Magnetic resonance behavior of normal and diseased lungs: spherical shell model simulations

The alveolar air-tissue interface affects the lung NMR signal, because it results in a susceptibility-induced magnetic field inhomogeneity. The air-tissue interface effect can be detected and quantified by measuring the difference signal (Delta) from a pair of NMR images obtained using temporally symmetric and asymmetric spin-echo sequences. The present study describes a multicompartment alveolar model (consisting of a collection of noninteracting spherical water shells) that simulates the behavior of Delta as a function of the level of lung inflation and can be used to predict the NMR response to various types of lung injury. The model was used to predict Delta as a function of the inflation level (with the assumption of sequential alveolar recruitment, partly parallel to distension) and to simulate pulmonary edema by deriving equations that describe Delta for a collection of spherical shells representing combinations of collapsed, flooded, and inflated alveoli. Our theoretical data were compared with those provided by other models and with experimental data obtained from the literature. Our results suggest that NMR Delta measurements can be used to study the mechanisms underlying the lung pressure-volume behavior, to characterize lung injury, and to assess the contributions of alveolar recruitment and distension to the lung volume changes in response to the application of positive airway pressure (e.g., positive end-expiratory pressure).

In vivo Hahn spin-echo decay (Hahn-T2) observation of regional changes in the time course of oleic acid lung injury

We studied the time course of changes in the Hahn spin-echo decay (Hahn-T2) in lungs of spontaneously breathing living rats at 1 hour, 3 hours, and 7 days following oleic acid injection. Motion artifacts were minimized by using the motion-insensitive interleaved rapid line scan (ILS) imaging technique. Prior to injury, the lungs exhibited two resolvable exponential Hahn-T2 components. One and 3 hours after injury the decay showed a regionally nonuniform behavior, which was fit with one, two, or three exponential components. The short and medium components increased at 1 and 3 hours after injection. The third, much longer, component is probably due to intraalveolar pulmonary edema. After 7 days the Hahn decay was similar to that observed before injury, probably reflecting resolution of the edema. Our data suggest that Hahn-T2 measurements can be used to characterize the time course and regional distribution of lung injury in living animals.

Modeling the nuclear magnetic resonance behavior of lung: from electrical engineering to critical care medicine

The present article reviews the basic principles of a new approach to the characterization of pulmonary disease. This approach is based on the unique nuclear magnetic resonance (NMR) properties of the lung and combines experimental measurements (using specially developed NMR techniques) with theoretical simulations. The NMR signal from inflated lungs decays very rapidly compared with the signal from completely collapsed (airless) lungs. This phenomenon is due to the presence of internal magnetic field inhomogeneity produced by the alveolar air-tissue interface (because air and water have different magnetic susceptibilities). The air-tissue interface effects can be detected and quantified by magnetic resonance imaging (MRI) techniques using temporally symmetric and asymmetric spin-echo sequences. Theoretical models developed to explain the internal (tissue-induced) magnetic field inhomogeneity in aerated lungs predict the NMR lung behavior as a function of various technical and physiological factors (e.g., the level of lung inflation) and simulate the effects of various lung disorders (in particular, pulmonary edema) on this behavior. Good agreement has been observed between the predictions obtained from the mathematical models and the results of experimental NMR measurements in normal and diseased lungs. Our theoretical and experimental data have important pathophysiological and clinical implications, especially with respect to the characterization of acute lung disease (e.g., pulmonary edema) and the management of critically ill patients.

Effects of endotoxin lung injury on NMR T2 relaxation

The effects of endotoxin injury on lung NMR relaxation times (T1, CPMG T2, and Hahn decay constant (Hahn T2)) were studied in excised unperfused rat lungs. Blinded histologic examination showed no clear-cut separation between endotoxin and control lungs. Morphometric lung tissue volume density and gravimetric lung water content did not differ significantly between the two groups. In contrast, the values of the fast, intermediate, and slow T2 components, obtained by multiexponential analysis of the CPMG decay curve, increased markedly after endotoxin administration, with minimal overlap between endotoxin and control values. The response of Hahn T2 was, in general, in the same direction as that of CPMG T2; however, Hahn T2 may be more affected by measurement errors and may be less sensitive to the presence of lung injury. T1 showed minimal changes after injury. The present data suggest that CPMG T2 measurements can consistently detect the presence of lung injury even when conventional histologic, morphometric, and gravimetric studies provide negative or equivocal results, and that the CMPG T2 method is superior, in this respect, to the Hahn decay method. T1 does not appear to be sensitive to lung injury in the absence of significant lung water accumulation.

Water self-diffusion measurements in excised rat lungs

Water self diffusion in excised rat lungs has been measured using pulsed-field-gradient (PFG) techniques. The apparent diffusion coefficient, Dapp, was measured from a plot of the magnetization M vs ga2 to be 4.0 x 10(-6) cm2/s in the limit of small gamma delta ga, where gamma is the gyromagnetic ratio, delta is the duration of the applied gradient pulses, and ga is the applied gradient strength. Dapp is independent of the diffusion time, t, for values of t between 18 and 106 ms. For larger values of gamma delta ga, an additional smaller value of the slope of M vs ga2 was observed, indicating the existence of other, more slowly dephasing spins. Variation of t revealed that the relative magnetization associated with the more slowly dephasing spins decreases as t is increased. In addition, the relative magnetization of the slowly dephasing spins decreases as the temperature, T, of the excised rat lung is increased. Slow exchange from the compartment of the more rapidly to that of the more slowly dephasing spins may explain some of the observed dependence of the relative magnetizations on t and T. Measurements of water self diffusion in rat lung at various levels of water content indicate a correlation between T2 components and diffusion components. A new technique that combines the PFG with the Carr-Purcell-Meiboom-Gill technique is presented. The application of this technique to excised rat lung confirms the correlation between T2 and diffusion components.

Extension of the Rorschach--Hazlewood theoretical model for spin-lattice relaxation in biological systems to low frequencies

The water-biopolymer cross-relaxation model, proposed by H. E. Rorschach and C. F. Hazlewood (RH) [J. Magn. Reson. 70, 79 (1986)], explains the Larmor frequency dependence of T1 in many biological systems. However, the RH theory fails at low Larmor frequencies. In this paper, a more general version of the RH theory has been developed. This theory is valid at all frequencies. Use of the new expression for the spin-lattice relaxation rate (1/T1), earlier published experimental data in H2O/D2O bovine serum albumin, which had been measured over a wide frequency range (10 kHz to 100 MHz), were fitted over the entire frequency range. The agreement between theory and the experimental data is excellent. Theoretical expressions for the rotating-frame spin-lattice relaxation rate (1/T1(rho)) were also obtained.

Application to rat lung of the extended Rorschach-Hazlewood model of spin-lattice relaxation

The spin-lattice relaxation time T1 was measured in excised degassed (airless) rat lungs over the frequency range 6.7 to 80.5 MHz. The observed frequency dependence was fitted successfully to the water-biopolymer cross-relaxation theory proposed by H. E. Rorschach and C. F. Hazlewood (RH) [J. Magn. Reson. 70, 79 (1986)]. The rotating frame spin-lattice relaxation time T1(rho) was also measured in rat lung fragments over the frequency range 0.56 to 5.6 kHz, and the observed frequency dependence was explained with an extension of the RH model. The agreement between the theory and the experimental data in both cases is good.

Comparison of calculated and experimental NMR spectral broadening for lung tissue

NMR lineshapes were calculated for a model of lung, and NMR proton spectra were measured for individual voxels in an excised inflated rat lung. NMR lines for parenchymal lung regions containing alveoli, alveolar ducts, and capillaries were calculated using a computer simulation of the NMR signal from a three-dimensional honeycomb-like structure, a collection of modified Wigner-Seitz cells. These cells were modified by rounding the corners and increasing the thickness of the boundaries to model various degrees of lung inflation and lung water. NMR lineshapes were also calculated for the central or nonparenchymal lung regions containing bronchi and large blood vessels. A comparison of theoretical lineshapes with those measured in individual voxels both in the parenchymal and in the central (largely nonparenchymal) regions in excised rat lungs at an inflation pressure of 30 cm of water shows excellent agreement. These results indicate that the NMR lineshape reflects the underlying lung geometry. This research constitutes the first calculations and measurements of NMR lineshapes in lung. The appendix describes a new method for calculating the magnetic field inside a weakly diamagnetic material of arbitrary shape placed in an otherwise uniform external magnetic field. This new method does not require either solution of simultaneous equations or evaluation of integral expressions.

Lung water measurement by nuclear magnetic resonance: correlation with morphometry

Estimates of lung water content obtained from nuclear magnetic resonance (NMR) and morphometric and gravimetric measurements were compared in normal and experimentally injured rats. Average lung water density (rho H2O) was measured by an NMR technique in excised unperfused rat lungs (20 normal lungs and 12 lungs with oleic acid-induced edema) at 0 (full passive deflation) and 30 cmH2O lung inflation pressure and in vivo (4 normal rats and 8 rats with lung injury induced by oleic acid or rapid saline infusion). The rho H2O values were compared with morphometric measurements of lung tissue volume density (Vv) obtained from the same lungs fixed at corresponding liquid-instillation pressures. A close correlation was observed between rho H2O and Vv in normal and injured excised lungs [correlation coefficient (r) = 0.910, P < 0.01]. In vivo rho H2O was also closely correlated with Vv (r = 0.897, P < 0.01). The correlation coefficients between rho H2O and gravimetric lung water content (LWGr) were lower in the excised lung group (r = 0.663 and 0.692, respectively, for rho H2O at 0 and 30 cmH2O lung inflation pressure, P < 0.01) than in the in vivo study (r = 0.857, P < 0.01). Our results indicate that NMR techniques, which are noninvasive and nondestructive, provide reliable estimates of lung water density and that the influence of lung inflation on rho H2O is important (compared with the effect of lung water accumulation in lung injury) only in the presence of deliberately induced very large variations in the lung inflation level.

Alveolar air/tissue interface and nuclear magnetic resonance behavior of normal and edematous lungs

The alveolar air/tissue interface markedly affects the NMR properties of lungs by causing an NMR signal loss as a result of internal (tissue-induced) magnetic field inhomogeneity. The signal loss can be measured as the difference in NMR signal intensity (difference signal delta) between a pair of images obtained using temporally symmetric and asymmetric spin-echo sequences. Previous data indicate that the difference signal measured at an asymmetry time of 6 ms (delta 6ms) is very low in degassed lungs and increases markedly with alveolar opening. Theoretically, the NMR behavior of edematous lungs is expected to differ from that of normal nondegassed lungs because alveolar flooding and collapse are equivalent to partial (regional) degassing. To test this prediction, we measured delta 6ms in normal and edematous (oleic acid-injured) excised unperfused rat lungs at 5, 10, 20, 30, and 0 (full passive deflation) cm H2O inflation pressure (PL). Lung volume changes were estimated from NMR lung water density (pH2O) measurements. In normal lungs, delta 6ms did not vary with PL. In edematous lungs delta 6ms was, as predicted, significantly lower than normal at 5 and 10 cm H2O PL but rose markedly (to about normal) as PL was further increased. Upon subsequent deflation from 30 to 0 cm H2O PL, delta 6ms did not vary significantly or decreased. On the basis of our theoretical models, the data could be interpreted as reflecting the loss of alveolar air/tissue interface as a result of alveolar flooding and the relative contributions of airspace recruitment and distension to the lung volume changes. Histologic and morphometric data obtained from the same lungs supported this interpretation.(ABSTRACT TRUNCATED AT 250 WORDS)

Nuclear magnetic resonance Hahn spin-echo decay (T2) in live rats with endotoxin lung injury

To determine the possibility of using nuclear magnetic resonance imaging to study experimentally induced lung injury, we measured in the lungs of spontaneously breathing living rats the time course of both the Hahn spin-echo decay (T2) and the proton density after endotoxin injury. In order to minimize artifacts arising from motions of the nearby chest wall and heart, we used a motion-insensitive technique (the interleaved line scan). A typical Hahn T2 measurement was obtained over a region of interest from a series of images each with a different echo time, which ranged from 16 to 110 ms. Lung water content was determined by integrating the proton density over the region of interest. The Hahn T2 and proton density were measured before and at 1, 3, 6, and 9 h after intravenous injection of endotoxin. The effects of the treatment administered before and after endotoxin injection were also evaluated. Endotoxin treatment caused lengthening of both fast (T2f) and slow (T2s) Hahn T2 components but had no significant effect on the proton density, consistent with the notion that endotoxin causes lung injury without significant lung water accumulation in rats. However, the methylprednisolone treatment prevented the lengthening of T2s but did not seem to have a significant effect on T2f. Our results suggest that NMR imaging can be used to detect and monitor experimental lung injury in intact living animals, even in the absence of variations of lung water content.

Comparison of in vivo and in vitro Hahn T2 measurements in rat lung

We compared in vivo and in vitro Hahn echo T2 measurements in rat lungs in both imaging and nonimaging modes. All measurements could be characterized by multiexponential functions consisting of either two or three exponentials. Essentially the same values of the time constants were observed for spontaneously breathing rats and for excised lungs.

Rapid line scan technique for artifact-free images of moving objects

Moving objects (e.g., heart, lung, chest wall, etc.) typically cause artifacts to appear in two-dimensional Fourier transform ("spin warp") images. The rapid line scan (RLS) technique is a simple inexpensive technique that can rapidly produce artifact-free images of moving objects, without requiring enormous magnetic field gradients or periodic motions. The basic concepts and potential industrial applications of the RLS technique are described.

Potential industrial applications of inhomogeneous broadening imaging

Variations in magnetic susceptibility at air-water interfaces can result in inhomogeneous broadening of the NMR line. By special asymmetrical imaging techniques, originally developed for lung imaging, images can be formed of only those molecules that experience this inhomogeneous broadening. The basic concepts and latest developments in inhomogeneous-broadening-imaging techniques are described. Potential industrial applications are also discussed.

Alveolar air-tissue interface and nuclear magnetic resonance behavior of lung

Inflated lungs are characterized by a short nuclear magnetic resonance (NMR) free induction decay (rapid disappearance of NMR signal), likely due to internal (tissue-induced) magnetic field inhomogeneity produced by the alveolar air-tissue interface. This phenomenon can also be detected using temporally symmetric and asymmetric NMR spin-echo sequences; these sequences generate a pair of NMR images from which a difference signal (delta) is obtained (reflecting the signal from lung water experiencing the air-tissue interface effect). We measured delta in normal excised rat lungs at inflation pressures of 0-30 cmH2O for asymmetry times (a) of 1-6 ms. Delta was low in degassed lungs and increased markedly with alveolar opening when measured at a = 6 ms (delta 6 ms); delta 6 ms varied little during the rest of the inflation-deflation cycle. Delta 1 ms (a = 1 ms) did not vary significantly on inflation and deflation. Measurements of delta at a = 3 and 5 ms generally lay between those of delta 1 ms and delta 6 ms. These findings, which are consistent with theoretical predictions, suggest that measurements of delta at appropriate asymmetry times are particularly sensitive to alveolar opening and may provide a means of distinguishing alveolar recruitment from alveolar distension in the pressure-volume behavior of the lung.

An in vivo NMR imaging determination of multiexponential Hahn T2 of normal lung

We describe the first in vivo imaging determination of normal lung tissue's multiexponential transverse magnetization decay. Normal spontaneously breathing rats were used for the measurements. To obtain motion-insensitive images, we used a modified line scan imaging technique which we call the interleaved line scan (ILS). The ILS overcomes the following difficulties associated with imaging lungs: low signal-to-noise ratio (S/N) due to lung's low proton density and short T2 decay, artifacts associated with cardiac and respiratory motion, and excessively long imaging times with conventional line scan techniques. Using the ILS, a 16-line 32-average image with an 8-s repetition time requires 4.3 min. From a series of 16 Hahn spin-echo images with echo times ranging from 16 to 90 ms, we obtained a two-component T2 decay for normal peripheral lung tissue. The measured fast and slow T2 components were 9.5 +/- 1.0 and 34 +/- 5.0 ms for the right lung and 9.0 +/- 1.5 and 32 +/- 4.5 for the left lung. The relative magnetization for the slow T2 component was 7.0 +/- 4.5% for the right lung and 10 +/- 3.0% for the left lung.

Regional effects of repetition time on NMR quantitation of water in normal and edematous lungs

It is well known that pulmonary edema is, in general, spatially nonuniform. Since the NMR spin-lattice relaxation time (T1) is increased by lung edema, the spatial distribution of T1 will be nonuniform. When the repetition time (TR) is short relative to the T1 of edematous lung, lung water content will be underestimated and this underestimation will be spatially nonuniform as well. Therefore, technical artifacts which are a complex function of lung edema and its spatial distribution are expected. We compared overall and regional (topographic) lung water density measurements obtained from living rats (with normal or edematous lungs) using repetition times of 2.0 and 6.2 s (at a magnetic field of 1 T), to quantify this uneven T1 effect for normal and edematous lungs. NMR measurements at TR = 2.0 s underestimated whole lung water density (-rho H2O) TR = 6.2 s) by an average of 7.2% in normal rats and 22.5% in rats with pulmonary edema. Regional -rho H2O underestimation (%delta-rho H2O) varied from 2.2 to 8.8% (groups means) in normal lungs and from 7.3 to 30.8% in edematous lungs. As a result, the interquartile range (of the voxel distribution as a function of rho H2O) underestimated the spatial nonuniformity of lung water density by 28.0% in edematous lungs, likely because of greater loss of NMR signal from high-water-density, long-T1 lung regions. Both %delta-rho H2O and T1 were significantly correlated with -rho H2O at TR = 6.2 s.(ABSTRACT TRUNCATED AT 250 WORDS)

Water proton NMR relaxation mechanisms in lung tissue

The NMR relaxation times T'2, T2, and T1 were measured in isolated rat lungs as functions of external magnetic field B0, temperature, and lung inflation. The observed linear dependence on B0 of the tissue-induced free induction decay rate (T'2)-1 provides independent confirmation of the air/water interface model of the lung. Furthermore, measurements of the Larmor frequency dependence of T1 are consistent with a spin-lattice relaxation rate of the form 1/T1 = A omega -1/2 + B as expected for the case in which the relaxation arises from water-biopolymer cross-relaxation, which should be proportional to the surface area of the lung. This prediction was verified by observations of an approximately linear dependence of 1/T1 on transpulmonary pressure and thus on the lung surface area.

Quantitative assessment of pulmonary edema by nuclear magnetic resonance methods

Considerable progress has been made in the application of nuclear magnetic resonance (NMR) imaging and nonimaging techniques to the quantitative assessment of pulmonary edema. NMR measurements offer the advantages of being noninvasive, relatively rapid, and easily repeatable. In addition, NMR imaging is suitable for the determination of lung water distribution. Studies of various animal models have shown that NMR techniques can adequately detect and quantify relative changes in lung water content and distribution in various types of experimental lung injury. Preliminary observations in humans suggest that NMR measurement of relative lung water changes in clinical pulmonary edema should be feasible. Although the application of NMR to the assessment of pulmonary edema appears to be very promising, it also poses significant problems that must be solved before it can be established as a standard experimental and clinical method.

Assessment of lung water distribution by nuclear magnetic resonance. A new method for quantifying and monitoring experimental lung injury

We have developed a new analytical method that uses nuclear magnetic resonance (NMR) imaging data to quantify lung water content and distribution. This new method generates a distribution of lung water density in which the fraction of voxels corresponding to a given water density is plotted on the vertical axis as a function of water density on the horizontal axis, thereby complementing the spatial information provided by the NMR image. We obtained reproducible lung water distribution data at comparable lung volumes in normal excised lungs and in intact living rats. In normal excised unperfused rat lungs, the distribution varied with the degree of inflation, but the changes were small compared with those associated with lung edema. The lung water density distribution changed markedly after induction of lung edema by intrabronchial saline instillation, intravenous oleic acid injection, and rapid intravenous saline infusion. Lung water density distribution data were well correlated (correlation coefficient = 0.948 for the excised lungs and 0.823 for the intact living rats) with gravimetric lung water measurements. The new analytical method is noninvasive, provides easily repeatable measurements, and is as sensitive as the gravimetric technique to lung water changes.

Determination of lung water content and distribution by nuclear magnetic resonance imaging

NMR imaging techniques are applicable to the assessment of lung water content and distribution because the NMR signal is, under certain conditions, proportional to tissue proton density. NMR imaging is noninvasive, easily repeatable, free from ionizing radiation, and particularly suitable for the assessment of spatial lung water distribution. Lung water content and distribution have been estimated in excised animal lungs and in intact dead or living animals, under normal conditions and in various types of experimental pulmonary edema. Excised human lungs and human subjects have also been studied. Published data indicate that measurements of lung water content by NMR imaging techniques are feasible. These techniques estimate lung water spatial distribution with satisfactory accuracy and excellent resolving power. The application of NMR imaging techniques poses several problems and limitations, but available data suggest that most of the problems can be solved. NMR imaging has the potential to become a powerful tool for lung water research. Prospects of clinical application are also encouraging; numerous applications can be foreseen, although lack of mobility of NMR imaging systems may be a significant limitation in critical care medicine.

Asymmetric spin echo sequences. A simple new method for obtaining NMR 1H spectral images

The nuclear magnetic resonance (NMR) signal decay produced by reversible tissue-induced dephasing of the magnetization components in the transverse plane (reversible tissue-induced dephasing) was measured and expressed as a function of a new transverse relaxation time T'2 (T2 prime) for samples of rat liver, retroperitoneal fat, inflated lung, and corn oil. Simple exponentials did not adequately describe the observed NMR signal decay. Inflated lung demonstrated the most rapid signal decay (T'2 = 4.8 ms) followed by retroperitoneal fat (T'2 = 16 ms). No reversible tissue-induced dephasing was observed in liver (T'2 immeasurably long). In tissues which contain both fat and water, the chemically shifted 1H resonance peaks from -OH and -CH-are in phase with symmetric spin echo sequences but out of phase with asymmetric sequences. The interference of these two peaks produces a beat pattern with asymmetric sequences. Subtraction images obtained from paired symmetric- and asymmetric-sequence images accurately (r = .96) reflect T'2 and can be used to indicate the presence of fat. In vivo subtraction images of ethionine-induced fatty rat livers were significantly different from similar in vivo images of normal rat livers (P less than .0005). Since for each pixel of a subtraction image, the magnitude of the difference signal should be approximately proportional to the ratio of hydroxyl and alkyl protons, this simple spin echo sequence modification may obviate the need for more time-consuming 3-dimensional Fourier transform proton chemical shift images.

A new nuclear magnetic resonance property of lung

Inflated lung has a nuclear magnetic resonance (NMR) free-induction decay (FID) which is short compared with that of collapsed lung and those of other body tissues. An almost identically short FID is obtained from a slurry of 5-micron alumina particles in water. Interfaces between air and water in lung and between alumina and water in the slurry appear to be the source of spatial internal magnetic inhomogeneities which produce NMR line broadening and the short FID. Paired images that included lung, taken with paired symmetric and asymmetric NMR spin-echo sequences, permit the generation of an image, by subtraction, of the lung isolated from surrounding tissue. These new lung images are neither proton density, T1 (spin-lattice relaxation time), nor T2 (spin-spin relaxation time) images. They complement current NMR images and provide information about regional lung inflation. This previously unrecognized NMR property of lung tissue has potential application in NMR imaging, in quantitative determination of lung water and its distribution, and in the quantitation of regional lung inflation.

Determination of lung water content and distribution by nuclear magnetic resonance

The present study was designed to determine the value of nuclear magnetic resonance (NMR) imaging as a technique for quantifying lung water distribution and to estimate the degree of spatial resolution achieved by this technique. The spatial distribution of water was determined in six small (0.76 ml) rat lung tissue specimens by an NMR line-scan technique. After NMR imaging, each lung specimen was frozen and subdivided into slices; the gravimetric lung water content for each lung slice was compared with the integrated NMR water content over the volume corresponding to the same lung slice. In each tissue specimen, NMR and gravimetric lung water values were significantly correlated; the correlation coefficient for the pooled data for all six lung specimens was 0.91 (P less than 0.01). In two lung specimens, NMR values tended to be slightly higher than the gravimetric values. The magnitude of the difference between NMR and gravimetric values was generally less than 20% and only occasionally exceeded 25%. Our results suggest that the NMR-imaging method provides satisfactory estimates of lung water content and its distribution; the resolving power of the technique is excellent, as shown by its ability to detect water content differences between lung tissue slices of volume as small as 0.076 ml.

Lung water quantitation by nuclear magnetic resonance imaging

Nuclear magnetic resonance imaging was used to determine quantitatively the water distribution of saline-filled and normal rat lungs in both isolated lung and in situ preparations. Regional lung edema was easily detected. Studies of an isolated lung fragment indicate an accuracy of better than 1 percent and images of H2O/D2O phantoms indicate an average error of 2.7 percent.