Purpose: To use a combination of helium-3 (3-He) magnetic resonance imaging (MRI) and proton single-shot fast spin echo (SSFSE) to compare ventilated lung volumes in groups of "healthy" smokers, smokers diagnosed with moderate chronic obstructive pulmonary disease (COPD), and never-smokers. Materials and Methods:All study participants were assessed with spirometry prior to imaging. 3-He images were collected during an arrested breath hold, after inhaling a mixture of 200 mL of hyperpolarized 3-He/800 mL of N 2 . Proton SSFSE images were acquired after inhaling 1 liter of room air. The ventilated volume for each study participant was calculated from the 3-He images, and a ratio was calculated to give a percentage ventilated lung volume.Results: Never-smokers exhibited a 90% mean ventilated volume. The mean ventilated lung volumes for healthy smokers and smokers diagnosed with COPD were 75.2% and 67.6%, respectively. No correlation with spirometry was demonstrated for either of the smoking groups. Conclusion:Combined 3-He/Proton SSFSE MRI of the lungs is a noninvasive method, using nonionizing radiation, which demonstrates ventilated airspaces and enables the calculation of ventilated lung volumes. This method appears to be sensitive to early obstructive changes in the lungs of smokers. HYPERPOLARIZED HELIUM-3 (3-He) magnetic resonance imaging (MRI) is an emerging technique, which has been shown to produce high-resolution images of ventilated human airspaces (1,2). Proton MRI of the lungs has long been regarded to be of limited use in lung imaging due to cardiac motion artifacts, low proton density, and large magnetic susceptibility gradients associated with lung tissue. However, with ongoing development of fast imaging sequences, proton MRI of the lungs is now enjoying somewhat of a renaissance (3,4). In this work single-shot fast spin echo (SSFSE) breathhold images of the lungs were used to calculate a thoracic volume (5), and 3-He ventilation images were used to calculate a ventilated volume (6 -8) for each study participant. A ratio of ventilated volume to thoracic volume was calculated to give a percentage ventilated lung volume. MATERIALS AND METHODSThe local research ethics committee gave approval, and written informed consent was obtained from each study participant. The lungs of 13 volunteers (5 male, 8 female; mean age ϭ 51; range ϭ 40 -62) and 5 patients with chronic obstructive pulmonary disease (COPD) (2 male, 3 female; mean age ϭ 53; range ϭ 47-61) were imaged using proton SSFSE MRI and 3-He MRI in the coronal plane. All participants were assessed with spirometry prior to imaging and assigned to three groups: eight healthy never-smokers, five "healthy" smokers with a smoking history of Ͼ10 pack years, and five smokers with moderate COPD as demonstrated by spirometry and clinical history. Thus, COPD was defined as a subject who is symptomatic (chronic cough and shortness of breath), and the spirometric indices are 30% Ͻ forced expiratory volume in one second (FEV1) Ͻ 80% of the predicted value in comb...

No abstract

A radial projection sliding-window sequence has been developed for imaging the rapid flow of 3 He gas in human lungs. The short echo time (TE) of the radial sequence lends itself to fast repetition times, and thus allows a rapid update in the image when it is reconstructed with a sliding window. Oversampling in the radial direction combined with angular undersampling can further reduce the time needed to acquire a complete image data set, without significantly compromising spatial resolution. Controlled flow phantom experiments using hyperpolarized 3 He gas exemplify the temporal resolution of the method. In vivo studies on three healthy volunteers, one patient with chronic obstructive pulmonary disease (COPD), and one patient with hemiparalysis of the right diaphragm demonstrate that it is possible to accurately resolve the passage of gas down the trachea and bronchi and into the peripheral lung. Hyperpolarized 3 He gas MRI has been shown to be effective in visualizing breath-hold images of ventilation in humans (1). With optical pumping techniques, polarization levels far in excess of those attainable at thermal equilibrium in a B 0 field of 1.5 T can be attained. Because this polarization is not constrained by processes of saturation recovery, imaging with very fast repetition times (TRs) at a high signal-to-noise ratio (SNR) is a realistic prospect in vivo. This has enabled the dynamic study of gas inhalation with repetitive-frame fast imaging techniques. The study of ventilation dynamics may provide insights into lung pathophysiology, including air-trapping in chronic obstructive pulmonary disease (COPD). Previous dynamic studies performed on human subjects have used low flip angle, short-TR spin warp gradient-echo sequences (2,3); gradient-echo EPI (4); and, most recently, interleaved spiral sequences (5). Single-shot EPI might appear to be the logical way to rapidly monitor the passage of inhaled gas, because a flip angle of 90°will convert all the polarization to transverse magnetization in one shot, and subsequent signal can therefore be equated to fresh influx of gas. However, diffusion attenuation constrains the spatial resolution attainable with EPI (4), and field inhomogeneity can severely distort the images in the coronal and sagittal planes. Salerno et al. (5) developed a 24-interleave spiral sequence for dynamic 3 He imaging that offers good spatial resolution, is robust to motion and susceptibility effects, and provides repeated sampling of central and outer kspace per RF excitation (view). This last feature means that fluoroscopic sliding-window reconstruction techniques (6) can be effectively applied, as dynamic contrast changes pertaining to gas flow dynamics are updated on each view. In a study of guinea pig lung ventilation, Viallon et al. (7) presented a radial projection cine sequence. This was used to sample k-space in a continuously revolving pattern. When combined with a sliding-window reconstruction this produced high-quality images with a fast pseudotemporal image refresh rate. The ...

A method for 3D volume-localized quantification of pO2 in the lungs is presented that uses repetitive frame 3D gradient-echo imaging of 3 He. The method was demonstrated by experiments on 3 He phantoms containing known concentrations of O 2 and in vivo on a group of three healthy human volunteers. The results were compared with those obtained by equivalent 2D thin-slice and 2D projection methodologies, and were found to be consistent with published results from the 2D projection methodologies (pO 2 ‫؍‬ 0.09 -0.18 bar). Studies performed on the same subject, on three separate occasions, demonstrated a repeatability of pO 2 measurement to within 14% using the 3D technique. Experimental differences between the 2D and 3D methods were substantiated with theoretical and numerical analyses of the signal decay, which took into account the effects of out-of-slice diffusion as a source of error in the thin-slice 2D experiments. It is shown that the 2D thin-slice technique systematically underestimates pO2 when there is significant gas diffusion (factor of 4 underestimate for D ‫؍‬ 0.9 cm MRI of hyperpolarized (HP)3 He gas can provide high spatial and temporal resolution images of gas ventilation in the lungs and airways, as well as useful functional information (1). Potentially one of the most useful functional parameters is an estimate of the partial pressure of oxygen (pO 2 ) in the lungs and the rate of oxygen extraction (2-6). Accurate spatially localized pO 2 measurement with 3 He MRI would be clinically useful as a means of assessing lung ventilation-perfusion (V/Q). When paramagnetic oxygen mixes with polarized 3 He in the lungs, a marked reduction in the 3 He T 1 arises through the electron-nuclear spin dipolar coupling. Saam et al. (7) quantified the dependence of the 3 He T 1 on oxygen concentration with phantom experiments. This O 2 dependence of the 3 He T 1 was used by Deninger et al. (2,3) for in vivo quantification of pO 2 in a series of breath-hold experiments on pigs and humans. These in vivo experiments used images acquired from whole-lung projections (very thick 2D slices). This was done to maximize the signal-to-noise ratio (SNR) and to circumvent diffusion of polarized gas out of the slice during the significant inter-image delay time (typically up to 7 s) as a possible source of change in the time course of the in vivo 3 He signal. Moller et al. (5) estimated the degree of partial volume mixing of polarization due to gas diffusion for a 2D thin-slice experiment in the guinea pig lung. Assuming a 24 s total acquisition time and an apparent diffusion coefficient of D ϭ 0.16 cm 2 s -1 , they predicted that only 5% of the in-slice magnetization would be exchanged between adjacent slices. In addition, the recently published results by Jalali et al. (6) from 2D pO 2 experiments in porcine lungs, with 2-cm-thick slices, are in fair agreement with the results from a 2D porcine study with 18-cm-thick slices (2). These results suggest that diffusion between slices may not be a considerable problem; however, a...

Images of hyperpolarized 3 He were acquired during breath-hold in four healthy volunteers with the use of an optimized 3D gradient-echo sequence. The images were compared with existing 2D gradient-echo methods. The average SNR from a 13-mm-thick slice in the peripheral lung was 1.4 times greater with 3D. In the airways the average SNR was 1.7 times greater with 3D. The higher SNR of 3D was particularly evident when regions of unimpeded gas diffusion, such as the major airways, were imaged with thin slices. This is because diffusion dephasing due to the slice-encoding gradient is minimized with a 3D sequence. The in vivo experimental findings were substantiated with experiments on phantoms of free gas, which showed more than four times the SNR with 3D compared to 2D. Most clinical studies of hyperpolarized (HP) gas ventilation have imaged the lungs using 2D, multislice, low-flipangle, gradient-echo sequences (1,2). A recent abstract (3) demonstrated impressive images of ventilation that were acquired with a 3D cylindrical k-space trajectory. The only other studies of 3D imaging with HP gas were conducted in animal models (4 -7). In this work, the factors that affect the SNR and spatial resolution of 3D gradient-echo sequences with rectilinear phase encoding were investigated. Experiments were performed on gas phantoms and healthy human subjects at breath-hold. The observed SNR and spatial resolution in the images were quantitatively compared with theory and the results from 2D experiments. THEORYDuring breath-hold, the transverse magnetization, M t (n), after the nth RF pulse is given by[1]Equation [1] accounts for polarization loss due to RF depletion by a pulse of flip angle , and also includes a longitudinal relaxation term exp(-nTR/T 1 ). The dominant contribution to the 3 He T 1 in vivo is from the dipolar interaction with paramagnetic oxygen (8). Spoiling of any residual transverse coherence is also assumed. The weighting of the SNR in the images (the k y ϭ 0 point) due to H(k y ) is thus given by sin( ). If this is considered with the diffusion attenuation effect of the slice encoding gradient, exp(-b s D) (9), and the inherent SNR weighting of 3D vs. 2D (10), then the theoretical SNR of 3D vs. 2D for HP gas is given byEquation [2] assumes an identical bandwidth, TR, and gas polarization for both 2D and 3D. Furthermore, a thin slice is assumed, and interslice polarization mixing from gas diffusion in 2D experiments is not included. The diffusion of gas out of the slice has previously been studied in phantom experiments (16,17). Assuming an infinite medium with no boundaries, the spatial mixing of gas between a single slice (the profile of which is denoted by f(z,t)) and its surroundings can be solved as a function of time from the 1D diffusion equation ‫ץ‬f/‫ץ‬t ϭ D͑‫ץ‬ 2 f/‫ץ‬z 2 ͒. The analytical solution is the convolution integral of the initial slice profile from the first view, f(z,0), with a Gaussian of variance given by 2 ϭ 2Dt. MATERIALS AND METHODS SimulationsComputer simulations were performed f...

Purpose: To demonstrate the feasibility of registering hyperpolarized helium-3 magnetic resonance images (3He-MRI) to X-ray computed tomography (CT) for functionally weighted intensity-modulated radiotherapy (IMRT) planning.Methods and Materials: Six patients with non–small-cell lung cancer underwent 3He ventilation MRI, which was fused with radiotherapy planning CT using rigid registration. Registration accuracy was assessed using an overlap coefficient, calculated as the proportion of the segmented 3He-MR volume (VMRI) that intersects the segmented CT lung volume expressed as a percentage of VMRI. For each patient, an IMRT plan that minimized the volume of total lung receiving a dose ≥20 Gy (V20) was compared with a plan that minimized the V20 to well-ventilated lung defined by the registered 3He-MRI.Results: The 3He-MRI and CT were registered with sufficient accuracy to enable functionally guided IMRT planning (median overlap, 89%; range, 72–97%). In comparison with the total lung IMRT plans, IMRT constrained with 3He-MRI reduced the V20 not only for the well-ventilated lung (median reduction, 3.1%; range, 0.4–5.1%; p = 0.028) but also for the total lung volume (median reduction, 1.6%; range, 0.2–3.7%; p = 0.028).Conclusions: Statistically significant improvements to IMRT plans are possible using functional information provided by 3He-MRI that has been registered to radiotherapy planning CT.

Purpose:To probe the variation of alveolar size in healthy lung tissue as a function of posture using diffusionweighted helium-3 hyperpolarized gas imaging. Materials and Methods:Measurements of the helium-3 apparent diffusion coefficient (ADC) were made on six healthy subjects. These were used to show the variation of alveolar size between the lowermost dependent regions of the lung compared to the uppermost regions of the lung in four postures: supine, prone, left-lateral decubitus, and right-lateral decubitus. Results:The distribution of acinar size in the lungs was found to be heterogeneous, and influenced by lung orientation. In nearly all postures, the ADC was significantly higher in the non-dependent uppermost regions of the lung compared to the dependent lowermost regions of the lung; the greatest variation was found in the left-lateral decubitus position. The difference in ADC between uppermost and lowermost regions was on average 0.012 cm 2 second -1 , which represents 20% of the average ADC value for the whole lung. A systematic decrease in ADC from the apex of the lung to the base was also found, which corresponds to an inherent gradient in alveolar size. Conclusion:The posture dependent variations in ADC were attributed to compression of the parenchyma under its own weight and the mass of the heart. OVER THE PAST DECADE, advances in polarizing technology have allowed imaging of the human lungs using the noble gases helium-3 and xenon-129 as inhaled contrast agents (1). A promising aspect of this technology is the ability to probe the lung microstructure with diffusion sensitive measurements using pulsed gradient sequences (2). Diffusion of the gas atoms in the presence of these gradients results in an attenuation of the magnetic resonance imaging (MRI) signal, which can be used to calculate the apparent diffusion coefficient (ADC). Helium has a very low atomic mass and consequently a very high diffusion coefficient. In the lungs, the dimensions of the acini are smaller than the average distances traversed by freely diffusing helium atoms during the time it takes to measure diffusion during a typical MRI experiment. This "restriction" in gas diffusion leads to a measured ADC that is much smaller in healthy lungs than for diffusion in free-space. Typical MRI measurements reveal that the helium-3 ADC is between 0.1 cm 2 second -1 and 0.2 cm 2 second -1 in healthy lungs (3-5), whereas in free-space it is 0.88 cm 2 second -1 when mixed with room air.The restricted nature of diffusion in the lungs means that the ADC can be used to probe the lung microstructure. For example, in emphysematous lung, destruction of alveolar infrastructure leads to an enlargement of the respiratory airways within the acini, which results in a significantly higher ADC measurement (4,5). Thus, the ADC's sensitivity to the confining length scales can provide an indirect measure of acinar size, despite the fact that the alveolar dimensions are microscopic compared to the size of image pixels. This property of the ADC can be exploi...

The purpose of this study was to compare hyperpolarized 3helium magnetic resonance imaging (3He MRI) of the lungs in adults with cystic fibrosis (CF) with high-resolution computed tomography (HRCT) and spirometry. Eight patients with stable CF prospectively underwent 3He MRI, HRCT, and spirometry within 1 week. Three-dimensional (3D) gradient-echo sequence was used during an 18-s breath-hold following inhalation of hyperpolarized 3He. Each lung was divided into six zones; 3He MRI was scored as percentage ventilation per lung zone. HRCT was scored using a modified Bhalla scoring system. Univariate (Spearman rank) and multivariate correlations were performed between 3He MRI, HRCT, and spirometry. Results are expressed as mean+/-SD (range). Spirometry is expressed as percent predicted. There were four men and four women, mean age = 31.9+/-9 (20-46). Mean forced expiratory volume in 1 s (FEV)1 = 52%+/-29 (27-93). Mean 3He MRI score = 74%+/-25 (55-100). Mean HRCT score = 48.8+/-24 (13.5-83). The correlation between 3He MRI and HRCT was strong (R = +/-0.89, p < 0.001). Bronchiectasis was the only independent predictor of 3He MRI; 3He MRI correlated better with FEV1 and forced vital capacity (FVC) (R = 0.86 and 0.93, p < 0.01, respectively) than HRCT (R = +/-0.72 and +/-0.81, p < 0.05, respectively). This study showed that 3He MRI correlates strongly with structural HRCT abnormalities and is a stronger correlate of spirometry than HRCT in CF.