a number of systematic errors. In this study, a general kinetic model for the ASL signal is described that can be used to assess these errors. With appropriate assumptions, the general model reduces to models that have been used previously to analyze ASL data, but the general model also provides a way to analyze the errors that result if these assumptions are not accurate. The model was used for an initial assessment of systematic errors due to the effects of variable transit delays from the tagging band to the imaging voxel, the effects of capillary/tissue exchange of water on the relaxation of the tag, and the effects of incomplete water extraction. In preliminary experiments with a human subject, the model provided a good description of pulsed ASL data during a simple sensorimotor activation task.

In vivo perfusion measurements in the human placenta using echo planar imaging at 0.5 T

This paper presents the first in vivo measurements of perfusion in the human placenta from 20 weeks gestational age until term, using the non-selective/selective inversion recovery echo-planar imaging sequence, in which data is alternately acquired following a selective and non-selective inversion pulse. Twenty pairs of images were collected, two each at the following inversion times: 20, 310, 610, 910, 1110, 1410, 1910, 2810, 3310, and 4510 ms with the sequence being repeated with a repetition time (TR) of 10 s. The results of these measurements were used to suggest the optimum sequence for future work in terms of the signal to noise ratio in the measured perfusion rate in a given measurement time. The sequence was also analyzed to determine the expected variability in the measurements. In normal pregnancies the average value of perfusion rate was found to be 176 (standard error = +/-24) ml/100 mg/min. (n = 16, standard deviation = 96 ml/100 mg/min). The expected variability in the measured parameters due to signal to noise ratio considerations alone was calculated to be 71%. For a maximum scanning time of 400 s, the optimum sequence for measuring placental perfusion was found to require 8 repetitions at each of 10 inversion times which were geometrically spaced (given by a(o), a(o)r, a(o)r2, a(o)r3, . . .), with a(o) = 850 ms, r = 1.073 and TR = 5 s, giving a pixel variability of 38%. Other timing schemes are recommended for measuring perfusion in other anatomical regions with different values of perfusion rate and longitudinal relaxation time.

Cocaine decreases cortical cerebral blood flow but does not obscure regional activation in functional magnetic resonance imaging in human subjects

The authors used functional magnetic resonance imaging (fMRI) to determine whether acute intravenous (i.v.) cocaine use would change global cerebral blood flow (CBF) or visual stimulation-induced functional activation. They used flow-sensitive alternating inversion recovery (FAIR) scan sequences to measure CBF and blood oxygen level-dependent (BOLD) sensitive T2* scan sequences during visual stimulation to measure neuronal activation before and after cocaine and saline infusions. Cocaine (0.6 mg/kg i.v. over 30 seconds) increased heart rate and mean blood pressure and decreased end tidal carbon dioxide (CO2). All measures returned to baseline by 2 hours, the interinfusion interval, and were unchanged by saline. Flow-sensitive alternating inversion recovery imaging demonstrated that cortical gray matter CBF was unchanged after saline infusion (-2.4 +/- 6.5%) but decreased (-14.1 +/- 8.5%) after cocaine infusion (n = 8, P < 0.01). No decreases were detected in white matter, nor were changes found comparing BOLD signal intensity in cortical gray matter immediately before cocaine infusion with that measured 10 minutes after infusion. Visual stimulation resulted in comparable BOLD signal increases in visual cortex in all conditions (before and after cocaine and saline infusion). Despite a small (14%) but significant decrease in global cortical gray matter CBF after acute cocaine infusion, specific regional increases in BOLD imaging, mediated by neurons, can be measured reliably.

Pulse sequences for steady-state saturation of flowing spins

It is useful to be able to suppress the NMR signal from spins in a flowing fluid, for example for "black-blood" visualization of blood vessels in vivo, for the suppression of flow artifacts, and for the estimation of tissue perfusion by continuous labeling of inflowing arterial spins. This work considers the flow of fluid through a region in which it is subjected to a train of saturation pulses. Computer simulations and in vitro measurements show that a train of equal-duration spoiler pulses produces less effective suppression than does a train of pulses of geometrically increasing duration. It is shown analytically that a long train of ideal equal-duration spoiler pulses converts initial magnetization (0, 0, M0) into a combination of longitudinal and transverse magnetization equal to 0. 29 (-M0, 0, M0) and is therefore unsatisfactory for continuous saturation.

On the characteristics of functional magnetic resonance imaging of the brain

In this review we discuss various recent topics that characterize functional magnetic resonance imaging (fMRI). These topics include a brief description of MRI image acquisition, how to cope with noise or signal fluctuation, the basis of fMRI signal changes, and the relation of MRI signal to neuronal events. Several observations of fMRI that show good correlation to the neurofunction are referred to. Temporal characteristics of fMRI signals and examples of how the feature of real time measurement is utilized are then described. The question of spatial resolution of fMRI, which must be dictated by the vascular structure serving the functional system, is discussed based on various fMRI observations. Finally, the advantage of fMRI mapping is shown in a few examples. Reviewing the vast number of recent fMRI application that have now been reported is beyond the scope of this article.

Magnetic resonance imaging of acute stroke

In the investigation of ischemic stroke, conventional structural magnetic resonance (MR) techniques (e.g., T1-weighted imaging, T2-weighted imaging, and proton density-weighted imaging) are valuable for the assessment of infarct extent and location beyond the first 12 to 24 hours after onset, and can be combined with MR angiography to noninvasively assess the intracranial and extracranial vasculature. However, during the critical first 6 to 12 hours, the probable period of greatest therapeutic opportunity, these methods do not adequately assess the extent and severity of ischemia. Recent developments in functional MR imaging are showing great promise for the detection of developing focal cerebral ischemic lesions within the first hours. These include (1) diffusion-weighted imaging, which provides physiologic information about the self-diffusion of water, thereby detecting one of the first elements in the pathophysiologic cascade leading to ischemic injury; and (2) perfusion imaging. The detection of acute intraparenchymal hemorrhagic stroke by susceptibility weighted MR has also been reported. In combination with MR angiography, these methods may allow the detection of the site, extent, mechanism, and tissue viability of acute stroke lesions in one imaging study. Imaging of cerebral metabolites with MR spectroscopy along with diffusion-weighted imaging and perfusion imaging may also provide new insights into ischemic stroke pathophysiology. In light of these advances in structural and functional MR, their potential uses in the study of the cerebral ischemic pathophysiology and in clinical practice are described, along with their advantages and limitations.

Non-invasive mapping of placental perfusion

BACKGROUND

We aim to develop a clinical technique for the non-invasive measurement of placental perfusion, to enable early detection of intrauterine growth restriction (IUGR). Pregnancies with this complication are characterised by low placental perfusion.

METHODS

We measured placental perfusion by means of perfusion-sensitive echoplanar imaging (EPI); a rapid method of making magnetic resonance images. Perfusion measurements were done on six healthy volunteers with normal pregnancies and nine with pregnancies complicated by IUGR. Perfusion maps were created to assess the relation between placental perfusion and fetal size at birth.

FINDINGS

Pregnancies complicated by IUGR differed significantly from normal pregnancies in patterns of perfusion within the placenta (p<0.0001, ANOVA). Subsequent analysis showed that the proportion of placentas with low perfusion rates was higher in the IUGR group than in the normal group. A significant correlation between areas of reduced placental perfusion and fetal size was demonstrated (p=0.041, Spearman's rank correlation).

INTERPRETATION

Non-invasive imaging of placental perfusion by means of EPI has potential as a clinical tool in assessing the dynamics of placental perfusion.

BASE imaging: a new spin labeling technique for measuring absolute perfusion changes

A new technique for magnetic resonance imaging of absolute perfusion changes that uses magnetically labeled tissue water proton spins as a freely diffusible tracer is described. It consists of unprepared basis (BA) images that serve as a reference and selective (SE) inversion prepared images that are sensitive to perfusion changes. In the present study, the BASE technique was applied to functional neuroimaging. BA and SE images were alternatingly and repeatedly acquired during periods of visual stimulation and control. Visual stimulation was achieved with an alternating black/white checkerboard operating at a frequency of 8 Hz. Maps of the absolute cerebral blood flow changes (deltaCBF) were calculated from the image intensities of the corresponding BA and SE images. The individual mean values of deltaCBF measured in five healthy volunteers ranged from 69 +/- 18 to 99 +/- 26 ml/min/100 g. Since the BASE technique does not require nonselective spin inversion, it can be used with small transmit/receive head coils (e.g., surface coils). In addition, the BASE technique is robust against a mismatch of the inversion and detection slice profiles.

Multislice imaging of quantitative cerebral perfusion with pulsed arterial spin labeling

A method is presented for multislice measurements of quantitative cerebral perfusion based on magnetic labeling of arterial spins. The method combines a pulsed arterial inversion, known as the FAIR (Flow-sensitive Alternating Inversion Recovery) experiment, with a fast spiral scan image acquisition. The short duration (22 ms) of the spiral data collection allows simultaneous measurement of up to 10 slices per labeling period, thus dramatically increasing efficiency compared to current single slice acquisition protocols. Investigation of labeling efficiency, suppression of unwanted signals from stationary as well as intraarterial spins, and the FAIR signal change as a function of inversion delay are presented. The assessment of quantitative cerebral blood flow (CBF) with the new technique is demonstrated and shown to require measurement of arterial transit time as well as suppression of intraarterial spin signals. CBF values measured on normal volunteers are consistent with results obtained from H2O15 positron emission tomography (PET) studies and other radioactive tracer approaches. In addition, the new method allows detection of activation-related perfusion changes in a finger-tapping experiment, with locations of activation corresponding well to those observed with blood oxygen level dependent (BOLD) fMRI.

Non-invasive temperature mapping using MRI: comparison of two methods based on chemical shift and T1-relaxation

PURPOSE

To implement and evaluate the accuracy of non-invasive temperature mapping using MRI methods based on the chemical shift (CS) and T1 relaxation in media of various heterogeneity during focal (laser) and external thermal energy deposition.

MATERIALS AND METHODS

All measurements were performed on a 1.5 T superconducting clinical scanner using the temperature dependence of the water proton chemical shift and the T1 relaxation time. Homogeneous gel and heterogeneous muscle phantoms were heated focally with a fiberoptic laser probe and externally of varying degree ex vivo by water circulating in a temperature range of 20-50 degrees C. Magnetic resonance imaging data were compared to simultaneously recorded fiberoptic temperature readings.

RESULTS

Both methods provided accurate results in homogeneous media (turkey) with better accuracy for the chemical shift method (CS:+/-1.5 degrees C, T1:+/-2.0 degrees C). In gel, the accuracy with the CS method was +/-0.6 degrees C. The accuracy decreased in heterogeneous media containing fat (T1:+/-3.5 degrees C, CS: +5 degrees C). In focal heating of turkey muscle, the accuracy was within 1.5 degrees C with the T1 method.

CONCLUSION

Temperature monitoring with the chemical shift provides better results in homogeneous media containing no fat. In fat tissue, the temperature calculation proved to be difficult.

Quantitative imaging of perfusion using a single subtraction (QUIPSS and QUIPSS II)

In the pulsed arterial spin labeling (ASL) techniques EPISTAR, PICORE, and FAIR, subtraction of two images in which inflowing blood is first tagged and then not tagged yields a qualitative map of perfusion. An important reason this map is not quantitative is that there is a spatially varying delay in the transit of blood from the tagging region to the imaging slice that cannot be measured from a single subtraction. We introduce here two modifications of pulsed ASL (QUIPSS and QUIPSS II) that avoid this problem by applying additional saturation pulses to control the time duration of the tagged bolus, rendering the technique relatively insensitive to transit delays and improving the quantitation of perfusion.

Quantitative measurements of cerebral blood flow in rats using the FAIR technique: correlation with previous iodoantipyrine autoradiographic studies

Flow-sensitive alternating inversion recovery (FAIR) is a recently introduced MRI technique for assessment of perfusion that uses blood water as an endogenous contrast agent. To characterize the FAIR signal dependency on spin tagging time (inversion time (TI)) and to validate FAIR for cerebral blood flow (CBF) quantification, studies were conducted on the rat brain at 9.4 T using a conventional gradient-recalled echo sequence. The T1 of cerebral cortex and blood was found to be 1.9 and 2.2 s, respectively, and was used for CBF calculations. At short TIs (<0.8 s), the FAIR signal originates largely from vascular components with fast flows, resulting in an overestimation of CBF. For TI > 1.5 s, the CBF calculated from FAIR is independent of the spin tagging time, suggesting that the observed FAIR signal originates predominantly from tissue/capillary components. CBF values measured by FAIR with TI of 2.0 s were found to be in good agreement with those measured by the iodoantipyrine technique with autoradiography in rats under the same conditions of anesthesia and arterial pCO2. The measured pCO2 index on the parietal cortex using the FAIR technique was 6.07 ml/100 g/min per mmHg, which compares well with the pCO2 index measured by other techniques. The FAIR technique was also able to detect the regional reduction in CBF produced by middle cerebral artery occlusion in rats.

Simultaneous multislice acquisition with arterial-flow tagging (SMART) using echo planar imaging (EPI)

Arterial spin tagging techniques have been used to image tissue perfusion in MR without contrast injection or ionizing radiation. Currently, spin tagging studies are performed primarily using single-slice imaging sequences, which are time consuming. This note reports a multislice echo-planar arterial spin tagging technique (Simultaneous Multislice Acquisition with aRterial-flow Tagging, or "SMART"). Multiband RF encoding (Hadamard) is used to provide simultaneous multislice acquisition capability for spin tagging techniques (such as echo planar imaging signal targeting with alternating radio frequency and flow-sensitive alternative inversion recovery). The method is illustrated with a two-slice pulse sequence that was implemented using the FAIR technique to generate two perfusion weighted images simultaneously. Compared with single-slice sequences, this two-slice sequence provided similar image quality, signal-to-noise ratio, and twice the spatial coverage compared with the single-slice technique within the same scan time.

Perfusion Magnetic Resonance Imaging to Assess Brain Tumor Responses to New Therapies

BACKGROUND: Although magnetic resonance imaging (MRI) is effective in detecting the location of intracranial tumors, new imaging techniques have been studied that may enhance the specificity for the prediction of histologic grade of tumor and for the distinction between recurrence and tumor necrosis associated with cancer therapy. METHODS: The authors review their experience and that of others on the use of perfusion magnetic resonance imaging to evaluate responses of brain tumors to new therapies. RESULTS: Functional imaging techniques that can distinguish tumor from normal brain tissue using physiological parameters. These new approaches provide maps of tumor perfusion to monitor the effects of novel compounds that restrict tumor angiogenesis. CONCLUSIONS: Perfusion MRI not only may be as effective as radionuclide-based techniques in sensitivity and specificity in assessing brain tumor responses to new therapies, but also may offer higher resolution and convenient co-registration with conventional MRI, as well as time- and cost-effectiveness. Further study is needed to determine the role of perfusion MRI in assessing brain tumor responses to new therapies.

Functional MRI using steady-state arterial water labeling

Flow-sensitive functional MRI (fMRI) was performed using steady-state arterial water labeling (SS-AWL). Arterial water labeling was accomplished by flow induced adiabatic fast passage. The signal intensity of the visual cortex in arterial water labeled images decreased by approximately 1.4% during visual stimulation of the brain. Acquisition of arterial water unlabeled and labeled images allows measurement of relative cerebral blood flow increase during brain activation. During visual stimulation, cerebral blood flow in the visual cortex increased by 17 to 35% as measured by SS-AWL. Quantitation of brain activation in terms of a physiological parameter using SS-AWL will facilitate comparative fMRI studies under different conditions.

Slice profile effects in adiabatic inversion: application to multislice perfusion imaging

Imperfections in the slice profile of the adiabatic inversion induced by relaxation effects are shown to cause signal variations in pulsed arterial tagging schemes on the order of magnitude of perfusion changes, and result in gross errors in perfusion quantitation. Significant improvement can be made with minor modifications to the inversion pulse which facilitate the acquisition of quantitative, multislice perfusion images, as demonstrated in both a phantom and a normal human volunteer.

Evidence for the exchange of arterial spin-labeled water with tissue water in rat brain from diffusion-sensitized measurements of perfusion

The extraction fraction of vascular water in rat brain is investigated by means of diffusion measurements of arterial spin labeled water at varying cerebral blood flow (CBF) values. The apparent diffusion coefficient (ADC) of the difference of the proton magnetization signal in the brain acquired with and without continuous arterial spin labeling is modeled to provide a measure of the amount of arterial water in tissue and vasculature and thus of the extraction fraction. The tissue and vascular portion of the arterial spin labeled water are differentiated based on their diffusion characteristics in a manner analogous to the intravoxel incoherent motion (IVIM) method. The amount of labeled arterial water that exchanges with tissue water is determined by estimating the fraction of the total signal that is associated with the slow-decaying component of a biexponential fit to the normalized difference signal between the magnetization of brain tissue acquired with and without arterial spin labeling. The results indicate that, at normal CBF (1.15 +/- 0.21 ml x g(-1) x min[-1]), about 90% of the arterial spin labeled water diffuses with an ADC of (1.21 +/- 0.37) x 10[-3] mm2 s[-1]), which is equal to tissue. At high CBF, an increasing fraction of the labeling water has a fast-pseudo-diffusion coefficient due to a decrease in water extraction fractions. The results also show that the contribution of vascular water to the measurement of perfusion by techniques that use endogenous water as a tracer can be efficiently eliminated by the use of diffusion sensitizing gradients with small effective b values (b approximately 20 s/mm2), enabling these techniques to monitor true changes in tissue perfusion.

Quantitation of regional cerebral blood flow increases during motor activation: a steady-state arterial spin tagging study

Steady-state arterial spin tagging MRI approaches were used to quantitate regional cerebral blood flow increases during finger tapping tasks in seven normal subjects. Statistically significant increases in cerebral blood flow were observed in the contralateral primary sensorimotor cortex in all seven subjects and in the supplementary motor area in five subjects. The intrinsic spatial resolution of the cerebral blood flow images was approximately 4 mm. If no spatial filtering was applied, the average increase in cerebral blood flow in the activated primary sensorimotor cortex was 60 +/- 10 cc/100 g/min (91 +/- 32%). If the images were filtered to a spatial resolution of 15 mm, the average increase in cerebral blood flow in the activated primary sensorimotor cortex was 23 +/- 7 cc/100 g/min (42 +/- 15%), in agreement with previously reported 133Xe and PET results.

Methods for assessing accuracy and reliability in functional MRI

In this paper, methods for assessing the accuracy and the reliability of functional magnetic resonance imaging techniques are presented. First, a modified receiver operating characteristic analysis is described for evaluating the accuracy of fMRI studies. With this modified approach, the true positives or the activated pixels are estimated based on highly averaged experimental data acquired with the same stimulation/task. Unlike ROC analysis based on simulated activation data, the present approach can be applied to experimentally acquired data without simplifying the activation related changes. To assess the reliability of fMRI studies, the kappa statistic was adopted for evaluating the overall agreement of functional activation maps from repeated experiments in individual subjects. To demonstrate the utility of these techniques, both the ROC analysis and the reliability assessment were applied to quantitatively evaluate the improvement in accuracy and reliability of a retrospective technique for physiological noise reduction in fMRI.

Multi-slice perfusion-based functional MRI using the FAIR technique: comparison of CBF and BOLD effects

Perfusion-weighted imaging techniques employing blood water protons as an endogenous tracer have poor temporal resolution because each image should be acquired with an adequate spin 'tagging' time. Thus, perfusion-based functional magnetic resonance imaging studies are typically performed on a single slice. To alleviate this problem, a multi-slice flow-sensitive alternating inversion recovery technique has been developed. Following a single inversion pulse and a delay time, multi-slice echo-planar images are acquired sequentially without any additional inter-image delay. Thus, the temporal resolution of multi-slice FAIR is almost identical to that of single slice techniques. The theoretical background for multi-slice FAIR is described in detail. The multi-slice FAIR technique has been successfully applied to obtain three-slice cerebral blood flow based functional images during motor tasks. The relative CBF change in the contralateral motor/sensory area during unilateral thumb-digit opposition is 45.0+/-12.2% (n=9), while the blood oxygenation level dependent signal change is 1.5+/-0.4 SD%. Relative changes of the oxygen consumption rate can be estimated from CBF and BOLD changes using FAIR. The BOLD signal change is not correlated with the relative CBF increase, and thus caution should be exercised when interpreting the BOLD change as a quantitative index of the CBF change, especially in inter-subject comparisons.

The efficiency of adiabatic inversion for perfusion imaging by arterial spin labeling

Velocity-driven adiabatic inversion is an attractive method for labeling arterial blood spins for quantitative perfusion imaging. To quantify perfusion and to optimize experimental parameters, an accurate estimate of labeling efficiency is required. We present theoretical and numerical methods to calculate the labeling efficiency over a wide range of experimental and physiologic parameters. The results are compared to experimental measurements in vivo. Inversion efficiency was found to be higher than previously assumed and relatively insensitive to flow velocity and the amplitude of the RF irradiation used for inversion. Assuming laminar flow, labeling efficiencies of greater than 90% are easily obtainable over a broad range of flow velocities. For applications where RF power deposition is a limiting factor such as at high field strengths, labeling efficiency can be maintained by reducing the labeling gradient. These results further illustrate the capability of adiabatic inversion to effectively label flowing blood.

Simultaneous assessment of flow and BOLD signals in resting-state functional connectivity maps

We have recently demonstrated using functional magnetic resonance imaging the presence of synchronous low-frequency fluctuations of signal intensities from the resting human brain that have a high degree of temporal correlation (p < 0.0001) both within and across the sensorimotor cortex. A statistically significant overlap between the resting-state functional connectivity map and the task-activation map due to bilateral finger tapping was obtained. Similar results have been obtained in the auditory and visual cortex. Because the pulse sequence used for collecting data was sensitive to blood flow and blood oxygenation, these low-frequency fluctuations of signal intensity may have arisen from variations of both. The objective of this study was simultaneously to determine the contribution of the blood oxygenation level signal and the flow signal to physiological fluctuations in the resting brain using the flow-sensitive alternating inversion recovery pulse sequence. In all subjects, the functional connectivity maps obtained from BOLD had a greater coincidence with task-activation maps than the corresponding functional connectivity maps obtained from blood-flow signals at the same level of statistical significance. Results of this study suggest that while variations in blood flow might contribute to functional connectivity maps, BOLD signals play a dominant role in the mechanism that gives rise to functional connectivity in the resting human brain.

Perfusion imaging by a flow-sensitive alternating inversion recovery (FAIR) technique: application to functional brain imaging

Perfusion is a crucial physiological parameter for tissue function. To obtain perfusion-weighted images and consequently to measure cerebral blood flow (CBF), a newly developed flow-sensitive alternating inversion recovery (FAIR) technique was used. Dependency of FAIR signal on inversion times (TI) was examined; signal is predominantly located in large vessels at short TI, whereas it is diffused into gray matter areas at longer TI. CBF of gray matter areas in the human brain is 71 +/- 15 SD ml/100 g/min (n = 6). In fMRI studies, micro- and macrovessel inflow contributions can be obtained by adjusting TIs. Signal changes in large vessel areas including the scalp were seen during finger opposition at a TI of 0.4 s; however, these were not observed at a longer TI of 1.4 s. To compare with commonly used BOLD and slice selective inversion recovery techniques, FAIR and BOLD images were acquired at the same time during unilateral finger opposition. Generally, activation sites determined by three techniques are consistent. However, activation of some areas can be detected only by FAIR, not by BOLD, suggesting that the oxygen consumption increase couples with the CBF change completely. Relative and absolute CBF changes in the contralateral motor cortex are 53 +/- 17% SD (n = 9) and 27 +/- 11 SD ml/100 g/min (n = 9), respectively.

Estimation of water extraction fractions in rat brain using magnetic resonance measurement of perfusion with arterial spin labeling

The model used for calculating perfusion by MRI techniques that use endogenous water as a tracer assumes that arterial water is a freely diffusible tracer. Evidence shows that this assumption is not valid in the brain at high blood flow rates, at which movement of water into and out of the microvasculature becomes limited by diffusion across the blood-brain barrier. In this work, the arterial spin-labeling technique is used to show that fraction of arterial water that is dependent on blood flow rate remains in the vasculature and does not exchange with brain tissue water. By using perfusion measurements without and with magnetization transfer (MT) effects, one can distinguish arterial label that exchanges into tissue because blood has much smaller MT than brain tissue. Using this technique, the extraction fraction for water is measured in the rat brain at various cerebral blood flow rates. At high flow rates (approximately 5 ml/g/min), the extraction fraction for water is found to be about 45% in rat brain. Disruption of the blood-brain barrier with D-mannitol caused an increase in the extraction fraction for water. It was possible to form an image related to the extraction fraction for water. The ability to estimate the amount of vascular water exchanging with tissue water by MRI may represent a noninvasive approach to detect the integrity of the blood-brain barrier.

Perfusion imaging by un-inverted flow-sensitive alternating inversion recovery (UNFAIR)

A new pulse sequence for estimating cerebral blood flow called UNFAIR, which uses a combination of sequential hyperbolic secant preparatory pulses, is introduced. This sequence is based on the same generalized conditions as previously introduced inversion recovery techniques except that the spins in the image slice of interest always have +z magnetization and the in-flowing spins are alternately inverted and uninverted. CBF-weighted images of rat brain under conditions of normocpnia and hypercapnia are presented and demonstrate the expected CBF response. A model describing the signal response to this pulse sequence is also presented and compared with in-vivo data acquired from gray and white matter.