To the Editor:
Sry-related High Mobility Group-box gene 10 (Sox10) is a member of the high-mobility-group domain family of transcription factors, which are widely expressed during embryonic development and in the adult organism.1 Sox10 is expressed in embryonic neural crest cells and plays a critical role in the differentiation of melanocytes.1–3 It is also expressed in melanocytes in the adult, as well as Schwann cells, myoepithelial cells, and oligodendroglial cells.2–4 Neoplasms derived from these lineages, including melanoma, are also positive.5 In recent years Sox10 has emerged as a diagnostically useful marker for melanoma, particularly desmoplastic melanoma which is usually negative for other markers such as Melan-A or HMB-45.6 Several recent studies have suggested that Sox10 may be of particular value in the distinction of desmoplastic melanoma from an excision scar, reporting no significant staining of fibroblasts or other spindled cells within the scar tissue.7,8
However, despite these reported staining patterns, we have anecdotally noted variable numbers of spindled cells within scar tissue which have stained positively for Sox10. As we have noted these in re-excision specimens from both melanocytic and nonmelanocytic lesions, we strongly suspected that they represented benign spindled cells. To confirm and quantitate our observation, we used digital image analysis to determine the extent of staining within scar tissue using 3 different Sox10 antibodies and compare it with that seen in examples of desmoplastic melanoma.
Nine examples of cutaneous scars from nonmelanocytic re-excisions were collected prospectively, and immunohistochemically stained for Sox10 using 3 different antibodies. The details of the antibodies, antigen retrieval method, and primary antibody incubation are outlined in Table 1. All immunohistochemical staining was performed using the Ventana Benchmark Ultra staining system with Optiview detection (Roche). For comparison, 6 examples of desmoplastic melanoma were retrieved from the archives of PathWest, QEII Medical Centre. These cases had all been previously staining for Sox10 as part of the routine diagnostic workup, and these archival immunohistochemical sections were used for this study. The diagnosis was confirmed by 2 dermatopathologists (NTH and BAW). They had all been stained using the same automated system, using both antibodies 1 and 2.
The immunohistochemically stained sections were scanned using Aperio ScanScope AT. The number of positively staining nuclei was then assessed digitally using Aperio ImageScope software and the Aperio IHC Nuclear Algorithm, as described previously.9 Briefly, areas of scarring or desmoplastic melanoma were selected for analysis, using the corresponding hematoxylin and eosin–stained sections as a guide. Nerves, follicles, and eccrine glands were manually removed from the area selected for analysis. For this study, only nuclei recorded as showing 2+ or 3+ positivity by the software were considered to be positive. Up to 4 selected areas were assessed per case. The total number of positive nuclei was then divided by the total area assessed (as measured by the imaging software) to produce the number of positive cells per square millimeter.
The results from the 3 different antibodies were compared using analysis of variation. The comparison between scars and desmoplastic melanoma was performed using a Student's t test. Statistical analysis was performed using a publicly available online statistical calculator (www.vasserstats.net).
The clinical details and Sox10 immunohistochemical staining results for the scars are summarized in Table 2. All these specimens represent re-excisions from nonmelanocytic skin malignancies such as squamous cell carcinoma and basal cell carcinoma. Most had originated from the head and neck area. Sox10-positive cells were identified within all scars, with the number of cells identified by digital imaging ranging from 2 to 117 cells per square millimeter. Two examples of scars, along with the Sox10-positive cells within them, are illustrated in Figures 1A–D. None of these cases had any morphological evidence of unexpected desmoplastic melanoma. Sox10-positive cells were seen with all 3 antibodies, and no statistically significant difference was identified between the number of positive cells identified between the 3 antibodies (P = 0.58). The mean number of positive cells within scar, across all cases and antibodies, was 41 cells per square millimeter.
As expected, the desmoplastic melanomas showed widespread, strong, nuclear Sox10 positivity. An example is illustrated in Figures 1E, F. The average number of positive cells within an area of desmoplastic melanoma was 1292 cells per square millimeter. To enable a valid statistical comparison between the numbers of positive cells within scar to within melanoma, we used the scar results using antibody 2 only (an average of 44 positive cells per square millimeter). This comparison revealed that the difference was statistically significant (Table 3, P < 0.001).
In the setting of a re-excision for melanoma, the distinction of fibroblasts (and other benign spindled cells) within scar tissue from residual desmoplastic melanoma can be challenging and can have significant clinical implications. While S100 immunohistochemistry is potentially useful in this scenario, the presence of S100-positive spindled cells is well recognized and is potentially confusing.10 Several recent studies have suggested that Sox10 may be more useful in these cases.7,8 Ramos-Herbeth et al8 examined the pattern of Sox10 immunohistochemistry in 26 melanomas or atypical melanocytic lesion excisions. Although melanoma cells exhibited strong Sox10 expression, spindled fibroblasts within the scar tissue were described as showing weak staining, with no staining reported in epithelioid histiocytes. Plaza et al examined 40 cases of desmoplastic melanoma and confirmed positive staining for desmoplastic melanoma in all cases. They also examined 24 scars as a control population, reporting negative staining for Sox10 in all of those cases.7
In contrast to these studies, here, we have demonstrated that Sox10-positive spindled cells are seen within scar tissue. There is significant variability in the numbers of positive cells seen; however, all cases which we assessed showed at least some positive cells. Indeed, in most cases, the number of cells could be sufficient to cause some level of diagnostic concern for an occult melanocytic proliferation, particularly in the context of the reports referenced above. The cause for the discrepancy between our findings and those reported by others is not clear. We considered that there may be differences associated with different antibodies; however, a comparison of 3 different commercially available products (including both polyclonal and monoclonal examples) revealed no difference in the staining pattern within scar tissue. One can only assume that there is a technical difference in the immunohistochemical methodology between the respective laboratories. In this regard, we note that both of the aforementioned studies used a different antibody to that used in this study.7,8
Delineating the exact nature of the Sox10-positive cells within the scar tissue was beyond the scope of this study; however, some inferences can be drawn from previous work. In particular, Trejo et al demonstrated the presence of spindled cells within scar tissue, a proportion of which expressed markers of early Schwann cell differentiation.11 They propose that proliferation of such cells might form part of the normal wound-healing process. It seems reasonable to propose that at least a proportion of the Sox10-positive cells which we have observed may represent similar Schwann cell precursors.
Despite our finding that Sox10-positive cells can be identified within scar tissue, our anecdotal experience was that the magnitude of this staining was far less than that seen within bona fide desmoplastic melanoma. Indeed, the digital imaging analysis confirmed that there is a significantly greater number of cells showing Sox10 positivity with desmoplastic melanomas than that in scars (Table 3, and compare Fig. 1F with Figs. 1B, D). Thus, we would argue that it remains a useful marker to detect residual desmoplastic melanomas in re-excision specimens, so long as the reporting pathologist is aware of the potential for smaller numbers of spindle cells to stain positively within otherwise benign scar tissue.
The authors thank Dr Carla Thomas for assistance with the scanning of slides and digital image analysis.
1. Wegner M. Secrets to a healthy Sox life: lessons for melanocytes. Pigment Cell Res. 2005;18:74–85.
2. Mollaaghababa R, Pavan WJ. The importance of having your SOX on: role of SOX10 in the development of neural crest-derived melanocytes and glia. Oncogene. 2003;22:3024–3034.
3. Bondurand N, Kobetz A, Pingault V, et al. Expression of the SOX10 gene during human development. FEBS Lett. 1998;432:168–172.
4. Mohamed A, Gonzalez RS, Lawson D, et al. SOX10 expression in malignant melanoma, carcinoma, and normal tissues. Appl Immunohistochem Mol Morphol. 2013;21:506–510.
5. Tacha D, Qi W, Ra S, et al. A newly developed mouse monoclonal SOX10 antibody is a highly sensitive and specific marker for malignant melanoma, including spindle cell and desmoplastic melanomas. Arch Pathol Lab Med. 2015;139:530–536.
6. Wood BA. Desmoplastic melanoma: recent advances and persisting challenges. Pathology. 2013;45:453–463.
7. Plaza JA, Bonneau P, Prieto V, et al. Desmoplastic melanoma: an updated immunohistochemical analysis of 40 cases with a proposal for an additional panel of stains for diagnosis. J Cutan Pathol. 2016;43:313–323.
8. Ramos-Herberth FI, Karamchandani J, Kim J, et al. SOX10 immunostaining distinguishes desmoplastic melanoma from excision scar. J Cut Pathol. 2010;37:944–952.
9. Harvey J, Thomas C, Wood B, et al. Practical issues concerning the implementation of Ki-67 proliferative index measurement in breast cancer reporting. Pathology. 2015;47:13–20.
10. Chorny JA, Barr RJ. S100-positive spindle cells in scars: a diagnostic pitfall in the re-excision of desmoplastic melanoma. Am J Dermatopathol. 2002;24:309–312.
11. Trejo O, Reed JA, Prieto VG. Atypical cells in human cutaneous re-excision scars for melanoma express p75NGFR, C56/N-CAM and GAP-43: evidence of early Schwann cell differentiation. J Cut Pathol. 2002;29:397–406.