| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1 Department of Diagnostic Imaging, St. Bartholomew's Hospital, West Smithfield,
London EC1A 7BE, United Kingdom.
2 Department of Endocrinology, St. Bartholomew's Hospital, London EC1A 7BE,
United Kingdom.
3 Academic Department of Radiology, St. Bartholomew's Hospital, Dominion House,
London EC1A 7BE, United Kingdom.
Received December 23, 2002;
accepted after revision March 10, 2003.
Presented in part at the annual meeting of the American Roentgen Ray
Society, Atlanta, April-May 2002.
Abstract
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MATERIALS AND METHODS. A retrospective study of CT scans of 28 patients with primary hyperaldosteronism was performed. The means of two observers' measurements of adrenal gland size were recorded and compared with published normal values. In addition, a radiologist experienced in adrenal imaging and unaware of the cause of the primary hyperaldosteronism diagnosed either bilateral adrenal hyperplasia or aldosterone-producing adenoma by visual inspection.
RESULTS. The adrenal glands in patients with bilateral adrenal hyperplasia were significantly (p < 0.05) larger than those in patients with aldosterone-producing adenoma or in healthy control subjects. A sensitivity of 100% was achieved when a mean limb width of greater than 3 mm was used to diagnose bilateral adrenal hyperplasia, and a specificity of 100% was achieved when the mean limb width was 5 mm or greater. Receiver operating characteristic curve analysis showed that the overall performance of the radiologist and the mean adrenal limb width in detecting bilateral adrenal hyperplasia were equivalent.
CONCLUSION. In patients with primary hyperaldosteronism, adrenal limb measurements on CT can aid in differentiating bilateral adrenal hyperplasia from aldosterone-producing adenoma because the adrenal glands in bilateral adrenal hyperplasia are larger.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
There is no CT criterion based on adrenal size to differentiate between adenoma and bilateral adrenal hyperplasia in patients with primary hyperaldosteronism. We previously documented normative adrenal gland sizes on CT [10] and used these data for comparison with adrenal size in patients with adrenal hypertrophy caused by adrenocorticotrophic hormone-dependent Cushing's syndrome [11]. However, adrenal gland sizes in primary hyperaldosteronism on CT are not well documented. One of the aims of this study was to determine whether adrenal gland size on CT is greater in patients with bilateral adrenal hyperplasia than in those with adenoma and compare these data with published values for healthy control subjects [10]. Furthermore, we wished to determine whether we could establish a CT criterion based on adrenal size to positively diagnose bilateral adrenal hyperplasia in an attempt to increase diagnostic accuracy in differentiating between bilateral adrenal hyperplasia and aldosterone-producing adenoma.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The size of the adrenal glands was measured independently by two radiologists who were unaware of the underlying cause for primary hyperaldosteronism. Adrenal gland size was measured using the technique previously described in a study for measuring normal adrenal glands on CT [10]. The measurements were performed on the 10-mm-collimated unenhanced CT images in order to compare these measurements with the published historic controls [10]. At the computer console, the reviewers measured the maximal adrenal gland body and limb widths for each gland (Fig. 1). The widths were measured at the widest point at any part not involved in an obvious nodule and perpendicular to the long axis of the body or limbs of the gland. The means of three measurements of each limb and of the body were recorded, and the average of the two observers' measurements was compared with normal control measurements [10]. The mean widths of both adrenal bodies and all the adrenal limbs were also compared in patients with aldosterone-producing adenomas and those with bilateral adrenal hyperplasia. In glands with aldosterone-producing adenomas, the maximum diameter of the adenomatous nodule was also recorded. If soft-copy images were not available, the hard-copy images together with their corresponding scales were scanned and converted into digital images using Photo-Paint software (version 5, Corel, Ottawa, ON, Canada) to enable measurements to be made electronically.
|
The 5-mm-thick contrast-enhanced CT scans were reviewed by a radiologist experienced in adrenal imaging and who was unaware of the cause of the primary hyperaldosteronism. Each gland was categorized either as being a normal or hyperplastic gland or as having an aldosterone-producing adenoma. The diagnosis of an aldosterone-producing adenoma was made only if a single nodule was visible with the remainder of the ipsilateral and contralateral glands appearing smooth and not enlarged. The radiologic confidence in the CT diagnosis of an adenoma or bilateral adrenal hyperplasia was scored on a 5-point scale with a score of 1 indicating a confident diagnosis of bilateral adrenal hyperplasia; a score of 2, probable bilateral adrenal hyperplasia; 3, possible bilateral adrenal hyperplasia; 4, probable aldosterone-producing adenoma; and 5, a confident diagnosis of an aldosterone-producing adenoma. These scores were subsequently used to perform receiver operating characteristic curve analysis for the radiologist's diagnosis for aldosterone-producing adenoma or bilateral adrenal hyperplasia.
Receiver operating characteristic curve analysis of the mean adrenal limb widths and the radiologist's diagnosis for bilateral adrenal hyperplasia by exclusion of an aldosterone-producing adenoma were compared. Statistical analysis was performed using the Statistical Package for the Social Sciences (version 9, SPSS, Chicago, IL). The Mann-Whitney U test for nonnormally distributed variables and the Student's two-sample t test for normally distributed variables were performed on various data sets, with statistical significance considered as a p value of less than 0.05. Tests for normality were performed using the Wilks-Shapiro test. Descriptive statistical values including sensitivity and specificity were also determined.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
|
A comparison of adrenal size in patients with primary hyperaldosteronism and the published values for adrenal size in healthy control subjects [10] is shown in Table 2. The Student's t test was used as only the mean, and the SD value was available from the published results. The adrenal glands in patients with bilateral adrenal hyperplasia were also larger than in healthy control subjects [10], with the difference between the limb widths being more significant than that between the body widths. A comparison of the adrenal gland sizes in aldosterone-producing adenoma and in normal controls showed that the only statistically significant result was the larger size of the left adrenal limbs in aldosterone-producing adenoma.
|
Receiver operating characteristic curve analysis (Fig. 3) showed that the overall performance of the radiologist's diagnosis of adenoma or bilateral adrenal hyperplasia and the mean adrenal limb width for the detection of bilateral adrenal hyperplasia were equivalent. The area under the receiver operating characteristic curve (Az) for adrenal limb thickness was 0.86 (standard error of the mean [SEM], 0.07), and the Az value for the radiologist's diagnosis was 0.91 (SEM, 0.06). A specificity of 100% was obtained when a mean adrenal limb width of 5 mm or more was used to diagnose bilateral adrenal hyperplasia, but the sensitivity was only 47%. A sensitivity of 100% was obtained when an adrenal limb width of greater than 3 mm was used, but the specificity was only 54%. When the exclusion of an aldosterone-producing adenoma was used to diagnose bilateral adrenal hyperplasia by the radiologist, a sensitivity of 93.3% and a specificity of 84.6% were obtained.
|
The adenomatous nodules ranged from 1.00 to 4.75 cm in diameter, with a mean of 2.20 cm and a median of 2.00 cm.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
After clinical and biochemical diagnosis of primary hyperaldosteronism, CT and MRI are used to differentiate aldosterone-producing adenomas from bilateral adrenal hyperplasia. Traditionally, this distinction is made by detecting an aldosterone-producing adenoma; hence, the diagnosis of bilateral adrenal hyperplasia is reached by exclusion. A high specificity for the detection of an adenoma is desirable because it will avert unnecessary surgery in patients with bilateral adrenal hyperplasia. However, as mentioned before, there is a wide variation in the literature in the diagnostic performance for the detection of an adenoma. Several reasons have been given for the lack of specificity for the detection of an adenoma, including the detection of a concomitant nonhyperfunctioning nodule, the presence of a dominant nodule in macronodular bilateral adrenal hyperplasia, and increased nodularity with age and hypertension [3, 14]. The lack of sensitivity, however, has been attributed to the small size of aldosterone-producing adenomas, which are generally less than 2.0 cm in maximal diameter [2, 6]. In our study, the adenomatous nodules ranged from 1.00 to 4.75 cm in diameter, with a mean of 2.20 cm and a median of 2.00 cm.
To our knowledge, no CT criterion based on adrenal size is available to differentiate an aldosterone-producing adenoma from bilateral adrenal hyperplasia. In light of the current findings showing that the adrenal gland limbs in bilateral adrenal hyperplasia are significantly greater than in aldosterone-producing adenoma, it follows that a CT criterion based on adrenal gland limb size could help differentiate the two entities from one another by positively identifying bilateral adrenal hyperplasia. Our study suggests an accurate test based on adrenal limb measurements for positively identifying bilateral adrenal hyperplasia. Receiver operating characteristic curve analysis shows that the CT size criterion has an overall diagnostic performance similar to that of the traditional method of diagnosing bilateral adrenal hyperplasia by the exclusion of an adenoma. A specificity of 100% was obtained for positively identifying bilateral adrenal hyperplasia when the mean limb width was 5 mm or larger. However, a high sensitivity for positively diagnosing bilateral adrenal hyperplasia is preferred, just as a high specificity is required in diagnosing an adenoma. The use of a test with a high sensitivity for positively diagnosing bilateral adrenal hyperplasia will avoid unnecessary surgery in patients with bilateral adrenal hyperplasia. A sensitivity of 100% is achieved when a mean limb width of greater than 3 mm is used to positively diagnose bilateral hyperplasia, although this criterion necessarily lowers the associated specificity of positively diagnosing bilateral hyperplasia to only 54%.
We suggest a diagnostic algorithm incorporating both adrenal size measurements and the detection of an adrenal nodule to differentiate aldosterone-producing adenoma from bilateral adrenal hyperplasia more accurately in patients with established primary hyperaldosteronism (Fig. 4) so that the use of adrenal venous sampling is limited to a smaller proportion of cases—only those for which the diagnosis remains equivocal. By virtue of its high specificity, this algorithm can be used to confidently positively diagnose bilateral adrenal hyperplasia when the mean limb width is 5 mm or greater (Fig. 5); likewise, bilateral adrenal hyperplasia is excluded when the mean limb width is 3 mm or less, due to the high sensitivity of positively diagnosing bilateral adrenal hyperplasia when the mean limb width is greater than 3 mm (Fig. 6). However, the difficulty in making a radiologic diagnosis of bilateral adrenal hyperplasia or adenoma arises when the mean limb width lies between 3 and 5 mm. If the mean limb width is between 3 and 5 mm and a nodule is seen, then the diagnosis of adenoma or nodular bilateral adrenal hyperplasia depends on the confidence of the radiologist in detecting the presence of other nodules in either gland. The correct distinction is especially important to avoid unnecessary surgery in patients with nodular bilateral adrenal hyperplasia. In our series, the four cases that fell into this category were all adenomas, and the two cases of nodular bilateral adrenal hyperplasia were correctly detected by the size criterion. Nevertheless, the radiologist can resort to adrenal venous sampling when the diagnosis is in doubt. When the mean limb width is between 3 and 5 mm and no nodule is seen, we suggest that adrenal venous sampling be used to confirm the diagnosis, which will usually be bilateral adrenal hyperplasia. In our series, eight cases of bilateral adrenal hyperplasia and one case of adenoma (Figs. 7A, and 7B) fell into this category.
|
|
|
|
|
Interestingly, the three cases that were mis-diagnosed by the radiologist (one false-negative and two false-positives for bilateral adrenal hyperplasia detected by excluding an aldosterone-producing adenoma) can be correctly diagnosed by applying the algorithm. In all these cases, the radiologist erred mainly on the assessment of whether the limbs were otherwise normal or hyperplastic rather than on the detection of the nodule. The mean limb width in the false-negative case (Figs. 8A, and 8B), where a 2.5-cm left adrenal nodule was seen, was more than 5 mm and therefore would have been correctly diagnosed as nodular bilateral adrenal hyperplasia using the algorithm. In both the false-positive cases, where a nodule was not seen in one case (Figs. 7A, and 7B), the mean adrenal limb width was between 3 and 5 mm and the algorithm suggests adrenal venous sampling to locate a presumed adenoma. These cases clearly illustrate the advantage of using adrenal gland limb size measurements over visual inspection in assessing for enlargement of the adrenal limbs.
|
|
When we applied the algorithm to our sample, a total of 13 of 28 patients would have required adrenal venous sampling because a large number of cases had mean limb widths between 3 and 5 mm. Nevertheless, the algorithm is useful. When a nodule is seen and the mean limb width is 5 mm or greater, an incorrect diagnosis of an adenoma (and unnecessary surgery) will be avoided. Conversely, with no visible nodule and a mean limb width of 3 mm or less, adrenal venous sampling to locate a presumed adenoma should be recommended. The algorithm highlights the value of limb width measurements in managing patients with primary hyperaldosteronism. Although all the cases in our series would have been correctly managed using our proposed diagnostic algorithm, further validation of the algorithm is required because the sample size is small.
There are other limitations to our study apart from the small sample size. First, our institution is a referral center for endocrine disorders: thus, our study population is skewed toward patients with more difficult diagnoses, with underrepresentation of the number of cases of adenomas as a cause for primary hyperaldosteronism. Second, histologic diagnosis of bilateral adrenal hyperplasia was not obtained routinely when there was a good response to appropriate medical treatment. Another notable limitation is the retrospective nature of the study with consequent introduction of methodologic variables such as different scanners, scanning parameters (milliampere-seconds, peak kilovoltage), and contrast medium enhancement regimens that were used. For the visual assessment of the glands, the radiologist was aware of this fact. However, slice thickness and partial volume effect would have a significant effect on adrenal gland size measurements; thus, in our study all the patients had 10-mm-collimated images available for measurements to be made. As with small lesions, accurate measurements of the adrenal gland parts on 10-mm-collimated images can be difficult in part because of partial volume effects. However, the coefficients of variation for inter- and intraobserver measuring of limb widths have been reported at 12% and 10%, respectively [11]. Indeed, with the advent of multidetector CT (MDCT), which will soon be the "norm," our absolute measurement results may differ from those acquired on these newer MDCT scanners because of the higher resolution of the images and less partial voluming artifact. However, the overall difference in size between the entities should remain the same. The visual assessment of the gland may also be different on MDCT; the glands may look larger and perhaps less regular. Radiologists will obviously have to "get their eye in" on the newer MDCT images, and with experience, results as good as, if not better, should be obtainable with MDCT.
Finally, the control measurements were obtained from a separate study [10], and therefore only limited clinical and statistical inferences can be made. However, we believe that comparison with this control group is valid because the technique of measurement was identical and data collection for the cohort in the control group was performed at our institution.
In conclusion, in primary hyperaldosteronism, the adrenal limbs in bilateral adrenal hyperplasia are significantly larger than those in either aldosterone-producing adenoma or in normal controls, as measured on CT, a finding not previously documented. We therefore propose a diagnostic algorithm based on both the adrenal limb size measurements and the detection of an adrenal nodule in an attempt to improve the differentiation between bilateral adrenal hyperplasia and adenoma on CT. Use of this algorithm may decrease the necessity for the cumbersome and more invasive adrenal venous sampling.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  P. T. Johnson, K. M. Horton, and E. K. Fishman Adrenal Imaging with MDCT: Nonneoplastic Disease Am. J. Roentgenol., October 1, 2009; 193(4): 1128 - 1135. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S A Whitley, V J Moyes, K M Park, A M Brooke, A B Grossman, S L Chew, A G Rockall, J P Monson, and R H Reznek The appearance of the adrenal glands on computed tomography in multiple endocrine neoplasia type 1 Eur. J. Endocrinol., December 1, 2008; 159(6): 819 - 824. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. M. Patel, R. K. Lingam, T. I. Beaconsfield, T. L. Tran, and B. Brown Role of Radiology in the Management of Primary Aldosteronism RadioGraphics, July 1, 2007; 27(4): 1145 - 1157. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
