Abstract and Introduction
Abstract
Objectives: We aim to determine molecular differences between Merkel cell polyomavirus (MCPyV)–negative Merkel cell carcinomas (MCCs) and neuroendocrine carcinomas (NECs).
Methods: Our study included 56 MCCs (28 MCPyV negative, 28 MCPyV positive) and 106 NECs (66 small cell NECs, 21 large cell NECs, and 19 poorly differentiated NECs) submitted for clinical molecular testing.
Results: APC, MAP3K1, NF1, PIK3CA, RB1, ROS1, and TSC1 mutations, in addition to high tumor mutational burden and UV signature, were frequently noted in MCPyV-negative MCC in comparison to small cell NEC and all NECs analyzed, while KRAS mutations were more frequently noted in large cell NEC and all NECs analyzed. Although not sensitive, the presence of either NF1 or PIK3CA is specific for MCPyV-negative MCC. The frequencies of KEAP1, STK11, and KRAS alterations were significantly higher in large cell NEC. Fusions were detected in 6.25% (6/96) of NECs yet in none of 45 analyzed MCCs.
Conclusions: High tumor mutational burden and UV signature, as well as the presence of NF1 and PIK3CA mutations, are supportive of MCPyV-negative MCC, whereas KEAP1, STK11, and KRAS mutations are supportive of NEC in the appropriate clinical context. Although rare, the presence of a gene fusion is supportive of NEC.
Introduction
Epithelial neuroendocrine neoplasms are broadly classified into either neuroendocrine tumors (NETs), including "carcinoid," or neuroendocrine carcinomas (NECs). NECs consist of a group of poorly differentiated and high-grade neuroendocrine neoplasms.[1,2] The latest World Health Organization classification identifies two subtypes of NEC: small cell NEC (SCNEC) and large cell NEC (LCNEC). Merkel cell carcinoma (MCC) represents a unique site-specific NEC.[1,2] These tumors are all characterized by their high-grade histologic appearance, growing predominantly in solid sheets with a high proliferation rate associated with necrosis, and apoptotic debris.[1,2]
Differentiating these neoplasms is imperative, as in the setting of metastasis the distinction of cutaneous MCC from other NECs has significant therapeutic implications. For example, while immunotherapy is the frontline treatment for advanced MCC, a combination of platinum and etoposide is the most widely used chemotherapy for small cell lung carcinoma (SCLC).[3,4]
Despite thorough clinical and imaging workup, the origin of a metastatic NEC remains unknown in some cases. In this setting, immunohistochemical stains can be of help. As NECs are epithelial neoplasms, a cytokeratin stain (keratin AE1/AE3 and/or CAM 5.2, MNF116) should be positive, although in a subset of NEC cases, cytokeratins may be only focally expressed, sometimes with a dotlike paranuclear pattern mimicking a staining pattern that is more characteristic of MCC. MCC classically exhibits dotlike positivity for keratin 20 and other keratins, including keratin AE1/AE3, CAM5.2, and MNF116.[5] The CM2B4 antibody that detects the Merkel cell polyomavirus (MCPyV) LT antigen can be useful in diagnosing MCPyV-positive cases. An immunohistochemical panel including keratin 20, neurofilament, SATB2, MCPyV, and TTF1 together with clinical findings and imaging studies would be helpful in the distinction of high-grade NEC from MCC.[6–8] The coexpression of keratin and neurofilament is typical of MCC and excludes small cell carcinoma.[6] In a study of 103 MCCs vs 70 extracutaneous NECs, Kervarrec et al[6] reported neurofilament, SATB2, and MCPyV to be the most discriminant markers. However, distinguishing poorly differentiated NEC from MCPyV-negative MCC remains challenging, as they can often have significant morphologic and immunohistochemical overlap.
In recent years, there has been an increased interest in using novel molecular methods to further understand and characterize NEC. For example, SCLC is characterized by nearly universal inactivation of the tumor suppressor gene tumor protein (TP53) and RB transcriptional corepressor (RB1). Recurrent genetic alterations have also been identified in LCNEC. Targeted next-generation sequencing (NGS) of 241 cancer genes of 45 LCNECs revealed alterations involving TP53 (78%), RB1 (38%), STK11 (33%), KEAP1 (31%), and KRAS (22%).[9] Interestingly, LCNEC could be grouped into two major subsets, SCLC-like and non-SCLC-like (NSCLC-like), based on their genomic profiles.[9] Fifty-six percent of LCNECs in a study by Rekhtman et al[10] exhibited NSCLC-like molecular profiles (lack of RB1 and TP53 coalteration and mutations of STK11, KRAS, KEAP1, and NFE2L2 with STK11/HRAS in 80% of the tumors), whereas 40% of LCNECs demonstrated a SCLC-like molecular profile (RB1 and TP53 coalteration, other SCLC-type alterations such as MYCL amplification, and complete absence of STK11 and KRAS mutations; MYCL, SOX2, FGFR1 amplifications; or PTEN mutations/loss).
Cutaneous MCC has two hypothesized pathogenetic pathways: UV-induced and clonal integration of MCPyV.[3] Specifically, MCPyV-negative MCCs have been found to exhibit a high tumor mutational burden (TMB), with frequent mutations in RB1 and TP53 along with mutations in genes involved in chromatin modification, JNK, and DNA damage pathways, whereas MCPyV-positive tumors demonstrate a lower TMB with none of the genetic alterations observed in the MCPyV-negative tumors.[3]
The diagnosis of MCPyV-positive tumors has recently been facilitated by CM2B4 immunohistochemistry; however, the distinction between MCPyV-negative MCC and NEC remains challenging in some cases. Since TP53 and RB1 mutations are signature molecular changes of UV-induced/MCPyV-negative MCC as well as SCNEC and a subset of LCNEC, detection of other molecular abnormalities is needed to separate the two entities in diagnostically challenging cases.[3] In this retrospective study, we aim to examine the results of our clinical NGS-based mutational analyses and targeted gene rearrangement testing to determine significant molecular differences that would be helpful in the clinical distinction of MCPyV-negative MCC vs NEC.
Am J Clin Pathol. 2023;159(4):395-406. © 2023 American Society for Clinical Pathology
