| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Genitourinary Imaging |
1 From the Department of Radiology, University of California San Francisco, Box 0628, M-372, 505 Parnassus Ave, San Francisco, CA 94143-0628. From the 2003 RSNA Annual Meeting. Received October 22, 2003; revision requested January 13, 2004; revision received June 3; accepted July 20. Address correspondence to B.M.Y. (e-mail: ben.yeh@radiology.ucsf.edu).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
MATERIALS AND METHODS: The study was approved by the committee on human research; written informed consent was not required. The authors retrospectively identified 197 patients with lung cancer who underwent serial chest or abdominal CT and did not have a focal adrenal mass at baseline CT. Two readers independently classified the morphologic features of each adrenal gland as normal, smoothly enlarged, or nodular at initial CT examination. They separately recorded the presence or absence of metastases to the adrenal glands (ie, any new focal adrenal mass) at final CT examination; a third independent reader arbitrated when interpretations were discordant (n = 11). Multivariate Cox proportional hazard models were used to assess for associations between baseline adrenal gland morphologic features and subsequent development of adrenal metastases.
RESULTS: At initial CT, reader 1 classified 253 (64%), 70 (18%), and 71 (18%) of the 394 adrenal glands and reader 2 classified 258 (65%), 45 (11%), and 91 (23%) of these glands as normal, smoothly enlarged, or nodular, respectively. The readers had moderate interobserver agreement regarding the classification of adrenal gland morphologic features (
= 0.54). Metastases subsequently developed in 13 adrenal glands in 11 patients. Cox proportional hazard models revealed no significant association between baseline adrenal gland morphologic features and subsequent development of adrenal metastases (P = .50 and P = .20 for readers 1 and 2, respectively).
CONCLUSION: In patients with lung cancer, smooth enlargement or nodularity of the adrenal glands at baseline CT is not associated with increased risk of subsequently developing adrenal metastases.
© RSNA, 2005
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
Despite these background issues, if such minor adrenal gland morphologic abnormalities are due to early metastatic disease, then we can assume that serial CT would demonstrate progression to frank metastatic disease over time. Accordingly, the key clinical question is relatively straightforward: Is a smoothly enlarged or diffusely nodular adrenal gland in a patient with lung cancer more likely to be a subsequent site of frank metastasis than is a normal adrenal gland in a patient with lung cancer? To our knowledge, no prior studies have addressed this question. Therefore, we undertook this study to determine the prognostic importance of minor morphologic abnormalities of the adrenal glands at CT in patients with lung cancer.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
Two hundred twenty-two patients were identified. The data on 25 patients with a focal adrenal mass—defined as a focal mass of any size in a smooth adrenal gland or a focal mass larger than 1 cm in diameter in a diffusely nodular adrenal gland—at baseline CT, as evaluated by either of the two readers (A.Q., B.M.Y.) in the study, were excluded from the primary analysis because such adrenal masses were considered to be suspicious for metastatic disease. The final, or primary, study population of 197 patients consisted of 98 women with a mean age of 63 years (range, 31–88 years) and 99 men with a mean age of 67 years (range, 43–91 years). Since the mean age of the men was significantly older than that of the women (P < .05), gender was included in all multivariate analyses. The mean follow-up time was 481 days (range, 2–1801 days; less than 6 months for 62 patients, 6 months to 1 year for 46 patients, longer than 1 year for 89 patients). Although two patients had mean follow-up times of less than 30 days, the length of follow-up times for all patients was accounted for in the statistical analysis of outcomes.
The histologic lung cancer type, cancer stage at the time of baseline CT, and form of treatment administered in conjunction with systemic chemotherapy in these patients were noted by one author (N.B.) on the basis of a review of all available histopathologic, radiologic, and medical records. The histologic cancer types in the study group were non–small cell in 165, small cell in 25, and unspecified in seven patients. The TNM cancer stage at presentation was I in 33, II in 11, III in 32, IV in 98, and unknown in 23 patients. One hundred thirty eight patients received systemic chemotherapy, and 59 did not.
The 25 patients whose data were excluded from our primary analysis owing to the presence of an adrenal mass at baseline CT were 13 men with a mean age of 65 years (range, 45–78 years) and 12 women with a mean age of 63 years (range, 39–81 years). There was no significant difference in age between the men and the women (P = .711, Student t test). The mean follow-up time for these patients was 222 days (range, 7–1210 days).
CT Technique
The initial and final CT examinations performed in the 197 patients in the study group, who underwent a total of 394 examinations, were conducted by using single–detector row (CT HiSpeed; GE Medical Systems, Milwaukee, Wis) (327 examinations) or four–detector row (LightSpeed; GE Medical Systems) (67 examinations) helical CT scanners. Three hundred sixty-seven examinations were performed after the intravenous administration of 150 mL of iohexol (Omnipaque 350; Nycomed Amersham, Princeton, NJ). Section collimation was 1.25–3.00 mm for six, 5 mm for 85, 7 mm for 296, and 8 mm for seven examinations. All images were contiguous. The examinations were performed to cover the chest (n = 258), chest and abdomen (n = 38), or abdomen (n = 98). The chest examinations extended from the thoracic inlet to below the diaphragm, including the adrenal glands, and were performed with a scanning delay of 40 seconds. The abdominal examinations extended from the diaphragm to below the iliac crests and were performed with a scanning delay of 70 seconds.
CT Image Interpretation
Two attending radiologists (A.Q., B.M.Y.) with, respectively, 5 and 3 years subspecialist experience in abdominal imaging, independently reviewed the images obtained at initial CT in all 197 patients on a picture archiving and communication system workstation (Impax; Agfa, Mortsel, Belgium). The readers were aware that the patients had lung cancer but were unaware of all other clinical and radiologic findings. They classified the appearance of each adrenal gland as normal, smoothly enlarged, nodular, or mass containing. Smooth enlargement was defined as a gland in which the thicknesses of both limbs were greater than 6 mm (7). Nodularity was defined as a multifocal surface irregularity without a dominant mass or a unifocal surface irregularity smaller than 1 cm in diameter. Adrenal mass was defined as a focal mass of any size in a smooth adrenal gland or a focal mass larger than 1 cm in diameter in a diffusely nodular adrenal gland.
At least 2 weeks after the initial CT image readings, the readers reviewed the images obtained at final CT in conjunction with those obtained at baseline CT and recorded the development or absence of metastases to the adrenal glands. Definite adrenal metastasis was defined as any new focal mass of any size in a smooth adrenal gland or a focal mass larger than 1 cm in diameter in a diffusely nodular adrenal gland. Discordant interpretations (n = 11) of the final CT images were arbitrated by a third independent reader (F.V.C.), who was an attending radiologist with 8 years subspecialist experience in abdominal imaging.
One patient underwent adrenalectomy for a histopathologically proved adrenal metastasis; hence, this adrenal gland was not present at follow-up CT. In this case, the histopathologic diagnosis was used as the standard of reference for purposes of outcome analysis.
For the 25 patients whose data were excluded from our primary analysis owing to the presence of an adrenal mass at baseline CT, follow-up CT images were assessed by one reader (B.M.Y.) for interval growth of the existing adrenal masses or for development of new masses. Interval growth was considered to have occurred if the adrenal mass size had increased measurably, by 30% in cross-sectional area, as determined on transverse CT images.
Statistical Analyses
The adrenal gland was the element of analysis in this study. Interobserver agreement regarding categorization of adrenal gland morphologic features was assessed by using statistics. Degree of observer agreement, expressed in statistic values, was graded as follows: values of 0–0.20 indicated slight agreement; 0.21–0.40, fair agreement; 0.41–0.60, moderate agreement; 0.61–0.80, substantial agreement; and 0.81–1.00, almost perfect agreement (8). The relationship between adrenal gland morphologic features and adrenal gland location on the right or left side was examined with the 
2 test. The relationships between adrenal gland morphologic features and the factors patient age, patient sex, histologic cancer type, cancer stage, and adrenal metastasis development were assessed by using generalized estimating equations that accounted for the presence of two adrenal glands in each patient.
The association between baseline adrenal gland morphologic features and subsequent development of adrenal metastases was examined by using survival-or-failure time analysis. This type of analysis enables a comparison of patient outcomes over a broad range of follow-up times and the inclusion of patients who have very short as well as very long follow-up times. With use of an alternative approach, such as that involving the criterion of no change in adrenal gland appearance for at least 12 months for determination of absent adrenal metastasis, patients who had aggressive disease and did not survive at least 12 months would have been excluded and the magnitude of metastasis development in patients who had late recurrences of malignancy may have been underestimated.
We used the Kaplan-Meier product-limit method rather than the life table method to analyze failure (development of metastases) data because this method enables one to calculate the survival function directly from the continuous failure times, and the Kaplan-Meier method is more robust because the resulting estimates do not depend on the grouping of data into a certain number of time intervals and analyses are not adversely affected by the inclusion of patients with very short follow-up times. If at the time of follow-up there are no metastases, these results represent censored observations because metastases may have developed after this time. All survival analysis methods, including the life table, survival distribution, and Kaplan-Meier methods, allow for such censored data.
To account for the effects of cancer stage, histologic cancer type, patient age, patient sex, type of treatment administered with systemic chemotherapy, and adrenal gland laterality on the risk of developing adrenal metastases, multivariate analysis with use of Cox proportional hazards models was performed (9). In these analyses, cancer stage, histologic cancer type, type of treatment administered with chemotherapy, patient sex, and adrenal gland laterality were categorical variables and patient age was a continuous variable. For all analyses, P values of .05 or less were considered to indicate statistical significance. One author (B.M.Y.) performed the statistical analyses by using Stata software, version 7.0 (Stata, College Station, Tex).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
= 0.54). The left-sided adrenal glands were more likely than the right-sided glands to be nodular (P < .001 for both readers), and nodular adrenal glands were more likely to be present with increasing age (P < .01 for reader 1, P < .005 for reader 2).
|
|
|
|
|
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
We believe that our findings support the use of a less ambiguous approach to reporting smooth enlargement or nodularity of the adrenal glands in patients with lung cancer: It seems reasonable to either describe such adrenal gland findings as unremarkable or indicating no focal masses or to describe them in association with a qualifier—stating, for example, that such changes do not appear to be associated with an increased risk of subsequently developing metastases. In our practice, we now prefer to use the term unremarkable, given the frequency of such minor morphologic abnormalities in the patients in our study: Readers 1 and 2, respectively, considered 141 (36%) and 136 (35%) of the 394 adrenal glands studied to be either nodular or smoothly enlarged at baseline CT.
Our study results raised an obvious question: What is the histopathologic basis of smooth enlargement or nodularity of the adrenal glands in patients with lung cancer, particularly given that such changes do not appear to be due to metastatic disease? Our results and previous reports provide possible answers: Smooth enlargement of the adrenal gland is a recognized finding in patients with malignancy and possibly represents a hyperplastic stress response to cancer (6,14). This phenomenon may have contributed to the adrenal gland enlargement in our patients, but we did not find a relationship between smooth enlargement and cancer stage, which might be expected on the basis of this putative mechanism. We found adrenal gland nodularity to be associated with older age. This suggests that such nodularity may be due to small adenomas since these tumors are known to be more common in older patients (15). Nodularity might also reflect hyperplasia. The distinction between adenomatous and hyperplastic change is probably not of clinical importance in patients without a known endocrine disorder. Although our patient population was not screened for biochemical evidence of endocrinopathy, it seems unlikely that 35%–36% of our patients had an undiagnosed adrenal endocrinopathy.
More puzzling is our finding that adrenal gland nodularity was more common on the left side. One could speculate that this finding is somehow related to differences in physiologic characteristics between the left and right adrenal glands, such as the so-called "nutcracker phenomenon" and the known increase in left renal vein pressure compared with the right renal vein pressure (16). As the left adrenal vein blood supply drains into the left renal vein, increased pressure might be transmitted to the left adrenal gland, perhaps resulting in an increased susceptibility to hyperplasia or adenomatous change. That said, we are unaware of literature that supports the theory that increased vascular pressure affects adrenal gland morphologic features. Alternatively, the explanation may be more prosaic: It is possible that the confined location of the right adrenal gland between the liver and the diaphragm reduces the visibility or frequency of such changes on the right side.
Our study had several limitations. First, it was confined to the retrospective examination of findings in patients with lung cancer, so our results may not be applicable to patients with other primary malignancies and minor morphologic abnormalities of the adrenal glands. Second, our study did not include the analysis of data from either histopathologic specimen examinations or additional CT or magnetic resonance imaging characterization of the adrenal glands to either determine the basis of the smooth enlargement or nodularity or confirm the development of metastases. That said, the CT diagnosis of adrenal metastasis based on the finding of a new adrenal mass is widely accepted in oncologic practice, and, thus, it would have been unrealistic to demand histopathologic proof in these patients.
Third, our requirement that all subjects whose data were recruited into the study had to have undergone serial CT at our institution may have introduced selection bias in the study population. For example, the large number of patients with stage IV disease (98 of 197 patients) was presumed to be due to the increased likelihood that such patients would receive chemotherapy and need to undergo serial imaging for evaluation of treatment response; however, the prevalence of stage IV disease in the general population of patients who present with lung cancer is estimated to be 40%–55% (17,18).
Fourth, the small number of patients (n = 11) with adrenal metastases at serial follow-up CT limited the power of our study: We found that the relative risk of frank metastasis developing in an adrenal gland with minor morphologic abnormalities at baseline CT in a patient with lung cancer is 0.05–5.50 times the risk of it developing in a normal-appearing adrenal gland. However, our results also indicate that the absolute risk of a patient with known lung cancer developing adrenal metastasis during follow-up is very low, irrespective of the presence of minor morphologic abnormalities of the adrenal glands at baseline. Our failure to find such an association in 197 patients who were followed up for a mean of 481 days and the 35%–36% frequency of baseline abnormalities suggest that any small association that might be revealed in a larger study would have little clinical importance and would not alter our assertion that raising the concern about adrenal metastasis on the basis of the presence of minor morphologic abnormalities is unhelpful.
Fifth, variability in the classification of adrenal gland morphologic features will undoubtedly occur in daily clinical practice, as reflected in our interobserver measure of agreement ( = 0.54). That said, neither reader found either diffuse enlargement or nodularity to be associated with increased risk of subsequently developing adrenal metastasis. This suggests that accurate classification of minor morphologic abnormalities of the adrenal glands is not clinically relevant for predicting the development of adrenal metastases.
In conclusion, in patients with lung cancer, neither smooth enlargement nor nodularity of the adrenal glands at baseline CT is associated with an increased risk of subsequently developing adrenal metastases. This observation suggests that these findings should not be interpreted as potential early signs of metastatic disease.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
Author contributions: Guarantor of integrity of entire study, B.M.Y.; study concepts, N.B., B.M.Y., F.V.C.; study design, N.B., B.M.Y., F.V.C., A.Q.; literature research, N.B., B.M.Y., F.V.C., A.Q.; clinical studies, F.V.C., B.M.Y., A.Q.; data acquisition, A.Q., B.M.Y., G.W.; data analysis/interpretation, B.M.Y., G.W., N.B., F.V.C.; statistical analysis, B.M.Y., G.W.; manuscript preparation, definition of intellectual content, editing, and final version approval, N.B., B.M.Y., F.V.C., R.S.B.; manuscript revision/review, N.B., B.M.Y., F.V.C., A.Q.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  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]
|
|
|
|
 |
|
 J. H. Song, F. S. Chaudhry, and W. W. Mayo-Smith The Incidental Adrenal Mass on CT: Prevalence of Adrenal Disease in 1,049 Consecutive Adrenal Masses in Patients with No Known Malignancy Am. J. Roentgenol., May 1, 2008; 190(5): 1163 - 1168. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
