| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Original Research |
1 Department of Radiology, Uludag University Medical Faculty, Gorukle Campus,
Bursa, Turkey 06141.
2 Department of Biostatistics, Uludag University Medical Faculty, Gorukle
Campus, Bursa, Turkey.
Received August 31, 2004;
accepted after revision December 16, 2004.
Address correspondence to G. Savci
(gsavci{at}uludag.edu.tr).
Abstract
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
SUBJECTS AND METHODS. Thirty-five patients with 42 adrenal masses (eight metastases and 34 nonfunctioning adenomas) underwent chemical shift MRI using a double-echo fast low-angle shot sequence. Subsequently, opposed-phase chemical shift MR images were subtracted from in-phase images. The subtraction images were assessed quantitatively and qualitatively. For quantitative assessment, the signal intensity values of the adrenal masses were measured by one investigator with manually defined regions of interest. Qualitative assessment of the subtraction images was performed independently by two investigators, who reported their confidence in diagnosing adenomas versus nonadenomas based on signal intensity of the adrenal masses on subtraction images.
RESULTS. The mean signal intensities were significantly different between adenomas and metastases on subtraction images (213 vs 18; p < 0.0001). There was no overlap in signal intensities between adenomas and metastatic tumors. The accuracy in distinguishing adenomas from metastatic tumors was 100% if the cutoff value of the signal intensity selected was 36-106. Quantitative results corresponding to 100% specificity were also observed, with similar sensitivity. No difference in interpretation between the two investigators occurred.
CONCLUSION. Chemical shift subtraction MRI provides a high confidence level in distinguishing adrenal adenomas from adrenal metastases. The image subtraction technique also facilitates quantitative and qualitative evaluation of adrenal masses in chemical shift MRI.
Keywords: abdomen • adrenal gland • genitourinary tract imaging • MRI • MRI technique
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Most adrenal adenomas contain large amounts of intracellular lipid, whereas metastases usually do not [3]. Therefore, several radiologic techniques attempting to distinguish adrenal adenomas from metastases based on this difference have been performed [4-11]. Chemical shift MRI is the most sensitive among them [4, 6-10]. In this technique, the signal intensity generated from water and lipid protons is additive on in-phase chemical shift images and is subtractive on opposed-phase chemical shift images. Therefore, adrenal adenomas show significant signal loss on opposed-phase images relative to in-phase images, contrary to adrenal metastases.
Despite its high sensitivity and specificity for differentiating adenomas from metastases, the chemical shift technique has some difficulties. Quantitative analysis of the chemical shift images requires calculations of ratios or other formulas, which may be cumbersome during daily radiologic practice. Qualitative analysis has much greater practical applicability than the quantitative approach. Nevertheless, using a visual analysis of opposed-phase versus in-phase images, the signal intensity decrease in adenomas is not always obvious and significant interpreter experience is required.
We hypothesized that if we subtracted the opposed-phase images from in-phase images, the signal intensity of those voxels that contain both lipid and water protons would double, so chemical shift subtraction MRI may improve the diagnostic certainty of adenomas. The purpose of the present study was to ascertain whether image subtraction offers any practical advantages over the use of chemical shift MRI alone in differentiation of adenomas from metastatic tumors by using qualitative and quantitative analyses.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Diagnosis of a metastasis was established by CT-guided percutaneous biopsy in the patient with renal cell carcinoma. The remaining seven metastases, six from lung carcinoma and one from stomach carcinoma, were assessed by rapid enlargement of the adrenal mass on serial imaging studies. Four of these patients had CT examinations obtained 7-12 months before the index examination that were used for follow-up data to assess the change in the adrenal glands. In those patients, the metastases were diagnosed with the documentation of a new adrenal mass that showed short-term growth. The size of the masses ranged from 20 to 42 mm. In the other two patients with three adrenal masses, the metastases were diagnosed by rapid enlargement of the adrenal mass on follow-up imaging studies performed 8-12 months from the index study. The growth rate of the masses ranged from 25 to 45 mm during the follow-up interval.
Stability in size or appearance on imaging follow-up from 12 to 31 months (mean, 20 months) was accepted as proof of diagnosis of adenoma for 22 adrenal masses of 17 patients. For eight of adenomas, the diagnosis was established by CT attenuation values less than or equal to 10 H (mean, –0.8 H; range, –6 to 10 H) [5]. In three patients with no primary malignant tumor, four incidental adrenal masses, for which pathologic proof or long-term imaging follow-up was not available, were diagnosed as adenoma based on the results of quantitative analysis of chemical shift MR images.
MRI Technique
All MRI examinations were performed on a 1.5-T unit (Magnetom Vision,
Siemens Medical Solutions) with a phased-array body coil. After localization
imaging, axial T2-weighted MRI and chemical shift imaging were performed at
the level of the adrenal mass. The axial T2-weighted images were obtained with
a HASTE sequence with an echo space of 4.4 msec, an effective TE of 90 msec,
an infinite TR, a 6-mm slice thickness, an intersection gap of 0.6 mm, a field
of view of 30
35 cm, and a 160 x 256 matrix. Chemical shift MRI was
performed using a double-echo fast low-angle shot (FLASH) sequence with a TR
of 128 msec, double TEs of 2.7 and 5.3 msec, a flip angle of 90°, 6-mm
slice thickness, a 192 x 256 matrix, a field of view of 3035 cm, an
intersection gap of 0.6 mm, and one signal acquisition. The spleen was imaged
in the same slice as the adrenal mass when possible. All scanning was done
while the patients held their breath after full expiration. The series of
subtraction images was obtained by subtracting series of opposed-phase images
from the series of in-phase images using the system's commercially available
software.
Methods of Evaluation
For each patient, signal intensity measurements were obtained from the
region of interest (ROI) in the adrenal mass and in the spleen on in-phase and
opposed-phase chemical shift images, and only in the adrenal mass on chemical
shift subtraction images. The ROI of the adrenal mass was chosen as large as
possible and avoiding inclusion of the edge of the mass. Cystic, necrotic, or
hemorrhagic components of the adrenal masses were excluded from the ROI
whenever possible. When calcification was determined in the adrenal masses on
CT images, the calcified components were also avoided to include within the
ROI. Signal intensities of the adrenal masses were measured at the same
position on in-phase, opposed-phase, and subtraction images.
All measurements were made by one investigator who was blinded to the
clinical data. Two quantitative parameters of signal changes between in-phase
and opposed-phase images were calculated from the measured signal intensities,
as described in previous reports
[4,
6-8,
10]. These parameters were
adrenal-to-spleen ratio and signal intensity index and were calculated as
follows:
 |
All subtraction images were reviewed independently by two investigators who were unaware of the clinical data and the results of quantitative assessment of chemical shift images. The investigators were asked to score the signal intensity of the adrenal masses on a three-point scale: 1 = hypointense (definitely nonadenoma), 2 = indeterminate, and 3 = hyperintense (definitely adenoma). Five months after the last index study, the investigators also reviewed the in-phase and opposed-phase chemical shift images independently. The relative signal intensity ratio of adrenal mass to spleen at opposed-phase imaging was compared with that at in-phase imaging using a five-point scale of certainty. The five-point scale for adenoma versus nonadenoma was scored as follows: 1 = definitely nonadenoma, 2 = probably nonadenoma, 3 = indeterminate, 4 = probably adenoma, and 5 = definitely adenoma.
Statistical Analysis
We compared the signal intensity ratios and the signal intensity
measurements of adenomas and metastatic tumors with a Mann-Whitney U
test. Receiver operating characteristic (ROC) curve analysis was performed for
the quantitative data for discriminating adenomas from metastases. Kappa
statistics were used as a measure of agreement between the independent
qualitative interpretations by two investigators. Interobserver agreement was
considered excellent if the kappa value was between 0.80 and 1.0
[12]. A p value of
less than 0.05 was regarded as statistically significant. Statistical analysis
was performed using Statistical Package for the Social Sciences version 10.0
(SPSS) for Windows (Microsoft).
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Results of quantitative assessment of the signal intensities of the adrenal masses on chemical shift subtraction MR images showed a mean of 213 ± 72.9 (range, 107-377) for adenomas and a mean of 18.3 ± 14 (range, 1-35) for metastases. There was no overlap in signal intensity between adenomas and metastatic tumors. The signal intensity of the adrenal adenomas was significantly higher than that of the metastases (p < 0.001). Figure 1 shows the scatterplot of the quantitative data. The ROC curve for the signal intensities for discriminating adenomas from metastases resulted in maximum area under the ROC curve (Az) value (Az = 1.0). The sensitivity, specificity, and accuracy for the diagnosis of adenoma were 100% if the cutoff value of the signal intensity selected was 36-106.
|
= 1.0). All adenomas
exhibited high signal intensity on subtraction images, consistent with lipid
content (Figs. 2A,
2B,
2C,
3A,
3B,
3C, and
3D), whereas all metastatic
tumors were hypointense on subtraction images (Figs.
4A,
4B, and
4C). The sensitivity,
specificity, and accuracy of chemical shift subtraction MRI using qualitative
analysis for differentiation of adenoma from metastases were 100%.
|
|
|
|
|
|
|
|
|
|
When qualitative assessment of chemical shift images by two investigators was evaluated at a threshold of 4 (probably adenoma) or 5 (definitely adenoma), the mean sensitivity, specificity, and accuracy values for the diagnosis of an adenoma were 99%, 100%, and 99%, respectively. At the highest threshold level (5, definitely adenoma), the mean specificity was 100%, with lower sensitivity (71%) and accuracy (76%). The kappa value for a definite or probable diagnosis of an adenoma was 0.93 and at the highest (definite) confidence of an adenoma was 0.81, indicating excellent interobserver agreement. When using a simple dichotomous adenoma/nonadenoma classification, one of the investigators correctly classified all masses as adenomas or nonadenomas. However, the other investigator classified an adenoma as indeterminate.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
We used double-echo chemical shift FLASH imaging. The double-echo sequence is performed during a single breath-hold. Corresponding in-phase and opposed-phase images are obtained from the same anatomic position. Therefore, it is possible to avoid slice misregistration. It is particularly important in a subtraction procedure, since imperfect slice coregistration can lead the subtraction image to be ambiguous.
In a quantitative analysis of subtraction images, we compared the signal intensity measurements of adrenal masses. We found a statistically significant difference in the signal intensity values of adrenal adenomas and metastases (mean values, 213 vs 18, respectively). The accuracy of the method in differentiating adrenal adenomas from metastases was 100% if the cutoff value of the signal intensity selected was 36-106. In comparison with the quantitative analysis of chemical shift MRI, the noteworthy advantage of chemical shift subtraction MRI is that there is no need for any type of calculation of ratios or formulas. In addition, this technique does not require measurement of reference organs, eliminating the problem of signal intensity variability produced by the reference tissues.
The qualitative assessment of the subtraction images showed results similar to those obtained with quantitative assessment. No overlap between adrenal adenomas and metastatic tumors was found. Besides, none of the adrenal masses was interpreted as indeterminate, which represents the high confidence level of the method. Although no significant difference was seen between the results of the qualitative assessment of the subtraction images and those of the chemical shift images, the subtraction technique simply facilitates the subjective interpretation. Therefore, it can be used in equivocal cases or by inexperienced observers to improve their accuracy.
In a qualitative analysis of chemical shift MRI, abdominal viscera are usually used as an internal reference [4, 6, 8, 13, 14]. It is also important to use the same window width and center values on both in-phase and opposed-phase images, since variable windowing values can vary the signal intensity of the adrenal gland [15]. However, with the subtraction technique, there is no need for reference tissue or for more caution for image windowing. After subtraction, visual conspicuity of the lipid content of the lesions was markedly increased. These results are important because visual analysis is easier to apply on a routine basis than is quantitative assessment.
Visual analysis can also compensate for motion artifacts, which can lead to misdiagnosis when using ROI measurements; however, it is not problematic for most examinations that are performed with the method defined. Nevertheless, susceptibility of the chemical shift MRI to motion is an important limitation of the technique. The lipid content in a lesion may be focal, which may be difficult to appreciate in ROI measurements applied over the whole lesion, in contrast to visual evaluations [16]. Subtraction images can facilitate the detection of focal lipid. The presence of lipid in an adrenal mass, even if it is focal, may markedly increase the likelihood of benignity, although some adrenal metastases from certain tumors (e.g., from hepatocellular carcinomas, renal cell carcinomas, or liposarcomas) rarely contain lipid [16]. Collision lesions (combined metastasis and adenoma) are also scarce. Adrenal adenomas may occasionally undergo degeneration and have a heterogeneous appearance. In our series, one adenoma 39 mm in diameter had a focal cystic region. The solid portion of the mass was big enough to avoid partial volume artifacts at the edge of the mass on opposed-phase or subtraction images, and displayed signal drop-off on opposed-phase images. The mass was unchanged in size and appearance during 14 months of follow-up. In a study by Gabriel et al. [17] of 18 patients with follow-up or pathologic proof, the adrenal masses that had heterogeneous suppression on opposed-phase images were adenoma.
Our study is limited by several factors. First, the number of adrenal metastases is small. More metastases studied with this method would be ideal before reaching definitive conclusions about chemical shift subtraction MRI. Second, some of the patients with adenomas did not undergo imaging follow-up. The diagnosis in these patients was based on the signal intensity ratios at chemical shift MRI, since the use of these criteria has been validated in several previous studies [4, 6-8, 10, 14]. However, we only invoked the criteria for incidental adrenal masses found in patients without a known extraadrenal primary malignancy. Third, we used only one MR scanner from one manufacturer in this study. The results of chemical shift subtraction MRI may vary among different manufacturers' scanners. Fourth, no equivocal lesion was present in our study group, such as lipid-poor adenoma. In a recent study, it was reported that lipid-poor adenomas with attenuation higher than 30 H on unenhanced CT are not well diagnosed with chemical shift MRI [18]. Unfortunately, the unenhanced CT attenuation values were not available for the majority of our cases because most abdominal CT scans are obtained with IV contrast media during daily radiologic practice. However, there was no overlap in signal intensity index between adenomas and metastases in our study group. Lipid-poor adenomas may cause some overlap with metastases [7]. The subtraction technique may improve the chemical shift imaging to differentiate lipid-poor adenomas. Nevertheless, further studies aimed specifically at evaluating the value of chemical shift subtraction MRI in differentiating lipid-poor adenomas from metastases are needed.
In conclusion, the results of our study showed that chemical shift subtraction MRI provides high confident differentiation of adrenal adenomas from metastases. The subtraction technique facilitates quantitative and qualitative assessment of adrenal masses in chemical shift MRI. In quantitative assessment, measurement of signal intensity of the mass alone on subtraction images without calculating some formulas is enough to identify adenomas. We found that visual assessment was as accurate as the quantitative measure. The subtraction technique improves the visual detection of lipid content of the adrenal masses. Therefore, qualitative assessment using the subtraction technique may attain wide-spread acceptability compared with quantitative assessment, which is dependent on factors that could vary from scanner to scanner.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  J. F. Faria, S. M. Goldman, J. Szejnfeld, H. Melo, C. Kater, P. Kenney, M. P. Huayllas, G. Demarchi, V. V. Francisco, C. Andreoni, et al. Adrenal Masses: Characterization with in Vivo Proton MR Spectroscopy Initial Experience Radiology, December 1, 2007; 245(3): 788 - 797. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
