| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Experimental Studies |
1 From the Depts of Radiology (M.M., A.R.N.B., B.M.M., E.S.R., N.P.), Urology (H.G.P., J.S.P., H.G.R.), Pathology (R.C.), Neonatology (K.R.B.), and Research Institute (K.M.P.), Children’s National Medical Center and the George Washington Univ School of Medicine, 111 Michigan Ave NW, Washington, DC 20010. From the 1997 RSNA scientific assembly. Received Feb 1, 2000; revision requested Mar 30; final revision received Jul 24; accepted Jul 26. Supported in part by the Board of Lady Visitors and the Discovery Fund of Children’s National Medical Center, Society for Pediatric Radiology, and Society of Uroradiology. Address correspondence to M.M. (e-mail: mmajd@cnmc.org).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
MATERIALS AND METHODS: Bilateral vesicoureteric reflux was surgically created in 35 piglets (70 kidneys). One week later, a liquid bacterial culture of Escherichia coli was injected into the bladder. Three days after induction of urinary infection, imaging studies were performed, and the kidneys were removed for histopathologic examination. SPECT images were obtained 2–3 hours after injection of 99mTc-DMSA. Transverse and coronal MR images were obtained with gadolinium-enhanced fast inversion recovery. Transverse CT images were obtained before and after injection of contrast agent. Power Doppler US was performed in longitudinal, transverse, and coronal planes. Each kidney was divided into three zones for correlation of findings.
RESULTS: Histopathologic examination revealed pyelonephritis in 102 zones in 38 kidneys. Sensitivity and specificity for detecting pyelonephritis in the kidneys were 92.1% and 93.8% for SPECT, 89.5% and 87.5% for MR imaging, 86.8% and 87.5% for CT, and 74.3% and 56.7% for US. Sensitivity and specificity for detecting pyelonephritis in the zones were 94.1% and 95.4% for SPECT, 91.2% and 92.6% for MR imaging, 88.2% and 93.5% for CT, and 56.6% and 81.4% for US. The pairwise comparison of these modalities showed no statistically significant difference among them except for US.
CONCLUSION: 99mTc-DMSA SPECT, spiral CT, and MR imaging appear to be equally sensitive and reliable for the detection of acute pyelonephritis; power Doppler US is significantly less accurate.
Index terms: Animals • Kidney, CT, 81.12114, 81.12115 • Kidney, MR, 81.121411, 81.121412, 81.121413, 81.12143 • Kidney, SPECT, 81.12171 • Kidney, US, 81.12984 • Nephritis, 81.212 • Ultrasound (US), comparative studies
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
Findings in previous experimental animal studies in which histopathology was used as the standard of reference have shown that both renal cortical scintigraphy with technetium-99m dimercaptosuccinic acid (DMSA) and MR imaging are highly sensitive and reliable for the detection of acute pyelonephritis (2–5). Nonspiral CT proved to be less reliable than 99mTc-DMSA scintigraphy (6), but the accuracy of spiral CT remained unknown. In a clinical study (7), gray-scale US was unreliable compared with 99mTc-DMSA scintigraphy. However, since experimental studies have shown that most pyelonephritic lesions are ischemic (8–11), it was anticipated that power Doppler US would greatly improve the sensitivity of US for the detection of acute pyelonephritis. The purpose of this study was to compare the sensitivity and specificity of four cross-sectional imaging techniques—99mTc-DMSA single photon emission CT (SPECT), spiral CT, MR imaging, and power Doppler US—for the detection and localization of acute pyelonephritis, directly and in the same experimental setting, by using strict histopathologic criteria as the standard of reference.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
Three days later, a jugular central venous access was established for the injection of imaging agents, and each piglet underwent all four imaging studies within a period of 2–3 hours. At the time of imaging, the piglets weighed 6–8 kg, approximately the size of a 1-year-old child. Four piglets underwent imaging at the same time, alternating among the four imaging modalities. Therefore, the order of the imaging studies varied among the piglets. After imaging, the kidneys were removed for histopathologic examination.
Anesthesia.—Anesthetic for reflux surgery and jugular cutdown was administered with intramuscular injection of ketamine hydrochloride (20 mg per kilogram of body weight), xylazine hydrochloride (1 mg/kg), and atropine sulfate (0.05 mg/kg). Deep anesthesia for reflux surgery was attained by using inhalation of 2%–3% halothane. Sedation for the imaging studies was achieved with intravenous injection of thiopental sodium (Pentothal [5–10 mg/kg]; Abbott Laboratories, North Chicago, Ill).
Intestinal gas reduction.—Three to six drops of 
-D-galactosidase enzyme (CurTail; Akpharma, Pleasantville, NJ) was added to the last meal before fasting for imaging.
Scintigraphic Technique
The SPECT technique for 99mTc-DMSA scintigraphy was used to make it comparable to the other cross-sectional imaging techniques. The images were obtained 2–3 hours after an intravenous injection of 99mTc-DMSA (Medi-Physics, Arlington Heights, Ill) in a dose of approximately 3.7 MBq/kg (100 µCi/kg). A state-of-the-art triple-detector gamma camera (Multispect; Siemens Medical Systems, Hoffman Estates, Ill) equipped with high-resolution collimators was used. The images were obtained with a 128 x 128 matrix by using a noncircular motion at 3° increments and 30 seconds of imaging time per stop. A total of 120 projectional images (40 images per detector) were obtained. The SPECT images were reconstructed in coronal, sagittal, and transverse planes by using a Butterworth filter with a frequency cutoff of 0.4 cycles per centimeter. The criterion for the diagnosis of acute pyelonephritis was subjective evidence of focal areas of decreased uptake seen with at least two projections. No attempt was made to quantify the severity of decreased uptake (Fig 1a).
|
|
During the intravenous administration of 0.1 mmol/kg of gadopentetate dimeglumine (Magnevist; Berlex Laboratories, Wayne, NJ), a repeat fast spoiled gradient-echo sequence was used (acquisition time, 45 seconds) to display the aorta, renal vessels, and perfused renal parenchyma. After the administration of the contrast agent, a fast inversion recovery sequence was used in the coronal and transverse planes. The acquisition was obtained with a coronal section thickness of 5–7 mm, intersection gap of 1 mm, repetition time of 2,000–2,500 msec, echo time of 17 msec, and inversion time of 160 msec, to image the kidneys in a single acquisition. Each fast inversion recovery sequence was used with three signals acquired. The total acquisition time was approximately 3 minutes and 40 seconds. Transverse T2-weighted fast spin-echo images were obtained with a section thickness of 10 mm. Images were reconstructed with a 256 x 256 matrix. In some cases, a delayed fast spoiled gradient-echo sequence was used for depiction of the high-signal-intensity collecting system produced by excretion of gadolinium.
An initial assessment of the timing of the maximum suppression of the normal renal parenchymal signal intensity by the contrast agent, as imaged with the fast spin-echo inversion recovery sequence, demonstrated a peak effect earlier than had been demonstrated previously by using the fast multiplanar inversion recovery sequence (5). Therefore, coronal fast spin-echo inversion recovery sequences were initiated approximately 1 minute after the administration of the contrast agent, immediately after the fast spoiled gradient-echo acquisition. Images were interpreted as positive for pyelonephritis when there was evidence of a well-defined focus of medium or high signal intensity in the renal cortex, which contrasted with the low signal intensity of normal renal cortical tissue on fast inversion recovery and/or T2-weighted fast spin-echo images (Fig 1b).
Spiral CT Technique
CT imaging (HiSpeed Advantage; GE Medical Systems) was performed at 120 kVp and 140 mAs. Transverse images (3 mm) were obtained by using 1:1 pitch before and after rapid injection of iothalamate meglumine (Conray 43; Mallinckrodt, St. Louis, Mo) in a dose of 3 mL/kg at a rate of 3 mL/sec with a power injector. Two sets of images were obtained 15 and 45 seconds after the beginning of contrast agent injection (ie, 15- and 45-second delay). The criterion for the diagnosis of acute pyelonephritis was a wedge-shaped, linear, or patchy area of decreased attenuation in the renal cortex (Fig 2). Striation in the enhanced cortex was also considered to represent acute pyelonephritis. The presence of swelling or bulge in the parenchymal outline was noted. Abscess was defined as a focal well-defined area of liquefaction with or without an enhancing rim.
|
|
The histopathologic criteria for the diagnosis of acute pyelonephritis were presence of intratubular neutrophils and interstitial infiltrates of mononuclear cells and neutrophils (Fig 4). The location and extent of each lesion were determined on the basis of combined gross and microscopic findings.
|
Statistical Analysis
Commercially available software (SAS/STAT software, version 6; SAS Institute, Cary, NC) was used for statistical analysis. The 
statistics and McNemar test were used to measure the agreement between findings at histopathologic examination with findings from each of the imaging modalities (12). The 95% CIs for the values were computed. The values were defined as follows: excellent agreement, > 0.75; fair to good agreement, = 0.40–0.75; poor agreement, < 0.40. Results of different imaging modalities for detection of histopathologically positive and negative kidneys and renal zones, assuming blinded and independent evaluation of zones, were also compared with each other by using the Fisher exact test (13). P values less than .05 were considered significant.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
|
|
statistics and McNemar test results demonstrate excellent agreement between histopathology and imaging results for 99mTc-DMSA SPECT, MR imaging, and spiral CT for the detection of acute pyelonephritic lesions in the kidneys and renal zones ( = 0.742–0.895). Agreement between power Doppler US and histopathology results was poor ( < 0.40). The McNemar test showed P values greater than .05 for the detection of affected kidneys and renal zones for SPECT, MR imaging, and CT. The power Doppler US P value was less than .05 for the detection of affected renal zones but greater than .05 for the detection of affected kidneys. However, it should be noted that false results at power Doppler US in two kidneys were not included in the analysis (Table 3).
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
Ischemia, as indirectly evidenced by elevated renal vein renin, has been shown to occur early in the inflammatory response to experimental pyelonephritis in primates; this has been attributed to intravascular granulocyte aggregation leading to arteriolar or capillary occlusion (9). In a more recent study (11) in which the microsphere technique was used, substantial focal ischemia was demonstrated in pyelonephritic lesions in piglets. Therefore, one or more of the following factors appear to be operational in the observed imaging findings: (a) focal decreased perfusion due to edema that causes vascular compression and/or intravascular granulocyte aggregation; (b) tubular obstruction due to accumulation of granulocytes and/or edema, which may result in focal decreased glomerular filtration and slow clearance of the filtrates; and (c) altered tubular cell membrane transport mechanism and cell death.
99mTc-DMSA Scintigraphy
Approximately 60% of the administered dose of 99mTc-DMSA is taken up by the proximal tubular cells, mainly through peritubular arterioles and some through filtration and tubular reabsorption (15). The remaining is filtered and excreted in the urine at low concentration. High-spatial-resolution images of the renal cortex can be obtained 2–3 hours after the administration of the tracer by using either planar or SPECT scintigraphic techniques. In the present study, we chose the SPECT technique to facilitate comparison with other cross-sectional imaging modalities.
Although there is a theoretic advantage to SPECT imaging, results of an experimental piglet study (4) have shown that planar imaging with pinhole magnification and SPECT imaging are almost equally sensitive and reliable for detecting acute pyelonephritis. Therefore, the sensitivity and specificity for 99mTc-DMSA SPECT observed in this study are probably achievable for planar imaging with a small-aperture (<4 mm) pinhole magnification. Decreased uptake of 99mTc-DMSA in the early phase of acute pyelonephritis appears to be primarily due to ischemia. In later stages, additional factors, such as tubular obstruction and resultant decreased filtration and reabsorption, and eventually cellular dysfunction and destruction, become operational (11).
MR Imaging
MR imaging has been shown to be highly effective in the diagnosis of both medical and surgical diseases of the kidney (16–20). More recently, MR imaging with the fast inversion recovery sequence with intravenous administration of gadopentetate dimeglumine has been shown to be a sensitive imaging method for the detection of acute pyelonephritis in the experimental pig model (5) and in a clinical study of acute pyelonephritis in children (21). This sequence markedly reduces the signal intensity of normal renal parenchyma and allows the pyelonephritic lesions to be seen as focal zones of medium or high signal intensity against the background low signal intensity of normal renal tissue (Figs 1b, 5b).
|
|
In our study, differences in the MR pulse sequences played a role in the overall tissue signal intensities, tissue contrast, sequence timing, and MR artifact production. The fast spin-echo inversion recovery sequence showed less signal suppression of the normal renal parenchyma, lower contrast between the inflamed and normal parenchyma, and increased motion artifact as compared with the fast multiplanar inversion recovery sequence. This may have had an effect on measured diagnostic accuracy, but our results represent the overall accuracy of blinded interpretation of the images as obtained. However, the contrast-enhanced inversion-recovery technique compares favorably with renal cortical scintigraphic technique. Tiny amounts of gas and urinary tract stones might best be depicted with CT (24). However, gas-forming infections are extremely rare in children, and, in the absence of urinary tract obstruction or metabolic disease, calculi are rare in the setting of pyelonephritis in children. The whole MR imaging examination is readily accomplished in a time frame suitable for a clinical imaging examination, although the method requires sedation in most young pediatric patients.
Spiral CT Scanning
The lesions of acute pyelonephritis typically appear as wedge-shaped, ill-defined, or striated areas of decreased attenuation on both conventional and spiral contrast-enhanced CT scans (24–26). All three patterns were seen in our study. The CT findings are primarily due to decreased perfusion and, to a lesser extent, tubular obstruction and slow clearance of the contrast agents from the involved regions. The striated nephrogram is thought to be due to obstructed tubules with intervening normal tubules (27,28).
With spiral CT, there is rapid acquisition of data. This allows the kidney to be imaged in multiple phases of intravenous contrast enhancement by varying the scanning delay time. Depending on the delay, three different phases can be imaged; the cortical phase, which yields the best corticomedullary differentiation; the parenchymal phase, in which both the cortex and medullary pyramids are enhanced; and the excretory phase (26).
The piglets in our study were imaged during the cortical phase (15-second delay) and during the parenchymal phase (45-second delay). The full extent of the lesions were better depicted during the parenchymal phase, which is similar to previously reported results (26,29,30). This observation is in keeping with the fact that acute pyelonephritis involves both the cortex and medulla. It was interesting to note that some lesions seen during the cortical phase exhibited a larger area of abnormality in the cortex during the parenchymal phase. There were even occasions on which the cortex appeared nearly normal during the cortical phase and then became abnormal during the parenchymal phase. This change in the size of the lesion or appearance of the cortical abnormality during the parenchymal phase was also noted by others (24,26). The reason for this is not clear; however, this further supports imaging of the kidney during the parenchymal phase for the diagnosis of acute pyelonephritis.
Power Doppler US
A clinical study (31) comparing power Doppler US and conventional CT in 30 children reported sensitivity for power Doppler US detection of pyelonephritis to be only slightly lower than that of CT. This study (31) had limitations, since uncooperative children were excluded, and there was no independent standard of reference for proof of pyelonephritic lesions. In another study (32) of 12 children with a clinical diagnosis of pyelonephritis, power Doppler US yielded a lower detection rate of 75% as compared with rates at scintigraphy or contrast-enhanced CT. In our study, power Doppler US was significantly less sensitive and specific than were the other imaging modalities.
The reasons for high false-positive and high false-negative results with power Doppler US are not clear. They may include insufficient ischemia and technical factors, such as interference from intestinal gas, breathing motion, and rib artifact. Although the piglets were sedated, heavy breathing motions and hiccups were sometimes a problem. Similarly, breathing motion in a crying infant or uncooperative child can be a major restricting factor.
Power Doppler US is based on the integrated Doppler power spectrum. This parameter is fundamentally different from the mean frequency shift. The frequency is determined by the velocity of the red blood cells, whereas the power depends on the number of red blood cells (volume of blood). Therefore, a plausible explanation for false-negative results at power Doppler US during the early phase of acute pyelonephritis is partial venous obstruction due to edema that leads to increased volume and sluggish flow of blood through the involved part of the kidney.
Advantages and Disadvantages of the Imaging Modalities for Clinical Use
Both MR imaging and power Doppler US have the advantage of not using ionizing radiation. They both allow evaluation of the perinephric space. However, power Doppler US has poor sensitivity and specificity, and routine use of MR imaging is less practical because of limited availability, need for sedation, and high cost. Spiral CT can be performed quickly and often without sedation. It is highly sensitive and specific. However, it requires rapid intravenous injection of a large volume of contrast agent and is associated with high radiation dose. The 99mTc-DMSA scintigraphic technique is highly sensitive and specific for the detection of acute pyelonephritis, and in our experience it can be performed without the use of sedation in more than 90% of infants and children. It is readily available and its cost is reasonable. More important, it allows qualitative and quantitative assessment of individual renal function that at the present time is not feasible with other imaging modalities. The major disadvantage of 99mTc-DMSA scintigraphy is the use of ionizing radiation. However, the radiation absorbed dose from 99mTc-DMSA scanning is lower than that from spiral CT scanning.
The estimated radiation dose to the kidneys (target organ) of a 1-year-old child from 99mTc-DMSA is 0.78 mGy/MBq administered dose (33). Our standard administered dose of 99mTc-DMSA in clinical studies is approximately 1.85 MBq/kg (50 µCi/kg), with a minimum dose of 18.5 MBq. Therefore, the estimated dose to the kidneys of a 1-year-old child is approximately 14.4 mGy, which is almost identical to the radiation dose to the entire abdomen from one spiral CT study. The radiation dose delivered to the abdomen of the piglets from spiral CT in our study was estimated by actual measurement of the dose by using a cylindric polymerized methyl methacrylate phantom, same spiral CT scanner, and same technical factors. The measurements showed a mean dose of 14.0 mGy ± 0.5 (SD) at the central location and 14.7 mGy ± 0.5 for the outer locations. The radiation dose from spiral CT would naturally be greater if a precontrast scan and more than one contrast-enhanced scan are obtained.
The limitation of this study is that all images from each modality were interpreted once by a single reader without evaluation of intra- and interobserver agreement.
In summary, the sensitivity and specificity of four cross-sectional imaging modalities for the detection of experimentally induced acute pyelonephritis were evaluated by using strict histopathologic criteria as the standard of reference. The pairwise comparison showed no significant difference in the sensitivity and specificity among 99mTc-DMSA SPECT, spiral CT, and MR imaging, but power Doppler US was significantly less reliable. However, with future advances in Doppler US, its role should be reexamined. At the present time, 99mTc-DMSA scintigraphy and spiral CT appear to be the most practical studies for routine clinical use.
Practical application: The porcine kidney is morphologically and physiologically similar to the human kidney. Furthermore, the piglets used in this study were approximately the size of a 1-year-old child. Therefore, the results of this study seem applicable to the diagnosis of acute pyelonephritis in infants and children. With the exception of power Doppler US results, the information gained from this experimental study also seems applicable to adults.
|
FOOTNOTES |
|---|
Author contributions: Guarantor of integrity of entire study, M.M.; study concepts and design, M.M.; definition of intellectual content, M.M.; literature research, M.M., B.M.M., A.R.N.B., E.S.R.; experimental studies, H.G.R., H.G.P., J.S.P., K.R.B.; data acquisition, M.M., B.M.M., E.S.R., A.R.N.B., R.C., N.P.; data analysis, M.M., B.M.M., A.R.N.B., E.S.R.; statistical analysis, K.M.P.; manuscript preparation, M.M., B.M.M., A.R.N.B., E.S.R.; manuscript editing, M.M.; manuscript review, E.S.R., H.G.R., M.M., B.M.M., A.R.N.B.; manuscript final version approval, M.M.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  W. D. Craig, B. J. Wagner, and M. D. Travis From the Archives of the AFIP: Pyelonephritis: Radiologic-Pathologic Review RadioGraphics, January 1, 2008; 28(1): 255 - 276. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Boubaker, J. O. Prior, J.-Y. Meuwly, and A. Bischof-Delaloye Radionuclide Investigations of the Urinary Tract in the Era of Multimodality Imaging J. Nucl. Med., November 1, 2006; 47(11): 1819 - 1836. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. H. Lowe, M. N. Patel, J. M. Gatti, and U. S. Alon Utility of Follow-up Renal Sonography in Children With Vesicoureteral Reflux and Normal Initial Sonogram Pediatrics, March 1, 2004; 113(3): 548 - 550. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Hitzel, A. Liard, P. Vera, A. Manrique, J.-F. Menard, and J.-N. Dacher Color and Power Doppler Sonography Versus DMSA Scintigraphy in Acute Pyelonephritis and in Prediction of Renal Scarring J. Nucl. Med., January 1, 2002; 43(1): 27 - 32. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S Mackenzie Controversies in the radiological investigation of paediatric urinary tract infection Imaging, December 1, 2001; 13(4): 285 - 294. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
