| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Genitourinary Imaging |
1 From the MR Research Division, Department of Radiology and Radiological Sciences, Uniformed Services University of the Health Sciences, 4301 Jones Bridge Rd, Bethesda, MD 20814-4799 (V.B.H., M.N.H., P.L.C.); and Department of Diagnostic Radiology, Warren G. Magnuson Clinical Center, National Institutes of Health, Bethesda, Md (V.B.H., S.F.A., M.N.H., P.L.C.). Received June 14, 2001; revision requested July 6; final revision received February 14, 2002; accepted March 12. Address correspondence to V.B.H. (e-mail: vho@usuhs.mil or vho@nih.gov).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
MATERIALS AND METHODS: Regions of interest were measured in 74 patients with renal lesions evaluated by means of dynamic contrast material–enhanced MR imaging with serial breath-hold spoiled gradient-echo acquisitions. Sensitivity for renal tumors and specificity for renal cysts were established by using percentage of enhancement thresholds that varied between 5% and 35%.
RESULTS: The mean percentage of enhancement at MR imaging for the 50 renal cysts was less than 5%; for the 50 renal tumors, it was 97% or higher. With use of a threshold percentage of enhancement of 15% and results obtained between 2 and 4 minutes after administration of contrast material, all malignancies (sensitivity for tumor, 100%) were diagnosed, and there were 6% or fewer false-positive tumor diagnoses. Lower thresholds resulted in unacceptably high false-positive rates (ie, cysts that appeared to enhance—pseudoenhancement), whereas higher threshold values (>20%) resulted in an unacceptably lower sensitivity for tumors.
CONCLUSION: The optimal percentage of enhancement threshold for distinguishing cysts from malignancies with the imaging technique prescribed was 15%, and the optimal timing for measurement was 2–4 minutes after administration of contrast material.
© RSNA, 2002
Index terms: Kidney, cysts, 81.311 • Kidney, MR, 81.121411, 81.121412, 81.121415, 81.121416, 81.12143 • Kidney neoplasms, 81.311, 81.32 • Magnetic resonance (MR), contrast enhancement, 81.12143
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Enhancement alone is seldom the sole criterion for characterization of a renal lesion, and it is always important to combine morphologic features of the lesion, such as homogeneity, wall thickening, and calcifications, with the degree of enhancement (5). Moreover, enhancement does not necessarily mean that the lesion is malignant, as oncocytomas and angiomyolipomas can also enhance. However, CT is accurate in helping to distinguish benign lesions, typically simple cysts, from more worrisome tumors of the kidney, namely renal cell carcinoma (1,3,5).
For many years, magnetic resonance (MR) imaging has been proposed for the evaluation of renal lesions, especially in cases in which ultrasonography (US) and/or CT results are not definitive (6–15). MR imaging is increasingly used as an alternative to CT especially when a patient has compromised renal function, is allergic to contrast material, or wishes to avoid exposure to ionizing radiation. MR imaging has several well-known advantages, including multiplanar imaging, MR angiography, and tissue characterization.
On the basis of experience with CT, contrast enhancement at MR imaging has also been used as a surrogate marker of malignancy. However, simple rules or thresholds, equivalent to those routinely used for the determination of contrast enhancement at CT, have not been established for MR imaging. Moreover, since signal intensity units at MR imaging are arbitrary and can vary substantially on the basis of pulse sequence, manufacturer, and individual patient variation, quantitative assessment can be difficult. The presence of renal lesion enhancement is thus often judged qualitatively by the interpreting radiologist by means of visually comparing images obtained before administration of contrast material with those obtained after contrast material administration.
Although several investigators (11–15) measured relative enhancement of renal lesions against that of adjacent renal parenchyma, these measurements are cumbersome to obtain and difficult to interpret, as the comparison, renal parenchyma itself, is enhancing (ie, changing) over time and is highly variable. The purpose of our study was to establish a simpler quantitative MR imaging contrast enhancement criterion for distinguishing cysts from solid renal lesions.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
All lesions determined to be cysts were incidental findings (ie, not a suggestive lesion that had generated the original indication for the performance of MR imaging) and fulfilled the traditional US and/or CT criteria for simple renal cyst (3,5). All solid lesions were removed surgically and confirmed to be tumors at histologic examination. To be included in this study, lesions had to be at least 1.0 cm in diameter to avoid partial volume artifact concerns, which could affect signal intensity measurements.
MR imaging was performed with a 1.5-T unit (Signa; GE Medical Systems, Waukesha, Wis). Since the images were obtained during a 4-year period, a number of software versions were used, but the basic pulse sequence was the same. Precontrast imaging consisted of T1-weighted spin-echo imaging and T2-weighted spin-echo imaging.
After this preliminary imaging, sequential spoiled gradient-echo images were obtained before and after the intravenous administration of 0.1 mmol of a gadolinium chelate contrast media (Omniscan, Nycomed Amersham Imaging, Princeton, NJ; Prohance, Bracco Diagnostics, Princeton, NJ; and Magnevist, Berlex Laboratories, Livingston, NJ) per kilogram of body weight. These coronal images had the following parameters: 70–133/3.6–4.2 (repetition time msec/echo time msec); flip angle, 60°–80°, matrix, 256 x 192; section thickness, 6–8 mm. No fat saturation was used for the dynamic imaging. The gadolinium chelate contrast medium was injected at a uniform rate of 1 mL/sec and was followed by a 10–20-mL saline flush injected at 1 mL/sec. Postcontrast serial breath-hold acquisitions were obtained at approximately 1-minute intervals for 5 minutes.
After completion of the dynamic imaging, static transverse imaging was performed by using a fat-suppressed T1-weighted spin-echo sequence with the following parameters: 500–600/10; section thickness, 5 mm; matrix size, 128 x 256.
Lesions were identified on each set of images, and the largest dimension was measured. The section for measuring the region of interest (ROI) was selected on the basis of optimum visualization of the lesion and the region of most enhancement. On an independent computer console, a circular ROI was selected within each lesion and measured over the series of acquisitions by two authors (S.F.A. and P.L.C.). ROIs were selected on the basis of an initial inspection of the full data set, and the ROI was placed on the most enhancing portion of the lesion.
ROI size varied from patient to patient but was at least 0.7 cm2 and was placed so that it was within the central two-thirds of the renal mass or enhancing portions of the mass. The same ROI size was used for all measurements within each patient data set. Lesion ROIs were expressed as a percentage of enhancement ratio by using the following formula: percentage of enhancement = (SIpost - SIpre)/SIpre x 100%, where SIpre is the precontrast signal intensity of the lesion and SIpost is the postcontrast signal intensity of the lesion.
The values for cysts and tumors were recorded separately and consolidated into five 1-minute intervals, starting with images obtained 1 minute after administration of contrast material. The mean and SD for the percentage of enhancement and actual signal intensity measurements for cysts and tumors were calculated for each interval. Analysis of variance for the ROI measurements was performed, and a P value less than .05 was considered to indicate a statistically significant difference. Statistical analyses were performed by using commercially available software (SigmaStat, version 2.0; Jandel, San Rafael, Calif and SAS PROC MIXED, version 8.0; SAS Institute, Cary, NC).
At each interval, the sensitivity for solid tumors and specificity for cysts were calculated and adjusted by using different thresholds that ranged from 5% enhancement to 35% enhancement in 5% intervals. A 5% threshold meant that any lesion that enhanced 5% or more was considered a solid mass and any lesion that enhanced less than 5% was considered a cyst. A similar analysis was conducted for thresholds of 10%, 15%, 20%, 25%, 30%, and 35% enhancement.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
|
|
|
The data are displayed graphically in Figure 5 and show that despite the good average results there was some variability in individual cases. Some signal intensity changes were noted in individual cysts, which are likely because of noise introduced during imaging, volume averaging, and motion. For instance, in one case, a 1.2-cm lesion appeared to enhance 9% at 1 minute, 18% at 2 minutes, 21% at 3 minutes, 17% at 4 minutes, and 9% at 5 minutes, emphasizing the value of multiple measurements in smaller lesions.
|
However, as demonstrated in Figure 5, there was great variability in the signal intensity enhancement of renal tumors on individual serial images. Of note is that the mean enhancement was maximal at 3 minutes and minimal at 1 minute after administration of contrast material. Again, variability was seen among particular lesions over time. One 1.5-cm lesion enhanced 17% at 1 minute, 21% at 2 minutes, 30% at 3 minutes, 21% at 4 minutes, and 11% at 5 minutes. A small papillary renal cancer was removed surgically in this case.
To determine the optimal enhancement threshold whereby tumors could be reliably distinguished from cysts, the threshold value for percentage of enhancement was varied from 5% to 35% in 5% increments. The effect on sensitivity and specificity at the five times is illustrated in the Table. Use of a threshold of 15%–20% gain in signal intensity 2–4 minutes after administration of contrast material resulted in a 100% sensitivity for detecting cancers, with a small false-positive rate of 0%–6%, as shown in Figure 5. Images obtained during the 1st and 5th minutes had a lower sensitivity at comparable specificity. For the other times, the specificity for cysts was unacceptably low for thresholds lower than15%, and the sensitivity for tumors was unacceptably low for thresholds higher than 20% for most times except 3 minutes. Thus, the optimal imaging was performed between 2 and 4 minutes after injection of contrast material, with use of a threshold of 15%–20% to distinguish cysts from tumors.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Authors of previous articles have focused on the mean difference between solid and cystic lesions. Our results for mean changes substantially agree with those of these previous reports. For instance, we observed a mean enhancement change of 1%–5% for cysts and a mean of 97%–118% change for solid renal tumors. This compares favorably with the results of Semelka et al (11), who found a 3.1% ± 3.5 change for cysts and an 87.9% ± 55.8 change for renal cancers, although the timing of this measurement may not have been similar. Theoretically, cysts should show no enhancement; however, minimal apparent enhancement or pseudoenhancement could result from motion artifacts, inherent noise, and partial volume averaging with adjacent enhancing normal renal parenchyma. The most problematic case is the minimally enhancing renal lesion. In this case, it is critical to understand the variability of signal intensity within nonenhancing lesions such as cysts to be confident that an observed change in enhancement is real.
With a percentage of enhancement threshold of 15%, our study results showed 100% sensitivity and greater than 94% specificity in helping to distinguish between cysts and solid renal masses on dynamic contrast-enhanced MR images. Given the malignant nature of renal cancer and the grave consequences of overlooking renal cancer, it is more important that a threshold be chosen that minimizes the possibility of a false-negative finding, even at the expense of including a few false-positive findings (ie, cysts that are mistaken for tumors). The threshold value of approximately 15%–20% meets these criteria. The use of the lower end of the range (15%) will ensure higher sensitivity and diminish the likelihood that a solid tumor would be missed.
One should not rely solely on enhancement characteristics, however, to decide whether a lesion is benign or malignant. Internal morphology, such as the presence of thickened walls, septa, and hemorrhage should also be considered. Moreover, it is important in such cases to obtain ROI measurements from the specific areas of abnormality rather than obtaining an ROI that encompasses the entire lesion. If the entire lesion is evaluated, the specific region that is enhancing may not be detected. Finally, it is important to remember that US and CT findings can be integrated with the MR imaging results to reach a final diagnosis.
Timing of imaging after injection of contrast material seems to be important for adequate differentiation of cysts and solid tumors with MR imaging. We found a range of 2–4 minutes after injection to be the optimal time to image renal lesions. Some solid lesions were insufficiently enhanced in the 1st minute after injection and/or had diminished enhancement at 5 minutes after injection. This result agrees with findings of Eilenberg et al (7) and Yamashita et al (15) who found peak contrast within tumor to be at 2 minutes, which is similar to our peak value.
This finding is fortunate in two respects. First, there appears to be some flexibility in the timing after injection. Second, there appears to be no reason to rapidly image after administration of contrast material. However, we believe it is advantageous to obtain a series of acquisitions, as this does not add substantially to the overall examination time or cost and is generally tolerated by patients. The performance of serial imaging allows for maximum enhancement of the renal lesion to be identified and improves diagnostic confidence for identification of the most prominently enhancing region within the lesion. Thus, we recommend at least three repeat studies in the same plane after administration of contrast material. As performed in this study, it is critical that imaging be performed by using the same sequence without retuning of the magnet. Retuning will reset receiver and transmitter gain settings and may lead to artifactual changes in signal intensity.
There are a number of limitations of our study. While we expect our criteria to be generally applicable, we have tested them on MR imagers from only one manufacturer. We fully expect that other units will perform similarly, provided that signal intensity versus 1/T1 approximately intercepts the origin and similar spoiled gradient-echo pulse sequence are used.
Another limitation is that MR imaging ROI data are inherently susceptible to artifacts from motion and to measurement variations due to spatial misregistration between breath-hold acquisitions. The most problematic lesions are those that are less than 1 cm in diameter. The technique described here is not effective for small lesions because of the section thickness (6–8 mm). Unfortunately, decreasing section thickness results in increases in noise levels, thus limiting the ability to obtain reliable ROI measurements. With improvements in the speed of three-dimensional pulse sequences, it is conceivable that these concerns may be minimized in the near future, as three-dimensional imaging would enable inherently higher resolution imaging.
In conclusion, cysts and solid lesions can be reliably distinguished at contrast-enhanced MR imaging. As a practical guideline, if a region of a lesion enhances more than 15% within 2–4 minutes after the intravenous administration of 0.1 mmol/kg of a gadolinium chelate contrast agent, it is highly likely to be a solid tumor. These criteria provide a framework for evaluation of renal masses that is similar to the criterion of a value greater than 10 HU used for renal masses at CT. Morphologic features of the renal lesion should always be considered, whether at CT or MR imaging, to avoid missing cystic renal masses with mural or septal enhancement. MR imaging both before and after administration of contrast material should be performed without retuning the magnet between acquisitions to avoid recalibrating gain settings and changing the dynamic range of the image. Multiple repeated measurements—we recommend three—will improve the confidence of the results without substantially lengthening imaging time. With these criteria and caveats, MR imaging should be an excellent method to aid in differentiating solid from cystic renal masses.
|
FOOTNOTES |
|---|
V.B.H. receives research and/or educational support from GE Medical Systems, Nycomed Amersham Imaging, Epix Medical, and Bracco Diagnostics.
Abbreviation: ROI = region of interest
Author contributions: Guarantors of integrity of entire study, V.B.H., P.L.C.; study concepts, V.B.H., P.L.C.; study design, all authors; literature research, V.B.H., S.F.A., M.N.H.; clinical studies, V.B.H., P.L.C., M.N.H.; data acquisition, S.F.A., P.L.C., V.B.H.; data analysis/interpretation, all authors; statistical analysis, all authors; manuscript preparation, all authors; manuscript definition of intellectual content, V.B.H., P.L.C.; manuscript editing, V.B.H., P.L.C., M.N.H.; manuscript revision/review and final version approval, V.B.H., P.L.C.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  J. Zhang, Y. Mazaheri Tehrani, L. Wang, N. M. Ishill, L. H. Schwartz, and H. Hricak Renal Masses: Characterization with Diffusion-weighted MR Imaging--A Preliminary Experience Radiology, May 1, 2008; 247(2): 458 - 464. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. P. Lokken, D. A. Gervais, R. S. Arellano, K. Tuncali, P. R. Morrison, S. Tatli, P. R. Mueller, and S. G. Silverman Inflammatory Nodules Mimic Applicator Track Seeding After Percutaneous Ablation of Renal Tumors Am. J. Roentgenol., October 1, 2007; 189(4): 845 - 848. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. G. Silverman, K. J. Mortele, K. Tuncali, M. Jinzaki, and E. S. Cibas Hyperattenuating Renal Masses: Etiologies, Pathogenesis, and Imaging Evaluation RadioGraphics, July 1, 2007; 27(4): 1131 - 1143. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. M. Israel and M. A. Bosniak How I Do It: Evaluating Renal Masses Radiology, August 1, 2005; 236(2): 441 - 450. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. S. Hartman, P. L. Choyke, and M. S. Hartman From the RSNA Refresher Courses: A Practical Approach to the Cystic Renal Mass RadioGraphics, October 1, 2004; 24(suppl_1): S101 - S115. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. Tuncali, E. vanSonnenberg, S. Shankar, K. J. Mortele, E. S. Cibas, and S. G. Silverman Evaluation of Patients Referred for Percutaneous Ablation of Renal Tumors: Importance of a Preprocedural Diagnosis Am. J. Roentgenol., September 1, 2004; 183(3): 575 - 582. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E. M. Hecht, G. M. Israel, G. A. Krinsky, W. Y. Hahn, D. C. Kim, I. Belitskaya-Levy, and V. S. Lee Renal Masses: Quantitative Analysis of Enhancement with Signal Intensity Measurements versus Qualitative Analysis of Enhancement with Image Subtraction for Diagnosing Malignancy at MR Imaging Radiology, August 1, 2004; 232(2): 373 - 378. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. M. Israel and M. A. Bosniak Follow-Up CT of Moderately Complex Cystic Lesions of the Kidney (Bosniak Category IIF) Am. J. Roentgenol., September 1, 2003; 181(3): 627 - 633. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
) and tumors (
) versus time. There is a small amount of overlap at the 15% threshold level (arrow) that is best used to distinguish the two groups. Note that there are some inherent variations when obtaining signal intensity measurements with this method whereby higher values may be seen on the precontrast images and yield an apparent negative enhancement. For practical purposes, these instances should be treated as showing no enhancement.