High Diagnostic Accuracy of Adrenal Core Biopsy: Results of the German and Austrian Adrenal Network Multicenter Trial in 220 Consecutive Patients
W. SAEGER, MD, M. FASSNACHT, MD, R. CHITA, MD,
G. PRAGER, MD, C. NIES, MD, K. LORENZ, MD, E. BÄRLEHNER, MD,
D. SIMON, MD, B. NIEDERLE, MD, F. BEUSCHLEIN, MD,
B. ALLOLIO, MD, AND M. REINCKE, MD, FOR THE GERMAN AND AUSTRIAN ADRENAL NETWORK
Incidentally detected adrenal tumors are a common finding dur- ing abdominal ultrasonography, computed tomography, and mag- netic resonance imaging. Although most of these lesions are benign adenomas, adrenocortical carcinomas and metastases constitute 5% to 10% of all tumors. Adrenal biopsy may be helpful, but its diagnos- tic value is controversial and disputed, and prospective studies have not yet been performed. Therefore, the diagnostic accuracy of adre- nal core biopsy was evaluated in a prospective multicenter study involving 8 surgical centers in Germany and Austria. A total of 220 biopsies from surgical specimens of the adrenal gland were punc- tured in an ex vivo approach and processed for pathohistologic diagnosis using paraffin sections, routine staining, and immunohisto- chemistry (keratin KL1, vimentin, S100 protein, chromogranin A, synaptophysin, neuron-specific enolase, D11, MiB-1, and p53 pro- tein). The evaluating pathologist was blinded for clinical data from the patients. A total of 89 adrenal adenomas (40.5%), 22 adrenal carcinomas (10.0%), 55 pheochromocytomas (25.0%), 15 metastases (6.8%), 16 adrenal hyperplasias (7.2%), and 23 other tumors (10.5%) were studied. Nine cases were excluded due to incomplete data (n = 2) or insufficient biopsy specimen (n = 7). In the remaining 211 tumors, compared with the final diagnoses of the surgical specimen,
Clinically silent adrenal tumors are frequently found incidentally by abdominal imaging procedures. The prevalence of these so-called adrenal “incidentalo- mas” is 1% for computed tomography (CT) series and up to 7% in autopsy studies, as recently summarized by a National Institutes of Health State of the Science conference.1 Most of these incidental lesions are adre- nal cortical nodules or hormonally inactive adenomas
bioptic diagnoses were absolutely correct in 76.8% of the cases, nearly correct in 13.2% of the cases, and incorrect in 10% of the cases. Pheochromocytomas were correctly diagnosed in 96% of the cases, cortical adenomas were correctly or nearly correctly reported in 91% of the cases, cortical carcinomas were correctly or nearly correctly reported in 76% of the cases, and metastases were correctly or nearly correctly reported in 77% of the cases. Of the 39 malignant lesions, only 4 were misclassified, 2 as benign and 2 as possibly malignant. This resulted in an overall sensitivity for malignancy of 94.6% and specificity of 95.3%. Our findings suggest that adrenal core biopsy is a useful method for identifying and classifying adrenal tumorous lesions if sufficient biopsy specimens can be obtained. However, in clinical practice it remains to be shown whether the benefits of biopsy outweigh the risks of the procedure. HUM PATHOL 34:180-186. Copyright 2003, Elsevier Science (USA). All rights re- served.
Key words: adrenal biopsy, adrenal adenoma, adrenal carcinoma, pheochromocytoma, metastasis.
Abbreviations: ACC, adrenocorticoid carcinoma; CT, computed tomography; PAS, periodic acid Schiff.
From the Institute of Pathology, Marienkrankenhaus, Hamburg, Germany; Department of Internal Medicine, Endocrinology, Univer- sity of Würzburg, Germany; Department of Surgery, University of Wien, Austria; Department of Surgery, University of Marburg, Ger- many; Department of Surgery, University of Halle, Germany; Depart- ment of Surgery, Klinikum Buch, Berlin, Germany; Department of Surgery, University of Düsseldorf, Germany; and Department of In- ternal Medicine, Endocrinology, University of Freiburg, Germany. Accepted for publication November 19, 2002.
Supported by a grant from Dr. Mildred Scheel-Stiftung and the Deutsche Forschungsgemeinschaft (Re 752/11-1 to M.R.).
Address correspondence and reprint requests to Dr. Wolf- gang Saeger, Institute of Pathology of the Marienkrankenhaus, Alfredstraße 9, D-22087 Hamburg, Germany.
Copyright 2003, Elsevier Science (USA). All rights reserved. 0046-8177/03/3402-0012$30.00/0 doi:10.1053/hupa.2003.24
that do not require surgical therapy; however, in some series the prevalence of malignancies, such as metasta- ses or adrenocortical carcinomas (ACCs), is reportedly as high as 10%.2-8
The poor prognosis for adrenal carcinoma makes early diagnosis very important. In this context, a reli- able histopathologic diagnosis from adrenal biopsy, by either fine-needle aspiration or adrenal core biopsy, is desirable but variously discussed because of its ques- tionable accuracy in light of the definite risks.9,10 Whereas sensitivity is high in adrenal metastasis,11 its accuracy is disputed in the context of primary adrenal tumors (adenomas and ACCs).1,2 On the other hand, adrenal biopsy has well-documented side effects, in- cluding pneumothorax, hemorrhage, needle track me- tastasis, and hypertensive crisis due to puncture of pheochromocytoma.12-14 However, large prospective studies comparing bioptic and histopathologic diag- noses obtained from the surgical specimen have not yet been reported, and most studies to date have relied on clinical follow-up for verification of the final diagnosis. Hence a controlled study for the diagnostic accuracy of adrenal biopsy would appear to be needed.
We performed a multicenter study of 211 adrenal
HIGH DIAGNOSTIC ACCURACY OF ADRENAL CORE BIOPSY (Saeger et al)
| All tumors | Adrenal adenoma | Adrenal carcinoma | Pheochro- mocytmoma | Metastasis | Others | |
|---|---|---|---|---|---|---|
| Age (years) | 50.1 (12-81) | 51.2 (18-81) | 51.0 (15-79) | 44.9 (12-75) | 58.0 (45-70) | 51.2 (19-79) |
| Sex | 130 F; 86 M | 63 F; 26 M | 11 F; 11 M | 28 F; 26 M | 6 F; 9 M | 22 F; 14 M |
| Diameter (cm) | 4.8 (2-25) | 3.2 (.4-9) | 10.1 (1.5-25) | 4.5 (1-9.5) | 5.1 (1-9) | 6 (.2-23) |
Abbreviations: F, female; M, male. NOTE. Nine of 220 tumors were excluded from further analysis (for details, see Material and Methods).
biopsies obtained ex vivo from surgically resected adre- nal lesions. Using this approach, we were able to com- pare the diagnosis from the biopsies with that of surgi- cal specimens.
MATERIALS AND METHODS Patients
A total of 220 surgical specimens were obtained during adrenalectomy. The specimens came from the following study centers: Wien, 73 specimens; Marburg, 36; Halle, 34; Würz- burg, 26; Berlin, 19; Freiburg, 16; Düsseldorf, 13; and Neubrandenburg, 3.
The fresh material was punctured 3 times with a biopsy needle normally used in CT-guided biopsy. Three biopsy cylinders, 1 mm in diameter and .8 to 2.0 cm long, were obtained, fixed in formalin, and sent to the study’s reference pathologist (W.S.), who was blinded to all clinical and bio- chemical data from the patients. Histopathologic analysis of the surgical specimens was performed locally in each center. Clinical data for the 220 tumors in 216 patients are given in Table 1. The diameter of the lesions ranged from 1 cm to 25 cm (Fig 1). Two biopsy specimens could not be compared with the surgical specimen and were excluded. In 7 biopsy specimens (1 adrenal adenoma, 1 adrenal carcinoma, 1 my- elolipoma, 1 adrenal hyperplasia, 1 metastasis, 1 lipoma, and 1 hematoma), the material was insufficient to allow evaluation of the lesion; these tumors were also excluded from further analysis, leaving a total of 211 tumors in the study.
The study protocol was approved by the local Ethics Committees of the Universities of Würzburg and Freiburg, and all of the patients consented to participate in the study.
60
50
number of tumors
40
30
20
10
0
< 2.0 cm
2.0 - 3.9 cm
4.0 - 5.9 cm
6.0 - 9.9 cm
> 10.0 cm
Histology
The biopsies were embedded in paraffin and processed in a routine manner using hematoxylin and eosin stain, pe- riodic acid-Schiff (PAS) reaction, and Masson’s trichrome stain. Furthermore, all specimens were stained using a panel of antibodies, including keratin KL 1 (monoclonal [mc], dilution 1:150; Immunotech, Marseille, France), vimentin (mc, 1:50; Dako, Glostrup, Denmark), S-100 protein (poly- clonal [pc], 1:500; Dako), neuron-specific enolase (mc, 1:100; Dako), synapthophysin (pc, 1:70; Dako), chromogranin A (mc, 1:150; Immunotech), D 11 (mc, 1:2; kindly provided by Dr. Beisiegel, University of Hamburg), Ki 67 (Mib-1) (mc, 1:60; Novocastra, Newcastle upon Tyne, UK) and p53 protein (mc, 1:50; Immunotech). For some biopsy specimens, addi- tional antibodies were necessary if soft tissue tumors or lym- phomas had to be identified or excluded.
In 164 of 211 tumors, the agreement between the biopsy specimen and the surgical specimen was identical or suffi- cient. In 47 cases, the biopsy and surgical specimens did not match. The slides or paraffin blocks of these tumors were sent to the reference pathologist, who made the final diagnoses.
Diagnostic Criteria
Cortical tumors were microscopically identified by the trabecular, alveolar or diffuse composition of clear or com- pact cells. Diagnostic algorithms15-18 (Table 2) were used for differentiating benign and malignant cortical tumors. Fur- thermore, a Ki-67 staining index of >5% in adrenocortical tumors was suggestive for an ACC. The structural diagnosis of pheochromocytoma was based on the typical Zellballen pat- tern. Metastases were diagnosed by their special adenoid or squamous pattern. In many cases, immunostainings were nec- essary for differentiating adrenocortical tumors from metas- tases, (especially in cases of renal or liver carcinoma) or from adrenomedullary tumors. The criteria used for these differ- ential diagnoses are listed in Table 3.
The following criteria were used to classify correspon- dence between the biopsy and surgical specimens:
1. Full agreement: Histopathologic diagnoses identical
2. Sufficient agreement: Origin of tissue and dignity of the biopsy specimen correctly identified but with mi- nor deviance (eg, adrenal adenoma versus adrenal nodulus or adenoma with myolipomatous foci versus myolipoma)
3. Incorrect diagnosis: Origin or dignity not correctly identified. ACCs fell into this category if they were misclassified as metastases.
Statistics
Sensitivity, specificity, and positive and negative predic- tive values were calculated according to standard algorithms.
| Criterion | Degree | Weiss17 | Hough et al15 | van Slooten et al16 |
|---|---|---|---|---|
| Nuclear atypia | Moderate to strong | 1 | .39 | 2.1 |
| Nuclear hyperchromasia | Moderate to strong | 2.6 | ||
| Nucleoli | Prominent | 4.1 | ||
| Mitoses | >5/50 HPF | 1 | ||
| >10/100 HPF | .69 | |||
| >2/10 HPF | 9.0 | |||
| Atypical mitoses | Present | 1 | ||
| Clear cells | <25% volume percentage | 1 | ||
| Architecture | Diffuse growth pattern | 1 | .92 | 1.6 |
| Veins | Tumor invasion | 1 | ||
| Sinus | Tumor invasion | 1 | .92+ | 3.3§ |
| Tumor capsule | Tumor invasion | 1 | .37 | |
| Necroses | Present | 1 | .69 | |
| Regressive changes | Strong | 5.7 | ||
| Fibrous bands | Present | 1.00 | ||
| Sum |
*Weiss18 modified according to Weiss et al17: 1-3, benign; ≥ 4, malignant; Hough et al15: . 17 ± .26, benign; 1 ± .58, indeterminate; 2.91 ±
.9, malignant; and van Slooten et al16: < 8, benign; ≥ 8, malignant. +Vessel invasion.
§Invasion of capsule and/or vessels.
“Necroses, bleeding, fibroses, and calcifications.
Abbreviation: HPF, high-power field.
RESULTS
Overall
Comparison of the histopathologic diagnoses of the 211 bioptic specimens with those of the surgical specimens showed full agreement in 76.8% and suffi- cient correspondence in 13.2%. The bioptic diagnosis proved to be incorrect in 21 cases (10%). The detailed results are given in Table 4. Only 2 of 39 proven ma- lignant lesions were misclassified by biopsy as a benign process, and 2 more were classified as possibly malig- nant. This translates to an overall sensitivity for malig- nancy of 94.6% and a specificity of 95.3%. The positive predictive value is 98.8%, and the negative predictive value is 81.4% (Table 5).
In 3 cases, the bioptic material was not represen- tative for the lesions, because the tumorous tissue with the main alteration was not included in the biopsy specimen. These cases included a benign nodule di- rectly beside a metastasis, the adjacent normal adrenal of a myelolipoma, and a pheochromocytoma attached to a ganglioneurinoma.
Adrenal Adenoma
Correct diagnosis was achieved in 65 of 85 cases (75.6%). The 13 cases (15.1%) with sufficient agree- ment were represented by 11 nodules, 1 myelolipoma, and 1 tumor classified as having a high probability of an adenoma. The 8 incorrect diagnoses (9.3%) were mostly based on necrotic tissue that led to the tentative diagnosis of malignancy (Fig 2) (n = 6). In 1 lipoma- tous adenoma and 1 pleomorphic adenoma, the dignity was classified as uncertain in the bioptic specimen. Overall sensitivity was 77.6%, and specificity was 93.3% (Table 5).
Adrenal Carcinoma
The diagnosis of ACC was correct in 16 out of 21 cases (76.2%). Five tumors (23.8%) were bioptically misclassified as metastases because D11 immunoreactiv- ity was negative and typical structures of cortical carci- nomas were missing (Fig 3). None of the adrenocortical carcinomas was classified as a benign lesion.
| Antigen | Adrenocortical tumor | Adrenomedullary tumor | Metastasis from renal carcinoma | Metastasis from liver carcinoma |
|---|---|---|---|---|
| Cytokeratins | 10%-36% | 0% | 95%-100% | 95%-100% |
| Epithelial membrane antigen | 0% | 0% | 100% | 0% |
| S-100 protein | 0% | Sustentacular cells | 0% | 0% |
| Chromogranin A | 0% | 100% | 0% | |
| Neuron-specific enolase | 30% | 100% | ||
| Synaptophysin | 40%-80% | 100% | ||
| D 11 | 90%-100% | 0% | 30% |
NOTE. Prevalence of immunoreactivities in adrenocortical and adrenomedullary tumors and metastases from renal or liver carcinomas as reported in the literature.28, 29
HIGH DIAGNOSTIC ACCURACY OF ADRENAL CORE BIOPSY (Saeger et al)
| Final diagnosis | Exact bioptic diagnosis | Sufficient bioptic diagnosis | Incorrect bioptic diagnosis | Biopsy not representative |
|---|---|---|---|---|
| Adrenocortical adenoma | 65 (76%) | 13 (15%) | 8 (9%) | |
| Adrenocortical carcinoma | 16 (76%) | 5 (24%) | ||
| Pheochromocytoma | 52 (96%) | 2 (4%) | 1* | |
| Adrenocortical hyperplasia | 6 (40%) | 8 (53%) | 1 (7%) | |
| Metastasis | 10 (77%) | 3 (23%) | 1+ | |
| Myelolipoma | 2 | 3 | 1§ | |
| Ganglioneuroma | 2 | 1 | ||
| Normal adrenal | 1 | 1 | ||
| Non-Hodgkin's lymphoma | 2 | |||
| Leiomyosarcoma | 2 | |||
| Adrenocortical tumor of uncertain dignity | 1 | |||
| Neurinoma | 1 | |||
| Liposarcoma | 1 | |||
| Hematoma | 1 | |||
| Schwannoma | 1 | |||
| Overall | 159 (75.4%) | 28 (13.2%) | 21 (10%) | 3 (1.4%) |
*Biopsy of a ganglioneurinoma beside a pheochromocytoma.
+Biopsy of a nodus beside a metastasis.
§Biopsy of the normal adjacent adrenal cortex.
Nodular Adrenal Hyperplasia
Six of 15 nodular adrenal hyperplasias, defined as a circumscribed increase in cortical cell number with slightly irregular architecture, were diagnosed exactly by biopsy (40%). In 8 lesions, sufficient agreement could be achieved, represented by adrenal adenoma (n = 6) or normal adrenal gland (n = 2). In 1 case, hyperplasia was misclassified as adrenal carcinoma due to extensive fibroses, lacking lipid-rich tumor cells and some necroses, which could not be differentiated from artifacts in the small specimens.
Metastasis
An unequivocal diagnosis of metastasis was achieved in 10 of 14 cases (76.9%). One case was in- correctly classified as a benign, partly cystic tumor of unclear discrimination but represented a metastasis or a direct invasion of a pancreatic island carcinoma. In 2 biopsy specimens, no definitive diagnosis could be achieved, because in 1 case the expression of D11 fa- vored the diagnosis of adrenal carcinoma but the struc- ture suggested a metastasis, and in the other case the inflammatory-like structure and the negative immuno- staining of 2 different pan-keratins appeared incompat-
ible with a metastasis of a carcinoma. In 1 case, a benign nodule directly beside a metastasis was biopsied.
Pheochromocytoma
The bioptic diagnosis of pheochromocytoma was generally easy because of the typical Zellballen pattern and immunostaining for chromogranin A (correct in 52 of 54 cases; 96.3%). Diagnosis was incorrect in 2 of 54 biopsies because of some small cell areas that led to the diagnosis of “neuroblastoma” and metastasis of small cell tumor. In 1 case a ganglioneuroma instead of the main lesion, a pheochromocytoma, was biopsied and correctly diagnosed.
Other Tumors
One tumor was classified as cortical neoplasm of unclear dignity by the biopsy and surgical specimens. Myelolipomas were correctly diagnosed by biopsy in 2 of 6 lesions. Three leasions were classified as adenomas with myelolipomatous metaplasia (n = 2) or nodules (n = 1). In 1 case, adjacent normal adrenal gland was biopsied and correctly diagnosed. Two non-Hodgkin’s lymphomas could not be identified by biopsy; a chronic
| Sensitivity | Specificity | Negative predictive value | Positive predictive value | |
|---|---|---|---|---|
| Overal dignity (benign vs. malignant) | 94.6% | 95.3% | 98.8% | 81.4% |
| Overal dignity (benign vs. malignant without 55 pheochromocytomas) | 94.6% | 93.9% | 98.2% | 83.3% |
| Adrenocortical adenoma | 77.6% | 93.3% | 85.4% | 89.2% |
| Adrenocortical carcinoma | 76.2% | 96.2% | 97.3% | 69.6% |
| Pheochromocytoma | 94.5% | 100% | 98.0% | 100% |
| Metastasis | 90.9% | 96.4% | 99.5% | 58.8% |
S
inflammation was diagnosed in 1 specimen, and a me- tastasis of a small cell carcinoma in the other.
The liposarcoma and the 2 leiomyosarcomas were correctly identified by biopsy. Two of 3 ganglioneuro- mas were correctly diagnosed; 1 was classified as a neurofibroma. One neurinoma was correctly demon- strated in the biopsy specimen. One schwannoma was classified as a neurofibroma, and 1 hematoma was found to be a necrosis.
Table 5 summarizes sensitivity and specificity val- ues for the different entities of the study.
DISCUSSION
Ultrasound- or CT-guided fine-needle aspiration cytology is a widely used and well-accepted method for the differential diagnosis of such tumors as thyroid nodules and solid hepatic and pancreatic tumors. In contrast, biopsy of adrenal tumors has been controver- sial and has not yet gained general acceptance by en- docrinologists, endocrine surgeons, and pathologists. The main reasons for this include documented mor- bidity and mortality due to hemorrhage, pneumotho- rax, abscess, needle track metastasis, and puncture of an unsuspected pheochromocytoma.12-14,19,20 The over- all complication rate ranges between 5% and 12%.8 Moreover, despite positive reports,9,13 there is serious doubt that adrenal biopsy can be useful in the differ- ential diagnosis of primary adrenocortical tumors (ie, adenomas and ACCs).1,2 The only widely accepted in- dication for adrenal biopsy is suspected metastasis from a known primary tumor in a patient in whom the result would change the therapeutic regimen.10 In this set- ting, adrenal biopsy, through either fine-needle aspira- tion cytology or adrenal core biopsy, yields high sensi- tivity and specificity (>80%) and overall accuracy of 90%. The technical success rate of the procedure with sufficient material is reportedly 70% to 80%. However, in most series, the final diagnosis depends on clinical
follow-up by CT or ultrasound, because surgery is per- formed in only a limited number of patients.
The aim of the present study was to evaluate pro- spectively the diagnostic accuracy of adrenal core bi- opsy in a large series of adrenal lesions. To circumvent uncertainties of the final diagnosis, we used an ex vivo approach that ideally results in full proof of histopa- thology through comparison with the surgical speci- men. To the best of our knowledge, this study is the first and only report that takes this approach. It shows that adrenal core biopsy has sufficient sensitivity and speci- ficity for classification of the lesions if rigid criteria for ACC in combination with immunohistochemistry are applied. In particular, the dignity of the lesion can be accurately classified by the methods described herein. However, our approach represents an ideal situation with respect to biospy quality and the amount of tissue obtained for study. It is certainly much more difficult to obtain a sufficient needle core biopsy specimen in deep-seated lesions in vivo. Because of this principal limitation, further clinical studies are needed to test the accuracy of adrenal biopsy in a clinical setting.
Compared to the diagnosis of pheochromocytoma in bioptic material, the diagnosis of cortical adenoma was more difficult, especially with respect to differenti- ation between hyperplasia and nodules. In this differ- ential diagnosis, the demonstration of irregular archi- tecture, slight cellular pleomorphism, enlarged nuclei, and increased nuclear chromatin is important for the diagnosis of adenoma.21-23 If these criteria are not present in biopsy samples, then differentiation is pre- sumably not possible. However, in a clinical context, this differentiation is not relevant.
Necroses are important features for the diagnosis of ACC15,16 but nonetheless led to an incorrect diagno- sis in 6 adenomas. This indicates that additional crite- ria, such as increased cell proliferation as assessed by the Ki-67 index and changes in cell structure, must be present in biopsy samples to enable identification of ACC. Malignancy of the lesions was correctly identified in all ACCs, but diagnosis of a metastasis was made in 6
of 21 tumors. Overall, our study demonstrates that adrenal core biopsy identifies ACC better than was previously assumed.
Interestingly, in a recent study, microdissection of fine-needle aspiration samples followed by polymerase chain reaction-based analysis for loss of heterozygosity of tumor-suppressor genes (eg, p53 and the von Hip- pel-Lindau gene) was shown to give excellent results in differentiating benign from malignant adrenocortical tumors.24 In the future, combination of those molecu- lar markers with conventional histopathology and im- munohistochemistry may ultimately further improve di- agnostic accuracy.
In the differential diagnosis of ACC and metasta- ses, the structure is most important. Thus, metastases of squamous carcinomas, adenocarcinomas, and small cell carcinomas can be easily differentiated from adrenal carcinomas, but metastases of large cell solid carcino- mas, clear cell carcinomas, and renal cell cancers are problematic. In such cases, immunostaining for keratin (mostly negative in adrenal carcinomas), epithelial membrane antigen (always negative in adrenal carcino- mas), or D11 (mostly positive in adrenal carcinomas; far less frequently positive in metastases) 25-27 proved helpful, but not sufficient in all cases.
Metastases were correctly diagnosed as “malig- nancy” in only 10 of 14 tissue specimens (76%). This percentage is lower than previously reported in the literature, but classical adrenal metastases (eg, oat cell carcinoma of the lung, breast cancer, renal cell carci- noma) were underrepresented in our surgical series. Normal and hyperplastic adrenals were mostly easily identified in biopsies. Only 1 case was problematic-an ACC was assumed by biopsy because of extensive fibro- ses and very few lipid-rich cortical cells. The surgical specimens revealed a hyperplastic adrenal.
In the present study, the reference pathologist was “blinded” with respect to patient age, sex, tumor size/ weight, and endocrine activity of the tumor. This made histologic classification in this study more difficult. For example, tumor size/weight is an important determi- nant for malignancy, with adrenocortical carcinoma having an average weight at diagnosis of 800 g and a diameter often exceeding 10 cm, whereas most benign adenomas have a diameter <4 cm. Also, the endocrine activity of a given tumor may be significant to the probability of malignancy. For example, a cortisol-pro- ducing adrenal tumor in childhood is very suggestive of an ACC, and androgen-producing tumors are generally suspicious of malignancy.6 This study proves that clas- sification of adrenal masses is possible even if clinical information is missing.
In summary, we conclude that adrenal biopsy spec- imens obtained ex vivo have a sensitivity of 95% and a specificity of 95% for malignancy. These figures remain nearly identical if pheochromocytoma-a tumor that is biopsied only accidentally-is excluded (Table 5). The bioptic identification of pheochromocytoma and corti- cal adenoma is unproblematic if differentiation be- tween adenomas and cortical nodules need not be made. ACC also can be diagnosed correctly by biopsy if
the diagnostic algorithm15-17,27 is consequently used. Immunostaining is often mandatory for the differenti- ation of ACC and metastases.27 The present study should stimulate further prospective studies in vivo with carefully selected patients to evaluate whether similar results can be obtained in a clinical setting. Those studies should also include prospective documentation of morbidity and mortality to evaluate the benefits and risks of the procedure.
Acknowledgment. Additional authors of this article are (in alphabetical order) S. Anders, University of Berlin; H. Dralle, University of Halle; M. Ernst, Neubrandenburg; P. E. Goretzki, University of Düssel- dorf; M. Rothmund, University of Marburg; W. Tim- mermann, University of Würzburg; and U. Wetterauer and U. Schöffel, University of Freiburg.
REFERENCES
1. National Institutes of Health: State-of-the-science statement: Management of the clinically inapparent adrenal mass (incidenta- loma), February 4-6, 2002; available at http://consensus.nih.gov/ta/ 021/021_statement.htm
2. Reincke M, Allolio B: Das nebenniereninzidentalom. Dtsch Ärztebl 92:A764-A770, 1995 (in German)
3. Osella G, Terzolo M, Boretta G, et al: Endocrine evaluation of incidentally discovered adrenal masses (incidentalomas). J Clin En- docrinol Metab 79:1532-1539, 1994
4. Kasperlik-Zaluska AA, Rosnolowska E, Slowinska-Srzednicka, et al: Incidentally discovered adrenal mass (incidentaloma): Investi- gation and management of 208 patients. Clin Endocrinol (Oxf) 46:29-37, 1997
5. Bastounis EA, Karayiannakis AJ, Anapliotou ML, et al: Inci- dentalomas of the adrenal gland: Diagnostic and therapeutic impli- cations. Am Surg 63:356-360, 1997
6. Latronico AC, Chrousos GP: Extensive personal experience: Adrenocortical tumors. J Clin Endocrinol Metab 82:1317-1324, 1997
7. 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 85:637-644, 2000
8. Kloos RT, Gross MD, Francis IR, et al: Incidentally discovered adrenal masses. Endocr Rev 16:460-484, 1995
9. Wu HH, Cramer HM, Kho J, et al: Fine-needle aspiration cytology of benign adrenal cortical nodules: A comparison of cyto- logic findings with those of primary and metastatic adrenal malignan- cies. Acta Cytol 42:1352-1358, 1998
10. Young WF Jr: Management approaches to adrenal inciden- talomas: A view from Rochester, Minnesota. Endocrinol Metab Clin North Am 29:159-185, 2000
11. Welch TJ, Sheedy PF 2nd, Stephens DH: Percutaneous ad- renal biopsy: Review of a 10-year experience. Radiology 193:341-344, 1994
12. Yankaskas BC, Staab EV, Craven MB, et al: Delayed compli- cations from fine-needle biopsies of solid masses of the abdomen. Invest Radiol 21:325-328, 1986
13. Lumachi F, Borsato S, Brandes AA, et al: Fine-needle aspi- ration cytology of adrenal masses in noncancer patients: Clinicora- diologic and histologic correlations in functioning and nonfunction- ing tumors. Cancer 93:323-329, 2001
14. McCorkell SJ, Niles NL: Fine-needle aspiration of catechol- amine-producing adrenal masses: A possibly fatal mistake. Am J Roentgenol 145:113-114, 1985
15. Hough AJ, Hollifield JW, Page DL, et al: Prognostic factors in adrenal cortical tumours. Am J Clin Pathol 72:390-399, 1979
16. van Slooten H, Schaberg A, Smeenk D, et al: Morphologic characteristics of benign and malignant adrenocortical tumors. Can- cer 55:766-773, 1985
17. Weiss LM, Medeiros LJ, Vickery AL: Pathologic features of
prognostic significance in adrenal cortical carcinoma. Am J Surg Pathol 13:202-206, 1989
18. Weiss LM: Comparative histologic study of 43 metastasizing and non-metastasizing adrenocortical tumors. Am J Surg Pathol 8:163-169, 1984
19. Casola G, Nicolet V, van Sonnenberg E, et al: Unsuspected pheochromocytoma: Risk of blood-pressure alterations during percu- taneous adrenal biopsy. Radiology 159:733-735, 1986
20. Habscheid W, Pfeiffer M, Demmrich J, et al: Puncture track metastasis after ultrasound-guided fine-needle puncture bi- opsy. A rare complication? Dtsch Med Wochenschr 115:212-215, 1990
21. McNicol AM, Laidler P: The adrenal gland and extra-adrenal paraganglia, in Lewis PD (ed): The Endocrine System. New York, NY, Churchill Livingstone, 1996, pp 131-186
22. Neville AM, O’Hare MJ: The Human Adrenal Cortex. New York, NY, Springer, 1982
23. Saeger W, Reinhard K, Reinhard C: Hyperplastic and tumor- ous lesions of the adrenals in an unselected autopsy series. Endocr Pathol 9:235-239, 1998
24. Abati A, Sanjuan X, Wilder A, et al: Utilization of microdis- section and the polymerase chain reaction for the diagnosis of adre- nal cortical carcinoma in fine-needle aspiration cytology. Cancer 87:231-237, 1999
25. Schröder S, Komminoth P, Padberg BC, et al: Morpholo- gische typisierung, dignitäts und prognosebeurteilung sowie ätiolo- gische einordnung adrenomedullärer und adrenokortikaler neo- plasien. Pathologe 16:307-314, 1995 (in German)
26. Schröder S, Niendorf A, Achilles E, et al: Immunocytochem- ical differential diagnosis of adrenocortical neoplasms using the monoclonal antibody D11. Virchows Arch A Pathol Anat Histopathol 417:89-96, 1990
27. Saeger W: Histopathological classification of adrenal tumors. Eur J Clin Invest 30(Suppl 3):58-62, 2000