Abstract

NF1 is mutated in 12-18% of cutaneous melanomas, with up to 46% of wild-type BRAF and NRAS melanomas found to harbor mutated NF1. NF1 mutations are more common in desmoplastic melanoma (DM), a rare histologic subtype. While DM is considered to be locally aggressive with a decreased risk of metastasis, non-DM NF1-mutated melanomas are associated with an increased risk of death and poor overall survival. Detection of NF1 mutations is technically challenging because of its large size and lack of mutational hotspots. Therefore, a more rapid, less labor-intensive, and inexpensive screening assay would be highly valuable for the identification of NF1-mutated melanomas. As the majority of NF1 mutations in cutaneous melanomas are truncating mutations, we hypothesize NF1 loss as detected by immunohistochemistry (IHC) can identify non-DM NF1 melanomas. In this study, we sought to validate the use of a commercially-available rabbit polyclonal anti-neurofibromin antibody in detection of loss-of-function NF1 mutations by IHC. We first identified a cohort of DM, a group known to be enriched in NF1 mutations and performed next-generation sequencing on five DM excision specimens. All 5 specimens harbored multiple truncating mutations in NF1. Two specimens harbored variants in BRAF with unknown significance. No specimens harbored NRAS mutations. Three specimens had an overall high mutational burden. IHC in these same specimens revealed relative or complete loss of neurofibromin staining in tumor cells exhibiting spindled morphology, while expression was preserved in the in situ component or in melanoma cells with a superficial spreading or nodular histologic subtype We will ultimately evaluate the performance of anti-neurofibromin IHC on a larger cohort of non-desmoplastic melanomas and compare it against NGS results to determine specificity/sensitivity and positive predictive value/negative predictive value.

Financial Disclosure:
No current or relevant financial relationships exist.