A Modern Assessment of Cancer Risk in Adrenal Incidentalomas Analysis of 2219 Patients
Bora Kahramangil, MD,* Emin Kose, MD,* Erick M. Remer, MD, t Jordan P. Reynolds, MD,} Robert Stein, MD, § Brian Rini, MD, | Allan Siperstein, MD,* and Eren Berber, MD* ☒
Objective: The aim of this study was to analyze the incidence of and risk factors for adrenocortical carcinoma (ACC) in adrenal incidentaloma (AI). Summary of Background Data: AI guidelines are based on data obtained with old-generation imaging and predominantly use tumor size to stratify risk for ACC. There is a need to analyze the incidence and risk factors from a contemporary series.
Methods: This is a retrospective review of 2219 AIs that were either surgically removed or nonoperatively monitored for ≥12 months between 2000 and 2017. Multivariate logistic regression was performed to define risk factors. ROC curves constructed to determine optimal size and attenuation cut-offs for ACC.
Results: 16.8% of AIs underwent upfront surgery and rest initial nonopera- tive management. Of conservatively managed patients, an additional 7.7% subsequently required adrenalectomy. Overall, ACC incidence in AI was 1.7%. ACC rates by size were 0.1%, 2.4%, and 19.5% for AIs of <4, 4 to 6, and >6 cm, respectively. The optimal size cut-off for ACC in AI was 4.6 cm. ACC risks by Hounsfield density were 0%, 0.5%, and 6.3% for lesions of <10, 10 to 20, and >20 HU, with an optimal cut-off of 20 HU to diagnose ACC. 15.5% of all AIs and 19.2% of ACCs were hormonally active. Male sex, large tumor size, high Hounsfield density, and >0.6 cm/year growth were indepen- dent risk factors for ACC.
Conclusion: This contemporary analysis demonstrates that ACC risk per size in AI is less than previously reported. Given these findings, modern manage- ment of AIs should not be based just on size, but a combination of thorough hormonal evaluation and imaging characteristics.
Keywords: adrenal incidentaloma, adrenocortical carcinoma, management (Ann Surg 2022;275:e238-e244)
A drenal incidentaloma (AI) is a common diagnosis, reported in 4.4% of patients undergoing abdominal computed tomography (CT). Its prevalence increases up to 10% in the elderly. With an aging population and the use of higher resolution scanners, the incidence of AI is likely to increase.1-4
Adrenocortical carcinoma (ACC) is the most concerning diagnosis that needs to be ruled out in a patient with AI. Although rare with an incidence of 2 per million,5 the diagnosis must not be missed, due to the grim prognosis of metastatic ACC.6 On the other
From the *Department of Endocrine Surgery, Cleveland Clinic, Cleveland, OH; +Imaging Institute, Cleveland Clinic, Cleveland, OH; ¿ Pathology and Labora- tory Medicine Institute, Cleveland Clinic, Cleveland, OH; §Department of Urology, Cleveland Clinic, Cleveland, OH; and [Department of Hematology and Medical Oncology, Cleveland Clinic, Cleveland, OH.
No financial support was received for this study.
The authors report no conflicts of interest.
Supplemental digital content is available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF versions of this article on the journal’s Web site (www.annalsofsurgery.com).
Copyright @ 2020 Wolters Kluwer Health, Inc. All rights reserved. ISSN: 0003-4932/20/27501-238
DOI: 10.1097/SLA.0000000000004048
hand, adrenalectomy is associated with a morbidity of 4.8% to 8.4%.7,8 Therefore, it is essential to balance the risks and potential benefits when selecting patients for adrenalectomy.
Paradigm shifts are occurring in the management of AIs with the “adrenalectomy by size recommendation in hormonally inactive tumors”9 being challenged4 and more conservative use of follow-up imaging studies recommended.1º In this era, it is essential to accu- rately redefine the risk of ACC based on clinical parameters. Currently, the management guidelines primarily rely on the literature that reported a 1.9% to 4.7% incidence of ACC and predicted a 2%, 6%, and 25% ACC risk for tumors <4cm, 4 to 6cm, and >6 cm, respectively.11-15 However, the references cited by these reviews are generally limited by small sample size, use of old generation, low- resolution scanners, and the inclusion of some purely surgical series, which could potentially overestimate the actual ACC risk.
The hypothesis of this study is that the true incidence of ACC based on size of AI is different in the current era. Therefore, the aim of this study is to perform a modern assessment of cancer risk in AI.
METHODS
This was an institutional review board-approved retrospective study of patients with AIs between 2000 and 2017. Patients with incidentalomas were identified through an Epic (Epic Systems, Verona, WI) database search. An AI was defined as a serendipitously discovered adrenal lesion on radiographic imaging, measuring >1 cm 1.
Inclusion criteria included a clearly defined AI per the above definition and either having undergone surgical resection and formal pathologic examination or at least 12 months of clinical and radio- logic follow-up for nonoperatively managed patients. Following patients were excluded from the study: patients with any known, active malignancy with a potential to metastasize to the adrenal gland, presence of signs or symptoms suggestive of hormonal hyperfunction, having undergone hormonal work-up before imaging and a question of an adrenal pathology already raised, and abdominal pain attributable to an adrenal mass. When a patient presented with vague abdominal pain not localized to the adrenal and had other findings on cross-sectional imaging which could explain the symp- toms (eg, small bowel obstruction, gallbladder pathology, and so on), identified adrenal lesions were considered incidentalomas.
Of nonoperatively managed lesions, those with <0.6 cm/year growth and no other signs of malignancy were presumed to be benign.16 All ACCs were confirmed pathologically.
Complete hormonal workup was defined as the evaluation of all 3 of the following axes: hypothalamicuitary-adrenal (cortisol) axis, renin-angiotensin-aldosterone axis, and catecholamine axis.1 Partial hormonal work-up was defined as evaluation of only 1 or 2 of these axes. Acceptable testing for hormonal activity included a clinically appropriate combination of serum adrenocorticotropic hormone, AM serum cortisol, midnight salivary cortisol, dexamethasone suppression test, and 24-hour urine cortisol for the hypothalamicuitary-adrenal axis; serum aldosterone and renin, as well as a 24-hour urinary aldosterone for the renin-angiotensin-aldosterone axis; and serum
Annals of Surgery . Volume 275, Number 1, January 2022
and/or urine catecholamine and metanephrine/normetanephrine levels for catecholamine axis.
In patients who had undergone a noncontrast computed tomography (CT), or a contrast CT with noncontrast views available, Hounsfield unit (HU) densities were obtained from radiology reports when specified, or otherwise manually calculated. In patients who underwent an adrenal protocol CT, a 15-minute washout was per- formed and an absolute washout value was calculated. A threshold of >60% was used as a cut-off to predict adrenal adenoma.
Univariate logistic regression was performed to determine the factors associated with an increased risk of ACC. Parameters with a P value of <0.1 on the univariate model were included in the multi- variate model. Receiver-operating characteristics (ROC) curves were constructed to determine the optimal size and HU density cut-offs to diagnose ACC in AI. A P value of <0.05 was considered statistically significant. Statistical analyses were performed using JMP version 14.3 (SAS Studio, Cary, NC).
RESULTS
Description of Cohort
A search with keyword “adrenal” at the Cleveland Clinic Epic (Epic Systems, Verona, WI) database yielded 14,495 results. From these 14,495 charts reviewed, 4791 patients with AIs were identified. Of these, 2219 patients met the study inclusion criteria.
Supplemental Table 1, http://links.lww.com/SLA/C236 summa- rizes demographic and clinical data. Mean size of AIs at diagnosis was 2.8 cm and noncontrast Hounsfield density 9.3 HU. 13.7% (n = 304) of AIs were heterogeneous, 2.7% (n = 59) irregular, and 0.8% (n = 18) invading the surrounding tissues on cross-sectional imaging.
AIs were most commonly detected by an abdominopelvic CT. 72.9% (n = 1617) of the patients had a noncontrast abdominal CT available to review. 58.7% (n = 1302) of patients had undergone a complete, 14.3% (n = 317) partial, and 27.0% (n = 600) no hormonal work-up. 17.2% (n = 382) of patients underwent surgery upfront and
82.8% (n = 1837) nonoperative monitoring. Of the initially non- surgically managed patients, 8.7% (n = 192) subsequently under- went adrenalectomy due to increase in tumor size (n = 109), development of hormonal activity (n = 63), patient preference (n = 18), and development of flank pain (n = 2). Final diagnosis was adrenocortical adenoma (n = 313, 14.1%), pheochromocytoma (n = 97, 4.4%), ACC (n = 38, 1.7%), other tumors (n = 127, 5.7%), and presumed benign pathology due to <0.6 cm/year growth (n = 1644, 74.1). Of the 38 pathologically confirmed ACCs in the cohort, 26 underwent complete hormonal evaluation preoperatively.
Clinical Behavior of AIs
Figure 1 summarizes the clinical features and management details of 1302 AIs with complete hormonal work-up. Overall, 15.7% (n = 204/1302) of AIs were hormonally active and 84.3% (n = 1098) inactive. Hormonal activity was related to pheochromocytoma (7.1%, n = 93), subclinical Cushing syndrome (5.0%, n = 65), primary hyperaldosteronism (3.2%, n = 41), hyperandrogenism (0.1%, n = 1), and concomitant cortisol and steroid secretion (0.3%, n =4). Of the 26 ACCs with complete preoperative hormonal evaluation, 19.2% (n = 5) were hormonally active, with an isolated cortisol secretion in 1 patient (3.8%) and concomitant secretion of cortisol and sex steroids in 4 (15.4%).
Of the 1098 hormonally inactive AIs, 16.8% (n = 184) required surgery upfront due to large size (n = 86), high HU density (n = 26), heterogeneity/irregularity/invasion on imaging (n = 20), simultaneous presence of more than one of these factors (n = 42), and patient preference (n = 10). The size range for these lesions was 1.5 to 10.0 cm. After an initial nonsurgical follow-up (median time to surgery 14 months), an additional 7.7% (n = 85) of AIs subsequently required surgery. 75.5% (n = 829) of hormonally inactive AIs never required surgery at a median follow-up of 45 months (range: 12-197 months) (size range: 1.0-6.3 cm). Of the nonoperated AIs, 76.1% (n = 631/829) remained stable in size and 23.9% (n = 198/829) grew at a rate of <0.6 cm/year. In follow-up of 1098 hormonally inactive
Pheochromocytoma (n=93) (7.1%)
Subclinical Cushing (n=65) (5.0%)
Total patients reviewed (n=14495)
Functional (n=204) (15.7%)
Hyperaldosteronism (n=41) (3.2%)
Incidentaloma (n=4791)
No or incomplete hormonal work-up - excluded (n=917)
Hyperandrogenism (n=1) (0.1%)
Meeting inclusion criteria (n=2219)
Complete hormonal work-up (n=1302)
Cortisol and androgen secretion (n=4) (0.3%)
Non-operative follow- up (n=914) (83.2%)
Adrenalectomy after non-surgical follow-up (n=85) (7.7%)
Non-functional (n=1098) (84.3%)
Stable size (n=631/829) (76.1%)
Upfront adrenalectomy (n=184) (16.8%)
Never required surgery (n=829) (75.5%)
<0.6cm/year growth (n=198/829) (23.9%)
| TABLE 1. A Comparison of Incidental Adrenocortical Carcinomas With Other Incidentally Detected Adrenal Tumors | |||
|---|---|---|---|
| Parameter | ACC (n = 38) | Other (n = 2181) | P |
| Male sex, count (%) | 24 of 38 (63.2) | 826 of 2181 (37.2) | 0.002 |
| Age, y, mean (SD) | 55.3 (16.5) | 60.2 (13.2) | 0.08 |
| BMI, kg/m2, mean (SD) | 30.5 (7.8) | 31.4 (7.8) | 0.52 |
| Imaging features | |||
| Largest dimension, cm, mean (SD) | 9.7 (5.2) | 2.7 (1.8) | <0.001 |
| Hounsfield density, mean (SD)* | 33.2 (9.2) | 9.0 (15.3) | <0.001 |
| Heterogeneity, count (%)* | 27 of 34 (79.4) | 277 of 2181 (12.7) | <0.001 |
| Irregularity, count (%)1 | 17 of 34 (50.0) | 42 of 2181 (1.9) | <0.001 |
| Invasion, count (%)1 | 10 of 34 (29.4) | 8 of 2181 (0.4) | <0.001 |
| At least 1 worrisome feature, count (%)+ | 28 of 34 (82.4) | 293 of 2181 (13.4) | <0.001 |
| 80.6 cm/year growth on follow-up, count (%)* | 6 of 8 (75.0) | 40 of 1829 (2.2) | <0.001 |
| Hormonal activity, count (%) $ | |||
| Nonfunctioning | 21 of 26 (80.8) | 1078 of 1276 (84.5) | 0.61 |
| Subclinical Cushing | 1 of 26 (3.8) | 63 of 1276 (4.9) | 0.80 |
| Primary hyperaldosteronism | — | 41 of 1276 (3.2) | — |
| Hyperandrogenism | — | 1 of 1276 (0.1) | — |
| Pheochromocytoma | — | 93 of 1276 (7.3) | — |
| Concomitant cortisol and sex steroid secretion | 4 of 26 (15.4) | — | — |
BMI, Body mass index; SD, Standard deviation.
*Includes patients who underwent noncontrast computed tomography at least once (n = 1064). +Tumor imaging features (heterogeneity, irregularity, invasion) could not be obtained for 4 ACCs. #Includes patients who were initially managed nonoperatively with or without subsequent surgery (n = 1837). §Includes patients who underwent complete hormonal work-up (n = 1302).
AIs, 4.5% (49/1098) were found to be secreting upon repeat hor- monal testing, prompting their surgical removal.
Of the 21 hormonally inactive ACCs, 16 underwent adrenalec- tomy upfront. Two additional patients underwent adrenalectomy after initial nonoperative management, due to increase in size at 16 and 19 months, respectively (respective growth rates: 2.5 cm/year and 0.3 cm/ year). One patient was initially found to have a 7-cm adrenal lesion, but refused surgery. Five years later, the patient returned with a 16-cm adrenal mass causing flank pain, which proved to be ACC upon surgical resection (growth rate: 1.8 cm/year). Another patient had a 3.5 cm adrenal mass with a HU density of 30, but refused surgery. Patient was lost to follow-up and presented 10 years later with a metastatic ACC, which was confirmed by core needle biopsy (growth rate: 1.8 cm/year). Finally, 1 hormonally inactive ACC, metastatic at presentation and managed with palliative chemotherapy, remained stable in size during the duration of chemotherapy.
All hormonally active ACCs were surgically removed, except for 1 in a patient with end-stage cardiomyopathy with a cortisol and sex steroid-secreting tumor diagnosed with core needle biopsy and managed palliatively (growth rate 1.1 cm/year). Lastly, 2 ACCs without hormonal work-up were surgically removed after 6 months of clinical follow-up due to growth in size (2 cm and 2.5 cm over 6 months, respectively). These lesions were excluded from growth rate calculation owing to <12 months of follow-up.
Comparison of Incidentally Detected ACCs to Other AIs
Patients with incidentally detected ACCs were more fre- quently males (63.2% vs 37.2% for ACC vs other pathologies; P = 0.002) (Table 1). ACCs were larger at presentation and had a higher HU density compared to other incidental adrenal lesions. In addition, ACCs were more frequently heterogeneous, irregular, and/or invading on cross-sectional imaging. When nonsurgically managed, ACCs more frequently grew at a rate of >0.6 cm/year (75.0% vs 2.2%, P < 0.001) [range for ACCs: 0.3-2.6cm/ year; range for non-ACCs: 0.009-3 cm/year (lesions which were surgically removed after a follow-up of <12 months were excluded)]. Of note, only 1 ACC grew at a rate of 0.3 cm/year. The growth rates
for the remaining ACCs were all ≥1.1 cm/year. Concomitant secre- tion of cortisol and sex steroids was only observed in ACCs.
Risk Factors for ACC in Different Clinical Scenarios
Supplemental Table 2, http://links.lww.com/SLA/C237 sum- marizes the factors associated with an increased risk of ACC in AI for different clinical scenarios. At the initial evaluation of all patients with AI, independent risk factors for ACC were male sex [odds ratio (OR) 4.36 (univariate OR was inadvertently quoted, correcting to multivariate OR)], tumor size [OR with each cm increase 1.41 (univariate OR was inadvertently quoted, correcting to multivariate OR)], and HU density [OR with each unit increase 1.17 (univariate OR was inadvertently quoted, correcting to multivariate OR)]. The same factors predicted ACC risk in a subset analysis of hormonally inactive tumors. Upon nonoperative follow-up of hormonally inac- tive AIs, larger tumor size (OR: 1.93) at presentation and >0.6 cm/ year growth (OR: 60.3) on follow-up were associated with an increased risk for ACC. High HU density approached statistical significance in this group.
ACC Risk by Size, Hounsfield Density, and Hormonal Activity
Table 2 summarizes the ACC risk by size, Hounsfield density, and hormonal activity. ACC rate in surgically removed AIs was 6.6%, with size-stratified rates of 0.8%, 3.5%, and 21.4% for <4 cm, 4 to 6 cm, and >6 cm, respectively. ACC rate in the entire cohort (including nonsurgically managed patients) was 1.7%. When strati- fied by size for the whole cohort, ACC rates were 0.1%, 2.4%, and 19.5% for lesions <4 cm, 4 to 6cm, and >6cm, respectively. On ROC curve, the optimal size cut-off that predicted ACC was 4.6 cm (sensitivity 92.1%, specificity 89.4%, area under curve 0.96) (Fig. 2).
ACC risk in AIs with Hounsfield densities of <10 HU, 10 to 20 HU, and >20 HU was 0%, 0.5%, and 6.3%, respectively. On ROC curve, the optimal cut-off that predicted ACC was 20 HU (sensitivity 94.1%, specificity 76.8%, area under curve 0.912).
ACC risks in AIs with isolated cortisol secretion and no hormonal activity were 1.5% and 1.9%, respectively. In the current series, 100% of the tumors with concomitant cortisol and sex steroid
@ 2020 Wolters Kluwer Health, Inc. All rights reserved.
| Parameter | No. of Patients | ACC No. of | Risk ACC |
|---|---|---|---|
| Risk by size (AIs reviewed on pathology specimens) * | |||
| <4 cm | 264 | 2 | 0.8% |
| 4-6 cm | 171 | 6 | 3.5% |
| >6 cm | 140 | 30 | 21.4% |
| All lesions | 575 | 38 | 6.6% |
| Risk by size (all AIs)* | |||
| <4 cm | 1815 | 2 | 0.1% |
| 4-6 cm | 250 | 6 | 2.4% |
| >6 cm | 154 | 30 | 19.5% |
| All lesions | 2219 | 38 | 1.7% |
| Risk by HU density | |||
| <10 HU | 883 | 0 | – |
| 10-20 HU | 399 | 2 | 0.5% |
| >20 HU | 336 | 21 | 6.3% |
| Risk by hormonal activityt | |||
| Nonfunctioning | 1098 | 21 | 1.9% |
| Isolated cortisol secretion | 65 | 1 | 1.5% |
| Isolated aldosterone secretion | 41 | — | — |
| Isolated sex steroid secretion | 1 | — | — |
| Isolated catecholamine secretion | 93 | — | — |
| Concomitant cortisol and sex steroid secretion | 4 | 4 | 100% |
*Three ACCs were diagnosed by core needle biopsy due to metastatic disease (n = 2) and the patient being a poor operative candidate (n = 1).
+Only includes tumors with complete hormonal evaluation. All of these tumors were clinically silent without signs and symptoms of hormonal hyperfunction.
secretion were ACCs. No ACC presented with isolated aldosterone secretion, isolated sex steroid secretion, or catecholamine secretion.
The Utility of CT Washout in the Risk Assessment of AIs
An adrenal protocol CT with washout calculation was per- formed in 202 AIs (Supplemental Table 3, http://links.lww.com/SLA/
C238). Overall, a CT washout of >60% had a 99% (n = 178/179) negative predictive value in ruling out malignancy, whereas a washout of <60% falsely predicted malignancy in 91% (21/23) patients.
DISCUSSION
This study demonstrates that the contemporary risk of ACC per size in AI is less than previously reported in the literature. In contrast to the numbers quoted in the literature,11 the risk of ACC increased at a size cut-off of >4.6 cm and HU density of ≥20.
There are discrepancies regarding the recommendations for management of AIs by different expert societies regarding an aggressive9 versus a more conservative approach.1º Furthermore, traditional algorithms are being challenged arguing that diagnostic studies and adrenalectomy may be being overutilized.4,15,17,18 We believe the present study will provide important data in an era when the management of adrenal tumors is evolving. To our knowledge, this study represents the largest experience to date on AI. By including both surgical and nonoperatively managed AIs evaluated with modern imaging technology, the risk of ACC was calculated.
The results of this study underline the importance of a complete hormonal evaluation in the work-up of an AI. Despite the absence of any signs or symptoms indicating hormonal hyperac- tivity, 15.5% of all AIs and 19.2% of ACCs were hormonally active upon comprehensive biochemical testing. Figure 3 depicts the pro- posed AI management algorithm based on the findings of this study and Figure 4 the performance of this approach based on data from the current series. The proposed management algorithm was able to channel all ACCs in this series to adrenalectomy. By moving the hormonal evaluation to the first step of the decision algorithm, almost 20% of all ACCs were appropriately triaged to surgery, irrespective of the imaging findings. The nature of the hormonal activity also appeared important in predicting the risk of ACC. All 4 of the AIs with concomitant secretion of cortisol and sex steroids in this series were ACCs. A lack of hormonal function or subclinical Cushing syndrome at presentation, on the contrary, could be observed in both benign AIs and ACCs with similar rates. The rate of hormonal function in AIs has been reported to range between 10% and 15% in the literature, which is in line with our results. 15,19
ROC Curve
ROC Curve
1.0
1.0
0.8
0.8
Sensitivity
0.6
Sensitivity
0.6
0.4
0.4
0.2
0.2
0.0
0.0
0.2
0.4
0.6
0.8
1.0
0.0
0.0
0.2
0.4
0.6
0.8
1.0
A
1 - Specificity
B
1 - Specificity
Adrenal incidentaloma
Hormonally active
Hormonally inactive
Adrenalectomy
≥4.6 cm, ≥20 HU and/or heterogeneous/ irregular/invading
<4.6 cm, <20 HU and no worrisome radiologic features
Adrenalectomy
Non-operative follow- up
When dealing with hormonally inactive AIs, imaging features become important. In agreement with the previously published literature,19 tumor size and Hounsfield density were independent predictors of ACC risk in the multivariate model. Of note, male sex was also associated with an increased risk of ACC in AI. When a subset analysis of non-operatively managed patients with hormonally inactive tumors was performed, >0.6 cm/year growth appeared to be a strong predictor of malignancy. To our knowledge, clinical sce- nario-specific (initial evaluation of all incidentalomas, initial evalu- ation of nonfunctioning incidentalomas, and during nonoperative follow-up of nonfunctioning incidentalomas) risk stratification of AIs is being reported for the first time in the literature.
The growth rate cut-off of 0.6 cm/year utilized in this study was previously reported.16 A recent report, on the contrary, suggested
ACCs with complete hormonal work-up (n=26)
Hormonally active (n=5)
Hormonally inactive (n=21)
Size <4.6 cm but Hounsfield density ≥20 HU (n=2)
Adrenalectomy
Size≥4.6 cm (n=19)
Adrenalectomy
a cut-off of 0.3 cm/year.20 In the current series, the growth rate of ACCs ranged between 0.3 and 2.6 cm/year and non-ACCs between 0.009 and 3.0 cm/year. Among 914 AIs which were initially non- surgically managed with or without subsequent surgery, only 2 ACCs grew at a rate <0.6 cm/year and 40 non-ACC AIs at a rate >0.6 cm/ year. The cut-off correctly predicted ACC in 96% (874/914) of AI cases.
In the 2002 NIH consensus statement, the ACC risks by size in AIs were reported as 2%, 6%, and 25% for lesions of <4cm, 4 to 6 cm, and >6 cm, respectively.11 Our results indicate that the actual rate is considerably lower. When only surgically resected AIs were taken into consideration, ACC risks for respective groups were 0.8%, 3.5%, and 21.4%. However, we believe that this is an inaccurate representation of the ACC prevalence, as patients with worrisome findings are preferentially channeled to surgery, resulting in a biased population. When all patients (surgically and non-operatively man- aged) were analyzed, ACC risks in the respective groups were 0.1%, 2.4%, and 19.5%. In the leading reviews of the literature, ACC prevalence in AI was estimated to be between 1.9 and 4.7%.12-15 In our series, ACC risk in all-comers with AI was 1.7%, indicating a general overestimation of the cancer risk in the literature. The optimal size cut-off to diagnose ACC was found to be 4.6 cm, which is larger than the 4 cm cut-off suggested by the National Italian Study Group19 and the 2009 AACE and AAES guidelines.9
It is established in the literature that AIs with a Hounsfield density of <10 HU and those with 10 to 15 HU without any other concerning imaging features are almost exclusively benign adeno- mas.21,22 In the current series, no ACC presented with a Hounsfield density of <10 HU. The rates of ACC in lesions with densities of 10 to 20 HU and >20 HU were 0.5% and 6.3%, respectively. The optimal Hounsfield density that predicted ACC was 20 HU, with a 94.1% sensitivity. The mean Hounsfield density of ACCs was 33.2 HU, which is similar to the previously reported value of 36.9 HU by Hamrahian et al.22
The current data on the utility of CT washout are mostly from the radiology literature reporting a high accuracy in differentiating benign adenomas from nonadenomatous lesions.22-24 In our
@ 2020 Wolters Kluwer Health, Inc. All rights reserved.
experience, a CT washout of >60% had an overall 99% negative predictive value in ruling out malignancy. However, a CT washout of <60% falsely predicted malignancy in 91% patients.
Magnetic resonance imaging (MRI) for AI is not routinely performed in our clinical practice and hence was not analyzed in this study. In the radiology literature, there has been an overall shift away from the use of MRI for adrenal lesions, with CT being considered as the first-line test, as it allows both Hounsfield density and washout measurements. MRI, on the contrary, can only assess for signal intensity drop and there is no agreed-upon MR washout test.25,26 In a more recent study, CT was found to be superior to chemical-shift MRI for adenomas with a Hounsfield density of >20 HU.27
Although it was reported that 8.0% to 21.3% of AIs increased in size and 45.8% to 90.2% remained stable during clinical follow- up,15 no studies to-date have specifically focused on hormonally inactive AIs. As shown in Figure 1, only 16.8% of hormonally inactive AIs required upfront and 7.7% subsequent surgical removal. Of the 75.5% that never required surgery, 76.1% remained stable in size and 23.9% grew at a rate <0.6 cm/year. When a hormonally inactive AI was nonsurgically followed up (n = 1098 in the current series), 4.5% were found to later gain hormonal activity. There are scant data in the literature on the acquisition hormonal activity for nonoperatively managed AIs.28
Although it was a risk factor in the univariate model, presence of at least 1 worrisome feature on cross-sectional imaging (hetero- geneity, irregularity, or invasion) did not retain statistical significance as a predictor of ACC in the multivariate model. In the literature, these features were reported to be associated with an increased risk for ACC.29 The present study dichotomized AIs as ACC versus others. Similar to previous findings, ACCs in the present study demonstrated a higher rate of heterogeneity, irregularity, and inva- sion. Non-ACC adrenal lesions also included pheochromocytoma and myelolipoma, which are known to be often heterogeneous,30, as well as benign adrenal adenomas which were removed due to imaging findings raising suspicion for a potential malignancy. Although imperfect in detecting malignancy, we still recommend surgical removal of lesions demonstrating heterogeneity, irregularity, and/or invasion on cross-sectional imaging, with the exception of clearly defined myelolipomas, which can be heterogeneous, but are readily differentiated from other lesions by their very low Hounsfield density.31 Given the benefit of surgically treating ACC at an early stage, we believe that such an aggressive surgical approach would be justified.
This study has several weaknesses due to its retrospective design. The imaging intervals were inconsistent in some patients due to interval scans being performed for other reasons. These studies, as well as the ones obtained beyond the recommended clinical follow- up, were taken into consideration to better reflect the long-term behavior of AIs. Due to the irregularity in imaging intervals, we did not attempt to define a new growth rate cut-off, but instead utilized a previously described one in the literature. Another weakness is that a testing for sex steroid secretion was performed upon an increased suspicion for ACC and not in all study patients. Although a concom- itant secretion of cortisol and sex steroids was found to be strongly associated with ACC, a selection bias resulting from selective testing cannot be excluded.
CONCLUSIONS
By including all-comers with surgically and nonsurgically managed AIs, this study redefined the risk of ACC per size. The results demonstrate that the actual risk of ACC is lower than previously reported and hence used to derive the guidelines in use today. Moving forward, we recommend revising the AI
guidelines to start with a full hormonal evaluation, with surgical removal of all hormonally active tumors. A significant number of ACCs will be detected at this step. Next, the imaging character- istics should be reviewed. According to the results of this study, AIs of ≥4.6 cm and/or Hounsfield density of >20 HU should be surgically removed due to the increased risk of malignancy. For lesions <4.6 cm and with a Hounsfield density of <20HU, surgical removal should also be considered in the presence of worrisome imaging features, such as heterogeneity, irregularity, and/or inva- sion. In the absence of all of these findings, an AI can be managed nonoperatively.
REFERENCES
1. Young WF Jr. Clinical practice. The incidentally discovered adrenal mass. N Engl J Med. 2007;356:601-610.
2. Bovio S, Cataldi A, Reimondo G, et al. Prevalence of adrenal incidentaloma in a contemporary computerized tomography series. J Endocrinol Invest. 2006;29:298-302.
3. Terzolo M, Stigliano A, Chiodini I, et al. AME position statement on adrenal incidentaloma. Eur J Endocrinol. 2011;164:851-870.
4. Birsen O, Akyuz M, Dural C, et al. A new risk stratification algorithm for the management of patients with adrenal incidentalomas. Surgery. 2014;156:959- 965.
5. Bilimoria KY, Shen WT, Elaraj D, et al. Adrenocortical carcinoma in the United States: treatment utilization and prognostic factors. Cancer. 2008;113:3130-3136.
6. Assie G, Antoni G, Tissier F, et al. Prognostic parameters of metastatic adrenocortical carcinoma. J Clin Endocrinol Metab. 2007;92:148-154.
7. Kahramangil B, Berber E. Comparison of posterior retroperitoneal and trans- abdominal lateral approaches in robotic adrenalectomy: an analysis of 200 cases. Surg Endosc. 2018;32:1984-1989.
8. Chen Y, Scholten A, Chomsky-Higgins K, et al. Risk factors associated with perioperative complications and prolonged length of stay after laparoscopic adrenalectomy. JAMA Surg. 2018;153:1036-1041.
9. Zeiger MA, Thompson GB, Duh QY, et al. American Association of Clinical Endocrinologists and American Association of Endocrine Surgeons Medical Guidelines for the Management of Adrenal Incidentalomas: executive sum- mary of recommendations. Endocr Pract. 2009;15:450-453.
10. Fassnacht M, Arlt W, Bancos I, et al. Management of adrenal incidentalomas: European Society of Endocrinology Clinical Practice Guideline in collabora- tion with the European Network for the Study of Adrenal Tumors. Eur J Endocrinol. 2016;175:G1-g34.
11. NIH state-of-the-science statement on management of the clinically inappar- ent adrenal mass (“incidentaloma”). NIH Consens State Sci Statements. 2002;19:1-25.
12. Mansmann G, Lau J, Balk E, et al. The clinically inapparent adrenal mass: update in diagnosis and management. Endocr Rev. 2004;25:309-340.
13. Barzon L, Sonino N, Fallo F, et al. Prevalence and natural history of adrenal incidentalomas. Eur J Endocrinol. 2003;149:273-285.
14. Young WF Jr. Management approaches to adrenal incidentalomas. A view from Rochester, Minnesota. Endocrinol Metab Clin North Am. 2000;29:159-185. x.
15. Cawood TJ, Hunt PJ, O’Shea D, et al. Recommended evaluation of adrenal incidentalomas is costly, has high false-positive rates and confers a risk of fatal cancer that is similar to the risk of the adrenal lesion becoming malignant; time for a rethink? Eur J Endocrinol. 2009;161:513-527.
16. Pantalone KM, Gopan T, Remer EM, et al. Change in adrenal mass size as a predictor of a malignant tumor. Endocr Pract. 2010;16:577-587.
17. Chomsky-Higgins K, Seib C, Rochefort H, et al. Less is more: cost-effective- ness analysis of surveillance strategies for small, nonfunctional, radiographi- cally benign adrenal incidentalomas. Surgery. 2018;163:197-204.
18. Song JH, Chaudhry FS, Mayo-Smith WW. The incidental indeterminate adrenal mass on CT (> 10 H) in patients without cancer: is further imaging necessary? Follow-up of 321 consecutive indeterminate adrenal masses. AJR Am J Roentgenol. 2007;189:1119-1123.
19. Mantero F, Terzolo M, Arnaldi G, et al. A survey on adrenal incidentaloma in Italy. Study Group on Adrenal Tumors of the Italian Society of Endocrinology. J Clin Endocrinol Metab. 2000;85:637-644.
20. Corwin MT, Navarro SM, Malik DG, et al. Differences in growth rate on CT of adrenal adenomas and malignant adrenal nodules. AJR Am J Roentgenol. 2019;213:632-636.
@ 2020 Wolters Kluwer Health, Inc. All rights reserved.
21. Grumbach MM, Biller BM, Braunstein GD, et al. Management of the clinically inapparent adrenal mass (“incidentaloma”). Ann Intern Med. 2003;138:424-429.
22. Hamrahian AH, Ioachimescu AG, Remer EM, et al. Clinical utility of non- contrast computed tomography attenuation value (hounsfield units) to differ- entiate adrenal adenomas/hyperplasias from nonadenomas: Cleveland Clinic experience. J Clin Endocrinol Metab. 2005;90:871-877.
23. Szolar DH, Korobkin M, Reittner P, et al. Adrenocortical carcinomas and adrenal pheochromocytomas: mass and enhancement loss evaluation at delayed contrast-enhanced CT. Radiology. 2005;234:479-485.
24. Pena CS, Boland GW, Hahn PF, et al. Characterization of indeterminate (lipid- poor) adrenal masses: use of washout characteristics at contrast-enhanced CT. Radiology. 2000;217:798-802.
25. Mayo-Smith WW, Song JH, Boland GL, et al. Management of incidental adrenal masses: a white paper of the ACR Incidental Findings Committee. J Am Coll Radiol. 2017;14:1038-1044.
26. Haider MA, Ghai S, Jhaveri K, et al. Chemical shift MR imaging of hyper- attenuating (>10 HU) adrenal masses: does it still have a role? Radiology. 2004;231:711-716.
27. Seo JM, Park BK, Park SY, et al. Characterization of lipid-poor adrenal adenoma: chemical-shift MRI and washout CT. AJR Am J Roentgenol. 2014;202:1043-1050.
28. Siren J, Tervahartiala P, Sivula A, et al. Natural course of adrenal incidentalomas: seven-year follow-up study. World J Surg. 2000;24: 579-582.
29. Dunnick NR, Korobkin M, Francis I. Adrenal radiology: distinguishing benign from malignant adrenal masses. AJR Am J Roentgenol. 1996;167: 861-867.
30. Blake MA, Kalra MK, Maher MM, et al. Pheochromocytoma: an imaging chameleon. Radiographics. 2004;24(suppl 1):S87-S99.
31. Rao P, Kenney PJ, Wagner BJ, et al. Imaging and pathologic features of myelolipoma. Radiographics. 1997;17:1373-1385.