Molecular Pathology of Adrenal Cortical Tumors: Separating Adenomas from Carcinomas

Thomas J. Giordano, MD, PHD

Abstract

Adrenal cortical carcinoma is a rare but interesting endocrine tumor. Its diagnosis is usu- ally straightforward using morphologic assessment and supplemental immunohistochem- istry. However, diagnostically challenging cases exist and pathologic evaluation would benefit from the availability of adjunctive molecular testing. Here, the relevant molecu- lar pathology of adrenal cortical tumors is reviewed with special reference to those meth- ods (e.g., DNA microarrays) that hold promise for improved diagnosis and prognosis, and prediction of therapeutic response.

Key Words: Adrenal cancer; genotyping; DNA microarray; molecular profiling.

Introduction

Adrenal cortical tumors (ACTs) are rela- tively rare with an incidence of one or two cases per million per year. Nonetheless, these tumors deserve attention for several important reasons. First, the incidence of adrenal tumors incidentally discovered during medical investigation (the so-called “adrenal incidentaloma”) has risen over the last decade owing to improved and more frequent imaging studies [1]. Second, these tumors are clinically and pathologically interesting due to their associated hor- monal [2] and genetic syndromes [3]. Finally, the available therapeutic choices for adrenal cortical carcinoma are quite lim- ited [4] and, thus, there is a significant clinical need for new, more effective, and less toxic therapies. Many of these aspects, as well as others, were discussed at an international meeting held in Ann Arbor in September 2003 [5].

Histopathology combined with immuno- histochemistry (IHC) of ACTs is often adequate to provide highly informed diag-

noses. However, there are diagnostically challenging cases in which adjunctive molecular methods would be useful. Furthermore, as the fields of surgical pathology and oncology becomes more molecularly oriented in general, it is both interesting and relevant to ask how reliably and accurately do molecular approaches separate benign and malignant ACTs. Here, the molecular pathology relevant to the distinction of benign and malignant ACTs is presented.

Department of Pathology, University of Michigan Medical School, Ann Arbor, Michigan.

Address correspondence to Dr. Thomas J. Giordano, Associate Professor, Department of Pathology, UH 2G332/0054, Ann Arbor, MI 48109-0054. E-mail: Giordano@umich.edu

Endocrine Pathology, vol. 17, no. 4, 355-364, Winter 2006 @ Copyright 2006 by Humana Press Inc. All rights of any nature whatsoever reserved. 1046-3976/1559-0097 (Online)/ 06/17:355-364/$30.00

Immunohistochemical Approaches

While molecular approaches are the focus of this review, it is also relevant to discuss the use of immunohistochemistry (IHC) as a diagnostic aid in the evaluation of ACTs, especially as many of the molecu- lar markers discovered by molecular pro- filing are likely to be converted in the future into IHC assays. Many of the studies per- formed to date have focused on rates of tumor cell proliferation, an approach

Table 1. Results of CGH Studies of Adrenal Cortical Carcinoma
Locus of gain or lossSidhu [14] (n = 13)Kjellman [13] (n = 8)Zhao [15] (n = 12)Dohna [12] (n = 14)
1p loss62%13%67%14%
2q loss31%38%42%0%
4p gain31%38%25%21%
4q gain31%50%0%7%
5p gain46%50%25%57%
5q gain38%50%33%50%
11q loss31%50%42%0%
12p gain38%13%17%43%
12q gain38%13%50%86%
17p loss54%50%8%0%
19 gain31%38%0%43%
19 loss23%0%0%0%
22 loss38%38%0%7%

supported by the subsequent molecular profiling studies. These studies revealed comparatively high expression of many proliferation-related genes in adrenal cor- tical carcinomas (ACCs) (see below). The mib1/Ki-67 labeling index was shown to possess diagnostic value in most studies [6- 8], as well as the corresponding topo- isomerase 2 alpha labeling index [9]. Both of these markers were subsequently shown to correlate at the protein and mRNA levels, providing further support for their use [10].

In addition to markers of proliferation, some studies have investigated the use of IGF2 protein as an IHC marker of ACC, given its important role in pathogenesis. Expression of IGF2 protein was associated with ACC in a statistically significant man- ner [11], prompting the authors to sug- gest that IGF2 IHC is a useful tool that could be added to the multifactorial evalu- ation of ACTs.

Genotyping Methods

As with other types of human tumors, ACTs acquire changes in their DNA (mutations) that manifest in the usual forms, ranging from gross chromosomal abnormalities, such as rearrangements, to

the most minimal type of change, single nucleotide substitutions (point mutation). Many types of mutations have been exam- ined in an effort to better understand the pathogenesis of ACTs, but it was also hoped that they could be exploited to develop robust and objective ways to dif- ferentiate benign [adrenal cortical adenoma (ACA)] and malignant tumors (ACC).

Comparative Genomic Hybridization

The experimental methods used to examine mutations in ACTs include “com- parative genomic hybridization” (CGH), which identifies large regions of chromo- somes that have either undergone deletion or amplification. By examining a large cohort of ACTs with CGH, it is possible to identify regions that have consistently been altered. Recurrent chromosomal alterations are likely to be related to patho- genesis and not simply a consequence of genomic instability. Thus, CGH poten- tially identifies regions that contain either oncogenes (regions of copy number gain) or tumor suppressor genes (regions of copy number loss). To date, there have been four CHG studies performed on adult ACTs [12-15], and the results are summarized in Table 1.

While CGH is a powerful chromosomal alteration discovery tool with the resulting CGH data in ACTs supporting a general model in which ACA progresses to ACC in a stepwise fashion, CGH studies have not yet yielded a reliable and accurate geno- type that can be clinically exploited to assist in the diagnosis of these tumors. Moreover, it is highly unlikely, especially given the promising prospects of assessing gene expression by microarrays (see below), that CGH will be clinically implemented for ACTs.

Single Marker Genotyping

One of the most successful approaches for the identification of cancer-related genes has been to isolate the genes respon- sible for familial cancer syndromes. There are two such syndromes in which ACC is a common manifestation; Beckwith- Wiedemann syndrome (BWS) [Online Mendelian Inheritance in Man (OMIM) entry 130650] and Li-Fraumeni syndrome (OMIM entry 151623). BWS is an over- growth disorder associated several tumor types. The mutations causing BWS have been tightly linked to the chromosomal region 11p15.5, which is a complex and imprinted region that contains several genes including H19, KIP2, and IGF2. Based on numerous studies, the role of IGF2 in sporadic and familial ACC has been well established [10,11,16-26]. IGF2 can undergo a variety of rearrangements, the most common being paternal isodisomy (loss of maternal allele and duplication of paternal allele), that result in increased IGF2 expression. These rearrangements are essentially restricted to ACCs [21], an observation that lends support to the notion of using these changes for clinical diagnosis, although I am not aware of a CLIA-approved laboratory that performs these tests for ACTs.

The Li-Fraumeni syndrome has been linked to mutations of the TP53 gene [27]. Many studies have examined TP53 muta- tions in ACTs with highly variable results [27-41]. Generally TP53 mutations were found in 20-67% of ACCs and were rare in ACAs. While the highest mutation fre- quency reported in ACCs was 67%, this is not high enough to be used clinically to accurately separate benign and malignant tumors. Loss of heterozygosity (LOH) studies hold the greatest diagnostic poten- tial [19]. However, the same is true for TP53 LOH genotyping in that there are no labs offering this testing for ACTs.

Other mutations have been associated with ACTs (reviewed in ref. 42) such as mutations of the MEN1 [43-45] and PRKARIA [46-49] genes, but they have not been demonstrated to have utility in separating ACAs from ACCs. Recent mutational work on the beta-catenin gene (CTNNB1) revealed similar mutations frequencies in ACAs and ACCs [50]. Thus, while interesting and important to patho- biology, there is no diagnostic utility in examining ACTs for CTNNB1 mutations.

Molecular Profiling Studies

One of the difficulties and limitations of using genotyping in a clinical setting is that several different types of mutations can inactivate (or activate) the same gene. Point mutations can be distributed across large genes (i.e., point mutations of MEN1) making their identification technically dif- ficult. Furthermore, it is also possible to mutate a single gene via distinct mutational mechanisms, thereby requiring different types of assays for the different types of mutations. For example, the most common type of BRAF mutation in papillary thy- roid carcinoma is a point mutation, but much less common translocations and

Fig. 1. Principal components analysis (PCA) and hierarchical clustering of gene expression in adrenal corti- cal tumors and normal adrenal cortex. Reproduced with permission from ref. 10.

A

70

First two principal components using 2913 probe-sets (mean>200, CV>0.5)

B

Normals

C7

C14

C19

Second Principal Component

☐ Adenoma

C14

50

Carcinoma

C11

C18

C33

C15

30

C8

C4

C17

10

C13

A22

A21

H29

-10

A28

C19

C13

A30

N6

N10

-30

N9

-70

-50

First Principal Component

-30

-10

10

30

C

6

First two principal components using 91 probe-sets (p <. 01, FC>3)

D

Normal

C13

C19

Second Principal Component

4

C13

☐ Adenoma

C14

Carcinoma

C17

2

C4

C18

C15

0

C14

C8

C33

C11

-2

C7

A22

A21

-4

A28

H29

A30

-6

C19

N10

N9

N6

-8

-10

-8

-6

-4

First Principal Component

-2

0

2

4

6

8

10

amplifications have also been reported [51- 53]. Thus, a method based on DNA sequencing for point mutations will fail to identify the other mutations. For this rea- son, as well as others, there is much excite- ment over high-throughput methods to examine gene expression in tissues [54-58]. The development of commercially avail- able DNA microarrays has permitted their use in a variety of clinicopathologic stud-

ies and the first such studies of adrenal cor- tical tumors have been published [10,59].

In a study from our laboratory [10], we used oligonucleotide arrays and a small cohort of normal adrenal cortex, ACAs, and ACCs to identify gene expression pro- files that robustly separated benign and ma- lignant tumors, including one low-grade tumor. Using a large group of variably expressed genes selected without reference

Steroidogenesis

IGF-2

319

320

317

319

81% = carcinomas (Weiss≥4)

316

90% = adenomas (Weiss≤3)

212

311

205

103

202

320

103

308

206

307

111

325

106

306

118

104

219

203

107

Molecular Diagnosis of Adrenal Cancer

204

217

315

117

324

110

303

113

304

208

314

109

302

105

312

218

313

210

305

207

322

307

321

204

201

104

318

101

301

115

213

114

102

203

202

323

93% = adenomas (Weiss≤3)

101

112

117

75% = carcinomas (Weiss≥4)

301

215

102

216

324 312

210

115

116

208

213

206

322

212

304

109

305

207

302

205

317

105

316

310

303

112

308

111

325

219

318

106

311

323

321

116

306

218

215

217

315

118

216

110

201

113

314

107

310

114

313

CYP21A2

INHA

CYP17

GPC3

S100B

TGFBR3

HSD 3B1

NME1

STAR

CYP11A

CYP11B1

PPM1A

RB1

CREM

KCN10T1

NST 1R

TGFB2

FGER1

TGEBR1

FGFR4

IGF 2

GAPD

Fig. 2. Hierarchical clustering of gene expression in adrenal cortical tumors. The top panel shows division of tumors into two clusters based on gene expression of genes related to /GF2 expression. The bottom panel shows a similar division of tumor based on expression of genes

related to steroidogenesis. Reproduced with permission from ref. 59.

to pathologic diagnosis, separation of ACCs from the other cortical tissues was observed. The one low-grade ACC (desig- nated C13) was intermediate in its classification. When a reduced gene list of 91 differentially expressed genes was used, C13 clearly segregated with the other ACCs. This was observed when both prin-

cipal component analysis (PCA) and hier- archical clustering was performed (Fig. 1). Consistent with the IGF2 genotyping work described above, IGF2 expression was greatly increased in the ACCs compared to the other tissues, along with many other genes related to increased proliferation. While limited to a small cohort of tumors,

this study revealed distinct gene expression profiles of ACAs and ACCs and derived a short yet robust list of differentially expressed genes.

A second microarray paper used a larger set of tumors but a much smaller set of 230 genes [59]. The real power of this study comes not from the depth of its array but from their ability to correlate gene expres- sion and patient outcome. Combining these variables, a 22-gene set was developed and shown to possess predictive diagnos- tic power (Fig. 2). In addition, a 14-gene set was shown to correlate with outcome. These gene expression signatures are excit- ing and await further validation.

Collectively, these DNA microarray studies demonstrated clear differences in gene expression between benign and malignant ACTs and strikingly illustrate the power of molecular profiling approaches for tumor classification and gene discovery.

Future Directions

Many vitally important questions remain unanswered. For instance, just as other endocrine tumors have biologically and molecularly distinct subgroups (e.g., papillary thyroid carcinoma), can ACC be classified into clinically meaning- ful groups? Stated more plainly: Is adrenal cancer one disease or several morphologi- cally related diseases? Furthermore, can molecular profiling studies identify addi- tional therapeutic targets? The existing studies suggest that increased IGF2 gene expression is the dominant cause of ACC and thus a valid therapeutic target. Is IGF2 the whole story or are other factors, yet undiscovered, of equal importance? Despite the rare nature of adrenal cancer, many groups are actively trying to address these questions.

It is quite difficult to predict the spe- cific path this field will take in the coming years, especially given the rapid evolution of high-throughput genomic and pro- teomic technologies to comprehensively examine various aspects of tumor cell biology. However, it is unmistakably clear that the disciplines of oncology and pathol- ogy are poised to undergo a transforma- tion in which more molecularly targeted therapies become available and more intelligent therapeutic choices are made, largely driven by molecular profiling-based assessments of individual patient’s tumors. It will be exciting to witness and partici- pate in this transformation as it is applied to the diagnosis, prognosis, and treatment of adrenal cortical carcinoma.

Acknowledgments

This work was originally presented at the Endocrine Pathology Society Compan- ion Meeting at the 2006 Annual Meeting of the United States Academy of Patholo- gists (USCAP). The author thanks the organizers for the opportunity to present and Donita Sanders for editorial assistance.

References

1. Bovio S, Cataldi A, Reimondo G, et al. Preva- lence of adrenal incidentaloma in a contem- porary computerized tomography series. J Endocrinol Invest 29:298-302, 2006.

2. Vierhapper H. Adrenocortical tumors: clini- cal symptoms and biochemical diagnosis. Eur J Radiol 41:88-94, 2002.

3. Stratakis CA. Genetics of adrenocortical tumors: gatekeepers, landscapers and conduc- tors in symphony. Trends Endocrinol Metab 14:404-410, 2003.

4. Kirschner LS. Emerging treatment strategies for adrenocortical carcinoma: a new hope. J Clin Endocrinol Metab 91:14-21, 2006.

5. Schteingart DE, Doherty GM, Gauger PG, et al. Management of patients with adrenal

cancer: recommendations of an international consensus conference. Endocr Relat Cancer 12:667-680, 2005.

6. McNicol AM, Struthers AL, Nolan CE, et al. Proliferation in adrenocortical tumors: corre- lation with clinical outcome and p53 status. Endocr Pathol 8:29-36, 1997.

7. Nakazumi H, Sasano H, Iino K, Ohashi Y, Orikasa S. Expression of cell cycle inhibitor p27 and Ki-67 in human adrenocortical neo- plasms. Mod Pathol 11:1165-1170, 1998.

8. Terzolo M, Boccuzzi A, Bovio S, et al. Immunohistochemical assessment of Ki-67 in the differential diagnosis of adrenocortical tumors. Urology 57:176-182, 2001.

9. Iino K, Sasano H, Yabuki N, et al. DNA topoisomerase II alpha and Ki-67 in human adrenocortical neoplasms: a possible marker of differentiation between adenomas and car- cinomas. Mod Pathol 10:901-907, 1997.

10. Giordano TJ, Thomas DG, Kuick R, et al. Distinct transcriptional profiles of adrenocor- tical tumors uncovered by DNA microarray analysis. Am J Pathol 162:521-531, 2003.

11. Erickson LA, Jin L, Sebo TJ, et al. Pathologic features and expression of insulin-like growth factor-2 in adrenocortical neoplasms. Endocr Pathol 12:429-435, 2001.

12. Dohna M, Reincke M, Mincheva A, et al. Adrenocortical carcinoma is characterized by a high frequency of chromosomal gains and high-level amplifications. Genes Chromo- somes Cancer 28:145-152, 2000.

13. Kjellman M, Kallioniemi OP, Karhu R, et al. Genetic aberrations in adrenocortical tumors detected using comparative genomic hybrid- ization correlate with tumor size and malig- nancy. Cancer Res 56:4219-4223, 1996.

14. Sidhu S, Marsh DJ, Theodosopoulos G, et al. Comparative genomic hybridization analysis of adrenocortical tumors. J Clin Endocrinol Metab 87:3467-3474, 2002.

15. Zhao J, Roth J, Bode-Lesniewska B, et al. Combined comparative genomic hybridiza- tion and genomic microarray for detection of gene amplifications in pulmonary artery inti- mal sarcomas and adrenocortical tumors. Genes Chromosomes Cancer 34:48-57, 2002.

16. Boulle N, Logie A, Gicquel C, et al. Increased levels of insulin-like growth factor II (IGF-II) and IGF-binding protein-2 are associated with malignancy in sporadic adrenocortical tumors.

J Clin Endocrinol Metab 83:1713-1720, 1998.

17. Fottner C, Hoeflich A, Wolf E, et al. Role of the insulin-like growth factor system in adrenocortical growth control and carcinogen- esis. Horm Metab Res 36:397-405, 2004.

18. Gao ZH, Suppola S, Liu J, et al. Association of H19 promoter methylation with the expression of H19 and IGF-II genes in adrenocortical tumors. J Clin Endocrinol Metab 87:1170-1176, 2002.

19. Gicquel C, Bertagna X, Gaston V, et al. Molecular markers and long-term recurrences in a large cohort of patients with sporadic adrenocortical tumors. Cancer Res 61:6762- 6767,2001.

20. Gicquel C, Bertagna X, Schneid H, et al. Rearrangements at the 11p15 locus and overexpression of insulin-like growth factor- II gene in sporadic adrenocortical tumors. J Clin Endocrinol Metab 78:1444-1453, 1994.

21. Gicquel C, Raffin-Sanson ML, Gaston V, et al. Structural and functional abnormalities at 11p15 are associated with the malignant phe- notype in sporadic adrenocortical tumors: study on a series of 82 tumors. J Clin Endocrinol Metab 82:2559-2565, 1997.

22. Ilvesmaki V, Kahri AI, Miettinen PJ, et al. Insulin-like growth factors (IGFs) and their receptors in adrenal tumors: high IGF-II expression in functional adrenocortical carci- nomas. J Clin Endocrinol Metab 77:852-858, 1993.

23. Liu J, Kahri AI, Heikkila P, et al. H19 and insulin-like growth factor-II gene expression in adrenal tumors and cultured adrenal cells. J Clin Endocrinol Metab 80:492-496, 1995.

24. Liu J, Kahri AI, Heikkila P, et al. Ribonucleic acid expression of the clustered imprinted genes, p57KIP2, insulin-like growth factor II, and H19, in adrenal tumors and cultured adrenal cells. J Clin Endocrinol Metab 82:1766-1771, 1997.

25. Logie A, Boulle N, Gaston V, et al. Autocrine role of IGF-II in proliferation of human adrenocortical carcinoma NCI H295R cell line. J Mol Endocrinol 23:23-32, 1999.

26. Weber MM, Fottner C, Wolf E. The role of the insulin-like growth factor system in adrenocortical tumourigenesis. Eur J Clin Invest 30(Suppl 3):69-75, 2000.

27. Sameshima Y, Tsunematsu Y, Watanabe S, et al. Detection of novel germ-line p53 mutations in diverse-cancer-prone families identified by selecting patients with childhood adrenocor- tical carcinoma. J Natl Cancer Inst 84:703- 707, 1992.

28. Barzon L, Chilosi M, Fallo F, et al. Molecular analysis of CDKN1C and TP53 in sporadic adrenal tumors. Eur J Endocrinol 145:207- 212, 2001.

29. Kobayashi H, Usui T, Fukata J, et al. Muta- tion analysis of Gsalpha, adrenocorticotropin receptor and p53 genes in Japanese patients with adrenocortical neoplasms: including a case of Gsalpha mutation. Endocr J 47:461- 466, 2000.

30. Latronico AC, Pinto EM, Domenice S, et al. An inherited mutation outside the highly con- served DNA-binding domain of the p53 tumor suppressor protein in children and adults with sporadic adrenocortical tumors. J Clin Endocrinol Metab 86:4970-4973, 2001.

31. Lin SR, Lee YJ, Tsai JH. Mutations of the p53 gene in human functional adrenal neoplasms. J Clin Endocrinol Metab 78:483-491, 1994.

32. Ohgaki H, Kleihues P, Heitz PU. p53 muta- tions in sporadic adrenocortical tumors. Int J Cancer 54:408-410, 1993.

33. Olivier M, Goldgar DE, Sodha N, et al. Li-Fraumeni and related syndromes: correla- tion between tumor type, family structure, and TP53 genotype. Cancer Res 63:6643-6650, 2003.

34. Pinto EM, Billerbeck AE, Fragoso MC, et al. Deletion mapping of chromosome 17 in benign and malignant adrenocortical tumors associated with the Arg337His mutation of the p53 tumor suppressor protein. J Clin Endocrinol Metab 90:2976-2981, 2005.

35. Pinto EM, Billerbeck AE, Villares MC, et al. Founder effect for the highly prevalent R337H mutation of tumor suppressor p53 in Brazil- ian patients with adrenocortical tumors. Arq Bras Endocrinol Metabol 48:647-650, 2004.

36. Reincke M, Karl M, Travis WH, et al. p53 mutations in human adrenocortical neo- plasms: immunohistochemical and molecular studies. J Clin Endocrinol Metab 78:790-794, 1994.

37. Reincke M, Wachenfeld C, Mora P, et al. p53 mutations in adrenal tumors: Caucasian

patients do not show the exon 4 “hot spot” found in Taiwan. J Clin Endocrinol Metab 81:3636-3638, 1996.

38. Sandrini F, Villani DP, Tucci S, et al. Inherit- ance of R337H p53 gene mutation in chil- dren with sporadic adrenocortical tumor. Horm Metab Res 37:231-235, 2005.

39. Sidhu S, Martin E, Gicquel C, et al. Muta- tion and methylation analysis of TP53 in adrenal carcinogenesis. Eur J Surg Oncol 31:549-554, 2005.

40. Varley JM, McGown G, Thorncroft M, et al. Are there low-penetrance TP53 alleles? Evi- dence from childhood adrenocortical tumors. Am J Hum Genet 65:995-1006, 1999.

41. Wagner J, Portwine C, Rabin K, et al. High frequency of germline p53 mutations in child- hood adrenocortical cancer. J Natl Cancer Inst 86:1707-1710, 1994.

42. Libe R, Bertherat J. Molecular genetics of adrenocortical tumours, from familial to spo- radic diseases. Eur J Endocrinol 153:477-487, 2005.

43. Heppner C, Reincke M, Agarwal SK, et al. MEN1 gene analysis in sporadic adrenocorti- cal neoplasms. J Clin Endocrinol Metab 84:216-219, 1999.

44. Schulte KM, Heinze M, Mengel M, et al. MEN I gene mutations in sporadic adrenal adenomas. Hum Genet 105:603-610, 1999.

45. Schulte KM, Mengel M, Heinze M, et al. Complete sequencing and messenger ribo- nucleic acid expression analysis of the MEN I gene in adrenal cancer. J Clin Endocrinol Metab 85:441-448, 2000.

46. Bertherat J, Groussin L, Sandrini F, et al. Molecular and functional analysis of PRKAR1A and its locus (17q22-24) in sporadic adreno- cortical tumors: 17q losses, somatic mutations, and protein kinase A expression and activity. Cancer Res 63:5308-5319, 2003.

47. Bossis I, Voutetakis A, Bei T, et al. Protein kinase A and its role in human neoplasia: the Carney complex paradigm. Endocr Relat Cancer 11:265-280, 2004.

48. Groussin L, Jullian E, Perlemoine K, et al. Mutations of the PRKAR1A gene in Cushing’s syndrome due to sporadic primary pigmented nodular adrenocortical disease. J Clin Endocrinol Metab 87:4324-4329, 2002.

49. Libe R, Mantovani G, Bondioni S, et al. Mutational analysis of PRKAR1A and Gs(alpha) in sporadic adrenocortical tumors.

Exp Clin Endocrinol Diabetes 113:248-251, 2005.

50. Tissier F, Cavard C, Groussin L, et al. Muta- tions of beta-catenin in adrenocortical tumors: activation of the Wnt signaling pathway is a frequent event in both benign and malignant adrenocortical tumors. Cancer Res 65:7622- 7627, 2005.

51. Ciampi R, Nikiforov YE. Alterations of the BRAF gene in thyroid tumors. Endocr Pathol 16:163-172, 2005.

52. Ciampi R, Zhu Z, Nikiforov YE. BRAF copy number gains in thyroid tumors detected by fluorescence in situ hybridization. Endocr Pathol 16:99-105, 2005.

53. Ciampi R, Knauf JA, Kerler R, et al. Onco- genic AKAP9-BRAF fusion is a novel mecha- nism of MAPK pathway activation in thyroid cancer. J Clin Invest 115:94-101, 2005.

54. Lakhani SR, Ashworth A. Microarray and his- topathological analysis of tumours: the future and the past? Nat Rev Cancer 1:151-157, 2001.

55. Mischel PS, Cloughesy TF, Nelson SF. DNA- microarray analysis of brain cancer: molecu- lar classification for therapy. Nat Rev Neurosci 5:782-792, 2004.

56. Snijders AM, Meijer GA, Brakenhoff RH, et al. Microarray techniques in pathology: tool or toy? Mol Pathol 53:289-294, 2000.

57. Alizadeh AA, Ross DT, Perou CM, et al. Towards a novel classification of human malignancies based on gene expression pat- terns. J Pathol 195:41-52, 2001.

58. Giordano TJ. Gene expression profiling of endocrine tumors using DNA microarrays: progress and promise. Endocr Pathol 14:107- 116, 2003.

59. de Fraipont F, El Atifi M, Cherradi N, et al. Gene expression profiling of human adreno- cortical tumors using complementary deox- yribonucleic acid microarrays identifies several candidate genes as markers of malignancy. J Clin Endocrinol Metab 90:1819-1829, 2005.