JOURNAL OF MAGNETIC RESONANCE IMAGING 11:215–222 (2000) Technical Note In Vivo Hahn Spin-Echo Decay (Hahn-T2) Observation of Regional Changes in the Time Course of Oleic Acid Lung Injury Sumie Shioya, MD,1,3* Rebecca Christman, BS,1 David C Ailion, PhD,1 Antonio G Cutillo, MD,2 K Craig Goodrich, BS,1 and Alan H Morris, MD4 We studied the time course of changes in the Hahn spinecho decay (Hahn-T2) in lungs of spontaneously breathing living rats at hour, hours, and 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 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 and hours after injection The third, much longer, component is probably due to intraalveolar pulmonary edema After 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 J Magn Reson Imaging 2000;11: 215–222 © 2000 Wiley-Liss, Inc Index terms: lung; pulmonary edema; Hahn-T2 INTRODUCTION MRI IS A POTENTIAL analytical tool for the noninvasive study of pulmonary disease In a previous in vivo study of endotoxin-induced lung injury, we observed an increase in the value of the Hahn spin-echo decay time constant (so-called Hahn-T2) but no corresponding increase in proton density (1) These data suggest that NMR techniques can detect not only pulmonary edema, but also other experimentally induced pathologic lung changes, including those not associated with changes Department of Physics, University of Utah, Salt Lake City, Utah 84112 Division of Respiratory, Critical Care, and Occupational Pulmonary Medicine, Department of Internal Medicine, University of Utah, Salt Lake City, Utah 84132-0001 Department of Internal Medicine, Tokai University School of Medicine, Isehara, Kanagawa 259-1193, Japan Pulmonary Division, Department of Internal Medicine, LSD Hospital, Salt Lake City, Utah 84143 Contract grant sponsor: National Institutes of Health; Contract grant numbers: HL31216 and CA44972 *Address reprint requests to: S.S., Department of Internal Medicine, Tokai University School of Medicine, Isehara, Kanagawa 259-1193, Japan Received May 27, 1999; Accepted September 17, 1999 © 2000 Wiley-Liss, Inc in the water content Therefore, combined Hahn-T2 and water content measurements may be useful in the characterization of lung injury Because of the sensitivity of the two-dimensional Fourier transform (2DFT) techniques (2) to motional artifacts, we developed a rapid version of the line scan technique [the rapid line scan (RLS)] (3), in which successive line scans in a planar image selectively involve previously unexcited regions and therefore can be obtained in rapid sequence All line scan techniques have a significant advantage with respect to motion artifacts in that, after the acquisition of data from a given line, subsequent motions of the spins in the line will have no further effect However, in the normal (i.e., nonrapid or conventional) line scan technique, successive excitations of adjacent lines involve spins in previously excited regions and therefore require waiting a time of order T1, resulting in distortion due to respiratory and cardiac motions that may occur during the long T1 interval The rapid line scan (or RLS technique) (3) results in significantly reduced sensitivity to such motions by utilizing diagonal selective excitation of successive lines, so that such excitation involves previously unexcited spins and thus does not require a time of order T1 between them Since this time can be as short as a few milliseconds, a related advantage of the RLS technique is that the data required to form an image of an entire plane can then be acquired much more rapidly Possible interference from spurious echoes generated from previously excited spins can be eliminated by the use of spoiler gradients (3) One disadvantage of line scan techniques, including RLS, is that the spatial resolution is poor, in order to avoid interference from spins near the edge of a previously excited line This problem can be reduced, with some improvement in the spatial resolution, by interleaving the excitation order of line scans In this interleaved rapid line scan (or ILS) technique, enough time elapses between the excitation of spatially adjacent lines in the image so that the previously excited spins have more time to dephase, thereby allowing successive lines to be more closely spaced (4) With this technique we obtained Hahn spin-echo decay curves from lungs of 215 216 Shioya et al Table Experimental Protocol* Hahn-T2 measurement No of animals Control (n ϭ 7) s (E H) s s 3 2 Injection (n ϭ 11) hour (n ϭ 4) 2 2 hours (n ϭ 9) s s (D H) days (n ϭ 3) s s (E H) s (E H) s (E H) s (E H) *E ϭ euthanasia, D ϭ natural death, H ϭ histology, s ϭ Hahn-T2 measurement, ϭ oleic acid injection normal living rats and resolved these curves into two exponential components (4) In a subsequent comparative study (5), we found no significant differences between the values of the Hahn-T2 components obtained, respectively, from live, spontaneously breathing rats and from excised rat lungs Furthermore, we measured the Hahn-T2 in whole excised rat lungs in both the imaging mode (using the interleaved RLS technique) and the conventional nonimaging mode and found no significant differences between the results with these two techniques In the present study, we investigated the in vivo Hahn-T2 changes in oleic acid-injured lungs at various times following the injury Our goal was to determine whether the Hahn-T2 measured by in vivo NMR imaging can monitor and characterize the time course of oleic acid lung injury This type of experimental lung injury is one of the most widely used animal models of increased permeability pulmonary edema and is of clinical interest because of its pathophysiological similarities (especially in the acute phase) with the acute respiratory distress syndrome (ARDS) (6) The oleic acid injury model has been used extensively to investigate the effects of pulmonary edema on lung fluid balance, hemodynamics, mechanical properties, and gas exchange as well as to test the response of lung injury to various therapeutic strategies (6) MATERIALS AND METHODS Experimental Protocol In all, 23 imaging experiments were performed using 14 female Sprague-Dawley rats (250 –300 g) Details of the experimental protocol are in Table The rats were anesthetized by intraperitoneal injection of sodium pentobarbital with an initial dose of 60 mg/kg During the time course study, an additional dose (30 mg/kg intramuscularly) was given hours after the initial dose The rats received a bolus of 0.12 mg/kg oleic acid through the tail vein Control data were obtained in seven normal rats In four of these seven rats, the Hahn-T2 was measured both before oleic acid injection and hours after injection After the acquisition of the 3-hour MRI data, one of the four rats was sacrificed for histologic examination, and three were allowed to survive in order to obtain measurements again days later In four additional rats, the Hahn-T2 was measured at hour after oleic acid injection Two of these four rats died after the 1-hour MRI measurements, and two were again studied hours after injury Another three rats were added for Hahn-T2 measurements at hours after injection and for histology examination A larger number of animals were studied at hours after injury because at this stage we observed the most marked changes in proton density and in the Hahn-T2 decay (Table 2) On the whole, histological data were Table Time Course of Hahn-T2 Response in Oleic Acid-Injured Lungs Control No of animals Components No of ROIs Proton density (%) T2 component Short (msec) Medium (msec) (Mo %) Long (msec) (Mo %) Time hr (14) 9Ϯ2 (2) 15 Ϯ 10 9Ϯ1 31 Ϯ (9 Ϯ 4) 30 Ϯ (100) hr (10) 15 Ϯ 13 Ϯ 62 Ϯ 22* (28 Ϯ 20)** (3) 17 Ϯ 12 Ϯ 40 Ϯ (26 Ϯ 12) 103 Ϯ 69 (9 Ϯ 6) (6) 37 Ϯ 15* 28 Ϯ (100) days (23) 30 Ϯ 13* (6) 27 Ϯ 9* 20 Ϯ 9* 64 Ϯ 27* (24 Ϯ 17)** 14 Ϯ 49 Ϯ 19 (19 Ϯ 17) 105 Ϯ 34 (6 Ϯ 5) (8) 13 Ϯ 9Ϯ1 36 Ϯ 14 (25 Ϯ 19)*** Data are expressed as mean Ϯ standard deviation ROI ϭ region of interest, proton density ϭ relative signal intensity of lung tissue at an echo time of 16 msec expressed as a percentage of the signal intensity of a water phantom, Mo ϭ relative contribution of medium and long T2 components to the total signal intensity *P Ͻ 0.001 **P Ͻ 0.01 ***P Ͻ 0.05, compared with the corresponding control values Hahn-T2 Measurements of Lung Injury obtained from three normal control rats and from two rats at hour, six rats at hours, and three rats at days after injury NMR Imaging The NMR imaging studies were performed using a 2.35 T, 33 cm bore superconducting magnet (Oxford Instruments, Oxford, England), operating at 0.94 T (40.10 MHz) A 5-cm-diameter high-pass birdcage coil (7) was used The images were generated by the motion-insensitive ILS technique, as described earlier (4) Because of its essential features (diagonal excitation of the 90° and 180° selected planes, use of spoiler gradients to dephase spurious echoes, and interleaved excitation order), the ILS technique overcomes several limitations associated with lung imaging, artifacts induced by respiratory and cardiac motion, long imaging time (required by the conventional line scan technique), and low signal-to-noise ratio (SNR) due to the low proton density of lungs, at the price of lower spatial resolution (3,4) In the present experiments, each ILS image consisted of 16 lines with 128 voxels in each line The lines were positioned so that all lines included some lung information (i.e., regions that contained only chest wall were not imaged to save imaging time) Using a raised cosine radiofrequency pulse of 1.4 msec total width, we obtained a spatial resolution of 1.5 ϫ 1.5 ϫ 2.5 mm per voxel Since lung tissue has a short T2, the pulse was made as narrow as possible to allow for a short echo time for the Hahn-T2 determination while still providing adequate spatial resolution Each line in an image was repeated and averaged 32 times to improve the SNR The repetition time was seconds, resulting in an acquisition time of 4.3 minutes for each image or 70 minutes for a series of 16 images During the imaging studies, the anesthetized animals breathed spontaneously Because of the reduced sensitivity of our imaging tech- 217 nique to motion artifacts, respiratory gating was not used Figure 1a provides an example of an ILS image obtained from a normal rat lung The image shows no motion artifacts even in the absence of cardiac or respiratory gating Figure 1b presents, for comparison, an MRI image of the same lung acquired using the 2DFT technique It should be noted that the lower spatial resolution of the ILS images compared with that achievable by other techniques is not a significant limitation for the present quantitative MRI studies These studies did not require the high spatial resolution needed for conventional descriptive MRI because they were designed to obtain quantitative data from relatively large lung regions (see below) Furthermore, the use of a larger voxel size for our MRI measurements offered the advantage of providing more NMR signal per voxel in less time Measurements of Hahn T2 Hahn-decay curves were obtained from a series of 16 –18 spin-echo images with echo times ranging from 16 to 250 msec The echo time intervals were –50 msec The Hahn-decay time constant (8), measured in the present study, is called Hahn T2 to distinguish it from the T2 measurements obtained from the CarrPurcell-Meiboom-Gill (CPMG) technique (9,10) In the presence of diffusion across inhomogeneous magnetic fields (11,12), the Hahn-decay time constant may be much shorter than the CPMG T2 (which reflects the “true” T2 if the 180° pulse spacing is sufficiently short) (13,14) In this case the Hahn decay is a measure of the rate of diffusion as well as the strength of the magnetic gradients, whereas the CPMG decay will reflect only the underlying T2 if the pulses are sufficiently closely spaced If, however, the pulses are selective and not particularly closely spaced, then it is possible that CPMG will also reflect diffusion However, the shortness of the underlying T2 in lung (4) along with the time Figure a: ILS image of a normal rat lung, consisting of 32 lines b: The image is compared with an image of the same lung obtained by the conventional two-dimensional Fourier transform (2DFT) technique Note the absence of motion artifacts in the ILS image (acquired without cardiac or respiratory gating) 218 Shioya et al Figure a: ILS image of an oleic acid-injured lung hours after injection Since signal intensity was regionally nonuniform, measurements were performed in five different regions of interests (ROI) We chose ROIs in the peripheral lung tissue to avoid signal contributions from large blood vessels and the chest wall For the measurement of both the Hahn-T2 decay and proton density, the ROIs were chosen so that the signal intensity within each ROI was approximately uniform b: Hahn-T2 decay curves obtained from five different ROIs (A–E) in the lung shown in Fig 2a The Hahn-T2 behavior is nonuniform even between regions having the same signal intensity Curves C and D consisted of three exponential components (Hahn-T2 value for each component: 16.6, 49.6, and 114 msec for C and 14.7, 81.1, and 158 msec for D), whereas A, B, and E consisted of two components (36.6 and 97.7 msec for A, 42.5 and 65.8 msec for B, and 13.9 and 39.6 msec for E) required for the external imaging gradients would make it considerably more difficult to perform CPMG measurements in the imaging mode in the lung For these reasons, we chose in this study to examine the response of the Hahn echo decay time constant (rather than CPMG T2) to oleic acid injury For each image, the magnetization was determined from regions of interest (ROIs) in the lung tissue As shown in Figs 2a and 3a, we chose ROIs in the peripheral lung tissue to avoid signal contributions from large blood vessels and the chest wall Each ROI consisted of 10 – 40 voxels and sampled a relatively large portion of the lung (ϳ1 cm2 cross-sectional area) Accordingly, it seems likely that the relatively small displacements (about mm) due to the breathing motion resulted in very small errors due to lack of repeatability of the exact position of each voxel in the lung Each ROI was used for a series of 16 –18 images required for the Hahn-T2 measurements For the normal control and 7-day measurements, two ROIs were chosen for each rat, one in the right and one in the left lung, except for two experiments performed days after oleic acid injury, in which three ROIs were selected (Fig 3a) At and hours after oleic acid injection, Hahn-T2 data were generally obtained from four ROIs in each rat, two in the right lung and two in the left, avoiding large vessels The use of two ROIs for each image acquired in the control studies was considered to be adequate because under these experimental conditions the Hahn-T2 behavior was regionally uniform (The Hahn-T2 decay curve was always characterized by two exponential components, as shown in the Results section and in Table 2.) In Figure a: ILS image days after oleic acid injection The signal intensity is now regionally more uniform b: Hahn-T2 decay curves obtained from three different ROIs (A–C) in the lung shown in Fig 3a The Hahn-T2 behavior is uniform, the decay curves consisting of two components (Hahn-T2 value for each component, A: 9.3 and 43.4 msec; B: 11.0 and 47 msec; C: 9.3 and 62.2 msec) Hahn-T2 Measurements of Lung Injury contrast, the Hahn-T2 behavior of the lungs studied at and hours after injury varied regionally (the decay curves being characterized by one, two, or three components) Therefore, a larger number of ROIs was required to obtain adequately representative data To determine the average signal intensity for an ROI, the real and imaginary parts of the complex signal were first averaged separately (15), and then the magnitude was calculated This procedure avoids errors due to noise rectification, which occur when signal and noise are of comparable magnitude Thus the complex averaged signal reflects the actual magnetization, since the complex-averaged noise should approach zero when the number of voxels is large Using this method, we were able to determine the longest exponential component of the Hahn-T2 decay curve with improved accuracy The transverse magnetization decay obtained by the Hahn spin-echo method was described as a multiexponential function, Mϭ A i exp͑Ϫk i t͒, where M denotes the actual magnetization at time t, kiϪ is the time constant, and Ai is the relative magnetization characterizing each different component i The Hahn decay curves from ROIs placed on the lung tissue were resolved into one, two, or three Hahn-T2 components (5) In the presence of three components, the third component was subtracted from the original data, and the subtracted data were then fit with two-exponent functions Measurement of Proton Density To detect and quantify the changes in lung water content due to oleic acid injury, we determined the proton density by measuring the complex-averaged magnetization over each of the ROIs used for the Hahn-T2 measurements The relative magnetization at an echo time of 16 msec for peripheral lung tissue was obtained from the lung signal intensity and expressed as a percentage of the signal intensity of a water phantom The water phantom consisted of a plexiglass cylinder (45 mm in diameter, which was close to the rat’s chest size) filled with water doped with CuSO4 to decrease T1 and thereby prevent saturation The data from the water phantom were obtained each time using the same coil just before acquiring the Hahn-T2 data from the rat lungs The T1 and T2 values for the water phantom were approximately 350 and 250 msec, respectively Our proton density data were not corrected for signal losses due to the T2 decay and therefore not quantify absolute lung water content Back-extrapolation of the Hahn-T2 decay curves to correct for T2 losses would cause significant errors because of the multiexponentiality of these curves and because magnetic field inhomogeneity has different effects on the signals from lung and phantom On the other hand, several studies have shown that uncorrected proton density measurements closely reflect relative changes in lung water content, as demonstrated by the close correlation observed between these measurements and lung water content or 219 lung tissue volume density values obtained by gravimetric or morphometric techniques (16 –18) Since the purpose of our proton density measurements was to document relative changes in lung water content, rather than measure absolute lung water content, our procedure was quite adequate for the present study Histology After completion of the imaging studies, lungs were resected en bloc The trachea was cannulated The lungs were fixed by instillation of a 10% formaldehyde solution at 20 cm H2O fixative pressure Light microscopy samples were prepared by hematoxylin-eosin staining Statistical Analysis The changes in Hahn-T2 and proton density measured at various times during the experiments were evaluated statistically by analysis of variance with the Scheffe’s multiple comparison test and by linear regression analysis (19) A value of P Ͻ 0.05 was considered significant RESULTS Normal control values for proton density and Hahn-T2 were obtained from 14 ROIs in seven normal rats prior to the injection of oleic acid SNR values for normal peripheral lung tissue were 100 and at echo times of 16 and 90 msec, respectively The Hahn-T2 decay for each ROI was fit with a biexponential curve as described in previous studies (1,4) There were no significant differences in the two Hahn-T2 components between the right and left lungs Signal intensity was uniform, with 9% Ϯ 2% (mean Ϯ standard deviation) of the signal intensity for pure water for all ROIs in the normal peripheral tissue (Table 2) In contrast to the normal control lungs, which showed uniform signal intensity, the signal intensity of the lung tissue after oleic acid injection was regionally nonuniform The spatial variations in lung signal intensity appeared hour after oleic acid injection Figure 2a shows an ILS image of oleic acid-injured lung acquired hours after injection As in the other lungs, the regions of interest used for the Hahn-T2 and proton-density measurements were selected so that signal intensity was relatively uniform within each ROI Hahn-T2 decay curves obtained from five ROIs (A–E) in the lungs shown in Fig 2a are presented, as an example, in Fig 2b As shown in Fig 2b, Hahn-T2 measurements obtained and hours after injection showed a regionally nonuniform behavior, characterized by one, two, or three T2 components One hour after injection, the Hahn-T2 decay was monoexponential for of 15 ROIs, biexponential for 10, and triexponential for Three hours after injection, the Hahn-T2 decay was monoexponential for of 35 ROIs, biexponential for 23, and triexponential for Seven days after oleic acid injury, the regional nonuniformity in signal intensity was no longer present, and the lungs appeared normal (Fig 3a) At this stage, 220 Shioya et al Figure Examples of monoexponential (a) and biexponential (b) Hahn-T2 decays with exponential component fit The curve in b corresponds to decay A in Fig 2b obtained from ROI (A) in Fig 2a E, original data; F, short component; Œ, medium component The dotted line in 4b, which is the reconstructed line from two exponential components, fits the experimental data well As explained in the text, in the triexponential Hahn-T2 decay curves the third component was subtracted from the original data, and the subtracted data were then fitted with a two-exponential function the Hahn-T2 decay curves obtained from different ROIs were similar, as illustrated in Fig 3b, which shows the decay curves obtained from ROIs (A–C) in the lungs presented in Fig 3a Monoexponential and triexponential Hahn-T2 decay curves were not detected: in all eight ROIs sampled at this stage, the decay curves were biexponential Figure 4a and b shows examples of monoexponential and biexponential Hahn-T2 decays with exponential component fit The biexponential decay shown in Fig 4b was obtained from ROI A in Fig 2a In the biexponential Hahn-T2 decays, the value of the short T2 component increased hours after the oleic acid injection (P Ͻ 0.01, compared with baseline and 1-hour values) (Table 2) The medium Hahn-T2 component increased at and hours after injection (P Ͻ 0.001, compared with the control) The monoexponential Hahn-T2 value was similar to the medium Hahn-T2 value for normal lung The short and medium components of the triex- ponential decays were not significantly different from the corresponding control values The proton density (Table 2) increased significantly at hours compared with the control and then returned to nearly the control value after days As shown in Table 2, at each time after injury (1 hour, hours, and days) there was no significant difference in proton density among the groups characterized, respectively, by a one-, two- or three-component Hahn-T2 decay However, at hours after oleic acid injection there was a correlation between the proton density and the short Hahn-T2 component (n ϭ 29, correlation coefficient: r ϭ 0.534; P Ͻ 0.005) and between the proton density and the medium Hahn-T2 component (n ϭ 35, r ϭ 0.360; P Ͻ 0.05) This correlation was not observed at hour and days Figure 5a and b presents microscopic views of the lung tissue hours and days after oleic acid injection One hour after injection, we observed diffusely con- Figure a: Microscopic view of a lung hours after oleic acid injection (ϫ25, hematoxylin-eosin staining) The lung tissue shows nonuniform lung injury characterized by massive pulmonary edema at the periphery b: Microscopic view of a lung days after oleic acid injection The lung shows disappearance of pulmonary edema with focal pulmonary fibrosis Hahn-T2 Measurements of Lung Injury gested alveolar capillaries, irregularly distributed alveolar edema, hemorrhage, and septal necrosis At hours (Fig 5a), capillary congestion, edema, and septal necrosis became more marked Seven days after injection (Fig 5b), all acute lesions disappeared, with the exception of mild edema Some fibrotic areas were disseminated throughout the lungs DISCUSSION Lung injury induced by oleic acid, an 18-carbon unsaturated fatty acid with a single double bond, is one of the most widely used animal models of acute permeability pulmonary edema and fatty embolism (6) The early stage of oleic acid-induced lung injury is characterized by the formation of thrombosis and cellular necrosis The presence of severe interstitial and alveolar pulmonary edema is a typical feature of the injury during the acute stage The repair stage is characterized by proliferation of type cells and fibrotic foci in the subpleural areas of the lung (20) The histologic changes observed in the present study are consistent with those previously reported for the acute and repair stages of oleic acid lung injury In two previous studies, investigators have attempted to characterize the acute stage of oleic acid-induced injury using MRI Schmidt et al (21) detected pulmonary edema induced by oleic acid in live rats and estimated T2 using two images acquired with repetition times of 2.0 seconds and echo times of 28 and 56 msec In another study, Phillips et al (22) estimated T2 in mechanically ventilated cats after oleic acid lung injury by fitting a semilogarithmic straight line to four echo data points In these studies, they only estimated changes in T2 Using the ILS imaging technique, we measured the multiexponential Hahn spin-echo decay using 16 –18 MR images in living rats Our data describe in detail the regional changes of oleic acid injury over the whole time course of the injury, including the late recovery stage In the present study, a regionally uneven Hahn-T2 was detected as early as hour after injection The Hahn-T2 at and hours after injection was nonuniform throughout the lung and was characterized by one, two, or three T2-component exponential decays This finding may reflect the typical spatial nonuniformity of oleic acid lung injury, which was detectable in our histologic sections However, the observed spatial differences in the Hahn-T2 decay are not due to variations in lung water content because proton density, which reflects lung water content (16,23), did not vary significantly between lung regions exhibiting nonuniform Hahn-T2 behavior As indicated above, proton density and Hahn-T2 measurements for each group (1 hour, hours, and days) were obtained from the same ROIs The present data differ from our previous observations of the Hahn-T2 behavior over the time course of endotoxin lung injury (1) In endotoxin-injured lung tissue, the Hahn-T2 decays were exclusively biexponential, and the changes in T2 were uniform throughout the lung at each measurement time (1) Seven days after oleic acid injury, the Hahn-T2 behavior was again similar to that observed before injection 221 The monoexponential Hahn decay in oleic acid-induced lung injury may reflect the presence of extremely severe tissue damage, such as the necrosis observed in the histologic sections Under these conditions, the environment for lung tissue water may become more uniform due to severe disruption of the tissue architecture The biexponential Hahn-T2 decays measured hour after injection show a significant increase in the medium T2 component with no corresponding increase in proton density This observation is of interest because it indicates that the early stage of lung injury prior to the development of overt pulmonary edema may be detectable by Hahn-T2 measurements Our data are in agreement with the results of previous in vitro and in vivo studies of endotoxin-injured lungs, which showed significant Hahn-T2 changes with no corresponding variations in lung water content (1,24) Proton density increased hours after oleic acidinduced injury This change was associated with an increase in the values of the two Hahn-T2 exponential components A third much longer component was observed in the Hahn-T2 decays obtained and hours after injury The value of this component did not change between and hours Similarly, in a previous in vitro study of oleic acid-injured lungs, we observed a third T2 component that had a value close to that measured in exudates obtained from the edematous lungs (25) In the histologic sections obtained from the injured lungs, irregularly distributed alveolar edema was observed hour after oleic acid injection and became more marked hours after injection Therefore, the third T2 component measured in vivo may reflect the development of a third water space, probably alveolar edema Regional changes in lung density in oleic acid-injury have been demonstrated by computed tomography (CT) Hedlund et al (26) reported that oleic acid infusion in dogs produced a patchy and predominantly peripheral increase in lung density 1– hours after infusion (26) In accordance with the results of conventional CT studies, our ILS images of oleic acid-injured lungs also showed regional variations in signal intensity The resolution of MR images obtained using the ILS technique is much lower than that of high-resolution CT (27) However, while the resolution achievable by the ILS technique is adequate for quantitative lung MR imaging, as discussed above, this technique provides other criteria, in addition to proton density, for the assessment of lung injury, for example, the measurements of Hahn-T2 decay Using the ILS technique, lung regions with the same NMR signal intensity can be further characterized on the basis of their Hahn-T2 behavior Because the time needed to acquire the Hahn data by the imaging strategy adopted for the present study was long (about 70 minutes), we studied the course of oleic acid injury at relatively wide time intervals (1 hour, hours, and days after injury) Our data (see, for example, Table 2) indicate that the severity of the oleic acid-induced injury differed substantially between the stages selected for the present study Therefore, we believe that the time resolution of our technique was adequate for the purposes of our study However, more rapid Hahn-T2 measurements would be desirable for practical applications and to monitor lung injury at 222 more closely spaced time intervals Substantial reduction in the imaging time can be obtained by using fewer but larger voxels (at the expense of spatial resolution), by reducing the repetition time somewhat (with corresponding loss of signal due to T1 effects), or by decreasing the number of averages (with corresponding loss of SNR) In conclusion, the present data suggest that in vivo Hahn-T2 measurements can detect and follow the time course and regional distribution of experimentally induced pulmonary edema This application further increases the potential of NMR as a noninvasive approach to the characterization of lung injury in intact living animals Furthermore, the present in vivo Hahn-T2 measurements represent an additional step toward the application of MRI techniques to the assessment of pulmonary edema in humans Presumably, the Hahn-T2 changes reported in the present study are not strictly specific of oleic acid lung injury but reflect the structural abnormalities detectable in pulmonary edema at comparable stages of evolution As discussed above, our data suggest that Hahn-T2 measurements in diseased lungs may provide independent information that complements the results of current measurements of lung water content Significant Hahn-T2 changes are detectable when conventional (proton density) NMR measurements of lung water content are still within normal range (Table 2) Therefore, MRI methods for measuring Hahn-T2 might facilitate the detection of edemagenic pulmonary disease at an early stage However, the application of T2 measurements to the detection and characterization of clinical pulmonary disease requires further investigation and, in particular, the implementation of faster imaging techniques REFERENCES Shioya S, Christman R, Ailion DC, Cutillo AG, Goodrich KC Nuclear magnetic resonance Hahn spin-echo decay [T2] in live rats with endotoxin lung injury Magn Reson Med 1993;29:441– 445 Edelstein WA, Hutchinson JMS, Johnson G, Redpath T Spin warp NMR imaging and applications to human whole-body imaging Phys Med Biol 1980;25:751–756 Ailion DC, Ganesan K, Case TA, Christman R Rapid line scan technique for artifact-free images of moving objects Magn Reson Imaging 1992;10:747–754 Shioya S, Christman R, Ailion DC An in vivo NMR imaging determination of multiexponential Hahn T2 of normal lung Magn Reson Med 1990;16:49 –56 Shioya S, Kolem H, Ailion DC, et al Comparison of in vivo and in vitro Hahn T2 measurements in rat lung Magn Reson Med 1992; 26:1– 6 Schuster DP ARDS: clinical lessons from the oleic acid model of acute lung injury Am J Respir Crit Care Med 1994;149:245–260 Hayes CE, Edelstein WA, Schenck JF, Mueller OM, Eash M An Shioya et al 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 efficient, highly homogeneous radiofrequency coil for whole-body NMR imaging at 1.5T J Magn Reson 1985;63:622– 628 Hahn EL Spin echoes Phys Rev 1950;80:580 –594 Carr HY, Purcell EM Effects of diffusion on free precession in nuclear magnetic resonance experiments Phys Rev 1954;94:630 – 638 Meiboom S, Gill D Modified spin-echo method for measuring nuclear relaxation times Rev Sci Instrum 1958;29:688 – 691 Case TA, Durney CH, Ailion DC, Cutillo AG, Morris AH A mathematical model of diamagnetic line broadening in lung tissue and similar heterogenous systems: calculations and measurements J Magn Reson 1987;73:304 –314 Laicher G, Ailion DC, Cutillo AG Water self-diffusion measurements in excised rat lungs J Magn Reson S B 1992;111:243–253 Shioya S, Haida M, Kurita D, et al Dependence of transverse relaxation time T2 of biologic tissues on the interpulse delay time in Carr-Purcell-Meiboom-Gill (CPMG) measurements Tokai J Exp Clin Med 1997;22:27–31 Ganesan K, Ailion D, Hackmann A, et al NMR relaxation and water self-diffusion mechanism in lung In: Cutillo AG, editor Application of magnetic resonance to the study of lung Armonk: Futura; 1996 p 115–139 Shioya S, Christman R, Case TA, Goodrich KC, Ailion DC Complex signal averaging to avoid errors in T2 determination Magn Reson Imaging 1989;7(Suppl 1):165 Cutillo AG, Morris AH, Ailion DC, et al Assessment of lung water distribution by nuclear magnetic resonance A new method for quantifying and monitoring experimental lung injury Am Rev Respir Dis 1988;137:1371–1378 Cutillo AG, Goodrich KC, Ganesan K, et al Lung water measurement by nuclear magnetic resonance: correlation with morphometry J Appl Physiol 1995;79:2163–2168 Cutillo AG, Ganesan K, Ailion DC, et al Determination of lung water content and distribution by NMR techniques (especially MRI) In: Cutillo AG, editor Application of magnetic resonance to the study of lung Armonk: Futura; 1996 p 205–226 Zar JH Multiple comparisons In: Biostatistical analysis, 2nd ed New York: Prentice-Hall; 1984 p 185–205 Derks CM, Jacobovitz-Derks D Embolic pneumopathy induced by oleic acid A systematic morphologic study Am J Pathol 1977;87: 143–158 Schmidt HC, Tsay D, Higgins CB Pulmonary edema: an MR study of permeability and hydrostatic types in animals Radiology 1986; 158:297–302 Phillips DM, Allen PS, Man SPF Assessment of temporal changes in pulmonary edema with NMR imaging J Appl Physiol 1989;66: 1197–1208 Cutillo AG, Morris AH, Blatter DD, et al Determination of lung water content and distribution by nuclear magnetic resonance J Appl Physiol 1984;57:583–588 Shioya S, Haida M, Tsuji C, et al T2 of endotoxin lung injury with and without methylprednisolone treatment Magn Reson Med 1990;15:201–210 Shioya S, Haida M, Fukuzaki M, et al Nuclear magnetic resonance study of lung water compartments in the rat Am J Physiol 1997; 272:L772–L778 Hedlund LW, Vock P, Effmann EL, Putman CE Morphology of oleic acid-induced lung injury Observations from computed tomography, specimen radiography, and histology Invest Radiol 1985;20: 2– Murata K, Herman PG, Khan A, et al Intralobular distribution of oleic acid-induced pulmonary edema in the pig Evaluation by high-resolution CT Invest Radiol 1989;24:647– 653 ... images in living rats Our data describe in detail the regional changes of oleic acid injury over the whole time course of the injury, including the late recovery stage In the present study, a regionally... study, we investigated the in vivo Hahn- T2 changes in oleic acid- injured lungs at various times following the injury Our goal was to determine whether the Hahn- T2 measured by in vivo NMR imaging can... previous observations of the Hahn- T2 behavior over the time course of endotoxin lung injury (1) In endotoxin-injured lung tissue, the Hahn- T2 decays were exclusively biexponential, and the changes in