Molecular pathology as a diagnostic aid in difficult to classify melanocytic tumors with spitzoid morphology

Abstract

Accurate classification of melanocytic proliferations has important implications for prognostic prediction, treatment and follow-up. Although most melanocytic proliferations can be accurately classified using clinical and pathological criteria, classification (specifically distinction between nevus and melanoma) can be challenging in a subset of cases, including those with spitzoid morphology. Genetic studies have shown that mutation profiles differ between primary melanoma subtypes and Spitz nevi. These differences may aid in distinguishing benign from malignant in some melanocytic tumors. Here, we present a selection of melanocytic proliferations with equivocal histopathological criteria, where genetic analysis was requested to help guide classification. In two out of four cases, the genetic results offered valuable insights allowing a definitive diagnosis, indicating the diagnostic value of mutation profiling in a real world routine clinical setting. Though histopathological assessment remains decisive in melanocytic proliferation classification, we recommend to include genetic profiling in cases of borderline or atypical lesion to support accurate classification.

Introduction

Although most melanocytic proliferations can be accurately diagnosed and classified based solely on clinical and pathological assessment, determining the malignant potential of some melanocytic tumors remains challenging for histopathologists. An example is the group of Spitz tumors (1), which encompasses the entire spectrum of benign to malignant spitzoid melanocytic tumors (2). While Spitz nevi are benign, tumors with a related morphology that have a poor prognosis are referred to as Spitz melanoma or spitzoid melanomas (2). Spitzoid morphology is characterized by epithelioid to spindled melanocytes with abundant homogenous cytoplasm and large vesicular nuclei. Spitz tumors often display epidermal hyperplasia and suprabasal intraepidermal melanocyte nests, a phenomenon known as pagetoid spread (2, 3). Many features of Spitz tumors can also be found in melanomas, which can make pathological differentiation challenging, especially in older patients (4). Tumors with spitzoid ambiguous histopathological features intermediate to benign Spitz nevi and malignant melanoma are frequently referred to as atypical Spitz tumors (AST) (3, 4). In cases where diagnosis by morphology and immunohistochemistry (IHC) alone is difficult, molecular assays may provide valuable additional information to aid classification. Accurately classifying a tumor as benign or malignant has enormous implications for the patient in terms of prognosis, treatment (including newly introduced adjuvant therapies) and follow-up.

The last decade brought a more detailed understanding of the genetics of malignant melanoma (MM). Melanoma can be classified based on driver mutations as 1. BRAF-mutant (50–60%), 2. RAS-mutant (20–30%), 3. NF1-mutant (10–15%) or 4. triple (BRAF, NRAS and NF1) wild-type (~10%) (5, 6). Many other mutations can occur, TERT promoter mutations are particularly frequent (50–80%) (7, 8). Genetic insights led to the development and introduction of targeted BRAF- and MEK-inhibitors and improved survival of melanoma patients (9–11) (12, 13).

The genetic profile of Spitz tumors differs from that of conventional (i.e. non-spitzoid) melanoma. Spitz tumors may either harbor activating HRAS mutations (14) or a range of translocations involving several genes including ALK, ROS1, RET, NTRK1, NTRK3, BRAF and MET (15–18). A subset of Spitz tumors also demonstrates BAP1 loss (19, 20). Furthermore, the presence of TERT promoter mutations has been associated with poorer prognosis in Spitz tumors (21).

Distinguishing between Spitz nevus or atypical Spitz tumor and melanoma can be difficult and is a diagnostic challenge regularly encountered by dermatopathologists. If an incorrect diagnosis is rendered, the tumor may be under- or over-treated, both with potentially serious long-term consequences for the patient. Assessment of the genetic profile of Spitz tumors is an ancillary aid with the potential to decrease the number of cases in which a definitive diagnosis cannot be made (22). Here we present selected cases from a routine dermatology practice in which genetic analysis was performed to help classify atypical Spitz tumors.

Results:

Case 1

A 27-year-old female patient presented to our skin cancer centre two months after complete resection of a pigmented skin lesion on the helix of the right ear. The patient’s medical history revealed chronic hepatitis C and several cutaneous haemangiomas. The patient reported that the lesion had been present since childhood and previously classified as a nevus based on clinical findings. Three months earlier, the skin lesion had started to grow rapidly. There was no bleeding, itching or ulceration. Histological analysis showed a spitzoid spindle-cell tumor with an epithelioid dermal component. Pleomorphic cells with enlarged nuclei were noted (Fig. 1A). Melan A staining confirmed asymmetry of the lesion including the junctional component (Fig. 1D). Ki-67 staining demonstrated increased dermal proliferative activity (Fig. 1C). Two experienced histopathologists had seen the case and recommended molecular diagnostic assessment to help differentiate between an atypical spitzoid melanocytic lesion and melanoma. Next generation sequencing (NGS, supplementary table 1) revealed BRAF^V600E and TERT promoter mutations, favouring diagnosis of a melanoma with a tumor thickness of 4.1 mm without ulceration. The patient was classified as high-risk melanoma pT4a, tumor stage IIB according to the 8^th edition of the American Joint Committee on Cancer (AJCC). The subsequently performed sentinel lymph node biopsy (SLNB) was negative for tumor cells. The tumour board recommended the inclusion into a clinical trial for adjuvant therapy versus close follow-up observation. The patient opted for follow-up observation according to the German melanoma guidelines, and showed no evidence of disease at second follow-up 6 months later.

Figure 1.

Histological findings showed an exophytic dermal-based tumor composed of spindle and epithelioid cells (A, bar represents 2 mm). The tumor cells are pleomorphic with enlarged nuclei; there are irregularly distributed junctional components with isolated breakdown of the overlying epidermis (B, bar represents 400 μm). Ki-67 staining shows increased proliferative activity in the junctional and dermal components present (C, bar represents 300 μm). Melan-A staining highlights the asymmetrical silhouette of the tumor and the junctional melanocytic component (D, bar represents 300 μm). The BRAF V600E (c.1799_1800 CA>TT) mutation and the TERT promoter mutation (Chr. 5: 1,295,250 C>T) determined by targeted next generation sequencing are shown at the top with a representative wild-type sequence at the bottom. The mutation site is highlighted by the black arrow (E). The tumor was diagnosed as melanoma.

Case 2

In a 42-year-old male patient, a suspicious lesion on the right lower leg was detected during regular skin cancer screening (Fig. 2A). An AST had been excised at age 13 (29 years earlier) from the same location. The lesion was excised and histological assessment revealed dermally located, predominantly nested melanocytes (Fig. 2B) with hyperpigmentation of the basal cell layer (Fig. 2C). The asymmetric tumor showed a proliferation of atypical melanocytic cells with pleomorphic nuclei in the basal layers of the epithelium, which were also located suprabasal. In the junctional zone, cells were lying in unevenly configured nests. The proliferation index with Ki-67 was focally discretely increased and individual mitotic figures detected. Cytomorphologically, predominantly epithelioid cells with pale grayish cytoplasm, nuclear size variation, and occasional hyperchromatic nuclei were noted (Fig. 2D). Definitive determination of the dignity of this melanocytic tumor based solely on histopathology was deemed challenging, and in the absence of conclusive pathological evidence to support spitzoid melanoma, a diagnosis of AST was favoured. NGS identified no known driver mutations. No translocations were identified in a subsequent screening for 509 known cancer relevant gene fusions, including those known to occur in spitzoid melanocytic lesions (supplementary table 2). As no genetic alterations were detected, the molecular results were of no further significance with regard to classifying the biological potential of the tumor. Due to the histological atypia and the difficulty in determining the likely biological behaviour of the tumour, it was recommended to regard this lesion as malignant and to manage it as a melanoma. The patient was informed of all possibilities and opted for SLNB as recommended by melanoma guidelines. Neither SLNB nor staging examination showed evidence of metastatic spread.

Figure 2.

Clinically, multiple brownish pigmented macules are visible (A, arrow indicates biopsy site). A dermally located, predominantly nested melanocytic cell population (B, bar represents 2 mm). The epidermis shows hyperpigmentation of the basal cell layer with no melanocytic cell proliferation (C, bar represents 400 μm). The dermal melanocytic cells extend into the subcorium and show no maturation to the deep. Cytomorphologically, predominantly epithelioid cells with pale gray cytoplasm, nuclear size variation and hyperchromatic nuclei (D, bar represents 80 μm).

Case 3

A 48-year-old patient presented with melanoma lymph node metastasis of the left groin. Clinical history revealed that a highly atypical spitzoid melanocytic lesion of the left lower leg had been diagnosed and excised 5 years earlier.

Histopathologically, the primary tumor of the lower leg demonstrated a predominantly dermal asymmetric melanocytic lesion with junctional nests and suprabasal intraepidermal melanocytic cells (Fig. 3A–B). Atypical melanocytic cells with pleomorphic hyperchromatic nuclei and prominent nucleoli were distributed in the corium as small nests and strands (Fig. 3C). Ulceration was not present. Melan A labelled junctional and dermal portions of the lesion (Fig. 3D). At the time of excision, diagnosis of a highly atypical Spitz tumor had been made and confirmed by second opinion from a highly respected histopathologic reference laboratory. The histology of the groin lymph node demonstrated unequivocal metastatic melanoma without extracapsular extension (Fig. 3E).

Figure 3.

The primary tumor from the lower leg showed a predominantly dermal asymmetric melanocytic lesion (A, bar represents 2 mm) with junctional nests and single intraepidermal melanocytic cells (B, bar represents 400 μm). No epidermal ulceration was present. The dermal tumor cells did not mature with depth; the small nests grew between dermal collagen fibers. Atypical melanocytic cells distributed in the corium as small nests and cords, partly epitheloid cells with variably sized hyperchromatic nuclei with prominent nucleoli were present (C, bar represents 80 μm). Melan A staining showed positivity of junctional and dermal portions of the leg tumor (D, bar represents 500 μm) and in in the lymph node metastasis (E, bar represents mm). NRAS Q61R (c.182 T>C) and TERT Chr. 5: 1295228 C>T promoter mutations determined by targeted next generation sequencing are shown at the top with a representative wild-type sequence at the bottom. The mutation site is highlighted by the black arrow. The tumor was diagnosed as a metastasized cutaneous melanoma (F).

To assess whether the leg tumor was in fact malignant and responsible for the current metastasis, the molecular profile of both the cutaneous tumor and the lymph node metastasis was analysed by NGS. NGS analysis identified NRAS Q61R and TERT promoter mutations in both the primary tumor (Fig. 3F) and the lymph node metastasis. Based on the mutation data, the histomorphology and the clinical course, the diagnosis of the leg tumor was revised to primary melanoma with subsequent lymph node metastasis. For treatment as stage IIIB melanoma (8^th edition AJCC), lymphadenectomy followed by radiotherapy was discussed by the tumor board, however declined by the patient. Adjuvant nivolumab immunotherapy initiated six weeks after diagnosis of advanced melanoma is ongoing without relapse to date.

Case 4

A melanocytic lesion was excised from the calf of a 4-year-old girl. Histopathological analysis identified a cellular melanocytic cell population abutting the epidermis and extending into the deep corium. The predominantly spindle-shaped dermal melanocytic cells were arranged in strands and larger nests. Focally, infiltration of lymphocytes and granulocytes were observed. Individual areas demonstrated an increased mitotic activity (up to 4/mm²). Parts of the tumor were almost solid (Fig. 5A-C). Differentiation between an AST and a spitzoid melanoma was challenging by histopathology and immunohistochemistry alone. Molecular analysis identified no classical melanoma driver mutations or translocations. Overall, the histological atypia and negative genetic findings best fit a diagnosis of atypical Spitz tumor. Due to concern about the histological atypia, however, treatment as a melanoma was recommended.

Discussion:

The presented cases illustrate the potential diagnostic value of mutation profiling and its limitations in a real world clinical setting. Especially in adult patients, melanocytic tumors with spitzoid morphology can pose diagnostic and management dilemmas. Not uncommonly, relevant clinical information, e.g. if the lesion arose de novo or if it has been present since childhood, is not reliably obtainable (in our experience, patients are often uncertain of how long the lesion was present or if it had changed). Pathological evaluation is the gold standard for diagnosis, but as described above, the overlap in pathological features between Spitz tumors and melanoma can make definitive classification difficult in some cases. In difficult cases, genetic analysis may be a helpful additional tool in classifying spitzoid tumors, as mutation profiles differ between primary melanomas and Spitz nevi.

When diagnostic questions arise, we usually perform genetic testing with an NGS panel encompassing 36 genes known recurrently altered in melanocytic tumors (all genes listed in the Supplemental Methods). This covers frequently mutated melanoma genes (i.e. BRAF, NRAS, NF1, the TERT promoter, etc.), and others, including GNAQ, GNA11, MAP2K1, CTNNB1, BAP1 and HRAS which are mutated in less common melanocytic proliferations, including blue nevi, deep penetrating nevi, Spitz nevi and other lesions (14, 19, 23, 24). Since this is a targeted panel with limited coverage of the genome, copy number alterations are not reliably detected. If we fail to identify a gene mutation profile compatible with the histopathological picture, we perform translocation analysis, using an NGS RNA sequencing assay with a pan-cancer translocation panel assessing 509 known fusions (Supplemental material and methods).

In AST, detection of a driver mutation in BRAF or NRAS strongly argues against a Spitz nevus. If the tumor also harbors a TERT promoter mutation, a diagnosis of melanoma or recommendation of treatment as melanoma should be considered. In two of our cases (cases 1 and 3) a diagnosis of melanoma was made based on the identification of a BRAF or NRAS mutation along with a TERT promoter mutation. In cases where a HRAS or BAP1 mutation is detected, a HRAS-mutant Spitz nevus should be considered (14), or a spitzoid tumor characterized by BAP1-loss (19). In tumors in which no activating mutations are detected via DNA NGS sequencing, one should consider testing for translocations frequent in Spitz tumors, such as those involving ALK, ROS1, RET, NTRK1, NTRK3, MET, BRAF, etc. (15). As there are many translocation candidates, some of which are probably not yet recognized, we recommend using a broad panel. The presence of Spitz translocations would argue for a Spitz tumor and in conjunction with patient age and histological features, a Spitz nevus can be considered (25). In a number of cases however, such as cases 2 and 4 presented above, a translocation may not be detected. Although the absence of detectable mutations or translocations does not permit definitive classification or prediction of a given tumor’s biologic potential, this may be potentially useful information in some situations, e.g. in a spitzoid tumor with atypical histological features suggestive of melanoma may be managed more conservatively in the absence of any melanoma-associated genetic alterations - whereas without this genetic information, a more aggressive treatment regimen with attendant comorbidities might be implemented.

The histopathological picture remains the gold standard for classifying melanocytic tumors and assessing their likely clinical/biological potentials. We do not perform genetic analysis for histologically unequivocal melanoma. The genetic assays are only used as an additional diagnostic aid in histologically ambiguous tumors in which (often due to the presence of overlapping morphological features that have been described in both benign and malignant tumors) pathological classification and determination of biological potential is not clear-cut. The cases in our study demonstrate that for some of these lesions, genetic analysis can be of great value in accurate diagnosis, prediction of likely clinical behaviour and institution of appropriate management. However, in other cases, additional genetic analyses are of limited value.

We recommend molecular analysis of all spitzoid neoplasms in which the biological potential cannot be definitively determined based on histological criteria. In such cases, genetic testing can aid in diagnosis, prognosis and selection of optimal treatment. The discovery of additional genetic alterations and refinement of genetic assays will likely continue to improve the value of genetic testing as additional tools in the management of patients with spitzoid tumors.

Conclusions:

• Mutation analysis can be a diagnostic aid in determining the dignity in some cases of atypical spitzoid melanocytic tumors

• In cases in which no gene mutations are identified: translocation analysis should be considered

• Suggested diagnostic procedure: histology + IHC ➜ gene mutation analysis ➜ translocation analysis ➜ possibly whole exome sequencing (WES)/whole genome sequencing (WGS)

Figure 4.

Cellular melanocytic cell population abutting the epidermis and extending into the deep corium is seen (A, bar represents 2 mm). In the overlying epidermis, there is hyperpigmentation of the basal cell layer with no significant melanocytic cell proliferation or ulceration (B, bar represents 200 μm). The dermal melanocytic cells are predominantly spindle cells and grow in strands and larger nests (B). Cytomorphologically, there are hyperchromatic nuclei with prominent nucleoli, and occasional deep mitoses (C, bar represents 200 μm).

Table 1:

Clinical and genetic characteristics

CASE	GENDER	AGE AT DIAGNOSIS	LOCALISATION PRIMARY TUMOR	DIAGNOSIS	BRAF	NRAS	TERT PROMOTER	PD-L1 STAINING	TRANSLOCATIONS
1	f	26	right ear helix	cutaneous melanoma	V600E	wt	250 C>T	PD-L1 negative (0 % tumor cells)
2	m	42	right popliteal fossa	atypical spitz tumor	wt	wt	wt	PD-L1 negative (0 % tumor cells)	no relevant translocations
3	m	48	left lower leg	cutaneous melanoma	wt	Q61R	228 C>T	PD-L1 negative (<1 % tumor cells)
4	f	4	left lower leg	atypical spitz tumor	wt	wt	wt	PD-L1 negative (<1 % tumor cells)	no relevant translocations

Abbreviations: f, female; m, male, MM, malignant melanoma; PD-L1, programmed death ligand-1; wt, wildtype; mut, mutated

Highlights:

• Mutation analysis diagnostic aid in determining dignity in some spitzoid tumors

• Where no gene mutations are identified: translocation analysis should be considered

• Molecular analysis aids in identifying malignant tumors to be treated as melanoma