| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Original Research |
1 Department of Medical Imaging, University Health Network and Mount Sinai
Hospital, Princess Margaret Hospital, 610 University Ave., Toronto, ON M5G
2M9, Canada.
2 Department of Diagnostic Radiology, Christie Hospital NHS Trust, Withington,
Manchester M20 4BX, United Kingdom.
3 Main X-ray Department, City General Hospital, Stoke-on-Trent ST4 6QG, United
Kingdom.
Received June 21, 2005;
accepted after revision September 29, 2005.
Address correspondence to K. S. Jhaveri
(kartik.jhaveri{at}uhn.on.ca).
Abstract
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MATERIALS AND METHODS. Thirty-nine adrenal masses that were indeterminate on CT were identified in 37 patients. Histogram analysis was performed on unenhanced CT from a region of interest (ROI) that recorded mean attenuation, number of pixels, and percentage of negative pixels. MR signal intensity drop between the in- and opposed-phase images was measured for the masses. Analyses to determine whether correlations existed among the mean CT attenuation, percentage of negative pixels, and MR signal intensity drop were performed. A final diagnosis was obtained by pathology results, follow-up of more than 6 months, or adrenal washout CT findings.
RESULTS. Negative pixels were present in 25 of 28 adenomas and nine
of 11 nonadenomas. A threshold of more than 10% negative pixels for the
diagnosis of adenoma provided a sensitivity of 46% and specificity of 100%. A
threshold of more than 20% MR signal intensity drop yielded a sensitivity of
71% and specificity of 100%. An increase in the percentage of negative pixels
was correlated with a decrease in mean CT attenuation. Using MRI, observers
characterized seven additional nodules as adenomas compared with CT histogram
analysis (McNemar test, 
2 = 5.1429; p = 0.023).
CONCLUSION. CT histogram analysis with a threshold of a 10% negative pixel presence increases sensitivity for the characterization of adenomas compared with analysis of the mean CT attenuation alone. The use of chemical shift MRI with a threshold of 20% signal intensity drop results in a higher sensitivity than CT histogram analysis.
Keywords: adenoma • adrenal nodules • chemical shift MRI • CT histogram analysis
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Washout CT currently has a high accuracy for the diagnosis of adenoma [1, 4] because it distinguishes adenoma from metastases on the basis of physiologic temporal differences in the enhancement and the exit of iodinated contrast material from the adrenal lesion. Adenomas show a percentage enhancement washout of at least 60% or greater or a relative percentage washout of 50% [5] on delayed contrast-enhanced CT. However, washout CT has some disadvantages. In a busy clinical practice, monitoring every CT scan obtained is not feasible. Hence, a separate patient visit may be needed after an indeterminate finding on an unenhanced scan. Washout CT also involves multiphasic scanning and the injection of IV contrast material. Recently, there has also been heightened concern about radiation from the use of CT with adrenal washout CT, including the three passes through the upper abdomen. Even though complications are infrequent, risks related to IV contrast material remain. Certainly, one could agree that a test avoiding the use of both radiation and IV contrast material would have better patient acceptance. An additional workflow-related issue is the need to obtain a delayed scan 15 minutes after contrast administration.
The role of chemical shift MRI in indeterminate adrenal masses has been suggested to be limited because of older studies showing a parallel correlation between out-of-phase signal intensity drop and unenhanced CT attenuation [6]. Results of those studies suggested that MRI is not helpful in cases in which CT attenuation fails to characterize adrenal lesions. However, we and other authors have recently published MRI results of hyperdense nodules (> 10 H) that are encouraging, with MRI characterizing lesions as adenomas even when the CT attenuation classifies them as indeterminate [7, 8].
Mean CT attenuation results in averaging of tissue density over the CT pixels. Tissue heterogeneity therefore may result in inadequate information being obtained about densities present in a smaller volume from a region of interest (ROI) where the mean CT attenuation is measured. CT histogram analysis provides objective insight into the varying CT densities and the number of pixels with these densities. CT histogram analysis therefore is more likely to be sensitive for lipid detection, represented as pixels with negative attenuation, particularly in lipid-poor adenomas, which are likely to have a mean CT attenuation of more than 10 H.
Recently, Bae et al. [9] proposed the use of CT histogram analysis to evaluate adrenal nodules. This technique has the advantage of involving no additional tests, radiation exposure, or repeat trips to the hospital for the patient if an adenoma is confirmed. Obviously, if the CT histogram analysis is indeterminate, other imaging tests such as adrenal washout CT will be necessary and will require an additional patient visit.
The purpose of this study was to compare CT histogram analysis and chemical shift MRI for differentiating adenomas from nonadenomas in the same cohort of adrenal nodules judged to be indeterminate on unenhanced CT (> 10 H).
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Fifty-one patients were identified, and their CT images were reviewed by one radiologist (fellowship-trained in abdominal imaging). Patients were excluded if the nodule was less than 8 mm in diameter, if the nodule had an unenhanced CT attenuation of less than 10 H, or if macroscopically fat was visible in the nodule on CT. Macroscopically visible fat in an adrenal lesion has a high probability for the diagnosis of myelolipoma and hence was not included in the study because characterization of the mass was considered to be adequate on the unenhanced study. Eight millimeters was chosen as a cutoff to avoid difficulty in measuring MR signal drop and avoid partial volume effects. An adrenal mass was considered indeterminate if it had an unenhanced CT attenuation of more than 10 H.
Lesions excluded were as follows: three due to small size (< 8 mm), four due to gross fat content, and seven due to lack of adequate follow-up. This left a study population of 37 patients (mean age, 61 years; age range, 31-77 years) with 39 adrenal masses for evaluation. One patient had metastasis in each side, and another had an adenoma on each side. Twenty patients were men (mean age, 64 years; age range, 33-77 years) and 17 were women (mean age, 56 years; age range, 31-67 years). Twenty (54%) of 37 patients had a known primary malignancy.
The following criteria were used as a reference standard to classify masses when histologic findings were not available: The diagnosis of adenoma was made if there was no interval growth over a period of more than 6 months or if absolute washout on delayed CT was greater than 60% (or both). The diagnosis of metastasis was made if the mass showed absolute growth (i.e., increase in diameter) in a patient with a known primary malignancy. The reference standard was follow-up in 34 masses, histologic diagnosis in four, and washout CT and interval follow-up in one.
Twenty-eight of the masses were adenomas, nine were metastases, one was an adrenocortical oncocytoma, and one was a pheochromocytoma. Nodules with a final diagnosis other than adenoma were categorized as nonadenoma. In the nine metastases, the primary tumor sites were the lung in eight (one of which was a pulmonary carcinoid tumor) and the cervix in one (a carcinoma). The mean diameter of the nonadenomas was 2.6 cm (range, 1.4-5.2 cm) and that of the adenomas was 2.2 cm (range, 1.0-5.2 cm). The sizes of the adenomas and nonadenomas were not significantly different (p > 0.4).
Imaging
Unenhanced adrenal CT was performed on a 4-MDCT scanner (LightSpeed, GE
Healthcare). CT scanning parameters included a slice collimation of 2.5 mm, HQ
(high-quality) mode, table speed of 7.5 mm/s, 50% overlap reconstruction, 120
kV, and 280-320 mA. A single-detector CT scanner (CTi, GE Healthcare) was used
with a 3- to 5-mm collimation, 50% overlap reconstruction, 120 kV, 200-320 mA,
and a pitch of 1:1. Of the 37 patients, 23 underwent MDCT and 14 underwent CT
on the single-detector scanner.
MRI was performed on a 1.5-T system (Echo-Speed LX, GE Healthcare). In- and opposed-phase MR images were obtained using the following parameters in the transverse and coronal planes in two consecutive breath-holds with a 2D spoiled gradient-echo sequence: TR range/first-echo TE range, second-echo TE range, 80-200/4.2-4.6, 2.1-2.3; band-width, 31.25-62.5 kHz; 1 signal acquired; section thickness, 5-7 mm; gap, 0 mm; field of view to cover the adrenals, 24-38 cm; matrix, 256 x 160-192; and flip angle, 75-90°. Dual-echo acquisition—that is, when both the in-phase and opposed-phase MR images were obtained during the same breath-hold— was used in 22 of 37 patients, and non-dual-echo acquisition was used in the remaining 15 patients.
CT Histogram Analysis
One observer with 7 years of CT experience performed the CT analysis. CT
images were evaluated on a postprocessing workstation (Advantage Windows
workstation [version 3.0], GE Healthcare). Commercially available software (CT
Perfusion, GE Healthcare) was used to obtain the histograms. For every adrenal
nodule, the slice showing the maximal cross-sectional area was chosen for
analysis. The long-axis and short-axis diameters of the lesion were measured
on that slice. An ROI was placed to include as much of the nodule as possible
while avoiding the outermost portion of the edges to prevent partial volume
effects. The mean attenuation over the ROI was recorded. A CT histogram was
obtained from the ROI using a software application on the workstation. A graph
of the number of pixels on the y-axis versus the pixel attenuation on
the x-axis was obtained from the ROI. Histogram analysis included
recording the total number of pixels, number of negative pixels (attenuation
< 0 H), and resultant percentage of negative pixels in each ROI.
CT Noise Estimation
CT noise was assessed by measuring SDs of the mean attenuations over an ROI
of identical size of the adrenal nodule and spleen to normalize differences in
the SDs associated with scanning technique and body habitus. The mean and
error in the relative noise (ratio of SD of adrenal to SD of spleen) for
adenomas and nonadenomas were computed.
MRI Interpretation
One observer with 8 years of experience in interpreting abdominal MR images
performed all quantitative MRI measurements. This observer was not the same
observer who performed the CT measurements and was unaware of the CT
attenuation measurements. ROIs were placed on each adrenal mass using the
largest ellipse possible while avoiding the edges where chemical shift
artifact was present. Identical ROIs were drawn on both in- and opposed-phase
MR images using the copy and paste functions of the workstation. A similar
method was used to measure the signal intensity of the kidney by avoiding
renal hilar fat and including the same amount of cortex and medulla. The
kidney was used instead of the spleen because some of our patients underwent
splenectomy and others had hemosiderosis, which limited our ability to use the
spleen as a control for signal intensity in all cases.
The percentage of signal intensity decrease was calculated using the
following formula [7], which
normalizes signal intensity to renal parenchyma,
 |
 |
Statistical Analysis
Mean attenuation values were compared using the Student's t test.
Pearson's correlation coefficients were used when assessing the CT attenuation
versus the percentage of negative pixel decrease and percentage of signal
intensity decrease. The McNemar test was used to compare CT histogram analysis
and chemical shift MRI. A p value of less than 0.05 was considered to
indicate a statistically significant difference.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
|
|
|
|
|
The relative noise in nonadenomas (1.257) was higher than that in adenomas (1.025) by 23%. This difference was statistically significant as determined by a simple Student's t test assuming a two-tailed distribution and allowing for unequal variance in the distributions (p = 0.006).
The mean percentage of signal intensity drop was 0.4% for nonadenomas versus 35.4% for adenomas (p < 0.001). A threshold of 20% signal intensity drop was used for the diagnosis of adenoma, and the 99.9% CI for mean percentage of signal intensity drop in nonadenomas was -10.8% to 11.6%. Thus, this threshold of 20% signal intensity drop was chosen retrospectively to maintain 100% specificity. The sensitivity of chemical shift MRI for adenomas was 71% (20 of 28 masses). When considering masses with an attenuation of between 10 and 30 H, the sensitivity for adenomas was 89% (17 of 19 masses). Table 1 shows the sensitivity when the mean CT attenuation was between 10 and 20 H, between 20 and 30 H, and > 30 H. The specificity for the diagnosis of adenoma was 100% (11 of 11 masses). There was an inverse correlation between CT attenuation and percentage of signal intensity decrease for adenomas (r = -0.69; p < 0.001; 95% CI, -0.85 to -0.42).
MRI allowed correct characterization of a greater number of nodules as adenoma (n = 7) than CT histogram analysis, which had classified those nodules in the indeterminate category. This resulted in a statistically significant McNemar test, p = 0.023. Overall, MRI had superior sensitivity to CT (Fig. 4 and Tables 1 and 2).
|
|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
In the present study, we show that by using a CT histogram-derived threshold of greater than 10% negative pixels, it is possible to diagnose an adenoma. To maintain 100% specificity, the sensitivity is compromised, resulting in a modest value of approximately 46%. However, when interpreting this modest sensitivity, one has to bear in mind that this is an additional gain achieved over unenhanced CT, from which we start with a sensitivity of 0% for this group of CT-indeterminate adrenal nodules.
The sensitivity for diagnosing an adenoma using a CT histogram-derived threshold of more than 10% negative pixels decreased as the unenhanced CT attenuation increased. The demonstration of a statistically significant relationship between increasing percentage of negative pixels and decreasing mean CT attenuation supported this relationship. This relationship is consistent with the fact that adenomas having abundant lipid have lower CT attenuation and therefore are likely to have a higher percentage of negative pixels. In fact, sensitivity dropped dramatically when the unenhanced CT attenuation was more than 20 H. This decline in sensitivity can be explained by two reasons. First, because of CT noise, we have elevated our threshold to diagnose an adenoma only when more than 10% negative pixels are present. Second, as the CT attenuation is rising, the percentage of negative pixels is decreasing, thus pushing these nodules under the threshold and yielding a nondiagnostic result.
In our study, a statistically significant relationship for diagnosing an adenoma by the simple presence of negative pixels in a lesion was not established. Therefore, we do not recommend that the mere presence of negative pixels be used to diagnose an adenoma. A previously published study reported no negative pixels in metastases [9]. The increased presence of negative pixels in nonadenomas or false-positives in our study may be attributed to noise. Quantitative CT noise estimations revealed statistically significant higher relative noise in the nonadenoma group than in the adenoma group. Currently, there are no quality assurance tests for histogram analysis, but the CT scanners used in our study did undergo routine calibration scans and phantom assessments to ensure standard CT attenuation values.
Because of prior studies that have shown a strong relationship between CT attenuation and percentage of signal intensity decrease on chemical shift MRI [1, 6, 11], it has been assumed that chemical shift MRI may be of little added value in the differentiation of adenoma from malignancy, particularly when attenuation is higher than 10 H. However, the number of adrenal masses with an attenuation higher than 10 H in those studies was limited, with only one to four lesions having an unenhanced CT attenuation of greater than 10 H. In our previously published study [2] using a threshold of more than 20% signal intensity drop between in- and opposed-phase chemical shift MR sequences, a sensitivity of 71% for diagnosing an adenoma was achieved while maintaining 100% specificity. In the present study, we show a higher sensitivity of 89% for adenoma with an attenuation of 10-30 H and 100% for adenoma with an attenuation of 10-20 H with a maintained specificity of 100%.
We also found in this study that MRI correctly characterized a greater number of nodules (n = 7) as adenomas than CT histogram analysis, yielding better overall sensitivity.
This study has several limitations. First, it is retrospective and there is potential for sample bias. The criteria of 10% negative pixels and 20% signal intensity drop were determined retrospectively. These numbers need to be validated as optimal threshold values in a prospective study.
Another limitation is that pathologic findings were used as a reference standard in a small percentage of patients; however, because of the ability of CT and MRI to definitively characterize adenomas, it is rare that patients with lesions diagnosed as benign will undergo surgery or biopsy in clinical practice. The minimum follow-up time in our study for the diagnosis of adenoma was 24 weeks, and it is possible that some adrenal metastases may have long doubling times. Some authors have used 1-year follow-up as a lower limit [12, 13]; however, many leading authors have used 6-month follow-up as a reference standard [5, 6, 14-16].
Our specificity was 100%, which should be viewed with skepticism. There were no collision lesions (combined metastases and adenoma), adrenocortical cancers, or renal cell carcinoma metastases in our study group. It is in those groups that one would expect specificity to be reduced. There is a paucity of data about the relative values of chemical shift MRI and washout CT for the diagnosis of adrenocortical carcinoma. Adrenocortical cancers may exhibit a signal intensity decrease on chemical shift MRI [17, 18] and may exhibit rapid washout on CT [15]; however, most are large heterogeneous masses [19, 20]. Such large heterogeneous lesions would be sampled for biopsy or resected regardless of the other chemical shift MRI or CT washout findings; thus, we think that this is not a major study limitation. Similarly, a history of renal cell carcinoma would lead to consideration of biopsy regardless of imaging findings because washout CT may not be accurate in this setting [15]. Collision lesions are rare. Inhomogeneity on CT would lead to consideration of biopsy in most of those cases as well.
CT histogram analysis is influenced by variations in CT image quality and noise levels. Because CT technology is constantly being upgraded, images from a multitude of different generation scanners from different vendors were included in this study. Also, the CT protocols changed over the years, with the most significant factor being slice collimation. In our study, many of these technical parameters were not standardized. One of the end results of this lack of standardization is the probable significant noise we encountered in this study. As a result, a number of nonadenomas showed negative pixels, although not to as high a percentage as adenomas. Nevertheless, this did compromise the study by resulting in a lowered sensitivity. However, we argue that despite high noise the results obtained using a higher threshold—to reduce the influence of noise-related false-positives—show satisfactory sensitivities in this group.
Because of the impact of noise on the results, multiple independent and preferably prospective studies at different institutions are needed to confirm these results before this technique can be recommended for main-stream clinical work. Based on the scanning techniques, individual institutions may also need to assess their individual threshold levels, although we believe those values may be close to the threshold values we found. The argument for this that most institutions follow the 10-H threshold on unenhanced CT despite using differing scanners and techniques. In this study, we do not question the current status of adrenal washout CT that is recommended rightly as the imaging protocol of choice to characterize an adrenal nodule. It would be of further clinical interest to directly assess CT histogram analysis and adrenal washout CT in a future study.
Chemical shift MRI has some limitations as an imaging technique. It is difficult to obtain images of adequate quality with a section thickness of less than 5 mm in a single breath-hold with 2D pulse sequences. Volume averaging with the signal intensity artifact on out-of-phase MR images along the adrenal margins has the potential to artificially reduce signal intensity, thus giving spurious high measurements of signal intensity decrease. To avoid this, we excluded all nodules smaller than 8 mm and were careful when performing measurements to avoid including adrenal margins. No nodule in our series was smaller than 10 mm. Three-dimensional acquisitions have the potential to improve the spatial resolution of MRI in this setting, and further studies are necessary to see whether this will help in the characterization of small adrenal masses.
In conclusion, CT histogram analysis using a threshold of 10% negative pixels has the potential to increase sensitivity for the characterization of adenomas compared with mean CT attenuation. Chemical shift MRI with a threshold of 20% signal intensity drop has added value in difficult cases and avoids the use of IV contrast material and exposure to additional ionizing radiation.
Acknowledgments
We thank George Tomlinson for statistical support and Jeffrey H. Siewerdsen
for CT noise assessment and appreciate their contributions to this
article.
|
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 Multidetector CT: Evidence-based Protocol Optimization and Interpretative Practice RadioGraphics, September 1, 2009; 29(5): 1319 - 1331. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. W. L. Boland, M. A. Blake, P. F. Hahn, and W. W. Mayo-Smith Incidental Adrenal Lesions: Principles, Techniques, and Algorithms for Imaging Characterization Radiology, December 1, 2008; 249(3): 756 - 775. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. M. Ho, E. K. Paulson, M. J. Brady, T. Z. Wong, and S. T. Schindera Lipid-Poor Adenomas on Unenhanced CT: Does Histogram Analysis Increase Sensitivity Compared with a Mean Attenuation Threshold? Am. J. Roentgenol., July 1, 2008; 191(1): 234 - 238. [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]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
