Metastatic Malignant Melanoma With Neuroendocrine Differentiation

Discussion and Conclusions

We present a patient initially diagnosed as having metastatic NEC of unknown primary as dictated by the neuroendocrine markers of both first generation (SYP/CD56) and second generation (INSM1 and ISLET1) observed at histopathological investigations from a tumor biopsy. Intriguingly, standard melanoma markers (Melan A, HMB45) were negative at initial presentation, but subsequently positive in following investigations of a separate metastatic lesion in which more tissue was available for immunohistochemical assessment. The latter investigation rendered a diagnosis of malignant melanoma with neuroendocrine differentiation, which is a rare manifestation in the clinical setting. The case exemplifies how immunohistochemical patterns and routine morphology can differ within metastatic deposits and serves as a lesson to cautiously interpret neuroendocrine markers in the clinical context.

Malignant melanoma with neuroendocrine differentiation is rarely encountered, and the bulk of scientific data available is derived from case reports and small case series.^[6,10–15] In the largest study to date, Romano and co-workers described SYP positivity in 10 out of 34 malignant melanomas, typically only in small subsets of tumor cells.^[14] Interestingly, neuroendocrine differentiation was particularly evident in epithelioid malignant melanomas, similar to our case. Other reports describe scattered immunoreactivity for SYP and CD56, but rarely CgA.^[10–13,15] In our case, the tumor also exhibited widespread CD56 immunoreactivity, focal nuclear ISLET1 staining adjoined by a puzzling cytoplasmic INSM1 pattern – all neuroendocrine markers, of which the latter two exhibit high specificity in previous publications.^[16,17] A possible differential diagnosis besides metastatic NEC includes metastatic Merkel cell carcinoma (MCC); however, the tumor was negative for both CK20 and CgA, and the Ki-67-index was much lower than usually observed in MCCs. Moreover, the combination of positivity for Melan A, HMB45, and SOX10 strongly argued in favor of malignant melanoma as opposed to metastatic tumors of other origin.^[18–21]

From a clinical perspective, neuroendocrine differentiation in disseminated malignant melanomas might confuse the responsible pathologist into rendering a diagnosis of metastatic NEC, as in our case. The overall histomorphology is suboptimal in core-needle biopsy material compared to routine sections of surgically excised material, which is clearly demonstrated in our case (Figure 1a, b and Figure 2b, c). In the excised material, the level of nuclear details is in fact enough to suspect a malignant melanoma – which is hardly the case by judging the hematoxylin and eosin (H&E) sections from the core-needle biopsy. Moreover, the immunohistochemical analyses varied to some extent when comparing the core-needle biopsy with the excised material, which could be due to differences in the sheer tumor area covered. Intriguingly, the heterogeneity of the tumor staining is perhaps best exemplified by the negative Melan A and HMB45 immunostaining results when assessing the chest wall core-needle biopsy (Figure 1g, h), as compared to the excised femoral head and neck displaying focal but distinct immunoreactivity for both markers (Figure 2e, f). However, given the widespread SOX10 immunoreactivity in the excised material (Figure 2g), a lesson for future cases might be to always include SOX10 immunohistochemistry when assessing metastatic tumors with equivocal immunohistochemical findings regarding neuroendocrine markers.

Given the shared embryologic origin and functional aspects of melanocytes, neurons, and neuroendocrine cells, the notion that these cell types might share expressional profiles regarding vesicle transportation (SYP) and cell-cell adhesion (CD56) is not entirely inappropriate. However, given the general lack of published literature regarding neuroendocrine differentiation in malignant melanoma, little is known regarding patient outcome compared to conventional malignant melanomas. Although the coupling to prognosis is unclear, a general awareness among histopathologists regarding the potential for non-NETs to present with neuroendocrine features is warranted in order to secure the correct diagnosis. Treatment-wise, if a diagnosis of malignant melanoma would have been put forward earlier, the administration of carboplatin and etoposide would, in our case, probably have been avoided in favor for standardized melanoma protocols, irrespectively of neuroendocrine differentiation or not.

We conclude that surgical pathologists should interpret neuroendocrine markers with caution, as numerous non-NET variants have been reported in the literature, thereby potentially acting as clinical mimickers – which might lead to suboptimal treatment regimes. This is especially true from biopsy material, which is often scarce and thereby limits the number of immunohistochemical stains that can be ordered.

J Med Case Reports. 2020;14(44) © 2020 BioMed Central, Ltd.