Loss of p16 expression and copy number changes of CDKN2A in a spectrum of spitzoid melanocytic lesions

Summary

Spitzoid melanocytic lesions, including Spitz nevi (benign), spitzoid melanoma (malignant), and borderline atypical Spitz tumors (ASTs), frequently present challenges for accurate diagnosis and prognosis. Evaluation for loss of the tumor suppressor p16, encoded by CDKN2A gene on chromosome 9p21.3, has been proposed to be useful for evaluation of spitzoid melanocytic lesions. However, reports on the utility of p16 immunohistochemistry for spitzoid lesions have been conflicting, and few studies have directly compared p16 immunohistochemistry with fluorescence in situ hybridization (FISH) for CDKN2A genomic status. We analyzed a spectrum of benign (n = 24), borderline (n = 27), and malignant (n = 19) spitzoid lesions for p16 protein expression by immunohistochemistry and CDKN2A copy number by FISH. Immunohistochemistry was evaluated by 2 scoring methods: H score and 2-tiered score (positive or negative for p16 loss). By immunohistochemistry, loss of p16 expression was not observed in Spitz nevi (0/24) but was seen in ASTs (7/27; 26%) and spitzoid melanomas (3/19; 16%). By H score, p16 expression was significantly higher in Spitz nevi relative to ASTs or spitzoid melanomas. Similarly, copy number aberrations of CDKN2A by FISH were absent in Spitz nevi but were found in 2 (9.5%) of 21 ASTs and 4 (33%) of 12 spitzoid melanomas. Our findings from this large cohort suggest that p16 aberrations are highly specific for borderline and malignant spitzoid neoplasms relative to Spitz nevi. Similar to ASTs, p16 loss in spitzoid melanomas may occur in the presence or absence of genomic CDKN2A loss.

Introduction

Spitzoid lesions represent a subclass of melanocytic lesions with distinctive cytomorphology and architectural findings and include benign Spitz nevi, atypical Spitz tumors (ASTs) with uncertain malignant potential, and spitzoid malignant melanomas (SMMs). Accurate classification of spitzoid lesions is challenging, and there is considerable interobserver variability, even among experts [1], [2], [3].

CDKN2A is a tumor suppressor gene on chromosome 9p21.3 encoding 2 protein products, p16 (p16INK4A) and p14 (p14ARF). p16 is a cyclin-dependent kinase inhibitor that facilitates G1 cell cycle arrest. Multiple lines of evidence support a tumor suppressor role for p16/CDKN2A in melanoma. CDKN2A undergoes deletion in up to 86% of melanomas, with homozygous loss in up to 18% [4], [5], [6]. Germ line mutation of CDKN2A is associated with familial melanoma [7]. Many studies have shown loss of p16 expression in conventional melanoma but not in benign melanocytic lesions [8], [9], [10]. For these reasons, evaluation of p16 protein expression and genomic status has gained widespread use in the setting of diagnostically challenging melanocytic lesions.

Spitzoid melanocytic neoplasms are associated with distinct oncogenic drivers relative to conventional melanocytic neoplasms, including activating HRAS mutations, BAP1 loss, and kinase fusions, whereas BRAF and NRAS mutations are infrequent [3], [11]. Spitz nevi also harbor distinctive copy number changes relative to ordinary melanocytic nevi, whereas ASTs and SMMs may harbor multiple copy number abnormalities that overlap with conventional melanoma [1], [12], [13]. These observations raise the question whether spitzoid malignancies harbor loss of p16 similar to conventional melanoma. Reports are mixed regarding diagnostic utility of p16 protein expression in spitzoid melanocytic neoplasms. Most studies find p16 loss to be concerning for a borderline or malignant spitzoid neoplasm [8], [9], [14], [15], [16]. However, a recent study found that 3 (17%) of 18 Spitz nevi also displayed p16 loss [15]. Hence, clarification is needed regarding the sensitivity and specificity of p16 loss in this setting. Moreover, no single study, to our knowledge, has directly compared the utility of p16 immunohistochemistry in the full spectrum of spitzoid neoplasms (nevi, melanomas, and ASTs).

Studies have also examined the diagnostic and prognostic importance of CDKN2A copy number status in spitzoid lesions. Homozygous loss of CDKN2A by fluorescence in situ hybridization (FISH) is associated with more aggressive behavior in spitzoid tumors, whereas heterozygous loss is not diagnostic of malignancy [16], [17], [18], [19]. Few studies have correlated p16 immunohistochemistry with CDKN2A copy number changes to assist diagnosticians in selecting FISH or immunohistochemistry as a diagnostic approach [16], and to our knowledge, the correlation between p16 expression and CDKN2A copy number status in SMMs has not been systematically examined.

To characterize the utility of p16 immunohistochemistry and genomic status in a range of spitzoid neoplasms, we examined a cohort of 70 spitzoid neoplasms including benign, borderline, and malignant lesions, representing the largest and most complete cohort yet assembled for analysis of p16 protein expression and CDKN2A copy number status by FISH in this context.