Dermatologic Therapy, Vol. 19, 2006, 40 – 49 Printed in the United States · All rights reserved

Copyright © Blackwell Publishing, Inc., 2006

DERMATOLOGIC THERAPY ISSN 1396-0296

Distinguishing melanocytic nevi from melanoma by DNA copy number changes: comparative genomic hybridization as a research and diagnostic tool

Blackwell Publishing, Ltd.

JÜRGEN BAUER*† & BORIS C. BASTIAN* *Departments of Dermatology and Pathology and the University of California at San Francisco Comprehensive Cancer, University of California at San Francisco, and †Department of Dermatology, Eberhard Karls University, Tübingen, Germany

ABSTRACT: Cancer typically results in loosened control over genomic integrity, resulting in alterations of the genome of cancer cells. Comparative genomic hybridization (CGH) is a method that can be used on DNA extracted from routinely fixed tissue to assess the entire genome for the presence of changes in DNA copy number. CGH analysis has revealed that melanoma differs from melanocytic nevi by the presence of frequent chromosomal aberrations. In contrast, melanocytic nevi typically show no chromosomal aberrations, or have a restricted set of alterations with basically no overlap to melanoma. These marked differences between aberration patterns in melanomas and melanocytic nevi can be exploited diagnostically to classify melanocytic tumors that are ambiguous based on histopathologic assessment. In addition to potential diagnostic applications, detailed analyses of recurrent aberrations can lead to the identification of genes relevant in melanocytic neoplasia. KEYWORDS: classification, comparative genomic hybridization, copy number changes, melanocytic nevi, melanoma

Introduction Cancer is a disease of the genome. It arises through genetic alterations in cells that are subject to further selection. During this evolutionary process, control over genomic integrity becomes progressively compromised, resulting in loss of tumor-suppressor genes and activation of oncogenes. As a result of these alterations, cancer cells differ substantially from their normal counterparts. Address correspondence and reprint requests to: Boris C. Bastian, MD, Comprehensive Cancer Center, University of California, San Francisco, Box 0808, San Francisco, CA 94143-0808, or email: [email protected].

40

In most solid cancers, these differences express themselves in structural alterations of chromosomes. Only a minority of solid cancers develop without significant chromosomal instability. The most prominent example is hereditary nonpolyposis colorectal cancer in which loss over genomic integrity manifests itself in the form of microsatellite instability caused by alterations of the mismatched repair system instead of frequent chromosomal rearrangements (1). In melanoma, genomic instability occurs predominantly on the chromosomal level with over 95% of primary melanomas showing gains or losses of portions of chromosomes (2,3). In contrast, the majority of melanocytic nevi appears to maintain control over genomic integrity

Distinguishing melanocytic nevi from melanoma

Comparative genomic hybridization was originally described in 1992 as a method to detect and map DNA sequence copy number changes throughout the genome onto a cytogenetic map supplied by metaphase chromosomes (FIG. 1, left) (4). It

consists of the simultaneous hybridization of two differentially labeled DNA populations from the same species. One DNA population is derived from a test tissue sample (e.g., a tumor) and the other is derived from a healthy donor and serves as a reference representing the normal genome. Both DNAs are labeled with different fluorochromes, and the mixture of both DNA populations is hybridized onto normal metaphase spreads (FIG. 2). These spreads are also obtained from a healthy donor from the same species and serve as a substrate for the hybridization and as a genetic map of the genome of the species of interest. During the hybridization, the two populations compete for their corresponding sequences on the substrate chromosomes. If the relative abundances of the tumor DNA and the reference DNA are equal, the relative ratio of tumor to reference fluorescence intensity for the respective genomic region equals 1. If the tumor has increases in copy number of a given region, the ratio of tumor to reference fluorescence intensity exceeds 1. This is

FIG. 1. Scheme of conventional and array comparative genomic hybridization (CGH). Total genomic DNA is isolated from test and reference samples, and labeled with green and red fluorochromes, respectively. The mixture of denatured test and reference DNA can be hybridized to normal metaphase chromosomes (left) or a microarray of mapped clones of genomic DNA (right). When hybridized to chromosomes, the resulting ratio of the fluorescence intensities of the two colors at a location on a chromosome is approximately proportional to the ratio of the copy numbers of the corresponding DNA sequences in the test and reference genomes. When hybridized to a microarray, the ratio of the fluorescence intensity for each array element provides the relative copy numbers for the genomic locus represented by the array element. The overall resolution of the array depends on the genomic distance between the clones and their lengths.

FIG. 2. Conventional comparative genomic hybridization (CGH) of a melanoma. The tumor DNA was labeled with a green fluorochrome and hybridized with a red-labeled normal reference DNA onto normal human chromosomes. Copy number gains appear green and losses red. The highly repetitive sequences of centromeres appear blue because they were blocked from hybridization by unlabeled blocking DNA, and the blue-appearing counter-stain used for chromosome identification can be seen.

and can be distinguished from melanoma by the absence of gains or losses of DNA (2). Comparative genomic hybridization (CGH) is a method that allows the genome-wide detection of DNA copy number changes. CGH compares a tumor genome to a normal genome and allows the detection and mapping of genomic aberrations that result in DNA copy number changes (4). This review describes the method and its application as a research and diagnostic tool for melanocytic neoplasia.

CGH as method for genome-wide tumor gene screening Conventional CGH

41

Bauer & Bastian

called a copy number gain if it results in a moderate increase of the ratio. If the copy number increase is more substantial, indicating that multiple copies of the respective region have accumulated, higher tumor to reference fluorescence-intensity ratios are observed and are called amplification. Amplifications typically affect relatively narrow regions of the genome that often contain potent oncogenes whose increased gene dosage has been selected for during tumor development. If the tumor genome has lost one or more copies of a portion of its genome, the ratio of tumor to reference fluorescence intensity at this region is smaller than 1, indicating a deletion. The full experimental protocol for CGH is slightly more complex as outlined previously. A third, unlabeled DNA population is needed to ascertain that repetitive regions that are scattered throughout the genome do not cross-hybridize and interfere with the measurement. This blocking DNA is highly enriched for repetitive regions and suppresses unwanted cross-hybridization between repetitive regions in the labeled DNA populations and the chromosomes that serve as substrate. Comparative genomic hybridization has revolutionized the cytogenetic analysis of solid tumors. Compared to conventional cytogenetics, CGH does not require culture of cells for karyotypic analysis and, most importantly, can be performed on archival tissue. It is important to realize that the DNA copy number measurement obtained with CGH represents an average of the entire cell population analyzed. Only the copy number alterations present in a substantial portion of the cells from which the DNA is derived are detected by the method. Depending on the type of aberration – amplifications can be detected most easily – the copy number change needs to be present in about 30–50% of the cells in order to be identifiable. Thus, aberrations found by CGH most likely are related to selection, whereas random alterations or alterations affecting only a minority of cells remain undetected. CGH only detects genomic aberrations that result in DNA copy number changes. Balanced translocations and point mutations are not detected. In some instances, the latter reveals themselves indirectly in a CGH measurement, if, e.g., a translocation breakpoint or a mutated oncogene gets duplicated or amplified. Also, individual copies of tumor suppressor genes frequently become disabled by different mechanisms, e.g., a mutation affecting one allele and a deletion affecting the other. Therefore, regions of copy number changes identified by CGH can help identify the location of cancer genes.

42

Array CGH Recently, a variation of conventional CGH was developed in which metaphase chromosomes are replaced by arrays of genomic bacterial artificial chromosome (BAC) clones (FIG. 1, right) (5,6). The hybridization protocol for array CGH is similar to that described previously for CGH on human metaphase chromosomes. Depending on the spacing of the BAC clones, microarrays can significantly improve resolution, permitting detection of small aberrations below the threshold of detection of chromosome CGH. The BAC clones of the array are linked by their genomic sequence to the human genome, allowing a precise determination of the boundaries of aberrations (7). Array CGH allows accurate quantification of DNA copy number variations over a wide dynamic range, including reliable detection of single-copy deletions and duplications (5). Presently, the maximum resolution of array CGH is approximately 1 megabase across the human genome, compared to conventional CGH, which is limited to about 10–20 megabases. BAC arrays are now becoming available that are assembled from an entire tiling path of clones, allowing gapless assessment of copy number changes throughout the entire genome (8).

CGH of melanocytic neoplasms Initial cytogenetic studies on melanoma were mostly performed on metastases and cell lines using conventional karyotypic analysis (9–12). Later, loss of heterozygosity (LOH) (13–17) and CGH studies (18–21) have shown that the vast majority of primary melanomas of the skin also show chromosomal aberrations. LOH analyzes detect variants of alleles, and by comparing a patient’s normal tissue with the corresponding tumor, one can determine whether in the tumor, the quantitative ratio between the alleles has shifted. Interestingly, if analyzed for DNA copy number changes with CGH, melanocytic nevi differ significantly in their chromosomal aberration patterns from melanoma (2,18,22). In a recently published study, investigating 132 melanomas and 54 benign nevi, more than 96% of the melanomas presented with chromosomal aberrations on CGH analysis (2). In contrast, only 13% of the benign nevi studied had aberrations. All seven nevi with aberrations were Spitz nevi, six of which revealed the aberration to be an isolated gain of the short arm of chromosome 11. All other types of benign acquired nevi, congenital nevi, and blue nevi showed no aberrations

Distinguishing melanocytic nevi from melanoma

in this series (2). Analysis of a larger series of 102 Spitz nevi using fluorescence in situ hybridization (FISH) found at least threefold gains of chromosome 11p in 12 cases (11.8%) (23). The 11p gain in Spitz nevus was also found by another group using array CGH to investigate Spitz nevi and melanoma (24). Spitz nevi that recur at the excision site were found to have a higher frequency of 11p gain (25). Proliferative nodules in large congenital melanocytic nevi can be clinically, as well histologically, worrisome. These nodules typically present as rapid growths in the first year of life. Four different histological patterns of secondary nodular proliferations in congenital melanocytic nevi during the neonatal period have been described (26): (a) simulators of superficial spreading melanoma, in which the epidermis and superficial dermis contain large epithelioid melanocytes, sometimes with pagetoid upward spread in the epidermis; (b) simulators of nodular melanoma, characterized by a nodular proliferation of large melanocytes with uniform nuclei in the dermis; (c) cases described as “proliferative neurocristic hamartoma,” characterized by a deep dermal or subcutaneous proliferation with a variety of forms of neural or mesenchymal differentiation; and (d) true melanoma, most of which show small blastlike melanocytes with hyperchromatic nuclei, scant cytoplasm, and a high mitotic rate. A CGH analysis of 29 congenital melanocytic nevi and associated benign or malignant proliferations showed no aberrations in typical congenital melanocytic nevi (n = 6), in congenital nevi with foci of increased cellularity (n = 4), and in congenital nevi with a benign proliferation, simulating superficial spreading melanoma in situ (n = 3) (27). Seven of nine cases of congenital melanocytic nevi with a proliferation-simulating nodular melanoma showed gain or loss of entire chromosomes exclusively. This array differs from melanoma, in which chromosomal aberrations typically affect fragments of chromosomes rather than entire chromosomes. In addition, there were differences in the pattern of chromosomal aberrations between nonmalignant nodular proliferations and melanoma. The six cases of melanoma arising in congenital nevi showed multiple cytogenetic aberrations in a pattern indistinguishable from that of melanomas not associated with congenital nevi (27). One case classified morphologically as proliferative neurocristic hamartoma showed aberrations very similar to melanoma. Together, these studies show that benign melanocytic nevi differ significantly from melanoma with regard to their pattern of chromosomal abnormalities.

Comparative genomic hybridization and other genetic analyses have also revealed that melanoma is not a homogenous disease but rather a collection of distinct subtypes. Melanomas on the glabrous skin of hands and feet or under the nails are characterized genetically by frequent narrow amplifications that occur very early in tumor progression and can be detected already at the in situ stage (28). An analysis of 132 cases also showed that melanomas on glabrous acral skin have significantly more aberrations involving chromosomes 5p, 11q, 12q, and 15, along with the frequent focused gene amplifications. Melanomas occurring on severely sun-damaged skin, commonly but not exclusively of the lentigo maligna type, showed markedly more frequent losses of chromosomes 17p and 13q (2,28). In another study, melanomas and chronically sun-damaged skin also showed significantly less frequent mutations in BRAF when compared to melanomas presenting on skin without signs of chronic sun damage. These marked differences in the frequency of mutations in BRAF, along with differences in patterns of chromosomal aberrations, strongly suggest that these two common types of melanoma are genetically and biologically distinct (29). CGH analysis of 14 sinonasal mucosal melanomas showed frequent copy number increases of chromosome arm 1q along with gains of 6p and 8q. These aberrations were found to be present in a significantly higher percentage compared to other melanoma types, suggesting that mucosal melanoma is another distinct tumor subtype (21). BRAF mutations are also rare in mucosal melanomas (29–31). Such marked differences in genetic composition of melanomas, highly dependent on anatomical location and sun-exposure pattern, indicate that potential therapeutic targets might vary among melanoma subtypes (2).

Application of CGH in clinical diagnosis of melanocytic tumors The observation of frequent chromosomal aberrations in melanoma, and a relative absence of aberrations in benign nevi, raises the possibility that chromosomal analysis could be exploited diagnostically in melanocytic lesions that are ambiguous based on current methods of assessment. Histopathology is the gold standard for diagnosis of melanocytic tumors. However, in a subset of cases, an unequivocal diagnosis may not be possible, based on histomorphologic features alone (32–36). Immunohistochemistry is usually

43

Bauer & Bastian

of limited value in these settings. Typical examples of cases in which diagnostic ambiguity arises include atypical Spitz nevi (37); persistent or recurrent melanocytic nevi (38); melanocytic nevi in acral, genital, or mammary line regions (39,40); mechanically irritated or recently sun-exposed nevi (41); and proliferative nodules in congenital nevi. Misdiagnosis of a melanoma as a benign melanocytic nevus, or a benign melanocytic nevus as a melanoma creates havoc with patient management decisions, resulting in under-treatment or overly aggressive treatment (42). The following two cases demonstrate how CGH provides additional information that may be helpful in diagnosing microscopically ambiguous cases. Case 1 A 26-year-old female presented with a pigmented lesion in the left inguinal crease, clinically diagnosed as an atypical nevus. Microscopic examination showed an asymmetrical compound proliferation that was comprised of two morphologically distinct populations of melanocytes (FIG. 3a,b). There was a relatively well-circumscribed compound proliferation of morphologically bland-appearing

melanocytes that were arranged in nests, and partially single cells along the junction and in the underlying dermis. The dermal aggregates of this portion became smaller with descent and showed diminishing cell size. This portion was consistent with a preexisting melanocytic nevus (FIG. 3c). In some of the sections, there was a nodular proliferation of significantly larger melanocytes with ample eosinophilic cytoplasm, large pleomorphic nuclei, and nuclear pseudo-inclusions (FIG. 3d). Some cells were pigmented. Rare mitotic figures were present. A Ki-67 stain showed a slightly increased proliferation rate in the atypical nodular proliferation. The differential diagnosis included a combined nevus (preexisting compound nevus and Spitz’s nevus) and a spitzoid melanoma developing in a compound melanocytic nevus. CGH analysis performed on the spitzoid proliferation showed no aberrations throughout the entire genome. This finding supports an interpretation of a benign melanocytic proliferation over a melanoma. Case 2 A 61-year-old male presented with a lesion on his upper back. It was described as long standing.

FIG. 3. Case 1: (a) Center of a relatively symmetrical compound proliferation of melanocytes with two morphologically distinct populations of cells. (b) This panel shows a close-up view with small, morphologically bland-appearing melanocytes immediately subjacent to the epidermis, and a second morphologically distinct population of larger melanocytes below. (c) Close-up view of the population of small, morphologically bland melanocytes. (d) Nodular proliferation of significantly larger melanocytes with large pleomorphic nuclei and scattered interspersed melanophages. Individual melanocytes show fine granular cytoplasmic pigmentation. CGH was performed on the nodular proliferation of large cells and did not show chromosomal aberrations. This lesion was interpreted as benign.

44

Distinguishing melanocytic nevi from melanoma

FIG. 4. (a) Case 2 shows an intradermal proliferation of melanocytes surrounded by fibrosis. (b) Periphery of the nodule with small, morphologically bland-appearing melanocytes arranged in small nests with adjacent fibrosis. (c) The center of the nodule is more cellular and shows rare mitotic figures (arrowhead). (d) Within the nodule, areas with larger cells and markedly atypical nuclei are present. This lesion was interpreted as melanoma.

It apparently changed in size over time, and clinically appeared to be a cyst. Histopathologic assessment showed a dermal proliferation of atypical melanocytes surrounded by fibrous tissue (FIG. 4a). There were two tumor portions, less cellular areas with cords, and strands of small round to oval cells (FIG. 4b), and a large nodular area with striking cellularity and considerably larger cells. In this zone of striking cellularity, some cells had bizarre nuclei with large eosinophilic nucleoli; other cells showed an eccentric, deeply eosinophilic cytoplasm (FIG. 4c,d). Even smaller cells in this portion of tissue had vesicular nuclei and large nucleoli.

Mitotic figures were scant, but some were present. A Ki-67 stain revealed a proliferation rate of about 5%, not a decisive finding. The differential diagnosis included a morphologically unusual melanoma arising in an atypical dermal nevus, versus a combined melanocytic nevus with a Spitz nevus component. As mentioned previously, a dermal melanoma associated with a nevus would be unexpected and very rare. Yet CGH showed a pattern of multiple chromosomal aberrations with gains and losses (FIG. 5). These findings are typical for melanoma, and finally, a melanoma was diagnosed.

45

Bauer & Bastian

FIG. 5. Chromosomal CGH analysis of case 2. Displayed are ideograms of 22 pairs of autosomes and two sex chromosomes with their respective Giemsa banding patterns. Green lines next to the chromosomes indicate regions of gains of copy number and red lines losses. The graphs next to the chromosomes display the average ratio of tumor and reference DNA signal intensity and standard deviations along the axis of the chromosome. The n indicates how many chromosomes were analyzed for each graph. Ratios higher than 1.2 (green shaded area) and lower than 0.8 (red shaded area) are regarded as aberrations. There are losses of the long arm of chromosome 6 as well as the entire chromosome 9. Additionally, there are gains of the short arm of chromosome 6 and the entire chromosome 8. We note that the chromosomal portions close to the tips of chromosomes as well as centromeric areas (hatched area in the chromosome ideograms) are excluded from the measurement because they can show unreliable ratio changes.

Application of CGH to identify tumor genes in melanocytic tumors In addition to yielding potentially helpful diagnostic information, the genetic data provided by CGH can also provide insight into biology. Although genetic changes probably occur randomly throughout the genome during tumor evolution, alterations that promote tumor growth are preferentially selected. As a consequence, the clonal genetic alterations detected by CGH can be used to identify cancer genes that have been activated

46

or inactivated by DNA copy number changes. Focused gene amplifications are particularly informative, as they typically affect small genomic regions. Similarly, DNA copy number decreases can indicate the location of tumor suppressor genes. By way of example of this concept, the recurrent copy number increases of chromosome 11p in a subset of Spitz nevi (22) implicated HRAS as a candidate oncogene on 11p. The role of HRAS has been confirmed by demonstrating that the majority of Spitz nevi with increased copy number of chromosome 11p showed activating mutations in HRAS (23). HRAS is part of the Ras/Raf/MEK/ERK

Distinguishing melanocytic nevi from melanoma

pathway, which is a signaling cascade that conveys growth signals from receptor tyrosine kinases to the nucleus and controls proliferation, as well as differentiation, survival, and apoptosis. A study of 102 Spitz nevi for 11p copy number increases using FISH on tissue arrays has shown copy number increases of at least threefold in 12 cases (23). Sequence analysis of HRAS showed frequent oncogenic mutations in 67% of cases with copy number increase, contrasting with rare HRAS mutations (5%) in Spitz nevi with normal HRAS copy numbers (p < 0.0001). Spitz nevi with 11p copy number increases tended to be larger, predominantly intradermal, with marked desmoplasia, and characteristic cytological features than Spitz nevi without the 11p gain. Proliferation rates in the majority of these cases were low to absent, and long-term follow-up indicated no progression in cases with the 11p gain and/or the mutation in HRAS (23). The presence of a powerful oncogenic mutation in any benign neoplasm raises the question of how cells with increased copy number of a mutated HRAS allele in a Spitz nevus can withstand full transformation to melanoma. An analysis of potential mechanisms that inhibit proliferation in the presence of HRAS activation has revealed that among a variety of cell cycle inhibitors, p16 was highly expressed in Spitz nevus cells with HRAS mutations or copy number increases (43). P16 executes an important checkpoint function downstream of the RAS/RAF/MEK/ERK pathway by inhibiting cyclin-dependent kinase 4 (CDK4). Mutations in either p16 or CDK4 that disable this checkpoint result in a melanoma-prone condition (44). The finding of strong expression of p16 protein in Spitz nevi with mutations or copy number increases of HRAS suggests that in the presence of intact p16, melanocytes can withstand the oncogenic effects of a powerful oncogene such as HRAS. A similar effect is observed in cell culture, in which cells undergo oncogene-induced senescence when an activated HRAS gene is introduced into a cell that has intact p16 (45). In melanoma, p16 is frequently lost during tumor progression, as evidenced by losses involving chromosome 9p as the most common aberration in melanoma. The absence of an intact G1/S checkpoint as a result of loss of function of p16 may thus allow cells to proliferate unchecked in response to a mutation in oncogenes such as RAS. The unchecked proliferation would then result in progressive telomere attrition with subsequent genomic instability and gains and losses of chromosomes (46). This model renders a plausible

explanation why typical melanocytic nevi and melanoma differ from each other with respect to absence or presence of chromosomal alterations. Melanomas have progressed through telomeric crisis, whereas typical melanocytic nevi are prevented from entering crisis by an intact checkpoint that prevents critical telomere shortening. Because telomeric crisis is thought to invariably result in karyotypic alterations, the presence of chromosomal aberrations may thus be exploited to determine a postcrisis situation, i.e., a cancerous state. As noted previously, recurrent gene amplifications can help identify oncogenes relevant in a particular cancer type. About 40% of acral melanomas show amplifications at 11q13, a locus that contains several oncogenes including cyclin D1, FGF3, and FGF4 (28). Comparison of the cyclin D1 gene copy number and protein expression in 137 invasive primary cutaneous melanomas using FISH and immunohistochemistry showed frequent amplification of cyclin D1 in acral melanoma (44.4%), and occasional amplification in lentigo maligna melanoma (10.5%) and superficial spreading melanoma (5.6%) (47). Cyclin D1 protein was over-expressed in all cases with amplifications, and in an additional 20% of cases without amplification. The essential role of cyclin D1 over-expression for cell growth and survival in melanoma has been validated using adenovirusmediated antisense treatment targeted to cyclin D1 in two melanoma cell lines. Both cell lines expressed high levels of cyclin D1, and one had an amplification of the gene (47). Antisense-mediated down-regulation of cyclin D1 induced apoptosis in both cell lines and led to significant tumor shrinkage of melanoma xenografts derived from these cell lines. Antisense-mediated down-regulation of cyclin D1 did not alter the growth of normal melanocytes. These findings strongly suggest that cyclin D1 can act as an oncogene in melanoma (47). Comparative genomic hybridization also has been useful in analyzing somatic alterations in cancers developing in mouse models of cancer (48,49). Copy number increases of cyclin D1 have also been detected in melanomas arising in p53 deficient mice, which expressed an activated HRAS gene in their melanocytes. In addition to frequent copy number increases of cyclin D1, these melanomas also showed recurrent copy number increases and overexpression of the c-myc oncogene on mouse chromosome 15 (50). In conclusion, comparative genomic hybridization is a powerful method to detect and map DNA

47

Bauer & Bastian

copy number changes in DNA extracted from archival tissue. The DNA copy number changes in melanoma and melanocytic nevi differ remarkably from each other. These differences can be exploited for diagnostic purposes as well as for discovery of genetic mechanisms in melanoma progression.

References 1. Lynch HT, de la Chapelle A. Genetic susceptibility to nonpolyposis colorectal cancer. J Med Genet 1999: 36: 801– 818. 2. Bastian BC, Olshen AB, LeBoit PE, Pinkel D. Classifying melanocytic tumors based on DNA copy number changes. Am J Pathol 2003: 163: 1765– 1770. 3. Bastian BC. Understanding the progression of melanocytic neoplasia using genomic analysis: from fields to cancer. Oncogene 2003: 22: 3081– 3086. 4. Kallioniemi A, Kallioniemi OP, Sudar D, et al. Comparative genomic hybridization for molecular cytogenetic analysis of solid tumors. Science 1992: 258: 818– 821. 5. Albertson DG, Ylstra B, Segraves R, et al. Quantitative mapping of amplicon structure by array CGH identifies CYP24 as a candidate oncogene. Nat Genet 2000: 25: 144–146. 6. Pinkel D, Segraves R, Sudar D, et al. High resolution analysis of DNA copy number variation using comparative genomic hybridization to microarrays. Nat Genet 1998: 20: 207 –211. 7. Cheung VG, Nowak N, Jang W, et al. Integration of cytogenetic landmarks into the draft sequence of the human genome. Nature 2001: 409: 953– 958. 8. Ishkanian AS, Malloff CA, Watson SK, et al. A tiling resolution DNA microarray with complete coverage of the human genome. Nat Genet 2004: 36: 299– 303. 9. Mertens F, Johansson B, Hoglund M, Mitelman F. Chromosomal imbalance maps of malignant solid tumors: a cytogenetic survey of 3185 neoplasms. Cancer Res 1997: 57: 2765 –2780. 10. Thompson FH, Emerson J, Olson S, et al. Cytogenetics of 158 patients with regional or disseminated melanoma. Subset analysis of near-diploid and simple karyotypes. Cancer Genet Cytogenet 1995: 83: 93–104. 11. Trent JM, Thompson FH, Ludwig C. Evidence for selection of homogeneously staining regions in a human melanoma cell line. Cancer Res 1984: 44: 233– 237. 12. Kakati S, Song SY, Sandberg AA. Chromosomes and causation of human cancer and leukemia. XXII. Karyotypic changes in malignant melanoma. Cancer 1977: 40: 1173– 1181. 13. Healy E, Rehman I, Angus B, Rees JL. Loss of heterozygosity in sporadic primary cutaneous melanoma. Genes Chromosomes Cancer 1995: 12: 152–156. 14. Takata M, Morita R, Takehara K. Clonal heterogeneity in sporadic melanomas as revealed by loss-of-heterozygosity analysis. Int J Cancer 2000: 85: 492– 497. 15. Isshiki K, Elder DE, Guerry D, Linnenbach AJ. Chromosome 10 allelic loss in malignant melanoma. Genes Chromosomes Cancer 1993: 8: 178–184. 16. Peris K, Keller G, Chimenti S, Amantea A, Kerl H, Hofler H. Microsatellite instability and loss of heterozygosity in melanoma. J Invest Dermatol 1995: 105: 625– 628. 17. Millikin D, Meese E, Vogelstein B, Witkowski C, Trent J. Loss of heterozygosity for loci on the long arm of chromosome 6

48

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

32. 33.

34. 35.

in human malignant melanoma. Cancer Res 1991: 51: 5449–5453. Bastian BC, LeBoit PE, Hamm H, Brocker EB, Pinkel D. Chromosomal gains and losses in primary cutaneous melanomas detected by comparative genomic hybridization. Cancer Res 1998: 58: 2170–2175. Wiltshire RN, Dennis TR, Sondak VK, Meltzer PS, Trent JM. Application of molecular cytogenetic techniques in a case study of human cutaneous metastatic melanoma. Cancer Genet Cytogenet 2001: 131: 97–103. Balazs M, Adam Z, Treszl A, Begany A, Hunyadi J, Adany R. Chromosomal imbalances in primary and metastatic melanomas revealed by comparative genomic hybridization. Cytometry 2001: 46: 222–232. van Dijk M, Sprenger S, Rombout P, et al. Distinct chromosomal aberrations in sinonasal mucosal melanoma as detected by comparative genomic hybridization. Genes Chromosomes Cancer 2003: 36: 151–158. Bastian BC, Wesselmann U, Pinkel D, LeBoit PE. Molecular cytogenetic analysis of Spitz nevi shows clear differences to melanoma. J Invest Dermatol 1999: 113: 1065–1069. Bastian BC, LeBoit PE, Pinkel D. Mutations and copy number increase of HRAS in Spitz nevi with distinctive histopathological features. Am J Pathol 2000: 157: 967–972. Harvell JD, Kohler S, Zhu S, Hernandez-Boussard T, Pollack JR, van de RM. High-resolution array-based comparative genomic hybridization for distinguishing paraffinembedded Spitz nevi and melanomas. Diagn Mol Pathol 2004: 13: 22–25. Harvell JD, Bastian BC, LeBoit PE. Persistent (recurrent) Spitz nevi: a histopathologic, immunohistochemical, and molecular pathologic study of 22 cases. Am J Surg Pathol 2002: 26: 654–661. Clark WH Jr, Elder DE, Guerry D. Dysplastic nevi and malignant melanoma. In: Farmer ER, Hood AF, eds. Pathology of the skin. New York: McGraw-Hill, 1990: 729 –735. Bastian BC, Xiong J, Frieden IJ, et al. Genetic changes in neoplasms arising in congenital melanocytic nevi: differences between nodular proliferations and melanomas. Am J Pathol 2002: 161: 1163–1169. Bastian BC, Kashani-Sabet M, Hamm H, et al. Gene amplifications characterize acral melanoma and permit the detection of occult tumor cells in the surrounding skin. Cancer Res 2000: 60: 1968–1973. Maldonado JL, Fridlyand J, Patel H, et al. Determinants of BRAF mutations in primary melanomas. J Natl Cancer Inst 2003: 95: 1878–1890. Cohen Y, Rosenbaum E, Begum S, et al. Exon 15 BRAF mutations are uncommon in melanomas arising in nonsunexposed sites. Clin Cancer Res 2004: 10: 3444–3447. Sasaki Y, Niu C, Makino R, et al. BRAF point mutations in primary melanoma show different prevalences by subtype. J Invest Dermatol 2004: 123: 177–183. Cook MG. Diagnostic discord with melanoma. J Pathol 1997: 182: 247–249. Corona R, Mele A, Amini M, et al. Interobserver variability on the histopathologic diagnosis of cutaneous melanoma and other pigmented skin lesions. J Clin Oncol 1996: 14: 1218–1223. Jackson R. Malignant melanoma: a review of 75 malpractice cases. Int J Dermatol 1997: 36: 497–498. Kempf W, Haeffner AC, Mueller B, Panizzon RG, Burg G. Experts and gold standards in dermatopathology: qualitative and quantitative analysis of the self-assessment slide seminar at the 17th colloquium of the International Society of Dermatopathology. Am J Dermatopathol 1998: 20: 478–482.

Distinguishing melanocytic nevi from melanoma 36. Veenhuizen KC, De Wit PE, Mooi WJ, Scheffer E, Verbeek AL, Ruiter DJ. Quality assessment by expert opinion in melanoma pathology: experience of the pathology panel of the Dutch Melanoma Working Party. J Pathol 1997: 182: 266–272. 37. Barnhill RL, Argenyi ZB, From L, et al. Atypical Spitz nevi/ tumors: lack of consensus for diagnosis, discrimination from melanoma, and prediction of outcome. Hum Pathol 1999: 30: 513– 520. 38. Hoang MP, Prieto VG, Burchette JL, Shea CR. Recurrent melanocytic nevus: a histologic and immunohistochemical evaluation. J Cutan Pathol 2001: 28: 400– 406. 39. Clark WH Jr, Hood AF, Tucker MA, Jampel RM. Atypical melanocytic nevi of the genital type with a discussion of reciprocal parenchymal-stromal interactions in the biology of neoplasia. Hum Pathol 1998: 29: S1– S24. 40. Ackerman AB, Cerroni L, Kerl H. Pitfalls in histopathologic diagnosis of malignant melanoma. Philadelphia: Lea & Febiger, 1994. 41. Tronnier M, Alexander M, Neitmann M, Brinckmann J, Wolff HH. Morphologische veränderungen in melanozytären nävi durch exogene faktoren. Hautarzt 2000: 51: 561–566. 42. LeBoit PE. Stimulants of malignant melanoma: a rogue’s gallery of melanocytic and nonmelanocytic imposters. Pathology 1994: 2: 195– 258.

43. Maldonado JL, Timmerman L, Fridlyand J, Bastian BC. Mechanisms of cell-cycle arrest in Spitz nevi with constitutive activation of the MAP-kinase pathway. Am J Pathol 2004: 164: 1783–1787. 44. Zuo L, Weger J, Yang Q, et al. Germline mutations in the p16INK4a binding domain of CDK4 in familial melanoma. Nat Genet 1996: 12: 97–99. 45. Serrano M, Lee H, Chin L, Cordon-Cardo C, Beach D, DePinho RA. Role of the INK4a locus in tumor suppression and cell mortality. Cell 1996: 85: 27–37. 46. Sherr CJ, DePinho RA. Cellular senescence: mitotic clock or culture shock? Cell 2000: 102: 407–410. 47. Sauter ER, Yeo UC, von Stemm A, et al. Cyclin D1 is a candidate oncogene in cutaneous melanoma. Cancer Res 2002: 62: 3200–3206. 48. Hodgson G, Hager JH, Volik S, et al. Genome scanning with array CGH delineates regional alterations in mouse islet carcinomas. Nat Genet 2001: 29: 459–464. 49. Snijders AM, Nowak NJ, Huey B, et al. Mapping segmental and sequence variations among laboratory mice using BAC array CGH. Genome Res 2005: 15: 302–311. 50. Bardeesy N, Bastian BC, Hezel A, Pinkel D, DePinho RA, Chin L. Dual inactivation of RB and p53 pathways in RASinduced melanomas. Mol Cell Biol 2001: 21: 2144–2153.

49