Assessment of Myocardial Viability by Intracellular 23Na Magnetic Resonance Imaging
Background— Because of rapid changes in myocardial intracellular Na+ (Na+i) during ischemia and reperfusion (R), 23Na magnetic resonance imaging (MRI) appears to be an ideal diagnostic modality for early detection of myocardial ischemia and viability. So far, cardiac 23Na MRI data are limited and mostly concerned with imaging of total Na+. For proper interpretation, imaging of both Na+i and extracellular Na+ is essential. In this study, we tested whether Na+i imaging can be used to assess viability after low-flow (LF) ischemia.
Methods and Results— Isolated rat hearts were subjected to LF (1%, 2%, or 3% of control coronary flow) and R. A shift reagent was used to separate Na+i and extracellular Na+ resonances. Acquisition-weighted 23Na chemical shift imaging (CSI) was alternated with 23Na MR spectroscopy. Already during control perfusion, Na+i could be clearly seen on the images. Na+i image intensity increased with increasing severity of ischemia. During R, Na+i image intensity remained highest in 1% LF hearts. Not only did we find very good correlations between Na+i image intensity at end-R and end-diastolic pressure (R=0.85, P<0.001) and recovery of the rate-pressure product (R=−0.88, P<0.001) at end-R, but most interestingly, also Na+i image intensity at end-LF was well correlated with end-diastolic pressure (R=0.78, P<0.01) and with recovery of the rate-pressure product (R=−0.81, P<0.01) at end-R. Furthermore, Na+i image intensity at end-LF was well correlated with creatine kinase release during R (R=0.79, P<0.05) as well as with infarct size (R=0.77, P<0.05).
Conclusions— These data indicate that 23Na CSI is a promising tool for the assessment of myocardial viability.
Received June 24, 2004; revision received August 19, 2004; accepted August 26, 2004.
Assessment of myocardial viability in patients with chronic coronary artery disease or acute and subacute myocardial infarction is clinically important for distinguishing stunned or hibernating myocardium from irreversibly injured myocardium.1 Patients may benefit from revascularization when viable tissue is present in the dysfunctional area of the myocardium.
Several clinical imaging modalities exist for assessment of viable myocardium, eg, 201Tl scintigraphy, positron emission tomography,2 stress echocardiography,3 and delayed contrast-enhanced magnetic resonance imaging (DCE-MRI),4 which have proven useful in the chronic situation. In the subacute situation, however, a reliable technique for the assessment of myocardial viability does not exist.
Previous research has shown that one of the best ways to determine myocardial viability is to evaluate sarcolemmal function, in particular, Na+,K+-ATPase function.4 Intracellular sodium (Na+i) is a sensitive marker of myocardial ischemia and Na+,K+-ATPase function. During the very first minutes of myocardial ischemia, Na+i already starts to rise because of inhibition of the sarcolemmal Na+,K+-ATPase and continued or increased influx through the Na+ channel,5 via the Na+-H+ exchanger,6 and possibly through other routes. This very early rise in Na+i makes it a very sensitive indicator of ischemia. Recovery of Na+,K+-ATPase function after an ischemic episode is essential for recovery of myocardial function. Reperfusion (R) of myocardium after a brief (20-minute) period of ischemia causes an immediate (within 5 seconds) reactivation of Na+,K+-ATPase function and a decrease in Na+i7, which suggests that Na+i is also a very sensitive indicator of myocardial viability.
23Na MRI is a noninvasive methodology for measuring in vivo Na+, but so far, studies using this technique are scarce and mostly concerned with imaging of total Na+.8 After acute ischemia and R, increased total Na+ MRI intensity was shown to be associated with nonviable tissue.9 However, depending on the severity and duration of the ischemic insult, information on both Na+i and extracellular Na+ (Na+e) may be crucial. For instance, when coronary flow is entirely absent, the total-Na+ signal intensity may not change, whereas Na+i signal intensity may increase, or, in the case of extracellular edema, Na+e signal intensity may increase because of a larger extracellular space while Na+i signal intensity may be unaltered.
Recently, Weidensteiner et al10 showed that in ischemic myocardium, an increase in [Na+]i can be detected by a 3D chemical-shift imaging (CSI) technique. The time required to measure one 3D dataset was 30 minutes. We have adapted a 2D CSI technique to measure the spatial distribution of Na+i, Na+e, and total Na+ in <5 minutes, making relevant successive measurements feasible. Preliminary experiments with this technique showed that Na+i imaging is superior to total-Na+ imaging in identifying the affected zone after complete coronary occlusion.11 Increased Na+i signal intensity was already apparent from 0 to 15 minutes of left ascending coronary artery ligation of an isolated heart, whereas total Na+ signal intensity had not changed.
Clinically, ischemia may represent a variety of residual flow conditions. Therefore, in the present study, cardiac images of isolated rat hearts were obtained during low-flow (LF) ischemia at various flow rates and subsequent R. The aims of this study were to explore the potential of Na+i imaging for the assessment of myocardial viability and to evaluate the added value of Na+i MRI over total Na+ MRI.
Preparation of Isolated, Perfused Rat Hearts
Male Wistar rats (300 to 350 g) were anesthetized with diethylether inhalation and treated with heparin (1000 U/kg IV). The heart was rapidly excised and placed in ice-cold Krebs-Henseleit (KH) buffer. The aorta was cannulated, the left ventricle was vented, and constant 76–mm Hg perfusion was initiated with KH at 37°C. The KH, containing (in mmol/L) 125 NaCl, 4.7 KCl, 1.0 MgCl2, 25 NaHCO3, 1.75 CaCl2, 0.5 EDTA, 10 d-glucose, and 0.5 pyruvate, was filtered before use through a 0.8-μm filter and was equilibrated with 95% O2 and 5% CO2, resulting in a pH of 7.4.
A balloon was inserted into the left ventricle and connected to a pressure transducer. Left ventricular end-diastolic pressure (LVEDP) was set to 7 mm Hg. Developed pressure was calculated as the difference between peak systolic pressure and LVEDP. All hearts were paced (5 Hz) with agar-wick electrodes connected to a stimulator. Hearts subjected to regional ischemia were instrumented with an apex cannula instead of an intraventricular balloon. The experimental protocol was in accord with the guidelines of the Committee for Animal Experiments of the University Medical Center, Utrecht, the Netherlands.
Hearts in a 20-mm-diameter nuclear MR (NMR) tube were lowered into a 9.4-T, vertical-bore magnet, which was interfaced with a Bruker AVANCE 400 DRX spectrometer (Bruker Biospin) equipped with a 1000-mT/m microimaging gradient accessory.
1H MRI of Isolated, Perfused Hearts
The 20-mm-diameter birdcage coil was tuned to 400.15 MHz. Gated long- and short-axis 1H 2D gradient-echo images (matrix size, 256×256; field of view, 20×20 mm; slice thickness, 2.5 mm; echo time, 7 ms; repetition time, 200 ms) were acquired at the beginning and the end of the protocol to serve as anatomic references (see Figure 1). The short-axis slice was taken through the middle of the left ventricle according to the long-axis image.
23Na MR spectroscopy and CSI of Isolated, Perfused Hearts
After 1H MRI, the 1H resonator was replaced by a quadrature-driven 23Na birdcage resonator, which was tuned to 105.85 MHz while the heart stayed in place. To separate Na+i and Na+e resonances, perfusion was switched to a KH that was modified by removal of EDTA and addition of the sodium salt of the shift reagent, thulium(III)1,4,7,10-tetraazacyclododecane-N,N′,N″,N‴- tetra(methylenephosphonate) (TmDOTP5−, 3.5 mmol/L). [Na+] was kept constant at 150 mmol/L by adjusting the amount of NaCl. To correct for Ca2+ binding by the shift reagent, the total Ca2+ added to the perfusate was increased to 3.42 mmol/L, which resulted in a free [Ca2+] of 0.85 mmol/L as measured by a Ca2+-sensitive electrode.
23Na MR spectra were acquired with a 5-kHz spectral width and 2048 data points. Typically, 128 free induction decays were accumulated after 90° pulses with a repetition time of 250 ms. Free induction decays were Fourier-transformed after gaussian multiplication. After polynomial baseline correction, the Na+i and reference Na+ peaks were integrated. All 23Na spectra were fully relaxed. Quantification was performed relative to the peak area of the reference resonance. The reference capillary contained a known amount of 250 mmol/L Na+ and 5 mmol/L TmDOTP5−. NMR visibility of all 23Na signals was assumed to be 100%.8
The 2D 23Na CSI dataset was acquired in 4 minutes, 18 seconds, with a repetition time of 27.4 ms (matrix size, 16×16; field of view, 20×20 mm; 256 complex data points in the spectral domain; spectral width, 5 kHz). The slice was selected with a 1000-μs sinc pulse and was centered on the axial 1H slice. The delay before acquisition was kept as short as possible and amounted to 307 μs. We used acquisition weighting according to a sine-bell function to improve the spatial response function and hence, minimize signal contamination between voxels.12 The number of acquisitions in the middle row of the symmetrical acquisition matrix was as follows: 0, 6, 24, 50, 78, 104, 122, 128, 128, 122, 104, 78, 50, 24, 6, and 0 (total number of acquisitions, 9408). The acquisition weighting also resulted in a higher signal-to-noise ratio in the same amount of time because of more efficient k-space sampling. The nominal resolution was 1.25 mm in the x and y directions and 5 mm in the z direction (voxel size, 7.8 μL). Image resolution was improved by zero-filling in the spatial domain (resulting in a 64×64 matrix) and gaussian filtering in the spectroscopic domain. To calculate the average image pixel intensity of the Na+ standard, the buffer, and myocardial Na+i and Na+e, appropriate regions were chosen on the 1H image and copied onto the respective 23Na images. Images with the highest average pixel intensity of each of the different Na+ resonances were chosen. The intensities were expressed as percentage of the intensity of the buffer.
Hearts were allowed to stabilize, and meanwhile, tuning and matching of the 1H resonator were optimized, as well as the static magnetic field. After 1H MRI, the 1H resonator was replaced by the 23Na birdcage resonator. The perfusate was switched to the TmDOTP5−-containing KH, tuning and matching of the 23Na resonator coil were performed, and homogeneity of the magnetic field was optimized. After 25 minutes of control perfusion, the constant-pressure perfusion was switched to constant-flow perfusion, and the flow was adjusted to 1%, 2%, or 3% of coronary flow during the control period (n=4 hearts per each LF group). After 60 minutes, R was started by switching the perfusion mode back to constant pressure for 45 minutes. 1H MRI, 23Na MR spectroscopy, and 23Na CSI measurements were carried out as depicted in Figure 1.
To investigate the spatial correlation between infarct and elevated Na+i image intensity, hearts (n=5) were subjected to regional ischemia for 40 minutes and R. To this end, a dual-perfusion cannula was used that allowed independent perfusion of the left and right coronary bed.13 During the stabilization period, the left side of the heart was perfused with a Gd-DTPA-BMA–containing perfusate (Omniscan, Amersham Health), and a T1-weighted 1H image was acquired to assess the area at risk. Total ischemia of only the left side of the heart was induced for 40 minutes (flow to the other side remained unaltered), followed by R. After 2 hours, the right side of the heart was perfused with methylene blue to determine the area at risk for histology. After that, the whole heart was perfused with triphenyl tetrazolium chloride (TTC) to stain the viable tissue, as described below.
Assessment of Viability
To assess myocardial viability, recovery of the rate-pressure product (RPP=left ventricular developed pressure×heart rate) and LVEDP were analyzed. In an additional group of experiments (1 or 2% LF and R; n=4 hearts per group), the hearts underwent the same protocol as before except that the reperfusion period was extended to 2 hours to allow reliable infarct size determination with TTC staining (Ytrehus et al14 and http://www.southalabama.edu/ishr/help/ttc/). After this period, hearts were perfused with 1% TTC added to the perfusate for 10 minutes at a constant perfusion flow of ≈2 mL/min. Hearts were then frozen and cut in five 2.5-mm-thick slices, which were immersed in 10% formalin for 10 minutes. The slices were photographed, and infarct size was determined by planimetry and expressed as a percentage of total myocardial area. The effluent of these hearts was collected during the first 45 minutes of R and analyzed for creatine kinase (CK) activity at 25°C (Roche Diagnostics GmbH).
The data are summarized as the mean±SEM. A repeated-measures ANOVA compared measurements within each group. Differences between groups were analyzed by 1-way ANOVA. The test between any 2 means used Fisher’s least protected significant difference test. Differences were declared statistically significant when P<0.05. Statistical computations were performed with SPSS software.
All groups of hearts displayed similar left ventricular pressures and thus, RPPs during control perfusion (Figure 2A). During LF ischemia, the time to onset of contracture and the time of maximal contracture were earliest in 1% LF hearts (Figure 2B). Also, the amplitude of contracture was larger in the 1% LF hearts than in the 2%and 3% LF hearts. In the 3% LF group, contracture during ischemia occurred in only 1 heart. An additional increase in LVEDP occurred during R, and LVEDP remained elevated during R in all groups. Recovery of the RPP at end R was significantly less in 1% LF hearts compared with 2% and 3% LF hearts: 8.1±3.3% versus 68±21% and 81±29% (P<0.05), respectively. Coronary flow rate was 48.9±3.0 mL · min−1 · g dry wt−1 during control perfusion. At end R, recovery of coronary flow rate was worst in 1% LF hearts (21±5% of control coronary flow rate) and was significantly different from the coronary flow rate in 2% and 3% LF hearts (40±7% and 49±10%, respectively, P<0.05).
23Na MR Spectroscopy
Stacked plots of 23Na MR spectra of isolated rat hearts subjected to 3% or 1% LF are shown in Figure 3A and 3B, respectively. The Na+e resonance shifted ≈2 ppm downfield as a result of interaction of Na+e with the shift reagent, leaving the Na+i resonance well separated. It is clearly shown that the increase in Na+i resonance was much more pronounced in the 1% than in the 3% LF hearts. After 60 minutes of LF ischemia, [Na+]i had increased to 219±46% in 3% LF hearts (NS compared with the preischemic value), to 357±35% in 2% LF hearts (P<0.05 compared with the preischemic value), and to 396±18% in 1% LF hearts (P<0.05 compared with the preischemic value; Figure 4). In comparison, in our laboratory, the [Na+]i increase after 20 and 30 minutes of total ischemia had been found to be ≈200%15 and 300%,6 respectively. Recovery of [Na+]i during R was worst in 1% LF hearts: the average [Na+]i remained highly elevated (321±41% at 40 minutes of R, P<0.05 compared with the preischemic value). In 2% and 3% LF hearts, [Na+]i decreased during R to 218±12% (NS compared with end-LF value) and to 169±56% (P<0.05 compared with end-LF value), respectively, but remained elevated.
23Na Chemical Shift Imaging
Short-axis Na+i and Na+e CS images, obtained in 4 minutes, 18 seconds, of 3 hearts subjected to different degrees of ischemia are shown in Figure 5. Na+i was already visible during control perfusion. It is clearly shown that Na+i image intensity increased with increasing severity of ischemia. During R, Na+i image intensity in the 1% LF hearts remained high, whereas that of the 2% LF hearts decreased but remained elevated. In the 3% LF group, 2 hearts showed no apparent changes in Na+i image intensity during the entire protocol (Figure 5), whereas the other 2 showed an increase in Na+i image intensity during ischemia. However, of these 2 hearts, only 1 showed recovery of Na+i image intensity during R. Quantitative analysis of Na+ images was done at 3 time slots: control perfusion, after 55 to 60 minutes of LF (end-LF), and after 35 to 40 minutes of R (end-R). Average total (cardiac) Na+ image intensity did not show significant changes during the protocol in all groups (Figure 6A), and Na+e image intensity was significantly lower at the end of R compared with control perfusion in the 1% LF group only (Figure 6B). Na+i image intensity, however, increased to 222±70% in the 3% LF group, to 291±47% in the 2% LF group, and to 394±25% in the 1% LF group (Figure 6C), findings that are consistent with the 23Na MR spectroscopy data. During R, Na+i image intensity decreased in all groups but remained highest in the 1% LF hearts (309±34% versus 171±18% in 2% LF hearts and 136±61% in 3% LF hearts).
23Nai MR Spectroscopy/CSI and Viability
There was a very good correlation between [Na+]i measured by MR spectroscopy and the intensity of Na+i on 23Na images at the 3 analyzed time points (control, end-LF, and end-R): R=0.91, P<0.01 (n=36). To evaluate whether changes in [Na+]i or 23Na image intensities were related to recovery of hemodynamic performance and viability, we looked at the values of parameters of individual hearts rather than averages of the different groups. Na+i image intensity at end-R was correlated very well with RPP at end-R (R=−0.85, P<0.001, Figure 7A) and with LVEDP at end-R (R=0.88, P<0.001, Figure 7B). More important, however, there was also a good correlation between Na+i image intensity at end-LF with LVEDP and with RPP at end-R (Figure 7C and 7D, R=0.78, P<0.01; and R=−0.81, P<0.05, respectively). This shows that Na+i image intensity can predict the ability of myocardial tissue to recover after ischemia. A similar very good correlation was found when 23Na MR spectroscopy data were analyzed: [Na+]i at end-R was correlated with LVEDP and inversely with RPP at end-R (R=0.93, P<0.001; and R=−0.91, P<0.001, respectively). Again, [Na+]i at end-LF was correlated very well with LVEDP (R=0.85, P<0.01) at end-R and inversely with RPP (R=−0.86, P<0.01) at end-R, emphasizing the predictive value of Na+i in the assessment of myocardial viability.
Infarct size measured by TTC staining was significantly smaller in the 2% LF group compared with the 1% LF group (15±5.8% versus 48±12%, P<0.05). The same was true for CK release during the first 45 minutes of R: 63±18 versus 132±25 IU/g dry wt (P<0.05). Infarct size was correlated very well with CK release (R=0.95, P<0.001). Importantly, Na+i image intensity at end-LF was well correlated with CK release during R (R=0.79, P<0.05) as well as with infarct size (R=0.77, P<0.05).
Figure 8 shows short-axis Na+i images of a heart subjected to acute regional ischemia (B–F) and R (G–J) and the corresponding T1-weighted 1H (A) and TTC (K) images. Na+i image intensity increased significantly during ischemia of the left side, whereas that of the right side remained unaltered. Total Na+ image intensity remained unaltered during the entire protocol in both sides of the heart. As shown in Figure 8J and 8K, a good spatial correlation between the unstained area on the TTC image and the enhanced signal area on the Na+i image is present. Interestingly, during R, Na+i image intensity in the peri-infarct zone diminished, whereas it remained highly elevated in the infarct core.
In this study, the value of intracellular 23Na MRI for assessment of myocardial viability was explored in rat heart models of global LF ischemia and total regional ischemia. We found a very good correlation between Na+i image intensity at end-R after LF ischemia and recovery of LVEDP and RPP at the same time. More important, however, there was also a good correlation between Na+i image intensity at end-LF, with recovery of LVEDP and RPP at end-R. This shows that Na+i image intensity can predict the ability of myocardial tissue to recover after ischemia. Furthermore, Na+i image intensity at end-LF was well correlated with CK release during R as well as with infarct size. From regional ischemia experiments, a good spatial correlation between Na+i image intensity and infarcted area was apparent. Total 23Na image intensities, however, did not change during ischemia and R, indicating that under these conditions, total 23Na MRI cannot be used to assess viability. This is the first study that shows the potential of intracellular 23Na MRI as a diagnostic modality for early detection of myocardial ischemia and viability.
23Na MRI of an isolated rat heart was first reported by DeLayre et al.16 The heart could be visualized because of the lack of signal from the ventricular wall compared with signal from the ventricular cavities and buffer surrounding the heart. The technique has been improved since then, and total 23Na imaging has been performed on the hearts of rats,17 rabbits,9 and dogs9,18,19 as well as humans.20,21
23Na CSI with the aid of a shift reagent makes it possible to independently analyze the spatial distribution of Na+i, Na+e, and total Na+ and therefore allows for a more precise understanding of pathophysiological changes in ion homeostasis during ischemia and R. In the regional ischemia experiments, Na+i imaging appeared superior to total Na+ imaging in identifying the affected zone after induction of ischemia, as we had also found before, after acute left coronary artery ligation of isolated rat hearts.11 The reason for this was that the occlusion was total, and total Na+ remained unchanged because of an apparently undeveloped collateral circulation, whereas Na+ flux into the intracellular compartment exceeded Na+ efflux as a result of ischemia. We anticipated that during LF ischemia, the situation might be totally different. Total Na+ intensity was expected to rise during LF ischemia, because [Na+]i will rise and [Na+]e will remain constant, in contrast to conditions of total occlusion, wherein Na+i image intensity will also rise, but at the expense of Na+e image intensity, resulting in no change of total Na+ image intensity. However, we observed no increase in total Na+ and a decrease in Na+e image intensities only in the group in which ischemia was most severe, indicating that in this group, the flow rate was insufficient to refresh the perfusate in the extracellular space. It is important to note that this decrease in total Na+ intensity was not beneficial, as might be falsely concluded if Na+i images would not have been available. The behavior of the Na+i image intensities during the protocol was as anticipated: the more severe the degree of ischemia, the larger the increase in [Na+]i and corresponding Na+i image intensity. This latter intensity appeared to be highly predictive for functional recovery during R. Ca2+ accumulation in the myocardial cell appears to play a central role in the damage that occurs during R of ischemic myocardium.22 One of the most important routes for Ca2+ influx during ischemia and in particular, during R is (reversed) Na+-Ca2+ exchange, and therefore, increased [Na+]i may underlie a significant portion of this damage.23
Thus, imaging of both Na+i and Na+e is essential for proper interpretation of total cardiac Na+. Furthermore, formation of edema can lead to increased total Na+ intensity (because of the extension of the extracellular space), whereas [Na+]i may not be increased, indicating that the cardiomyocytes are still viable.20 In other words, the discriminating power of the CSI technique is much higher (see the Table) compared with total 23Na and other imaging techniques.
In cardiac tissue, the concentration and NMR sensitivity of the 23Na nucleus are much lower than that of 1H, leading to a 20 000 to 30 000 times lower signal compared with that of 1H in water. This disadvantage can partially be overcome by using larger voxel sizes, longer acquisition times, and faster pulse sequences. The latter can be used because of the shorter relaxation times of 23Na. Using a 16×16 matrix and acquisition weighting, we were able to acquire a full CSI dataset (256 phase-encoding steps) with a reasonable signal-to-noise ratio of the Na+i image in 4 minutes, 18 seconds. The disadvantage of the acquisition weighting is a slightly worse spatial resolution, but this is outweighed by the resulting improvement of the spatial response function, which prevents Gibbs ringing, and a timesaving effect.
The intensity of the (intracellular or extracellular) Na+ signal depends on (1) the Na+ content of the tissue, (2) MR visibility, and (3) the size of the intracellular and extracellular compartments. With regard to (1), myocardial ischemia results in an increase in [Na+]i because of continued or enhanced Na+ influx and reduced Na+ pumping.24 Suppose that during LF ischemia [Na+]i increases by 100% (from 10 to 20 mmol/L) and [Na+]e remains constant (150 mmol/L). Then, with relative sizes of the intracellular and extracellular space of 75% and 25%, respectively,25 Na+i image intensity doubles but total Na+ image intensity will increase by only ≈17%. This shows that Na+i imaging is much more sensitive than total Na+ imaging and might explain why we did not find any significant changes in total Na+ intensity during our protocol.
With regard to (2), because of transverse (T2) relaxation during the 1000-μs pulse and the 307-μs postpulse delay in our CSI pulse sequence, signal loss occurs, which, based on our previous studies,15 we calculated to result in a reduction of MR visibility of ≈11%. In the MR spectroscopy experiments, the loss was <1%, resulting in a visibility of >99%. However, the correlation coefficients calculated from the CSI data were on the same order as those calculated from the spectroscopy data, indicating that neither the signal loss resulting from the unavoidable delay in the pulse sequence nor the use of magnitude images for the determination of Na+i image intensity affected our results. Furthermore, the 23Na CSI sequence used on 4% agar solution with different Na+ concentrations (20, 50, 100, 150, and 200 mmol/L NaCl) showed an excellent correlation between Na+ concentration and 23Na image intensity (R=0.996, P<0.001, data not shown).
With regard to (3), when the sizes of intracellular and extracellular space change as a result of an intervention, 23Na image intensity will be affected. Either or both the intracellular and extracellular spaces may change as a result of edema and contracture. However, the size of the ventricles on the 1H images before and after the protocol never differed by >10%. Apart from assuring that under all circumstances, the region selected for Na+ image intensity calculation was well within the outer contour of the left ventricle, no correction for swelling or contraction was applied.
18F-2-deoxyglucose positron emission tomography2 and 201Tl scintigraphy26 are being used clinically to assess myocardial viability, and, like 23Na MRI, test sarcolemmal function with a comparable spatial resolution. However, these techniques involve appreciable radiation exposure for the patient, in particular, in serial examinations. In addition, 23Na MRI has the advantage that it can easily be combined with 1H MRI for evaluation of wall-motion abnormalities and cardiac function.
DCE-MRI is rapidly becoming the new “gold standard” for assessment of myocardial viability in the chronic situation.27 However, it remains unclear how the difference in contrast is established. Furthermore, especially in the (sub)acute phase, signal enhancement due to edema in the peri-infarct zone can lead to overestimation of the affected area.28–30 In contrast to 23Na MRI, no information on the (patho)physiological state of the myocardium is obtained with DCE-MRI. The lower spatial resolution of 23Na MRI can be partially overcome by using 1H MR images as anatomic references.
So far, only total Na+ MRI has been applied in humans. Intracellular Na+ imaging requires a shift reagent suitable for use in humans, which is currently unavailable. With the development of magnets with higher magnetic field strengths, total acquisition time can be reduced, spatial resolution can be increased, or both. Extension of the technique to 3 spatial dimensions is highly desirable.
This is the first study that shows the feasibility of 23Na MR CSI for visualization of Na+i under (patho)physiological conditions. These data demonstrate that 23Nai CSI at 9.4 T can be used to assess viability in isolated, perfused rat heart models of ischemia. The results of this study may contribute to the development of a new clinical diagnostic imaging modality for early detection of ischemia and viability, which may also help in evaluation of the efficacy of pharmacological therapies.
A grant from the Netherlands Organization for Scientific Research supported this work (ZonMW 902-16-190). We thank Dr Markus von Kienlin (Roche, Basel, Switzerland) for helpful discussions and Dr Dieter Gross and Dr Thomas Oerther (Bruker Biospin, Rheinstetten, Germany) for help with the development of the CSI pulse sequence and providing the 23Na quadrature birdcage resonator.
Bonow RO. Identification of viable myocardium. Circulation. 1996; 94: 2674–2680.
Rochitte CE, Kim RJ, Hillenbrand HB, Chen EL, Lima JA. Microvascular integrity and the time course of myocardial sodium accumulation after acute infarction. Circ Res. 2000; 87: 648–655.
Kim RJ, Lima JA, Chen EL, Reeder SB, Klocke FJ, Zerhouni EA, Judd RM. Fast 23Na magnetic resonance imaging of acute reperfused myocardial infarction: potential to assess myocardial viability. Circulation. 1997; 95: 1877–1885.
Van Emous JG, Nederhoff MGJ, Van Echteld CJA. Intracellular sodium imaging in acute coronary ligation. Proc Intl Soc Mag Reson Med. 2001; 9: 1832. Abstract.
DeLayre JL, Ingwall JS, Malloy C, Fossel ET. Gated sodium-23 nuclear magnetic resonance images of an isolated perfused working rat heart. Science. 1981; 212: 935–936.
Kim RJ, Chen EL, Lima JA, Judd RM. Myocardial Gd-DTPA kinetics determine MRI contrast enhancement and reflect the extent and severity of myocardial injury after acute reperfused infarction. Circulation. 1996; 94: 3318–3326.
Polimeni PI. Extracellular space and ionic distribution in rat ventricle. Am J Physiol. 1974; 227: 676–683.
Perrone-Filardi P, Pace L, Prastaro M, Squame F, Betocchi S, Soricelli A, Piscione F, Indolfi C, Crisci T, Salvatore M, Chiariello M. Assessment of myocardial viability in patients with chronic coronary artery disease: rest-4-hour-24-hour 201Tl tomography versus dobutamine echocardiography. Circulation. 1996; 94: 2712–2719.
Saeed M, Lund G, Wendland MF, Bremerich J, Weinmann H, Higgins CB. Magnetic resonance characterization of the peri-infarction zone of reperfused myocardial infarction with necrosis-specific and extracellular nonspecific contrast media. Circulation. 2001; 103: 871–876.