Artur Zembowicz, MD, PhD; Sung-Eun Yang, MD; Antonios Kafanas, MD; Stephen R. Lyle, MD, PhD

Correlation Between Histologic Assessment and Fluorescence In Situ Hybridization Using MelanoSITE in Evaluation of Histologically Ambiguous Melanocytic Lesions Artur Zembowicz, MD, PhD; Sung-Eun Yang, MD; Antonios Kafanas, MD; Stephen R. Lyle, MD, PhD N Context. The 4-probe, multicolor, fluorescence in situ hybridization (FISH) panel targeting chromosomes 6 and 11 has shown promising sensitivity and specificity in distinguishing between benign nevi and malignant. Only a few studies have assessed the potential utility of FISH in classification of histologically ambiguous melanocytic lesions. In the United States, this assay is exclusively licensed to NeoGenomics Laboratories (Irvine, California), which provides the technical component and has developed an innovative service (MelanoSITE) allowing pathologists to interpret FISH results using a dedicated Web portal. Thus far, use of MelanoSITE as a diagnostic adjunct in the diagnosis of melanocytic lesions has not, to our knowledge, been reported in the literature. Objective. To analyze 1.5 years of experience with the MelanoSITE FISH assay in the evaluation of histologically ambiguous lesions in the context Recent advances in understanding the molecular pathogenesis of melanocytic proliferations have revealed many genetic differences between benign nevi and that could serve as potential targets for developing molecular diagnostic tests. Most benign nevi appear to be driven by point mutations in selected oncogenes but only exceptionally rarely show gross chromosomal abnormalities. 1 In contrast, tumor progression from a nevus to Accepted for publication February 6, 2012. Published as an Early Online Release April 5, 2012. From the Department of Pathology, DermatopathologyConsultations.com, Harvard Vanguard Medical Associates, Boston, Massachusetts (Drs Zembowicz and Lyle); the Department of Pathology, Lahey Clinic, Burlington, Massachusetts (Dr Zembowicz); the Department of Pathology and Laboratory Medicine, Tufts Medical Center, Boston (Dr Yang); the Department of Pathology, Hospital General of Serres, Serres, Greece (Dr Kafanas); and the Department of Pathology, University of Massachusetts Medical School, Worcester (Dr Lyle). Dr Zembowicz serves as a clinical consultant to NeoGenomics Laboratories (Irvine, California). In this capacity, he received honoraria for second-opinion consultations on cases submitted for FISH. The authors have no other relevant financial interest in the products or companies described in this article. Reprints: Artur Zembowicz, MD, PhD, Department of Pathology, DermatopathologyConsultations.com, Harvard Vanguard Medical Associates, 6th Floor, 133 Brookline Ave, Boston, MA 02215 (e-mail: dr.z@dermatopathologyconsultations.com). of second opinion and routine dermatopathology practice. Design. A prospective histologic/fish correlation study of 140 cases. Results. Twenty-seven percent of abnormal FISH results were false-positive results because of tetraploidy. After correcting for known false-positive results, all lesions considered atypical nevi showed normal FISH signals. Abnormal FISH signals were reported in 30% of lesions considered histologically borderline and in 48% of lesions in which a diagnosis of was favored. Conclusions. Four-probe, multicolor FISH results for correlate with the microscopic assessments of histologically ambiguous lesions. Pathologists using MelanoSITE must be aware of the high rate of false-positive results from tetraploidy. (Arch Pathol Lab Med. 2012;136:1571 1579; doi: 10.5858/arpa.2011-0673-OA) is associated with chromosomal instability resulting in gains, amplifications, and/or losses of specific chromosomal material, which can be detected by genetic techniques. 2,3 These findings laid the foundation for the development of the first 4-probe fluorescence in situ hybridization (FISH) test for, currently commercialized by Abbott Molecular (Abbott Park, Illinois). This test was a subject of our recent comprehensive review. 4 The FISH probes used in this assay target ras responsive element binding protein 1 (, 6p25), myeloblastosis viral oncogene homolog (MYB, 6q23), Cen6 (centromere 6) and cyclin D1 (CCND1, 11q13) loci. 5 Validation studies performed in academic centers determined basic protocols and signal cutoffs. 5 They demonstrated that 4-probe FISH for can achieve sensitivity and specificity as high as 86.7% and 95.4%, respectively, in distinguishing unequivocal from benign nevi. 6 Several proof-of-principle studies showed potential applications of FISH to solve a variety of diagnostic dilemmas in the evaluation of melanocytic tumors, including differentiating blue nevus-like metastasis from blue nevus, mitotically active nevus from nevoid, and dysplastic nevus from superficial spreading. 7 12 Fluorescence in situ hybridization test abnormalities characteristic of Arch Pathol Lab Med Vol 136, December 2012 FISH in Ambiguous Melanocytic Lesions Zembowicz et al 1571

were identified in lentiginous supporting this entity as a variant of. 12 These studies compared unequivocal benign and malignant counterparts; the signal cutoffs established in the validation studies showed similar test performance in most, but not all, diagnostic situations. For instance, using those cutoffs, FISH test results are abnormal in only 50% of desmoplastic s. 13 Early retrospective studies have shown correlation between the metastatic behavior among histologically ambiguous melanocytic lesions and the FISH results. 5,14 Outside of the United States, the Abbott Molecular FISH assay is available as a diagnostic kit, and the test can be performed by pathologists in-house if the technical expertise and equipment are available. In the United States, at the time of this writing, this FISH test is only performed at academic centers originally involved in the development of the test (University of California, San Francisco; Northwestern University, Chicago, Illinois; and Memorial Sloan-Kettering Cancer Center, New York City, New York). NeoGenomics Laboratories (Irvine, California) is the only commercial laboratory licensed to perform the technical component of the test. However, pathologists can interpret FISH results themselves using a Web-based portal for analysis and reporting FISH results. This service is marketed as MelanoSITE. We began using MelanoSITE in our practices in August 2010, and since then, at the time of this writing, we have examined 140 histologically ambiguous melanocytic lesions. The objective of this article is to present the analysis of this material and to assess our experience with FISH using MelanoSITE critically. MATERIALS AND METHODS The study was approved by the institutional review board at Lahey Clinic, Burlington, Massachusetts. The cases were collected in 3 different practice settings: 89 of the total 140 cases (63%) with histologically ambiguous melanocytic lesions were secondopinion consultations sent to Dr Zembowicz (A.Z. DermatopathologyConsultations.com, Boston, Massachusetts). Some aspects of this practice setting were recently reported. 15 The remaining cases were selected from routine accessions, including 15 cases (11%) from Lahey Clinic (Burlington, Massachusetts), a tertiary academic medical center serving New England rural and suburban communities, and 36 cases (26%) from Harvard Vanguard Medical Associates (Boston), a large multispecialty, outpatient medical practice with a large dermatology practice serving mostly urban and affluent suburban communities. With the exception of 2 lesions that were earlier biopsies of a recurrent lesion and were reexamined retrospectively, FISH was requested as a part of a routine diagnostic workup of cases. The FISH test was initiated by A.Z. or S.R.L. after reviewing all the available, routine histologic material or, in rare cases, was requested by a submitting pathologist. In all cases, preliminary diagnoses were rendered before sending a tissue for FISH analysis. In most cases, the histologic differential diagnosis was between Spitz nevus and spitzoid or between atypical nevus (such as variants of deep penetrating/clonal/inverted type-a nevus, pigmented spindle cell nevus, atypical genital nevus, atypical conjunctival nevus showing cytologic atypia, mitotic activity, limited maturation, or proliferative nodules, among others) and nevoid (Table 1). In some cases, diagnostic considerations included dysplastic nevus and superficial spreading. In 2 cases, FISH was performed to confirm a suspected but histologically challenging diagnosis of nonsarcomatoid desmoplastic. At the time of preliminary diagnosis, based on histologic features, each lesion was designated as likely representing an atypical nevus (favoring a benign type), a truly borderline lesion, or a. These designations reflect the most likely interpretation if FISH was not performed. After FISH was interpreted, the lesions were reexamined, and the final diagnoses were rendered, taking into account all available information, including the FISH results. Fluorescence In Situ Hybridization Fluorescence in situ hybridization was performed at NeoGenomics Laboratories as described in the company s materials and a recent report. 16 Briefly, when FISH was considered, 8 additional, 5-mm sections were cut. Two of those sections were stained with hematoxylin-eosin. The area of tumor to analyze was selected on the stained sections and sent together with the unstained sections to NeoGenomics Laboratories for analysis. After hybridization, the results were analyzed and quantified using MetaSystems (Boston) virtual slide imaging system. Each case was subjected to a quality-control process involving review of the data by a second NeoGenomics Laboratories technician. After quality-control signoff, data from the automatic data analysis and a panel of 4 nuclei selected by NeoGenomics Laboratories technicians were presented to the signing-out pathologist via the MelanoSITE portal. In cases showing abnormal or borderline results, and in cases where there was a discrepancy between FISH results and preliminary histologic interpretation, the original FISH data and images were reviewed by A.Z. In some cases, original FISH slides were also examined with a fluorescence microscope. Per NeoGenomics Laboratories recommendations, 16 FISH was considered abnormal if at least one of the following criteria was met: (1) more than 16% of the nuclei showed more than 2 signals for (red probe), (2) the ratio of red signal () to aqua signal (centromere 6) greater than 1 was observed in more than 53% of the nuclei, (3) the ratio of yellow signal (MYB) to aqua signal (centromere 6) less than 1 was observed in more than 42% of the nuclei, and (4) more than 19% of the nuclei showed more than 2 signals for CCND1 (green probe). With these cutoffs, MelanoSITE achieved 83.8% sensitivity and 98.1% specificity for detection of unequivocal in validation studies. 16 Tetraploidy The 4-probe FISH does not allow for definitive assessment of tetraploidy in cells. However, as recently discussed, 4,6,17 increased signals in all 4 probes are highly suggestive of balanced chromosomal duplications. Therefore, all cases with abnormal FISH were scrutinized for the presence of nuclei with increased FISH signals with all 4 probes. Using previously determined guidelines, 6,17 cells showing 3 or 4 copies of the targeted chromosomal loci with probes 6p25 (), 6q23 (MYB), and 11q13 (CCND1) were considered to likely be tetraploid and were rejected from the numeric analysis. Observation of 3 signals is sufficient because signals off the plane of section are often not visualized, particularly in large, spitzoid or epithelioid cells. Statistical Analysis Statistical significance between variables was assessed by x 2 test with a post hoc Cramer V coefficient test to determine the strength of association. The Cramer V coefficient test result is close to 0 when there is little association between the variables; values close to 1 indicate the highest association. Values greater than.1 are generally thought a good minimum threshold for suggesting there is a substantive relationship between two variables. RESULTS Use of Melanoma FISH in Different Practice Settings The 89 second-opinion consultations sent for FISH represented 2.3% of all cases reported at DermatopathologyConsultations.com during the period of the study. The test was used even less in the routine practice settings. The FISH test was deemed potentially helpful in 0.23% of cases examined at an outpatient dermatopathology prac- 1572 Arch Pathol Lab Med Vol 136, December 2012 FISH in Ambiguous Melanocytic Lesions Zembowicz et al

Histologic Category/ Subset Cases, No. (% of Total) Table 1. Cases, No. (% of Diagnostic Category) tice at Harvard Vanguard Medical Associates and in only 0.1% of cases seen at Lahey Clinic, a tertiary medical center dermatology practice. Histopathology/FISH Correlation Out of the 140 cases examined, at the time of preliminary diagnosis, a benign lesion (atypical nevus) was favored in 66 cases (47%). Malignant was favored in 27 cases (19%). In 47 cases (34%), a lesion was considered truly borderline. In the histologically benign (atypical nevus) group, the initial MelanoSITE results were normal in 59 cases (89%). A representative lesion with a normal FISH test result is illustrated in Figure 1, A through D. MelanoSITE results were abnormal in 7 cases (11%). On manual rereview, 5 of the 7 cases (71%) were determined to be false-positive results because of tetraploidy, and 2 cases (29%) were rejected because of poor hybridization and/or the small number of nuclei enumerated. Thus, in none of the Histologic/Fluorescence In Situ Hybridization (FISH) Correlation a Abnormal MelanoSITE. CEN MelanoSITE Results, No. (%) MYB, CEN CCND1 Multiple Tetraploidy Final FISH Interpretation Spitzoid tumor 51 (36) 14 (28) 4 (1) 1 (1) 8 (2) 5 (4) 5 4 (29) 10 (19) 18 (35) 2 (11) 1 (1) 0 (0) 0 (0) 2 (1) 1 2 (100) 0 (0) Borderline 28 (55) 8 (29) 3 (1) 0 (0) 5 (0) 4 (2) 2 1 (12) 8 (29) Favor 5 (10) 4 (80) 1 (1) 1 (1) 3 (2) 1 (1) 2 1 (25) 3 (60) Atypical nevus versus nevoid 54 (39) 14 (26) 3 (2) 1 (1) 3 (0) 3 (1) 4 5 (36) 9 (17) 32 (59) 4 b (13) 2 (1) 0 (0) 2 (0) 1 (1) 1 2 (50) b 0 b (0) Borderline 15 (28) 8 (53) 4 (2) 0 (0) 0 (0) 5 (2) 2 2 (20) 6 (40) Favor 7 (13) 2 (29) 1 (1) 1 (1) 1 (0) 0 (0) 1 0 (0) 2 (29) Dysplastic nevus versus 33 (24) 10 (30) 8 (4) 3 (3) 0 (0) 2 (1) 4 2 (20) 8 (24) 16 (49) 1 (6) 1 (1) 0 (0) 0 (0) 1 (1) 1 1 (100) 0 (0) Borderline 4 (12) 2 (50) 2 (0) 0 (0) 0 (0) 0 (0) 0 1 (25) 1 (25) Favor 13 (39) 7 (54) 5 (3) 3 (3) 0 (0) 1 (0) 3 0 (0) 7 (54) Suspected 2 (1) 1 (50) 0 (0) 0 (0) 1 (0) 0 (0) 0 0 (0) 1 (50) desmoplastic Total 140 (100) 39 (28) 15 (7) 5 (5) 12 (2) 10 (6) 13 10 (26) 29 (21) a The table summarizes MelanoSITE FISH results for each histologic diagnostic category (Spitzoid tumor, atypical nevus versus nevoid, dysplastic nevus versus, and suspected desmoplastic ) and their subsets (favor benign, borderline, and favor ). The columns indicate the following: number, number of lesions in a particular diagnostic category or subset, with percentages of total or corresponding diagnostic category shown in parentheses; abnormal MelanoSITE: number of lesions reported by MelanoSITE as abnormal, with percentages of corresponding diagnostic category/subset or total shown in parentheses;, number of lesions with abnormal MelanoSITE based on FISH criterion 1, which is the percentage of cells showing.2 signals with probe equaling.16%, with number of lesions in which this abnormality was associated with other abnormalities shown in parentheses;.cen, number of lesions with abnormal MelanoSITE based on FISH criterion 2, which is the percentage of cells showing more signals with than with centromere 6 probes equaling.53%, with the number of lesions in which this abnormality was associated with other abnormalities shown in parentheses; MYB,CEN, number of lesions with abnormal MelanoSITE based on FISH criterion 3, which is the percentage of cells showing fewer signals with myeloblastosis viral oncogene homologue (MYB) equaling.42%, with number of lesions in which this abnormality was associated with other abnormalities is shown in parentheses; CCND1, number of lesions with abnormal MelanoSITE based on FISH criterion 4, which is the percentage of cells showing.2 signals with cyclin D1 (CCND1) equaling.19%, with number of lesions in which this abnormality was associated with other abnormalities shown in parentheses; multiple, number of lesions with abnormal MelanoSITE based on.1 FISH criterion; tetraploidy, number of lesions with false-positive MelanoSITE because of tetraploidy, with the percentage of lesions in reference to the number of lesions with abnormal initial MelanoSITE shown in parentheses; final FISH interpretation, number of lesions with abnormal FISH after correcting for false-positives because of tetraploidy or being rejected because of poor hybridization or low number of nuclei counted during retrospective analysis. b There were 2 cases in this category that were found to be false-positive because of poor hybridization or the numbers of nuclei were too low to be enumerated. histologically benign lesions (0%) was the final FISH interpretation reported as an abnormal result. In the histologically borderline group, the MelanoSITE results were initially reported as abnormal in 18 of the 47 cases (38%). Of those 18 lesions, 4 (22%) showed features consistent with tetraploidy, so the percentage of cases in this group with an abnormal final FISH interpretation was 30% (14 of 47). After post-fish review, 10 of those 18 cases (56%) were reported as likely malignant. The remaining 4 of the 18 cases (22%) with an initial abnormal FISH result were interpreted as borderline melanocytic tumors of uncertain malignant potential with abnormal FISH test results. Preliminary diagnosis was not changed in cases with normal FISH results or with false-positive FISH results. An example of a borderline, atypical Spitz tumor with a false-positive FISH test result because of tetraploidy is illustrated in Figure 2, A through D. An example of an atypical conjunctival nevus with a falsepositive FISH result is illustrated in Figure 3, A through D. Arch Pathol Lab Med Vol 136, December 2012 FISH in Ambiguous Melanocytic Lesions Zembowicz et al 1573

Figure 1. A C, A second-opinion consultation of a skin-colored lesion from a 26-year-old woman submitted for evaluation of possible. D, A representative composite image of the nucleus showing the normal 2 signals with target ras responsive element binding protein 1 (, red), cyclin D1 (CCND1, green), myeloblastosis viral oncogene homologue (MYB, yellow) and centromere 6 (aqua) probes. In this case, preliminary diagnosis favored a benign Spitz nevus (hematoxylin-eosin, original magnifications 340 [A], 3100 [B], and 3200 [C]; fluorescence in situ hybridization, original magnification 3400 [D]). An example of an atypical Spitz tumor with an abnormal FISH result is illustrated in Figure 4, A through D. When the preliminary report favored a malignant diagnosis, the MelanoSITE result was abnormal in 14 of the 27 cases (52%). Only 1 of those 14 lesions (7%) was considered a likely false-positive from tetraploidy. That lesion was discussed and illustrated in a recent review on FISH.4(figure 2) Thus, after correcting for the false-positive case, the final FISH result was interpreted as abnormal in 13 of the 27 cases (48%). A representative case of nevoid with an abnormal FISH test result is illustrated in Figure 5, A through D. We did not change a diagnosis based on a negative FISH result for any of the lesions. A 2 3 3-way x2 analysis revealed a statistically significant difference (P,.001) between FISH results (normal versus abnormal) and a designation of the lesion into one of the 3 diagnostic categories (1, favor benign diagnosis; 2, favor borderline diagnosis; 3, favor a diagnosis of ) based on histologic examination before the FISH test was performed. In addition, a Cramer V posttest result, a measure of the strength of an association, showed a value of.49, indicating strong association. 1574 Arch Pathol Lab Med Vol 136, December 2012 Table 1 provides detailed histopathologic/fish correlation data. For each major diagnostic category of histologically ambiguous lesion (spitzoid tumors, lesions raising a differential diagnosis of an atypical nevus versus nevoid, and dysplastic nevus versus superficial spreading ), Table 1 shows the initial MelanoSITE results and the final FISH result interpretation after correcting for false-positive results. The number of lesions in different subgroups is too small for meaningful, comprehensive statistical analysis of the subsets. However, certain trends can be noted. MYB1 loss appears particularly common in spitzoid neoplasms and is rare in dysplastic nevi, whereas the opposite seems to be true for. False-Positive Results From Tetraploidy Out of 39 cases with abnormal, initial MelanoSITE results, 2 (5%) were rejected because of the poor quality of the hybridization and few enumerated nuclei. In retrospect, those cases should have been rejected at the technical quality-control step. Among the remaining 37 cases, 10 (27%) were considered to be falsepositive results because of tetraploidy. Table 1 shows FISH in Ambiguous Melanocytic Lesions Zembowicz et al

Figure 2. A C, Biopsy of a pigmented lesion from the back of a 24-year-old woman. D, A representative composite image of the nucleus showing 4 signals with ras responsive element binding protein 1 (, red), 2 signals with cyclin D1 (CCND1, green), 4 signals with myeloblastosis viral oncogene homologue (MYB, yellow), and 4 signals with centromere 6 (aqua) probes. The pattern of signal increases with all 4 probes is consistent with tetraploidy. Seeing fewer than 4 signals in a tetraploid nucleus is normal because some fluorescence in situ hybridization (FISH) probe-binding sites are expected to be out of the plane of the section. Preliminary diagnosis favored an atypical Spitz tumor, a borderline tumor (hematoxylin-eosin, original magnifications 340 [A], 3200 [B], and 3400 [C]; FISH, original magnification 3400 [D]. the number of tetraploid cases in different histologic diagnostic categories. Table 2 shows the details of those cases. DISCUSSION Histologic diagnosis of melanocytic proliferations is inherently difficult because it requires integration of multiple architectural and cytologic criteria. No single feature is diagnostic of a particular entity. Furthermore, similar features can be found in both benign nevi and s. Therefore, in everyday practice, even experienced pathologists cannot reproducibly classify a small subset of melanocytic proliferations into established nosologic categories and, in some cases, may not be able to accurately predict prospectively the biologic behavior of the tumor. Such difficult lesions are often reported descriptively as a melanocytic tumor of uncertain malignant potential, severely atypical melanocytic proliferation, borderline melanocytic tumor, nevomelanocytic tumors of undetermined risk, or other similar terms. We have recently reviewed the problem of a difficult melanocytic lesion. 18 As discussed in the introduction and in our recent review, 4 the Abbott Molecular 4-probe FISH test targeting ras responsive element binding protein 1 (, 6p25), myeloblastosis viral oncogene homologue (MYB, 6q23), centromere 6, and cyclin D1 (CCND1, 11q13) probes 5 is the first commercially available FISH probe set showing promise as a diagnostic adjunct in evaluation of melanocytic lesions. In validation studies using diagnostically unequivocal teaching sets, it discriminated between benign nevi and malignant with sensitivity and specificity as high as 86.7% and 95.4% (University of California, San Francisco/Northwestern University data) 6 and 83.8% and 98.11% (MelanoSITE). 16 The Abbott Molecular FISH test was also validated in critical differential diagnoses concerning histologically unequivocal benign and malignant blue nevi, nevoid, dysplastic nevi, desmoplastic, lentiginous, nevi with atypical epithelioid cell component, and Spitz nevi. 6 12,17,19 Obviously, a FISH test is not needed for histologically unequivocal cases. The only current justification for use of Arch Pathol Lab Med Vol 136, December 2012 FISH in Ambiguous Melanocytic Lesions Zembowicz et al 1575

Figure 3. A C, Biopsy of a conjunctival pigmented lesion from a 9-year-old boy. D, A representative composite image of the nucleus of a large cell showing the normal 3 signals with ras responsive element binding protein 1 (, red), 4 signals with cyclin D1 (CCND1, green), 3 signals with myeloblastosis viral oncogene homologue (MYB, yellow), and 3 signals with centromere 6 (aqua) probes. The pattern of signal increases with all 4 probes is consistent with tetraploidy. Seeing fewer than 4 signals in a tetraploid nucleus is normal because some fluorescence in situ hybridization (FISH) probe-binding sites are expected to be out of the plane of the section. Preliminary diagnosis favored benign juvenile conjunctival nevus22 (hematoxylin-eosin, original magnifications 340 [A], 3100 [B], and 3200 [C]; FISH, original magnification 3400 [D]). this expensive test in a diagnostic setting is to employ it as a diagnostic adjunct to help classify histologically ambiguous lesions. Early retrospective studies showed correlation between metastatic behavior among histologically ambiguous melanocytic lesions and FISH results.5,14,20 However, applying data generated from FISH validation studies to routine diagnosis of challenging lesions is not straightforward. By definition, there are no teaching sets of ambiguous melanocytic lesions. Therefore, it is impossible to establish signal cutoffs and sensitivity/specificity profiles for the FISH assay in this critical application as precisely as it was possible in validation studies of unequivocal lesions. Unfortunately, at this point, progress in this area has to be made by learning from histopathologic/fish correlation studies similar to the one reported here. We present our experience with 140 cases of histologically ambiguous melanocytic lesions evaluated with a FISH test in a setting of second-opinion consultation and routine outpatient and hospital-based practice. At the time of this writing, this is, to our knowledge, the first independent study evaluating MelanoSITE. The data reported here were gathered prospectively during diagnostic workup of diagnostically challenging 1576 Arch Pathol Lab Med Vol 136, December 2012 lesions. The FISH test was performed only on lesions that were considered severely atypical and/or diagnostically ambiguous by the senior author (A.Z.), S.R.L., or a submitting pathologist. We found that only a small percentage of cases (2.3% in the second-opinion consultation practice and 0.1 to 0.2% in the routine practice) may benefit from an additional FISH test. Our data indicate that when applied to histologically ambiguous lesions, MelanoSITE has a high rate of falsepositive results, which can be attributed to mischaracterization of tetraploid cells. This critical issue is discussed below. However, after correcting for known false-positives, we observed good correlation between histologic classification and FISH findings. None of the lesions thought to represent atypical nevus showed abnormal FISH results, whereas 48% of lesions favored to represent had abnormal FISH results. Consistently, FISH produced abnormal results in 30% of histologically borderline lesions. The lack of FISH-positive results among atypical nevi can be interpreted as validating the current morphologic criteria used to identify atypical nevi. Although experts with substantial experience will rarely have to rely on a FISH in Ambiguous Melanocytic Lesions Zembowicz et al

Figure 4. A C, Atypical nevus from the lower back of a 56-year-old man. D, A representative composite image of the nucleus showing 1 signal with myeloblastosis viral oncogene homologue (MYB, yellow) probe and 2 signals with centromere 6 (aqua) probe. Fluorescence in situ hybridization (FISH) signals are consistent with the loss of the MYB locus. In this case, preliminary diagnosis was atypical Spitz tumor, a borderline lesion (hematoxylin-eosin, original magnifications 340 [A], 3100 [B], and 3200 [C]; FISH, original magnification 3400 [D]). FISH test in this setting, we find that a normal FISH result can provide additional reassurance in challenging lesions. The percentage (48%) of FISH-positive histologically ambiguous lesions favored to represent unusual was close to the lower range (50% 88%) of that reported for conventional.6 12,17,19 This is reassuring regarding the performance of MelanoSITE. We believe that, in this context, abnormal FISH results clinch the diagnosis, and the lesion can be reported as an outright. However, the diagnostic importance of a negative FISH result is limited because a significant portion of the unequivocal s (12% 50%) are expected to produce normal FISH results. Thus, we did not change an interpretation based on a normal FISH result in any of our cases. The FISH test most affected the final interpretation of lesions in the borderline group. However, the best practice for the use of the FISH test in this setting is the most controversial because the optimal signal cutoffs in these cases of borderline melanocytic lesions cannot be established. We reclassified 10 lesions with abnormal FISH results as likely. Yet, we felt that in 4 lesions, histologic features were insufficiently atypical to warrant Arch Pathol Lab Med Vol 136, December 2012 an outright diagnosis of. Those cases were reported as atypical melanocytic tumor of uncertain malignant potential with abnormal FISH test results. This conservative approach reflects the authors philosophy that we should remain skeptical about this emerging test and use it only as an adjunct to, but not in lieu of, timetested histologic interpretations. This is particularly true in the context of a test performed at an outside laboratory when the pathologist is not able to examine the entire slide firsthand to assess qualitative features, such as the consistency of the FISH findings in distinct areas, and must rely solely on automated enumeration. We focused on using FISH as a diagnostic adjunct; however, recent evidence also indicates that, in conventional, the frequency of abnormal FISH results correlates with prognosis.21 Consistently, FISH results are abnormal in almost all metastatic s and aggressive nodular s and in only 50% of more-indolent desmoplastic. Thus, it may be reasonable to use FISH results as a prognostic indicator rather than a diagnostic tool to better predict the risk of malignant behavior of histologically ambiguous lesions, without necessarily reclassifying them. We believe that analysis of our cohort after additional years of follow-up will likely provide us FISH in Ambiguous Melanocytic Lesions Zembowicz et al 1577

Figure 5. A C, A recurrent, pigmented lesion from the back of 59-year-old woman, showing abnormal fluorescence in situ hybridization (FISH) test results. D, A representative composite image of the nucleus showing 3 signals with ras responsive element binding protein 1 (, red) and 2 signals with centromere 6 (aqua) probes. The numbers (2, 3, 2, 2) indicate the number of signals with cyclin D1 (green), (red), myeloblastosis viral oncogene homologue (MYB, yellow), and centromere 6 (aqua) probes. The FISH signals are consistent with amplification of the locus, which is consistent with preliminary diagnosis in this case that favored nevoid (hematoxylin-eosin, original magnifications 340 [A], 3100 [B], and 3200 [C]; FISH, original magnification 3400 [D]). with a better understanding of the role of a FISH result in the context of histologically ambiguous lesions. We found an unacceptably high (26%) rate of falsepositive results because of tetraploidy. Pathologists must Table 2. Case No. 1 2 3 4 5 6 7 8 9 10 1578 be aware of this phenomenon to avoid overdiagnosing benign nevi as based on FISH results. We were able to avoid this diagnostic pitfall only because of our emphasis on clinicopathologic correlation and because we Details of False-Positive Cases Because of Tetraploidy Histologic Category/Subset Spitzoid tumor Borderline Favor Atypical nevus versus nevoid Borderline Borderline Dysplastic nevus versus Borderline Arch Pathol Lab Med Vol 136, December 2012 Age, y/sex Site MelanoSITE Abnormality 25/F 31/F 21/F 17/F Back Leg Neck Ankle and CCND1 CCND1 CCND1 and CCND1 9/M 3/M 19/M 48/F Conjunctiva Ear Neck Nose and CCND1 and CCND1 21/M 50/F Conjunctiva Back and CCND1 FISH in Ambiguous Melanocytic Lesions Zembowicz et al

had access to the original FISH data, which enabled us to manually reevaluate the cases, to reject tetraploid cells, and to recalculate the results. The problem of tetraploidy in FISH results was first noted by Dr Gerami s group, 6,17 who reported 10% and 14% tetraploidy rates in Spitz nevi and combined nevi with an epithelioid cell component, respectively. The false-positive rate because of tetraploidy was higher in our study. We suspect that this may be due in part to lower signal cutoffs for (19%, criterion 1) and CCND1 (16%, criterion 2) used for MelanoSITE compared with those used by Gerami et al 6,17 (29% and 38%, respectively). Because of lower cutoffs, fewer tetraploid cells were needed to affect FISH signal counts sufficiently enough to satisfy MelanoSITE criteria 1 and 4. This should not be interpreted as a basis to advocate that pathologists should apply the Northwestern/University of California, San Francisco, cutoffs when using MelanoSITE. The signal cutoffs are empirically determined values and are expected to differ among laboratories because of slightly different hybridization conditions and, most important, manual (Northwestern University/ University of California) versus automatic (NeoGenomics Laboratories) method of selection of nuclei and analysis. 4 Overdiagnosing based on a false-positive FISH result would be a consequential mistake for the patient. Therefore, pathologists using the FISH test must be aware of this phenomenon. Tetraploid cells can be found in benign, borderline, and malignant lesions, so this feature is not useful diagnostically. A significant number of tetraploid cells simply invalidate FISH test results. MelanoSITE should include a process to alert pathologists to the possibility of tetraploidy and to give pathologists the tools to make appropriate assessment and/or to develop algorithms that would identify and reject suspected tetraploid cells during automated analysis. In addition, readers should be aware that there were 2 cases in which either too few nuclei were counted or poor hybridization was detected in our study, situations that may also lead to false-positive results. Although we believe that the MelanoSITE approach is a significant innovation because it allows pathologists not having access to FISH-capable laboratories to offer the test, we have several suggestions for improvement. Pathologists using MelanoSITE rely only on automated enumeration supervised by technicians from NeoGenomics Laboratories, without the ability to examine the actual FISH test slide or a digitized version thereof. In several controversial cases, we requested slides be sent back to us, and we examined them ourselves under a fluorescence microscope. We found out that the ability to examine the whole slide enhances the histologic/fish correlation and allowed us to assess the consistency of the FISH result abnormalities in different areas of the tumor. Obviously, having the ability to superimpose FISH slides with routine sections would be a plus, but just having the ability to examine the digital images of fluorescence slides would be a great improvement to the interface. In summary, our data show a correlation between the histologic classification of histologically ambiguous melanocytic lesions and MelanoSITE FISH results. Our experience indicates that MelanoSITE results can be used judiciously as an adjunct to the evaluation of challenging melanocytic lesions by experts. However, low sensitivity of the FISH test and the problem of falsepositives because of tetraploidy underscore the importance of clinicopathologic correlation and expertise. Thus, we are skeptical that the FISH test is ready for widespread use for the evaluation of challenging melanocytic lesions in a general community-practice setting. We acknowledge the excellent technical assistance of Patrice Green, BA (DermatopathologyConsultations.com). Reference 1. Bastian BC, Olshen AB, LeBoit PE, Pinkel D. Classifying melanocytic tumors based on DNA copy number changes. Am J Pathol. 2003;163(5):1765 1770. 2. Albertson DG, Collins C, McCormick F, Gray JW. Chromosome aberrations in solid tumors. Nat Genet. 2003;34(4):369 376. 3. Bastian BC. Understanding the progression of melanocytic neoplasia using genomic analysis: from fields to cancer. Oncogene. 2003 May;22(20):3081 3086. 4. Gerami P, Zembowicz A. Update on fluorescence in situ hybridization in : state of the art. Arch Pathol Lab Med. 2011;135(7):830 837. 5. Gerami P, Jewell SS, Morrison LE, et al. Fluorescence in situ hybridization (FISH) as an ancillary diagnostic tool in the diagnosis of [Erratum appears in Am J Surg Pathol. 2010;34(5):688]. Am J Surg Pathol. 2009;33(8):1146 1156. 6. Isaac AK, Lertsburapa T, Pathria MJ, Martini M, Guitart J, Gerami P. Polyploidy in Spitz nevi: a not uncommon karyotypic abnormality identifiable by fluorescence in situ hybridization. Am J Dermatopathol. 2010;32(2):144 148. 7. Bastian BC. Understanding the progression of melanocytic neoplasia using genomic analysis: from fields to cancer [Review] [36 refs]. Oncogene. 2003; 22(20):3081 3086. 8. Pouryazdanparast P, Newman M, Mafee M, Haghighat Z, Guitart J, Gerami P. Distinguishing epithelioid blue nevus from blue nevus like cutaneous metastasis using fluorescence in situ hybridization. Am J Surg Pathol. 2009;33(9):1396 1400. 9. Newman MD, Lertsburapa T, Mirzabeigi M, Mafee M, Guitart J, Gerami P. Fluorescence in situ hybridization as a tool for microstaging in malignant. Mod Pathol. 2009;22(8):989 995. 10. Gerami P, Wass A, Mafee M, Fang Y, Pulitzer MP, Busam KJ. Fluorescence in situ hybridization for distinguishing nevoid s from mitotically active nevi. Am J Surg Pathol. 2009;33(12):1783 1788. 11. Gerami P, Barnhill RL, Beilfuss BA, LeBoit P, Schneider P, Guitart J. Superficial melanocytic neoplasms with pagetoid melanocytosis: a study of interobserver concordance and correlation with FISH. Am J Surg Pathol. 2010; 34(6):816 821. 12. Newman MD, Mirzabeigi M, Gerami P. Chromosomal copy number changes supporting the classification of lentiginous junctional of the elderly as a subtype of. Mod Pathol. 2009;22(9):1258 1262. 13. Gerami P, Beilfuss B, Haghighat Z, Fang Y, Jhanwar S, Busam KJ. Fluorescence in situ hybridization as an ancillary method for the distinction of desmoplastic s from sclerosing melanocytic nevi. J Cutan Pathol. 2011;38(4):239 334. 14. Vergier B, Prochazkowa-Carlotti M, de la Fouchardière A, et al. Fluorescence in situ hybridization, a diagnostic aid in ambiguous melanocytic tumors: European study of 113 cases. Mod Pathol. 2011;24(5):613 623. 15. Zembowicz A, Ahmad A, Lyle SR. A comprehensive analysis of a Webbased dermatopathology second opinion consultation practice. Arch Pathol Lab Med. 2011;135(3):379 383. 16. Moore MW, Gasparini R. FISH as an effective diagnostic tool for the management of challenging melanocytic lesions. Diagn Pathol. 2011;6:76. doi:10.1186/1746-1596-6-76. 17. Pouryazdanparast P, Haghighat Z, Beilfuss BA, Guitart J, Gerami P. Melanocytic nevi with an atypical epithelioid cell component: clinical, histopathologic, and fluorescence in situ hybridization findings. Am J Surg Pathol. 2011;35(9):1405 1412. 18. Zembowicz A, Scolyer RA. Nevus/melanocytoma/: an emerging paradigm in classification of melanocytic neoplasms? Arch Pathol Lab Med. 2011,135:300 306. 19. Pouryazdanparast P, Newman M, Mafee M, Guitart J, Gerami P. Malignant with monster cells showing massive cyclin D1 amplification. Am J Dermatopathol. 2009;31(4):402 403. 20. Massi D, Cesinaro AM, Tomasini C, et al. Atypical Spitzoid melanocytic tumors: a morphological, mutational, and FISH analysis. J Am Acad Dermatol. 2011;64(5):919 935. 21. North JP, Vetto JT, Murali R, White KP, White CR, Jr, Bastian BC. Assessment of copy number status of chromosomes 6 and 11 by FISH provides independent prognostic information in primary. Am J Surg Pathol. 2011 9;35(8):1146 1150. 22. Thiagalingam S, Johnson MM, Colby KA, Zembowicz A. Juvenile conjunctival nevus: clinicopathologic analysis of 33 cases. Am J Surg Pathol. 2008;32(3):399 406. Arch Pathol Lab Med Vol 136, December 2012 FISH in Ambiguous Melanocytic Lesions Zembowicz et al 1579