Abdominal Radiology
CrossMark
Interobserver agreement in distinguishing large adrenal adenomas and adrenocortical carcinomas on computed tomography
Aaron J. Thomas,1 Mouhammed A. Habra,2 Priya R. Bhosale,3 Aliya A. Qayyum,3 Kareem Ahmed,4 Rafael Vicens,5 Khaled M. Elsayes3
1Department of Radiology, Baylor College of Medicine, Houston, TX, USA
2Department of Endocrine Neoplasia and Hormonal Disorders, The University of Texas MD Anderson Cancer Center, Houston, TX, USA
3Department of Diagnostic Radiology, The University of Texas MD Anderson Cancer Center, Houston, TX, USA
4Department of Radiology, Thomas Jefferson University Hospital, Philadelphia, PA, USA
5Department of Radiology, Hospital Auxilio Mutuo, San Juan, PR, USA
Abstract
Purpose: Large adrenal masses pose a diagnostic dilemma. The purpose of this study was twofold: first, to assess the degree of interobserver agreement in evaluating the morphology of pathologically proven adrenal adenomas and adrenocortical carcinomas larger than 4 cm in diameter; and second, to identify morpho- logic characteristics that correlated with the pathologic diagnosis.
Materials and methods: For this blinded, retrospective study, we collected cases of 25 adrenal adenomas and 33 adrenocortical carcinomas measuring larger than 4 cm. Two radiologists evaluated morphologic characteristics of the lesions on CT. Interobserver agreement was evaluated using kappa statistics, and the correlation of imaging characteristics with the diagnosis was evaluated using a logistic regression model.
Results: We found the highest interobserver agreement in the assessment of precontrast attenuation (A = 0.81) as well as substantial agreement in determining the shape and the presence of calcifications (A = 0.69 and 0.74, respectively). Readers agreed less often regarding the presence of fat (A = 0.48), as well as regarding the presence of necrosis, heterogeneity, and the overall impression (A = 0.15, 0.24, and 0.26, respectively). CT characteristics correlated with benignity included round shape (p = 0.02), an overall radiologic impression of a
benign lesion (p < 0.0001), the presence of fat (p = 0.01), and a precontrast attenuation of less than 10 Hounsfield units (p < 0.0001). The latter two of these characteristics were highly specific for benign pathology (93% and 100%, respectively).
Conclusion: Our study suggests that CT has the ability to consistently identify characteristics significantly corre- lated with benign vs. malignant adrenal tumors.
Key words: Adrenal cortex neoplasms-Adrenal glands-Adrenocortical adenoma-Adrenocortical carcinoma-Adrenal gland neoplasms
The adrenal glands can be affected by a wide spectrum of pathologies. Most adrenal lesions are benign and dis- covered incidentally on imaging; approximately 4% of the population will have incidental lesions measuring 1 cm or larger on CT [1]. Most lesions that are smaller than 4 cm can be readily identified as adrenal adenomas on an adrenal protocol CT on the basis of their pre- contrast attenuation and washout patterns [2].
The diagnosis of some larger adrenal lesions can be suggested by characteristically benign imaging findings, such as in myelolipomas or simple cysts. In other cases, biochemical or clinical data can suggest diagnoses such as pheochromocytoma or metastases. However, two primary considerations of lesions without these features are adrenocortical carcinoma and atypical adrenal ade- noma, which may be large and lack expected imaging features of adenomas [3].
Adrenocortical carcinoma is a rare tumor, affecting less than one individual per million in the United States [4]. It most commonly occurs in the first and the fourth ☒ decades of life, and it is slightly more common in women than in men [5-7]. Overall, approximately 60% of these tumors are hormonally active, but this proportion declines significantly in older populations [8]. Hormonal activity is also inversely related to the size of the lesion at presentation, as small hormonally active tumors are of- ten clinically apparent earlier in the disease course [9, 10]; common presentations include Cushing syndrome and virilization [11].
In patients without prior malignancy, size alone was a highly sensitive indicator of adrenocortical carcinoma in one study, with 90% sensitivity but only 24% specificity in one series using a 4-cm cutoff [12]. The likelihood of adrenocortical carcinoma increases as the size increases. Approximately 6% of all adrenal tumors that have a size range of 4-6 cm are malignant, and the probability of malignancy increases to 25% for lesions larger than 6 cm [13].
Morphologic features can also be useful for charac- terizing these lesions. Necrosis, irregular margins, heterogeneous appearance, and the absence of fat have all been associated with malignancy [14-16]. Adreno- cortical carcinomas also calcify in approximately 30% of cases [17]. Tumor morphology may more reliably indi- cate adrenocortical carcinoma than contrast washout, owing to the heterogeneity within a given lesion [18]. Nonetheless, adrenocortical carcinomas usually display a washout pattern similar to that of other malignant adrenal processes [19].
In contrast to adrenocortical carcinomas, adrenal adenomas are the most common adrenal tumors [20], and they are usually small and not hormonally active [21]. When smaller than 4 cm, approximately 70% of adenomas contain intracytoplasmic lipids and measure less than 10 HU on noncontract CT. Thus, they can confidently be considered benign [22, 23]. In the lipid- poor minority, a benign washout pattern can also be diagnostic of a benign lesion [24]. Unfortunately, atypi- cal adenomas can be large and heterogeneous, with areas of hemorrhage, calcification, and cystic degeneration, and in these cases, there can be significant imaging overlap with adrenocortical carcinomas (Fig. 1) [25].
Because no specific imaging features can definitively identify adrenal malignancies, and a substantial degree of subjectivity affects radiologists’ descriptions of suspi- cious features, larger masses are often followed with se- rial examinations or surgically excised. Guidelines for the management from the American Association of Clinical Endocrinologists and American Association of Endo- crine Surgeons and the European Society of Endocrinology advocate consideration of surgical resec-
tion of adrenal lesions greater than 4 cm [26-28], al- though guidelines from the American College of Radiology state that lesions with a precontrast attenua- tion of less than ten HU may be considered benign regardless of size [29].
Adrenalectomies are being performed at an increasing rate in the United States, owing in part to the more widespread use of abdominal imaging and the conse- quent detection of incidental adrenal lesions [30]. Despite increasing use of laparoscopic surgical techniques, which are generally associated with fewer complications than are open adrenal surgeries, the overall rate of complica- tions has recently increased; major perioperative com- plications still occur in more than 7% of patients. One reason for this increase is likely that surgeries increas- ingly involve older patients with more comorbidities [31]. In addition, an adrenalectomy costs, on average, approximately $10,000 [32]. Therefore, identifying imaging findings that allow radiologists to confidently distinguish large adrenal adenomas from adrenocortical carcinomas would help avert these complications and provide significant savings to the healthcare system.
The purpose of this retrospective study was twofold: first, to assess the degree of interobserver agreement between radiologists in evaluating the morphology of pathologically proven adrenal adenomas and adreno- cortical carcinomas larger than 4 cm in diameter; and second, to identify morphologic characteristics that cor- related with the pathologic diagnosis.
Materials and methods
Patient selection and data collection
This retrospective study was approved by our institu- tional review board and was compliant with the Health Insurance and Portability and Accountability Act. We retrospectively reviewed pathology records at our insti- tution to identify patients who had undergone resection or biopsy of adrenal adenomas or adrenocortical carci- nomas from 2004 to 2016. We identified 58 patients whose preoperative abdominal CT imaging studies showed a lesion with a maximum diameter in the axial plane of larger than 4 cm. Both precontrast and post- contrast CT images were available for 57 of the 58 in- cluded patients; 1 patient had only a single portal venous phase series available. Thirty of the studies were obtained using a dedicated adrenal mass protocol with a precon- trast scan, a 60 s delayed phase, and a 10-15 min delayed phase. The electronic medical record was used to collect patient demographic information and to assess hormone status of each tumor.
Assessment of interobserver agreement
Two radiologists with fellowship training in abdominal imaging and over 10 years experience in abdominal imaging (designated here as A and B) were blinded to the final results of pathologic examination of the resected tumors, and each reader independently assessed mor- phologic features of each lesion on preoperative CT images.
Several of the features, such as shape, margin, and heterogeneity, were determined subjectively by each reader and placed into predetermined categories. The readers classified the shape of each lesion as round, lobulated, or infiltrative (Fig. 2), and the margin was classified as either circumscribed or ill-defined. The le- sion attenuation was described as homogeneous or heterogeneous on contrast-enhanced images. Precontrast attenuation was divided into greater or less than 10 Hounsfield units (HU) by measuring a solid-appearing, enhancing portion of the tumor (Fig. 3). The exact re- gion and size to be measured was left to the discretion of the radiologist. Suspected necrosis, indicated by the presence of regions of nonenhancing low attenuation within the tumor, was reported as present or absent (Fig. 4). In 6 cases, one of the interpreting radiologists stated that they could not determine the presence or absence of necrosis, and these were excluded in the analysis of this feature. The presence of calcification and the presence of fat, determined by an attenuation of less than - 20 HU (Fig. 5), were reported as well. A sub- jective radiologic impression of benign, indeterminate, or malignant was reported on the basis of these factors and the radiologist’s overall impression.
The precontrast attenuation and greatest axial diameter of the tumors were also independently mea- sured for each type of tumor. When a tumor was heterogeneous, the attenuation of the most solid-ap- pearing portion was measured. The absolute percentage washout-obtained by dividing the difference between portal venous phase and delayed phase attenuation by the difference between the portal venous phase and
noncontrast attenuation-was also calculated for those studies that were obtained following adrenal mass pro- tocol with delayed images, which included 16 of the adrenal adenomas and 14 of the adrenocortical carci- nomas.
Statistical analyses
Demographic and clinical characteristics such as hor- mone status of patients with adrenal adenomas and those with adrenocortical carcinomas were compared using a Student t test or Chi squared test, as appropriate. Fre- quencies and percentages were calculated for each imaging feature for each reader. Kappa statistics were used to assess the degree of agreement between the two readers. A single patient’s study was unable to be opened by reader B for technical reasons, and was excluded from this portion of the analysis. Logistic regression with the generalized estimating equations method [33] was used to test for correlations between the above imaging features and pathology results because this method accounts for the correlation of scores for the same patient. A Tukey- Kramer adjustment was used to control the overall type I error rate for multiple pairwise comparisons. All tests were 2-sided, and p values of 0.05 or less were considered statistically significant. Statistical analyses were per- formed using SAS version 9 (SAS Institute, Cary, NC).
Results
Patient and imaging characteristics
The pathology reports identified 33 (57%) of the tumors as adrenocortical carcinomas and 25 (43%) as adrenal adenomas. The median age (± standard deviation) of the patients with adrenal adenoma was 58 ± 10.5 years, and the median age of the patients with adrenocortical car- cinoma was 51 ± 14.6 years. Of the 25 patients with adrenal adenoma, 13 were males and 12 females; of the 33 patients with adrenocortical carcinoma, 15 were males and 18 were females. We found no significant differences in age (p = 0.2) or sex (p = 0.62) between patients with benign and malignant lesions. Hormone activity was more common in patients with adrenocortical carcino- mas (45%) than in those with adenomas (36%), but this difference was not statistically significant (p = 0.47). In both groups, cortisol was the most common hormone produced, followed by androgens and aldosterone. The precontrast attenuation was significantly lower for the adrenal adenomas than for the adrenocortical carcino- mas (p < 0.001), although there was significant overlap. The attenuation of adenomas ranged from - 12.3 to 44.7 HU, with a median attenuation of 21.3 HU. The atten- uation of adrenocortical carcinomas ranged from 18.6 to 53.2 HU, with median of 37.8 HU. Absolute percentage washout of contrast was slightly higher in the adrenal adenoma group with median of 57% compared to med- ian of 51% in the carcinoma group. However, no statis- tically significant difference was seen (p = 0.09), and there was significant overlap, with 44% of the adenomas displaying “indeterminate” washout of less than 60%, and 36% of the adrenocortical carcinomas displaying a “benign” washout pattern of greater than 60%.
A. J. Thomas et al .: Interobserver agreement in distinguishing large adrenal adenomas
In addition, the adrenocortical carcinomas were sig- nificantly larger than the adenomas (p < 0.001). The median diameter (± standard deviation) of adrenocorti- cal carcinomas was 8.2 ± 2.1 cm, with a range of 4.1-12.1 cm, and adenomas measured 4.8 ± 1.5 cm, with a range of 4.0-9.9 cm.
Interobserver agreement
The readers’ assessments of each tumor imaging char- acteristic are detailed in Table 1, and the interobserver agreement is shown in Table 2. Although various inter- pretations of kappa scores exist, Table 3 demonstrates a common categorization of kappa statistics [34]. Sub- stantial agreement between readers was found in several domains, including assessment of precontrast attenua- tion (A = 0.81), shape (A = 0.69), and the presence of calcifications (A = 0.74). The readers agreed somewhat less often regarding the presence of fat (A = 0.48). The lowest levels of agreement were observed for the items regarding the presence of necrosis (A = 0.15), the radi- ologic impression (A = 0.26), the degree of heterogene- ity of the lesions (A = 0.24), and margin (A = - 0.02).
The latter two of these characteristics had very high percentage agreement (97% and 89%), respectively, but because the vast majority of both benign and malignant lesions were classified as circumscribed and heteroge- neous, kappa does not accurately convey agreement when data are heavily concentrated in one group. Also, the clinical utility of these features to discriminate lesions would be low given that the vast majority of the lesions appear similar.
Correlation of imaging features with pathology
Several of the above characteristics were also signifi- cantly correlated to malignant or benign pathology. Low precontrast attenuation was highly correlated with be- nign pathology (p < 0.0001), and all of the lesions rated as less than 10 HU in precontrast attenuation were be- nign in our sample, despite the fact that of these lesions, 50% were stated to have some degree of necrosis and 95% were heterogeneous. Fat-containing lesions were signifi- cantly more likely to be benign than were those without visible fat [odds ratio (OR) 0.17; 95% confidence interval (CI) 0.04-0.68; p = 0.01]. Lesions with a round shape were significantly more likely than those with a lobulated shape to be benign (OR 0.31; 95% CI 0.11-0.92; p = 0.02). In addition, lesions with a radiologic impression of benign were significantly more likely to represent benign adenomas than those rated by the radiologist as malignant (OR 0.14; 95% CI 0.06-0.34; p < 0.0001). A relatively small proportion of lesions (3-9% for each reader) were rated indeterminate for their radiologic impression, and these were not significantly correlated with either benign or malignant pathologic
diagnosis, although the majority were found to be malignant (71% overall). The presence of calcifications, degree of heterogeneity, margin, and the presence of necrosis were also not correlated with pathologic diag- nosis. The average sensitivity and specificity of the characteristics that were significantly correlated are given in Table 4.
Discussion
Our study identified several CT characteristics of adrenal tumors, which are consistently reported by radiologists and determined which of these characteristics correlated with pathologic diagnosis. The most predictive charac- teristic was precontrast attenuation. Lesions that were classified as < 10 HU in attenuation by the readers were all benign, and the lowest precontrast attenuation for an adrenocortical carcinoma was measured at 18.6 HU. Low precontrast attenuation has been demonstrated to be specific for benign adenomas [35, 36] due to the fact that adrenal adenomas often display abundant intracy- toplasmic lipid [37]. The presence of fat on CT was correlated to a benign pathologic diagnosis in our study with a specificity of 93%, and there was moderate interobserver agreement regarding its presence or ab- sence. However, macroscopic fat, likely representing myelipomatous metaplasia, is not common within ade- nomas, and has been reported in other entities as well, such as adrenocortical carcinoma and pheochromocy- toma [38]. Readers substantially agreed on the evaluation of shape, and a round shape was also significantly cor- related with benign pathology. Finally, lesions rated as likely benign by the radiologists were significantly more likely to be adenomas than those rated malignant, but agreement between radiologists on this dimension was only fair, indicating a high degree of subjectivity in this determination.
The importance of the above morphologic charac- teristics is underscored by the fact that the contrast washout pattern, which has been demonstrated to be accurate for distinguishing between small benign and malignant adrenal lesions [39], was not significantly dif- ferent between the two groups, although our study may be underpowered to detect small differences between the groups. The adrenocortical carcinomas were also gener- ally larger than adenomas, but there was significant overlap between the two groups, and this was not found to be a specific finding.
Although previous studies have stated that subjective findings such as significant heterogeneity, necrosis, or calcification on CT suggest adrenocortical carcinoma [8, 40, 41], none of these factors was significantly correlated to the pathologic diagnosis in our study, which may indicate that these are less-reliable indicators. Limita- tions of our study include the retrospective nature of our analysis and the limited sample size. In addition, the
| Reader | Characteristic | Total | Pathologic diagnosis | |||
|---|---|---|---|---|---|---|
| Adrenal adenoma | Adrenocortical carcinoma | |||||
| N | N | % | N | % | ||
| A | Shape | |||||
| Round | 31 | 17 | 55 | 14 | 45 | |
| Lobulated | 27 | 8 | 30 | 19 | 70 | |
| Infiltrative | 0 | 0 | 0 | 0 | 0 | |
| Margin | ||||||
| Circumscribed | 57 | 24 | 42 | 33 | 58 | |
| Ill-defined | 1 | 1 | 100 | 0 | 0 | |
| Heterogeneity | ||||||
| Homogeneous | 4 | 2 | 50 | 2 | 50 | |
| Heterogeneous | 54 | 23 | 43 | 31 | 57 | |
| Necrosis | ||||||
| Negative | 23 | 14 | 61 | 9 | 39 | |
| Positive | 29 | 9 | 31 | 20 | 69 | |
| Precontrast attenuation (HU) | ||||||
| < 10 HU | 12 | 12 | 100 | 0 | 0 | |
| > 10 HU | 45 | 13 | 29 | 32 | 71 | |
| Calcifications | ||||||
| No | 43 | 19 | 44 | 24 | 56 | |
| Yes | 15 | 6 | 40 | 9 | 60 | |
| Fat | ||||||
| No | 45 | 14 | 31 | 31 | 69 | |
| Yes | 13 | 11 | 85 | 2 | 15 | |
| Radiologic impression | ||||||
| Benign | 37 | 22 | 59 | 15 | 41 | |
| Indeterminate | 5 | 1 | 20 | 4 | 80 | |
| Malignant | 16 | 2 | 13 | 14 | 88 | |
| B | Shape | |||||
| Round | 30 | 17 | 57 | 13 | 43 | |
| Lobulated | 26 | 7 | 27 | 19 | 73 | |
| Infiltrative | 1 | 0 | 0 | 1 | 100 | |
| Margin | ||||||
| Circumscribed | 56 | 24 | 43 | 32 | 57 | |
| Ill-defined | 1 | 0 | 0 | 1 | 100 | |
| Heterogeneity | ||||||
| Homogeneous | 3 | 2 | 67 | 1 | 33 | |
| Heterogeneous | 54 | 22 | 41 | 32 | 59 | |
| Necrosis | ||||||
| Negative | 4 | 2 | 50 | 2 | 50 | |
| Positive | 53 | 22 | 42 | 31 | 58 | |
| Precontrast attenuation | ||||||
| < 10 HU | 8 | 8 | 100 | 0 | 0 | |
| > 10 HU | 48 | 16 | 33 | 32 | 67 | |
| Calcifications | ||||||
| No | 40 | 16 | 40 | 24 | 60 | |
| Yes | 17 | 8 | 47 | 9 | 53 | |
| Fat | ||||||
| No | 49 | 19 | 39 | 30 | 61 | |
| Yes | 8 | 5 | 63 | 3 | 38 | |
| Radiologic impression | ||||||
| Benign | 18 | 14 | 78 | 4 | 22 | |
| Indeterminate | 2 | 1 | 50 | 1 | 50 | |
| Malignant | 37 | 9 | 24 | 28 | 76 | |
Of the 58 lesions analyzed, 33 (57%) were carcinoma and 25 (43%) were adenoma. For some CT characteristics, fewer than 58 lesions were evaluable
presence of only two potential diagnoses, as well as the relative subjectivity of some of the characteristics iden- tified may limit generalizability of our data to other populations.
In conclusion, our study identified morphologic CT characteristics that correlate with the pathologic diag- nosis for large adenomas and adrenocortical carcinomas,
including precontrast attenuation, shape, and the pres- ence of fat, which can be determined by different radi- ologists with moderate or greater consistency. Two of these, attenuation and the presence of fat were both highly specific for benign pathology, although the sen- sitivity for each was low, and many lesions therefore remain indeterminate by CT. Attenuation was both more
A. J. Thomas et al .: Interobserver agreement in distinguishing large adrenal adenomas
| Characteristic | Kappa | 95% confidence interval |
|---|---|---|
| Precontrast attenuation | 0.81 | 0.61-1.00 |
| Calcifications | 0.74 | 0.54-0.93 |
| Shape | 0.69 | 0.50-0.87 |
| Fat | 0.48 | 0.20-0.76 |
| Radiologic impression | 0.26 | 0.08-0.44 |
| Heterogeneity | 0.24 | - 0.21-0.69 |
| Necrosis | 0.15 | 0.00-0.31 |
| Margin | - 0.02 | - 0.04-0.007 |
| Kappa | Interpretation |
|---|---|
| < 0 | Poor agreement |
| 0.01-0.20 | Slight agreement |
| 0.21-0.40 | Fair agreement |
| 0.41-0.60 | Moderate agreement |
| 0.61-0.80 | Substantial agreement |
| 0.81-1.00 | Almost perfect agreement |
| Characteristic | Specificity (%) | Sensitivity (%) |
|---|---|---|
| Round shape | 57 | 69 |
| Presence of fat | 93 | 33 |
| Precontrast attenuation < 10 HU | 100 | 41 |
| Benign radiologic impression | 67 | 77 |
specific and had higher interobserver agreement, indi- cating it may be the best indicator of the underlying pathology. Further study is needed to uncover additional imaging findings that can confidently and consistently predict a benign diagnosis, potentially allowing patients to undergo conservative management and avoid the cost and potential complications of surgery. This may require the use of advanced CT imaging techniques that have demonstrated promise, including dual-energy or perfu- sion studies, as well as magnetic resonance imaging or PET/CT to achieve an acceptably low risk of malignancy to avoid unnecessary surgery.
Compliance with ethical standards
Conflict of interest The authors declare that they have no conflict of interest.
Ethical approval All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards. This article does not contain any studies with animals per- formed by any of the authors.
Informed consent Informed consent was obtained from all individual participants included in the study.
References
1. Bovio S, Cataldi A, Reimondo G, et al. (2006) Prevalence of adrenal incidentaloma in a contemporary computerized tomogra- phy series. J Endocrinol Invest 29(4):298-302. https://doi.org/10. 1007/BF03344099
2. Caoili EM, Korobkin M, Francis IR, Cohan RH, Dunnick NR (2000) Delayed enhanced CT of lipid-poor adrenal adenomas. AJR Am J Roentgenol 175(5):1411-1415. https://doi.org/10.2214/ajr.17 5.5.1751411
3. Blake MA, Holalkere NS, Boland GW (2008) Imaging techniques for adrenal lesion characterization. Radiol Clin North Am 46(1):65-78. https://doi.org/10.1016/j.rcl.2008.01.003
4. Bilimoria KY, Shen WT, Elaraj D, et al. (2008) Adrenocortical carcinoma in the United States: treatment utilization and prog- nostic factors. Cancer 113(11):3130-3136. https://doi.org/10.1002/c ncr.23886
5. Wajchenberg BL, Albergaria Pereira MA, et al. (2000) Adreno- cortical carcinoma: clinical and laboratory observations. Cancer 88(4):711-736
6. Wooten MD, King DK (1993) Adrenal cortical carcinoma. Epi- demiology and treatment with mitotane and a review of the liter- ature. Cancer 72(11):3145-3155
7. Ayala-Ramirez M, Jasim S, et al. (2013) Adrenocortical carcinoma: clinical outcomes and prognosis of 330 patients at a tertiary care center. Eur J Endocrinol 169(6):891-899. https://doi.org/10.1530/ EJE-13-0519
8. Bharwani N, Rockall AG, Sahdev A, et al. (2011) Adrenocortical carcinoma: the range of appearances on CT and MRI. AJR Am J Roentgenol 196(6):W706-714. https://doi.org/10.2214/AJR.10.554 0
9. Koschker AC, Fassnacht M, Hahner S, Weismann D, Allolio B (2006) Adrenocortical carcinoma - improving patient care by establishing new structures. Exp Clin Endocrinol Diabetes 114(2):45-51. https://doi.org/10.1055/s-2006-923808
10. Fishman EK, Deutch BM, Hartman DS, et al. (1987) Primary adrenocortical carcinoma: CT evaluation with clinical correlation. AJR Am J Roentgenol 148(3):531-535. https://doi.org/10.2214/ajr. 148.3.531
11. Icard P, Goudet P, Charpenay C, et al. (2001) Adrenocortical carcinomas: surgical trends and results of a 253-patient series from the French Association of Endocrine Surgeons study group. World J Surg 25(7):891-897
12. Angeli A, Osella G, Alì A, Terzolo M (1997) Adrenal inciden- taloma: an overview of clinical and epidemiological data from the National Italian Study Group. Horm Res 47(4-6):279-283
13. Mansmann G, Lau J, Balk E, et al. (2004) The clinically inapparent adrenal mass: update in diagnosis and management. Endocr Rev 25(2):309-340. https://doi.org/10.1210/er.2002-0031
14. Boland GW, Blake MA, Hahn PF, Mayo-Smith WW (2008) Inci- dental adrenal lesions: principles, techniques, and algorithms for imaging characterization. Radiology 249(3):756-775. https://doi.o rg/10.1148/radiol.2493070976
15. Dunnick NR, Heaston D, Halvorsen R, Moore AV, Korobkin M (1982) CT appearance of adrenal cortical carcinoma. J Comput Assisted Tomogr 6(5):978-982
16. Malayeri AA, Zaheer A, Fishman EK, Macura KJ (2013) Adrenal masses: contemporary imaging characterization. J Comput Assisted Tomogr 37(4):528-542. https://doi.org/10.1097/RCT.0b013e31828 b690d
17. Elsayes KM, Emad-Eldin S, Morani AC, Jensen CT (2017) Prac- tical Approach to Adrenal Imaging. Radiol Clin North Am 55(2):279-301. https://doi.org/10.1016/j.rcl.2016.10.005
18. Johnson PT, Horton KM, Fishman EK (2009) Adrenal mass imaging with multidetector CT: pathologic conditions, pearls, and pitfalls. Radiographics 29(5):1333-1351. https://doi.org/10.1148/rg. 295095027
19. Szolar DH, Korobkin M, Reittner P, et al. (2005) Adrenocortical carcinomas and adrenal pheochromocytomas: mass and enhance- ment loss evaluation at delayed contrast-enhanced CT. Radiology 234(2):479-485. https://doi.org/10.1148/radiol.2342031876
20. Song JH, Chaudhry FS, Mayo-Smith WW (2008) The incidental adrenal mass on CT: prevalence of adrenal disease in 1,049 con- secutive adrenal masses in patients with no known malignancy.
A. J. Thomas et al .: Interobserver agreement in distinguishing large adrenal adenomas
AJR Am J Roentgenol 190(5):1163-1168. https://doi.org/10.2214/ AJR.07.2799
21. Dunnick NR, Korobkin M (2002) Imaging of adrenal inciden- talomas: current status. AJR Am J Roentgenol 179(3):559-568. h ttps://doi.org/10.2214/ajr.179.3.1790559
22. Korobkin M, Giordano TJ, Brodeur FJ, et al. (1996) Adrenal adenomas: relationship between histologic lipid and CT and MR findings. Radiology 200(3):743-747. https://doi.org/10.1148/radiol ogy.200.3.8756925
23. Boland GW, Lee MJ, Gazelle GS, et al. (1998) Characterization of adrenal masses using unenhanced CT: an analysis of the CT liter- ature. AJR Am J Roentgenol 171(1):201-204. https://doi.org/10. 2214/ajr.171.1.9648789
24. Johnson PT, Horton KM, Fishman EK (2009) Adrenal imaging with multidetector CT: evidence-based protocol optimization and interpretative practice. Radiographics 29(5):1319-1331. https://doi. org/10.1148/rg.295095026
25. Lattin GE, Sturgill ED, Tujo CA, et al. (2014) From the radiologic pathology archives: adrenal tumors and tumor-like conditions in the adult: radiologic-pathologic correlation. Radiographics 34(3):805-829. https://doi.org/10.1148/rg.343130127
26. Berland LL, Silverman SG, Gore RM, et al. (2010) Managing incidental findings on abdominal CT: white paper of the ACR incidental findings committee. J Am Coll Radiol 7(10):754-773. https://doi.org/10.1016/j.jacr.2010.06.013
27. Zeiger MA, Thompson GB, Duh QY, et al. (2009) The American association of clinical endocrinologists and American association of endocrine surgeons medical guidelines for the management of adrenal incidentalomas. Endocr Pract 15(Suppl 1):1-20. https://doi. org/10.4158/EP.15.S1.1
28. Fassnacht M, Arlt W, Bancos I, et al. (2016) Management of adrenal incidentalomas: european society of endocrinology clinical practice guideline in collaboration with the European Network for the study of adrenal tumors. Eur J Endocrinol 175(2):G1-G34. https://doi.org/10.1530/EJE-16-0467
29. Mayo-Smith WW, Song JH, Boland GL, et al. (2017) Management of incidental adrenal masses: a white paper of the acr incidental findings committee. J Am Coll Radiol 14(8):1038-1044. https://doi. org/10.1016/j.jacr.2017.05.001
30. Gallagher SF, Wahi M, Haines KL, et al. (2007) Trends in adrenalectomy rates, indications, and physician volume: a statewide
analysis of 1816 adrenalectomies. Surgery 142(6):1011-1021. https://doi.org/10.1016/j.surg.2007.09.024
31. Murphy MM, Witkowski ER, Ng SC, et al. (2010) Trends in adrenalectomy: a recent national review. Surgical endoscopy 24(10):2518-2526. https://doi.org/10.1007/s00464-010-0996-z
32. Monn MF, Calaway AC, Mellon MJ, et al. (2015) Changing USA national trends for adrenalectomy: the influence of surgeon and technique. BJU Int 115(2):288-294. https://doi.org/10.1111/bju.12 747
33. Liang K-Y, Zeger SL (1986) Longitudinal data analysis using generalized linear models. Biometrika 73(1):13-22
34. Landis JR, Koch GG (1977) The measurement of observer agree- ment for categorical data. Biometrics 33(1):159-174
35. Korobkin M, Brodeur FJ, Yutzy GG, et al. (1996) Differentiation of adrenal adenomas from nonadenomas using CT attenuation values. AJR Am J Roentgenol 166(3):531-536. https://doi.org/10. 2214/ajr.166.3.8623622
36. Lee MJ, Hahn PF, Papanicolaou N, et al. (1991) Benign and malignant adrenal masses: CT distinction with attenuation coeffi- cients, size, and observer analysis. Radiology 179(2):415-418. http s://doi.org/10.1148/radiology.179.2.2014283
37. Elsayes KM, Mukundan G, Narra VR, et al. (2004) Adrenal masses: mr imaging features with pathologic correlation. Radio- graphics 24(Suppl 1):S73-86. https://doi.org/10.1148/rg.24si045514
38. Adam SZ, Nikolaidis P, Horowitz JM, et al. (2016) Chemical shift MR imaging of the adrenal gland: principles, pitfalls, and appli- cations. Radiographics 36(2):414 432. https://doi.org/10.1148/rg.2 016150139
39. Peña CS, Boland GW, Hahn PF, Lee MJ, Mueller PR (2000) Characterization of indeterminate (lipid-poor) adrenal masses: use of washout characteristics at contrast-enhanced CT. Radiology 217(3):798-802. https://doi.org/10.1148/radiology.217.3.r00dc2979 8
40. Taffel M, Haji-Momenian S, Nikolaidis P, Miller FH (2012) Adrenal imaging: a comprehensive review. Radiol Clin North Am 50(2):219-243. https://doi.org/10.1016/j.rcl.2012.02.009
41. Hussain S, Belldegrun A, Seltzer SE, et al. (1985) Differentiation of malignant from benign adrenal masses: predictive indices on com- puted tomography. AJR Am J Roentgenol 144(1):61-65. https://doi.org/10.2214/ajr.144.1.61