Contrast-Enhanced MR Imaging of Experimental Acute Tubular Necrosis

Copyright C Acta Radiologica 2001

Acta Radiologica 42 (2001) 74–79 Printed in Denmark ¡ All rights reserved

AC TA R A D I O L O G I C A ISSN 0284-1851

CONTRAST-ENHANCED MR IMAGING OF EXPERIMENTAL ACUTE TUBULAR NECROSIS B. D1, M. F. B1, G. K1, N. V1, K. M1 and T. B-G2 1

Department of Radiology and Medical Imaging, Centre Hospitalier Universitaire, Nantes, France and 2Department of Clinical Pharmacology, University Hospital, Uppsala, Sweden.

Abstract

Purpose: To demonstrate the involvement of the various renal structures in acute tubular necrosis (ATN). Material and Methods: In 15 rats, using a T1-/T2-weighted sequence, either gadodiamide alone, or gadodiamide in combination with sprodiamide (a susceptibility agent) were used to enhance the various anatomical substrates of the kidney. The results were compared to those of pathological verification. Results: Experimentally induced ATN of the rat kidney causes profound changes in the medulla, leaving the cortex largely intact. The difference between the normal cortex and the partially necrotic outer medulla, on the one hand, and the papillary region, was significantly enhanced with the combination, whereas a larger region composed of the inner and outer medulla was enhanced after the gadolinium chelate alone. Conclusion: The results varied considerably between the two procedures; the double contrast demonstrated a clear difference between the inner and outer medulla, and the gadolinium chelate alone demonstrated a clear difference between the medulla and the cortex. These results demonstrated a clear difference in the compartmentalization between the inner and outer medullary regions, providing complementary information about the pathological condition of the kidney.

The issue of tubular necrosis is often discussed in kidney transplantation and in renal failure consecutive to the use of nephrotoxic drugs. Various imaging techniques have been proposed (1, 7, 15, 16), but only an invasive methodology with renal biopsy can readily assess this condition, a procedure that may have to be performed repeatedly to address the problem. Formulated dysprosium (Dy-DTPA-BMA, sprodiamide injection) acts as a compartment-dependent, signal-reducing contrast agent (10, 11, 14, 17), combining high magnetic moment with low relaxivities (6). For the purpose of double contrast, the low effect on T1 is of utmost importance. A Dy chelate in a compartmentalized area will induce a strong T2* effect at a dosage of 1 mmol/ kg b.w., resulting in substantial signal loss. How74

Key words: Kidney, tubular necrosis; MR imaging, contrast media; gadodiamide; sprodiamide; experimental. Correspondence: Benoıˆt Dupas, Service de radiologie et d’imagerie me´dicale, Hoˆtel-Dieu, Centre Hospitalier Universitaire de Nantes, 1 Place Alexis Ricordeau, FR-44093 Nantes Cedex 01, France. FAX π33 24 00 84 107. Accepted for publication 12 July 2000.

ever, in a region with cellular membrane disruption, the resulting homogeneous distribution of the contrast agent minimizes this effect (5, 18). Additionally, the low R1 results in only minor alterations in signal intensity (SI). In non-compartmentalized regions, T1 shortening induced by a gadolinium complex will result in a strong increase of the SI. The same shortening of T1 in an area with an extremely short T2* due to a Dy chelate, i.e. in a compartmentalized region, will only have minimal effect on the SI. Thus, the additional short T1 and T2 will have little further effect on the SI, the total effect remaining that of an extremely short T2*, i.e. the regional SI will be very low. The simultaneous administration of gadodiamide and sprodiamide injection has been shown to

ACUTE TUBULAR NECROSIS

enhance the difference between normal and necrotic tissue in models of myocardial infarction (8, 12, 19), between cells with intact and disrupted cellular membrane in an in vitro system (5), and in induced hepatic necrosis (4). The aim of this study was to investigate if double contrast, used with a sequence capable of taking advantage of both the shortening of T1 and of T2, could be useful in the diagnosis of tubular necrosis in the rat, by identifying the region of cellular membrane disruption. The results were compared with those obtained using gadodiamide alone in an in vivo model. Material and Methods

Gadodiamide injection, formulated Gd-DTPABMA (gadolinium diethylenetriamine pentaacetic-acid bis-methylamide, Omniscan, Nycomed Amersham) is a Gd-based contrast agent, normally distributed in the extracellular space (ECS). The magnetic moment of Gd is 7.94 Bohr magnetons. The relaxivities are similar to those of other currently available ECS agents; namely an r1 of 4.1 sª1 mmolª1/liter and an r2 of 4.86 s–1 mmolª1/liter in water at pH 5.8, at 37æC and at 0.47 T (6). Sprodiamide injection (formulated Dy-DTPABMA, Nycomed Amersham) is based on an identical chelate, but the Gd ion has been substituted with Dy (13). Dy has a magnetic moment of 10.63 Bohr magnetons. The relaxivities of this complex, measured at the same conditions as above, are: r1Ω 0.11 sª1 mmolª1/liter, and r2Ω0.12 sª1 mmolª1/ liter (6). The dose of gadodiamide injection was 0.2 ml, corresponding to approximately 0.3 mmol/kg b.w. In the double contrast experiment, sprodiamide injection at a dose of 1.2 ml, equivalent to a 1.8–2 mmol/kg dosage, was injected simultaneously with the gadodiamide injection (7 rats), or after approximately 10 min delay (allowing for an extra set of images after gadodiamide alone, in 5 rats). Contrast administration was performed as a bolus injection through a catheter in the penal vein. The standard 0.5 mmol/ml solution was used for both agents. The ethical requirements of the University Hospital of Nantes for the handling of experimental animals were followed. Fifteen locally bred, male, brown rats (BD IX), weighing approximately 300 g were anesthetized by an intramuscular injection of 0.3 ml of a solution based on a combination of 4 ml Ketalar and 2 ml Xylazine in 4 ml phosphate-buffered saline. Additional injections were performed when needed. At the inferior extremity of the last costal arc, a

1.5- to 2-cm-long anterior incision was made. Depending on the preference of the surgeons, either the left or the right kidney was exteriorized, and the renal pediculi dissected, without separation of the artery and the vein. A ‘‘bulldog’’ clamp was applied to the renal pediculi for 60–75 min, then removed to allow for reperfusion of the kidney. This procedure induces tubular necrosis in close to 100% of the cases (9). After the 60–75 min, the incision was closed in two separate planes. Following surgery the rats were maintained in individual heated cages. After 24 h, the same anesthetic procedure was used, imaging was performed and the rats were then sacrificed. A pathological verification was performed for comparison to the imaging results. Imaging was performed in the head coil of a 1.5 T Siemens Magnetom Vision system. In order to maximize the potential of both sprodiamide (T2* effect) and gadodiamide (T1 effect), injections a mixed T1- and T2-weighted sequence was used (18). A spin-echo (SE) sequence, with a repetition time of 600 ms and an echo time of 50 ms (SELESTRA, spin-echo, long echo, short TR acquisition), was used. A complete pre- and postdouble-contrast study was performed in 12 of the 15 rats. Gadodiamide injection and sprodiamide injection were administered simultaneously in 7 of these rats. In 5 rats, the gadodiamide injection was administered approximately 10 min prior to the sprodiamide injection to allow for an extra set of images with the gadodiamide alone. The extra images from these animals together with the images from the remaining 3 rats (that received only gadodiamide injection), also served as comparative controls, providing common T1-weighted images. Pathological verification was performed in all 15 rats. In the 8 rats imaged after gadodiamide injection, the mixed T1- and T2-weighted sequence SE 600/ 50 was performed 4 min after contrast injection. In addition, an ordinary T1-weighted SE sequence (600/14) was performed. Both axial and frontal images were acquired, and the measurements were performed on the better of the two projections. For the pathological verification, slices were performed either along the long axis, or as a crosssection. The slice thickness was 3 mm and imaging was performed with two acquisitions. A field-ofview of 201¿230 mm and a 224¿512 matrix gave a voxel size of 0.9¿0.45¿3 mm. The imaging process was initiated before and approximately 4–10 min after the injection of the contrast agents. The SI was measured in regions of interest defined by the operator. The SI of the normal renal parenchyma and of the high-intensity papillary re75

B. DUPAS ET AL.

Table 1

Table 2

SI before contrast enhancement

SI after Gd administrations, SELESTRA sequence

Kidney Normal. SE 600/50, nΩ3 ATN. SE 600/50, nΩ15

Cortex

Medulla

Sinus

316.7∫22 297.3∫34.4

236.3∫10.5 No CMD

Not seen 278.8∫37.1

Kidney Normal. SE 600/50, nΩ8 ATN. SE 600/50, nΩ8

Cortex

Medulla

Sinus

71.7∫12.7 350.4∫75.1

No CMD 548.2∫81.7

424.8∫59.9 782.6∫114.1

SIΩsignal intensity. CMDΩcortico-medullar differentiation.

Table 3

gion were recorded, and the extent of the high SI region was measured. The mean and the standard deviations of the SI were calculated and used to illustrate the differences between the various regions of the kidney. Pathological verification: Immediately after the imaging procedure the animals were sacrificed and the kidneys removed and placed in a formalin solution before a standard pathological verification was performed, after hemoxylin and eosin staining. Results

SI after Gd administration, T1-weighted sequence Kidney Normal. SE 600/14, nΩ8 ATN. SE 600/14, nΩ8

Cortex

Medulla

Sinus

1024∫56 1076∫41

No CMD No CMD

363∫64 Not seen

region, with a significantly higher SI in the entire medullary region, demonstrating a clear CMD. The renal sinus was visualized as a very high SI area (Fig. 2a, b). Using the T1-weighted sequence, the expected increase in SI of the entire kidney was observed in

Before contrast administration (Table 1): The mixed T1-/T2-weighted sequences could demonstrate a slight difference between the cortex and the medulla in the normal kidney, whereas the kidney with inflicted acute tubular necrosis (ATN) had a homogeneous SI. There were only minor differences in the SI between the normal and the pathological kidney (Fig. 1). Gadodiamide injection alone (Tables 2 and 3): Using the T1-/T2-weighted sequence the normal kidney demonstrated a loss of SI, with no corticomedullary differentiation (CMD). In the sinus, on the other hand, the SI was markedly increased. The kidney suffering from ATN demonstrated close to pre-injection value of the SI in the cortical

Fig. 1. Rat kidneys before contrast administration, SE 600/50, axial image. Normal right kidney and experimental ATN of the left kidney. Slight CMD in the normal kidney and no CMD in the ATN kidney.

76

Fig. 2. Comparison between axial SE 600/50 and SE 600/14 sequences after gadodiamide injection. a) SE 600/50 shows marked reduction in SI of the normal parenchyma (right kidney) with high SI of the renal pelvis (»). The left kidney with ATN demonstrates CMD and hyperintense signal in the medullar region. b) SE 600/14 image: increased SI after gadodiamide injection observed in both kidneys. There is no CMD in the left kidney (with ATN). Note the low SI in the renal sinus (») of the normal right kidney.

ACUTE TUBULAR NECROSIS

Table 4 SI after GdπDy administration Kidney Normal SE 600/50, nΩ12 ATN. SE 600/50, nΩ12

Fig. 3. Axial SE 600/50 images after double contrast of gadodiamide injection and sprodiamide injection. a) Low SI of the normal right kidney and high signal in the renal pelvis (»). b) Left kidney (same as in Fig. 2b) with ATN: there is a marked reduction in SI of the cortical and outer medulla, whereas a high signal is seen in the inner medulla (»), corresponding to the papillary region. Note the very high SI in the renal pelvis (➤).

Cortexπ Int. medulla ext. medulla 50.6∫5.9 38.9∫6.3

Sinus

No CMD 1061.2∫60 269.2∫63.5 856.1∫247.1

both kidneys, but without any CMD. A reduction of the SI in the renal sinus was also observed in the normal kidney (Fig. 2c). Double contrast administration (Table 4): The normal kidney displayed a low SI contrasting with a high SI in the collecting tubes, as has been demonstrated earlier (3) (Fig. 3a). In the pathological kidney the administration of the combination of sprodiamide and gadodiamide injections resulted in an 85% reduction in the signal of the cortical kidney parenchyma and the outer medulla. The inner medulla, corresponding to the papillary region, displayed a strong enhancement, as was observed in the renal pelvis, using the short TR, long TE sequence (Fig. 3b). Pathological correlation: The affected kidney was compared to the contralateral normal kidney. After 60–75 min of ischemia, and subsequently 24

Fig. 4. Pathological correlation. a) Sagittal section of normal kidney. b) In the kidney with ATN, four zones are identified: 1. cortex, 2. outer stripe of outer medulla, 3. inner stripe of outer medulla, 4. inner medulla. Note renal pelvis (*). c) Normal outer medulla (¿200). d) ATN: outer medulla with tubular epithelium showing disseminated patchy necrosis (¿200). e) Normal inner medulla (¿400). f) ATN: inner medulla with cellular edema and apical cytoplasmic protrusion into tubular lumen without necrosis (¿400).

77

B. DUPAS ET AL.

narrow, without any signs of dilatation, with the various structures, as the urothelium, remaining normal (Fig. 5). Discussion

Fig. 5. Histologic sections (¿25) of the papillary region (inner medulla). a) Normal tubular lumen. b) ATN with substantially decreased tubular lumen. Note the normal pelvis (*).

h reperfusion, the following four zones were readily identified relatively homogeneously throughout the material (Fig. 4a, b): The cortex was normal, and equivalent to the opposite side. The outer medulla showed regional necrosis. Macroscopically, the pathological kidney demonstrated two different zones of the outer medulla, the outer and inner stripes (2). These, however, could not be identified by MR imaging. Microscopically both the inner and the outer stripes showed the same alterations. Some 50% of the cells displayed pycnotic nuclei, the capillaries were congested, and the tubular lumen was filled with necrotic debris. The remaining cells appeared normal (Fig. 4c, d). In the inner medulla or papillary region, a general cellular turgescence could be observed. The tubular lumen was totally obstructed by the turgescence of the bording cells. The mean cell volume was estimated to be 2–4 times that of normal cellular population, but there were no direct signs of necrosis (nuclear involvement). The capillaries were congested, containing red blood cells (Fig. 4e, f). Finally, the cavities were normally 78

Using a mixed T1- and T2-weighted sequence (SELESTRA) a marked differentiation of the inner and outer medulla was observed after double contrast administration, corresponding to two markedly different consequences of the ischemia-reperfusion process, as shown by the anatomic correlation. This sequence has been shown useful together with sprodiamide alone in tumor necrosis (18) and in a double contrast study in rats demonstrating liver necrosis (4). Using the same sequence, after gadodiamide alone, a cortico-medullary differentiation, also of diagnostic value, was evident, whereas this was not observed using a common T1-weighted sequence. The imaging parameters used were based on the need to take advantage of the different effects of gadodiamide and sprodiamide. At the concentrations used for this study, gadodiamide enhanced the SI (with the exception of the renal cortex). On the other hand, as sprodiamide does not cause any significant T1 shortening, it only caused a decrease in SI in regions where it is compartmentalized. Consequently, there was a signal enhancement in noncompartmentalized tissues due to gadodiamide, and a signal reduction in compartmentalized tissues due to then overwhelming T2* effect of sprodiamide. The CMD was only observed after gadodiamide alone, using the mixed sequence, and not after a true T1-weighted sequence. This is due to the T2 effect of Gd being enhanced in a region where the concentration of the contrast agent is at the limit where the T2 effect becomes predominant. When assessing the kidneys structure by structure by microscopy, the following emerged: The cortex showed no abnormality. The normal and pathologic kidney both showed identical effects of contrast enhancements, i.e. after the Gd chelate alone the dipole-dipole T2 effect and the T2* effect, were predominant. After the combination the additional T2* effect of the Dy chelate reduced the SI of the cortex even further. For the medulla, a significant variation in the effects of the Gd chelate alone and the combination was evident. In the outer medulla, with a high proportion of dead cells, the Gd chelate alone was, as expected, responsible for a clear increase of the SI, and, using the mixed sequence, displayed a clear CMD. In the medulla, where the concentration of Gd is lower compared to the cortex, the T1 effect is dominating.

ACUTE TUBULAR NECROSIS

The combination of the Dy and the Gd chelates was responsible for a strong loss of SI in this region. It has been shown that the T2* effect of the combination of a Dy chelate and a Gd chelate will only dominate the SI behavior of a tissue if there is large proportion of living cells (or cells with a normal cellular membrane) (5). In order to comply with accepted theories about Dy chelate effects (10), a strong signal-reducing effect implies that the remaining cells then seen in the outer medulla, either are normal (i.e. have a cellular membrane capable of excluding the Dy chelate) or that the microenvironment is rigid so as to prevent a homogeneous distribution of the Dy chelate. In the inner medulla, the Gd chelate alone expressed an identical effect as in the outer medulla, i.e. a strong increase in SI. The combination of the Gd and Dy chelates also resulted in a marked SI increase. This region was shown microscopically to consist largely of cells with a volume 3 to 4 times the normal, i.e. remarkable intracellular edematous changes in the tissue. The lack of T2* effect must then indicate that the cellular membrane in these cells would allow a diffusion of the Dy and Gd chelates, allowing for a close to identical distribution of these agents in the intra- and extracelluar compartments. In conclusion, an MR series, involving first a Gd chelate, then a double contrast effect by adding sprodiamide to the gadodiamide injection, demonstrated first a CMD, then a clear differentiation between the inner medulla with an intracellular edema, and the outer medulla with diffuse cellular necrosis. This procedure should be capable of diagnosing ATN and, for example, potentially differentiate this condition from acute rejection. ACKNOWLEDGEMENTS This work was financed by a grant from the INSERM. Nycomed Imaging, Oslo, Norway, delivered the contrast agents free of charge.

REFERENCES 1. C M. J., A P. H., K H. L. et al.: Acute tubular necrosis. Use of gadolinium-DTPA and fast MR imaging to evaluate renal function the rabbit. JCAT 11 (1987), 488. 2. C W. L.: Adult kidney. In: Histology for pathologists, 1st edn., p. 677. Edited by S. S. Sternberg. Raven Press, New York 1992. 3. D B., B-G T., M K. & M R. N.: Synergistic effects of relaxation and susceptibility in

4. 5.

6.

7. 8.

9.

10. 11.

12. 13. 14.

15.

16.

17.

18.

19.

differentiation between compartmentalised and non-compartmentalised tissues. Invest. Radiol. 33 (1998), 268. D B., B-G T., N M. F. & M K.: Delineation of liver necrosis using double-contrast enhancement. JMRI 7 (1997), 472. E A., B-G T., N F. & H A.: Combination of gadolinium and dysprosium chelates as a cellular integrity marker in MR imaging. Acta Radiol. 36 (1995), 41. F S., J C., F A.-K., G D. & K J.: Lanthanide-based susceptibility contrast agents. Assessment of the magnetic properties. JMRI 7 (1997), 251. F J. A., C P. L., G M. E. et al.: Gadolinium-DTPA enhanced dynamic MR imaging in the evaluation of cisplatinum nephrotoxicity. JCAT 13 (1989), 448. G J.-F. H., W M. F., S M., L K., D N. & H C. B.: Identification of myocardial cell death in reperfused myocardial injury using dual mechanisms of contrast-enhancing MRI. Acad. Radiol. 1 (1994), 319. J P., H B. O., R D. A., B C. S., M V. C. & T J.: An experimental model for assessment of renal recovery from warm ischemia. Transplantation 35 (1983), 198. K R. P., Z J. & G J. C.: Intravascular susceptibility contrast mechanisms in tissues. Magn. Reson. Med. 31 (1994), 9. M M., V Z., A H. et al.: Comparison of Gd- and Dy-chelates for T2* contrast-enhanced imaging. In: Workshop on contrast-enhanced magnetic resonance, p. 131. Edited by R. Brasch. Soc. Magn. Reson. Med., Napa, CA 1991. N S., W G., E A. et al.: Doublecontrast MR imaging of reperfused porcine myocardial infarction. Acta Radiol. 37 (1996), 27. R S. M. & W A.: Chelates of gadolinium and dysprosium as contrast agents for MR imaging. J. Magn. Reson. Imaging 3 (1993), 167. S M., W M. F., M T. & H C. B.: Reperfused myocardial infarction on T1- and susceptibility-enhanced MRI. Evidence for loss of compartmentalization of contrast media. Magn. Reson. Med. 31 (1994), 31. V V. S., B Y., C´  O. et al.: Detection of zonal renal ischemia with contrast enhanced MR imaging with a macromolecular blood pool contrast agent. JMRI 2 (1992), 311. V V. S., B Y., M M. E. & B R. C.: Magnetic resonance imaging enhanced with a macromolecular contrast agent. Detection of the zonal renal ischemia. Invest. Radiol. 26 (1991), S131. V A., R B. R., B J. W. et al.: Dynamic imaging with lanthanide chelates in the normal brain. Contrast due to magnetic susceptibility effects. Magn. Reson. Med. 6 (1998), 164. W C., S A., E A., B-G T., H A. & A H.: Dysprosium-enhanced MR imaging for tumor tissue characterization. An experimental study in a human xenograft model. Acta Radiol. 38 (1997), 281. W M., M H. J., E A., W G., W A. & H A.: Double-contrast-enhanced MR imaging of experimental myocardial infarction in the pig. Acta Radiol. 34 (1993), 64.

79