| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Genitourinary Imaging |
1 From the Departments of Radiology (J.K.K., S.H.K., Y.J.J., K.S.C.) and Urology (H.A., C.S.K., H.P.), Asan Medical Center, University of Ulsan, 388-1 Poongnap-dong, Songpa-gu, Seoul, 138-736, South Korea; and Department of Radiology, Pusan National University Hospital, Pusan, South Korea (J.W.L., S.K.). Received January 25, 2005; revision requested March 29; revision received May 18; final version accepted June 21. Address correspondence to J.K.K. (e-mail: rialto{at}amc.seoul.kr).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
Materials and Methods: Institutional review board approval and informed consent were obtained. Double-echo GRE chemical shift MR imaging was performed in 55 patients (29 men and 26 women; mean age, 49 years ± 14 [standard deviation]) with 55 renal tumors, including 37 (67%) pathologically proved tumors (23 renal cell carcinomas, nine AMLs, two oncocytomas, two lymphomas, and one reninoma) and 18 (33%) clinically diagnosed tumors (17 AMLs and one indeterminate malignancy). All tumors showed no intratumoral fat and had homogeneous enhancement and a prolonged or gradual enhancement pattern on biphasic helical computed tomographic scans. Signal intensity was measured in the renal tumor and spleen on in-phase and opposed-phase images. The signal intensity index and tumor-to-spleen ratio in AMLs and non-AMLs were calculated and compared with the Student t test. Receiver operating characteristic (ROC) analysis was performed to evaluate the diagnostic accuracy of the signal intensity index and tumor-to-spleen ratio and to extract the optimal cut-off values in the differentiation of AMLs and non-AMLs.
Results: The signal intensity index and tumor-to-spleen ratio were different between AMLs (42% ± 11 and –43% ± 17, respectively) and non-AMLs (5% ± 14 and –4% ± 16, respectively) (P < .001). The area under the ROC curve was 0.975 for the signal intensity index and 0.952 for the tumor-to-spleen ratio. For differentiation of AMLs from non-AMLs, sensitivity and specificity were (a) 96% and 93%, respectively, with a signal intensity index of 25% and (b) 88% and 97%, respectively, with a tumor-to-spleen ratio of –32%.
Conclusion: Double-echo GRE chemical shift MR imaging can be used to differentiate AML with minimal fat from other renal neoplasms.
© RSNA, 2006
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
Gradient-echo (GRE) chemical shift magnetic resonance (MR) imaging is a useful technique in the identification of a small amount of fat (6,7). This method has been widely used to differentiate adrenal adenoma from other neoplasms by quantifying the fat content (8–13). Moreover, the use of double-echo GRE chemical shift MR imaging can be used to resolve section misregistration between in-phase and opposed-phase images because both images are acquired during a single breath hold (12). Accordingly, one would anticipate that this MR technique may be useful in the diagnosis of AML with minimal fat.
The purpose of our study, therefore, was to prospectively evaluate the diagnostic performance of double-echo GRE chemical shift MR imaging in the differentiation of AML with minimal fat from other renal neoplasms by using pathologic analysis or follow-up as the reference standard.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
Patient Selection and Tumor
This study was prospectively designed, and between February 2001 and December 2004, patients were included on the basis of biphasic multi–detector row helical CT findings. Biphasic multi–detector row CT included unenhanced, corticomedullary, and nephrographic and/or excretory phase examinations; the scanning delay was 30 seconds for corticomedullary phase examinations and 120–150 seconds for nephrographic and/or excretory phase examinations. The section thickness in our CT protocol was 5.0 mm for unenhanced examinations, 2.5 mm for corticomedullary phase examinations, and 5.0 mm for nephrographic and/or excretory phase examinations.
The inclusion criteria for this study were as follows: Unenhanced CT scans showed no detectable fat in renal tumors. Renal tumors showed homogeneous enhancement and a gradual or prolonged enhancement pattern over time. Craniocaudal diameter of renal tumors was equal to or greater than 12 mm. According to a recent report (1), the heterogeneous and early washout pattern seen on biphasic helical CT scans strongly favored a diagnosis of RCC. In this study, we attempted to evaluate renal tumors without enhancement characteristics favoring RCC. Thus, renal tumors with homogeneous enhancement and a prolonged or gradual enhancement pattern over time were included in this study.
The CT enhancement pattern over time was determined by measuring the attenuation value of tumors with reference to a recent report (1). An early washout pattern was considered to be present when a tumor showed peak enhancement in the corticomedullary phase and demonstrated a washout of at least 20 HU in the early excretory phase. A gradual enhancement pattern was considered to be present when the tumor attenuation value in the early excretory phase was at least 20 HU greater than it was in the corticomedullary phase. A prolonged enhancement pattern was considered to be present when the difference in tumor attenuation between the corticomedullary and early excretory phases ranged from –20 to 20 HU.
The criterion of a craniocaudal diameter of 12 mm was used to avoid partial volume averaging artifacts because the section thickness of transverse MR imaging was 6 mm in this study.
During the study period, two staff radiologists (J.K.K. and K.S.C., with at least 10 years of experience in kidney CT) in the genitourinary division of the radiology department of Asan Medical Center selected candidates in a consensus fashion according to the inclusion criteria. Consequently, 55 renal tumors were collected from 55 patients (29 men, 26 women; age range, 23–82 years; mean age, 49 years ± 14 [standard deviation]).
Pathologic diagnoses were made by means of tumor resection for 37 (67%) of 55 renal tumors, including 23 RCCs, nine AMLs, two oncocytomas, two lymphomas, and one reninoma. In the 18 renal tumors in which pathologic diagnoses were not available, the following criteria were applied for classification: (a) A tumor was regarded as an AML if there was no tumor growth for at least 15 months on follow-up CT images and the signal intensity index on double-echo GRE chemical shift MR images was equal to or greater than 40% (n = 7). This value (signal intensity index of 40%) was derived from past study data. The signal intensity index of 40% corresponded to a fat fraction of 30% in a phantom with a similar T1 value in the adrenal gland (12). Furthermore, the signal intensity index of 40% was twice as great as a cutoff value of 20%, which yielded a specificity of 100% in the diagnosis of adenoma in patients with hyperattenuating (>10 HU) adrenal masses (9). (b) A tumor was considered to be an AML when there was no tumor growth for at least 15 months and two or more obvious AMLs in the same kidney were either visible on CT scans (n = 6) or met the clinical criteria for tuberous sclerosis (n = 4). (c) A tumor was diagnosed as an indeterminate malignancy if it grew during a period of at least 15 months (n = 1). The mean follow-up period for AMLs was 23 months (range, 15–30 months). As a result, the 55 renal tumors were divided into the AML group (n = 26) or the non-AML group (n = 29). The mean value of the greatest tumor diameter was 2.1 cm (range, 1.3–3.0 cm) in the AML group and 2.2 cm (range, 1.4–3.0 cm) in the non-AML group. In the final 55 patients in our study, two patients (one with RCC and one with AML) were included in both our previous study (1) and our current study.
MR Imaging Technique
MR imaging was performed by using a 1.5-T unit (Vision; Siemens Medical Systems, Erlangen, Germany) with a phased-array body coil. After localizer images were acquired, double-echo GRE chemical shift MR images were obtained during a single breath hold in the transverse plane covering the entire area of the kidneys and the lower part of the spleen. The imaging parameters were as follows: repetition time msec/echo time msec, 152/2.7 for opposed-phase imaging and 152/5.3 for in-phase imaging; one signal acquired; flip angle, 80°; section thickness, 6 mm; no intersection gap; field of view, 30–35 cm; and matrix size, 128 x 256.
MR Image Analysis
For quantitative measurement of intratumoral fat content, two staff radiologists (J.K.K. and K.S.C., with at least 10 years of experience in kidney MR imaging) independently measured signal intensity in both the renal tumor and the spleen in each patient. A round or elliptical region of interest (ROI) cursor was placed over a renal tumor while attempting to avoid including the tumor margin within the ROI. The area of the ROI was 0.4–3.9 cm2, and its location and size were constant between in-phase and opposed-phase images. To minimize the effect of partial volume averaging from the surrounding renal parenchyma or perirenal fat, the observers tried to place the ROIs near the center of the tumors. After measuring the signal intensity of the renal tumors, the signal intensity of the spleen that was considered to be appropriate for use as the reference signal intensity was measured. In each patient, the size of the ROI in the spleen was the same as that in the renal tumor, and the area, location, and size of the ROI were constant between in-phase and opposed-phase images.
The mean signal intensity values were calculated from the signal intensity values independently measured by the two observers. Then, on the basis of the mean signal intensity values, the signal intensity index and the tumor-to-spleen ratio were calculated. The signal intensity index was calculated as follows: [(TSIin – TSIopp)/(TSIin)] x 100, where TSIin is tumor signal intensity on in-phase images and TSIopp is tumor signal intensity on opposed-phase images. The tumor-to-spleen ratio was calculated as follows: {[(TSIopp/SSIopp)/(TSIin/SSIin)] – 1} x 100, where SSIin is spleen signal intensity on in-phase images and SSIopp is spleen signal intensity on opposed-phase images.
Statistical Analysis
Between the AML and non-AML groups, the male-to-female ratio was compared by using the Fisher exact test, and patients' ages were compared by using the Student t test.
The signal intensity values measured in the renal tumor and spleen on both in-phase and opposed-phase images were compared between the two observers with the paired t test.
The signal intensity index and tumor-to-spleen ratio were compared between the AML and non-AML groups by using the Student t test. To evaluate the diagnostic performance of these two parameters in the differentiation of AML from other neoplasms, receiver operating characteristic (ROC) analysis was performed. From this analysis, the optimal cutoff values were extracted; these values showed the best separation (minimal false-negative and false-positive results) between the two groups. With these values, we then calculated the sensitivity, specificity, and positive and negative predictive values for differentiating AML from other neoplasms.
All statistical comparisons between the AML and non-AML groups were performed by using statistical software (SPSS, version 12.0.0; SPSS, Chicago, Ill). Receiver operating characteristics analysis was performed by using an analysis program (ROCKIT, version 1.9B; Charles E. Metz, University of Chicago, Chicago, Ill). A P value of less than .05 was considered to indicate a statistically significant difference.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
Signal Intensity at In-Phase and Opposed-Phase Imaging
The signal intensities of the renal tumors and spleen were independently measured by the two observers (Table). There was no significant difference (P > .05) in the signal intensity of the renal tumor or spleen at either in-phase or opposed-phase imaging between the two observers.
|
In the non-AML group, the mean signal intensity value between the two observers at in-phase imaging was 275 au ± 79 (range, 139–450 au) in non-AML tumors and 305 au ± 76 (range, 208–520 au) in the spleen, whereas the mean signal intensity value at opposed-phase imaging was 265 au ± 98 (range, 113–532 au) in non-AML tumors and 303 au ± 73 (range, 183–496 au) in the spleen.
Signal Intensity Index
The signal intensity index was 42% ± 11 (range, 12%–62%) in the AML group (Fig 1) and 5% ± 14 (range, –22% to 38%) in the non-AML group (Fig 2); this parameter was significantly different between the two groups (P < .001). The tumor-to-spleen ratio was also significantly different (P < .001) between the AML (mean, –43% ± 17; range, –68% to 19%) (Fig 1) and non-AML (mean, –4% ± 16; range, –36% to 29%) (Fig 2) groups.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
Double-echo GRE chemical shift MR imaging was specific in the diagnosis of AML with minimal fat. The specificity of this technique was (a) 93% with a signal intensity index of 25% and (b) 97% with a tumor-to-spleen ratio of –32%.
There may be uncertainty as to whether signal intensity index or tumor-to-spleen ratio is better in the differentiation of AMLs with minimal fat and non-AMLs. There seems to be no significant difference in the performance of these parameters. We suggest that to prevent the delay of treatment of probable malignant tumors, surgical removal of the tumor is necessary in any case with a disagreement between these two parameters.
We believe our study has some advantages over prior studies. First, our study was prospective, whereas previous studies—to our knowledge—were retrospective (1–6). Second, overlap of MR imaging findings between AML and non-AML groups was smaller than overlap in prior studies (1). Third, two observers evaluated MR images independently, and reliability between the two observers was statistically evaluated. Fourth, compared with contrast material–enhanced CT findings in previous studies (1,4) that could not be used to quantify the intratumoral fat content and only gave information regarding the contrast enhancement pattern and tumor morphologic features, our method of measuring signal intensity at chemical shift MR imaging seems to be more objective as a quantitative evaluation of the intratumoral fat content.
A major problem occurring in studies that use ROI measurement is the partial volume averaging artifact. To minimize the effect of volume averaging, the section thickness should be no more than half the diameter of the mass being evaluated. In this study, the section thickness was 6 mm, which was the minimal thickness available on the MR unit. Thus, we could evaluate only those tumors with a craniocaudal diameter equal to or greater than 12 mm.
Another concern related to ROI measurement in this study was the so-called phase cancellation artifact, which decreases the signal intensity in voxels around renal tumors. This artifact is caused by water protons and lipid protons in the retroperitoneal fat along the edges of the kidneys (7,8). To reduce the effect of this artifact, we placed ROIs over the renal tumors while attempting to avoid including the tumor margin within the ROIs. It is still uncertain whether this intrinsic problem could have been totally neglected in small tumors; therefore, this is a potential limitation of our study.
An important limitation of our study was the small number of pathologically proved AMLs (n = 9). Because surgical removal of renal tumors was not desired by patients or urologists for tumors with a substantial loss of signal intensity at opposed-phase imaging during the study period, histopathologic confirmation was not available for many clinically diagnosed tumors. This situation caused an undeniable limitation: that we used the signal intensity index as a parameter for clinical diagnosis of seven AMLs, the usefulness of which was evaluated for differentiating AMLs from non-AMLs. In addition, although we attempted to apply strict criteria for the clinical diagnosis of AML, there remained a possibility that some of the clinically diagnosed AMLs were actually non-AML tumors.
A question may arise concerning the usefulness of detecting intratumoral fat in the differentiation of AMLs from other neoplasms, as it has been shown that non-AMLs—such as RCC and oncocytoma—can have intratumoral fat (14–20). Our results also showed that a considerable number of non-AML tumors could be presumed to have intratumoral fat because their signal intensity indexes were greater than 0% and their tumor-to-spleen ratios were less than 0%. However, our results also showed that the intratumoral fat content was significantly greater in the AML group than in the non-AML group and that sensitivity and specificity were high in the differentiation of AMLs from non-AML tumors. Thus, we suggest that quantification of the intratumoral fat content at double-echo GRE chemical shift MR imaging is useful in the differentiation of AMLs with minimal fat from other neoplasms.
Our study may be subject to criticism regarding study design, although this criticism is understandable given ethical concerns and other considerations. Classification of a tumor as benign disease, particularly AML, was made on the basis of no tumor growth for at least 15 months; however, this does not necessarily mean that such a tumor is not a RCC.
Although double-echo GRE chemical shift MR imaging revealed satisfactory diagnostic performance in the differentiation of AML with minimal fat from other neoplasms in this study, it bears repeating that the use of this technique does not enable a confirmative diagnosis but can lead to alternate methods of disease management. Because differentiation between benign and malignant diseases should be as accurate as possible, we propose that an AML with minimal fat diagnosed at chemical shift MR imaging be closely followed up with CT or MR imaging for 2 years after detection.
In conclusion, our prospective study shows that double-echo GRE chemical shift MR imaging can be used to differentiate AML with minimal fat from other renal neoplasms.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
Abbreviations: AML = angiomyolipoma • au = arbitrary units • GRE = gradient echo • RCC = renal cell carcinoma • ROC = receiver operating characteristic • ROI = region of interest
Authors stated no financial relationship to disclose.
Author contributions: Guarantors of integrity of entire study, all authors; study concepts/study design or data acquisition or data analysis/interpretation, all authors; manuscript drafting or manuscript revision for important intellectual content, all authors; manuscript final version approval, all authors; literature research, J.K.K.; clinical studies, J.K.K., S.H.K., H.A., H.P., J.W.L., S.K., K.S.C.; statistical analysis, J.K.K., S.H.K.; and manuscript editing, all authors
|
References |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  M. R. Oliva, J. N. Glickman, K. H. Zou, S. Y. Teo, K. J. Mortele, M. S. Rocha, and S. G. Silverman Renal Cell Carcinoma: T1 and T2 Signal Intensity Characteristics of Papillary and Clear Cell Types Correlated with Pathology Am. J. Roentgenol., June 1, 2009; 192(6): 1524 - 1530. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 I. Pedrosa, M. R. Sun, M. Spencer, E. M. Genega, A. F. Olumi, W. C. Dewolf, and N. M. Rofsky MR Imaging of Renal Masses: Correlation with Findings at Surgery and Pathologic Analysis RadioGraphics, July 1, 2008; 28(4): 985 - 1003. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 O. A. Catalano, A. E. Samir, D. V. Sahani, and P. F. Hahn Pixel Distribution Analysis: Can It be Used to Distinguish Clear Cell Carcinomas from Angiomyolipomas with Minimal Fat? Radiology, June 1, 2008; 247(3): 738 - 746. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L R Williams, M J Oldale, and A J Bradley Imaging renal masses and staging renal tumours Imaging, March 1, 2008; 20(1): 73 - 86. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. R. Prasad, V. R. Surabhi, C. O. Menias, A. A. Raut, and K. N. Chintapalli Benign Renal Neoplasms in Adults: Cross-Sectional Imaging Findings Am. J. Roentgenol., January 1, 2008; 190(1): 158 - 164. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Y. Kim, J. K. Kim, N. Kim, and K.-S. Cho CT Histogram Analysis: Differentiation of Angiomyolipoma without Visible Fat from Renal Cell Carcinoma at CT Imaging Radiology, December 19, 2007; (2007) 2462061312. [Abstract] [Full Text]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
