Soft Tissue Sarcoma Mimicking Melanoma: A Systematic Review

1. Introduction

The 2020 World Health Organization (WHO) Classification of Soft Tissue Tumors indicates sarcomas as rare tumors that are further subclassified in approximately 70 subtypes, each characterized by a distinct morphology, that often translates into a specific clinical behavior as well as into specific therapeutic approaches [1,2]. They can occur anywhere in the body, affecting the extremities in 50% of cases, the trunk and retroperitoneum in 40%, and the head and neck in 10% [3].

Of note, some of these neoplasms can be confused with cutaneous melanoma since they share similar clinical, histological, and immunohistochemical features [4]. In addition, they are rare tumors that are often overlooked by the clinician who does not pose the diagnostic suspicion to the pathologist [2,5,6]. Therefore, health care specialists may encounter some difficulties in differentiating between sarcoma and melanoma during the diagnostic process. For example, clear cell sarcoma (CCS) clinically presents as a deep and small (<5 cm) soft tissue lesion, often juxtaposed with tendons, fascia, or aponeurosis, which may mistakenly suggest some forms of melanoma such as acral melanoma, nodular melanoma, or amelanotic melanoma [7,8]. Moreover, CCS shows a phenotype identical to that of conventional melanoma, characterized by strong expression of S100 protein in 100% of cases and variable expression of HMB-45, Melan-A, and MiTF. As a matter of fact, CCS can be genetically differentiated from melanoma due to some peculiarities, including i) the typical reciprocal translocation t(12;22) (q13;q12) that gives rise to the EWSR1-ATF1 oncogene and ii) the absence of BRAF/NRAS mutations that can often characterize melanoma [4]. Other examples include malignant peripheral nerve sheath tumor (MPNST) and Kaposi's sarcoma. MPNST usually arises from peripheral nerves and may be associated with patients with neurofibromatosis type 1 (NF1). The clinical presentation involves the development of a painful and/or rapidly expanding mass with associated neurological deficits. The biological behavior of MPNSTs has been described as unpredictable, and the differential diagnosis includes several tumors, particularly spindle cell/desmoplastic melanoma, which serves as the main differential because of the higher incidence, remarkably similar morphology, and overlapping immunochemical markers [9,10]. Kaposi's sarcoma is typical of immunocompromised patients and may also present as a single papular skin lesion clinically mimicking melanoma, from which can be distinguished by histologic and immunohistochemical appearance [11].

The clinical presentation of such neoplasm may mislead the physician, with potential implications on therapeutic strategy and prognosis. For example, surgical widening of the margins varies according to the neoplasm under treatment (melanoma or sarcoma), chemo-radiotherapy could be offered to sarcoma, but melanoma patients could benefit from a different first-line approach such as immunotherapy or target therapy or both,and sentinel lymph node biopsy is routinely performed in the diagnostic workup of melanoma but is still under debate for soft tissue sarcoma [12,13,14].

Awareness of the similarities between sarcoma and melanoma, and the ability to recognize the two entities play a crucial role in patient care. However, a clinician may have little-to-no direct experience on this matter and may retrieve only limited information from a case report or a small case series.

This systematic review aimed at summarizing available evidence on cases of sarcoma that were initially diagnosed as malignant melanoma, to help clinicians in the diagnostic process, to improve patient care.

2. Materials and Methods

2.1. Study design

This is a systematic review of case series and case reports describing cases of sarcoma that were initially diagnosed as malignant melanoma. The review was conducted according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines [15]. The review protocol was registered in PROSPERO (CRD42023403882).

2.2. Search strategy

We systematically searched MEDLINE/PubMed, EMBASE, and SCOPUS to detect eligible studies. The search strategy was conducted without language restrictions through March 2023. In PubMed, the following search strategy was used: sarcoma mimicking melanoma OR sarcoma resembling melanoma OR melanoma-like. The search strategy was tailored to conform to the other electronic sources. The lists from each source were joined and the duplicates were removed. Two investigators (FC, AD) separately evaluated titles and abstracts of the records and removed those unfit outside the scope of the review. The full-text of all potentially eligible records were examined to dismiss those not fulfilling the inclusion criteria. Finally, the reference lists of included records were hand-searched to detect further studies of interest. Any disagreement was solved by consensus with a third investigator (PDF). Studies not including human subjects were excluded. No language restrictions were applied.

2.3. Data collection

Two investigators (FC, AD) independently extracted relevant data from the included articles. For each article, study features, patient characteristics, tumor information and outcome measures were collected. A third investigator (ABF) checked the extracted data. Any inconsistency was solved by consensus.

2.4. Assessment of the quality of included studies

The quality of the included studies was assessed according to eight criteria: i) clear criteria for inclusion of the patient(s), ii) valid methods for identification of the initial condition, iii) valid methods for identification of the final condition, iv) in a case series, consecutive inclusion of patients, v) clear reporting of demographics, vi) clear reporting of clinical information, vii) reporting of the time of the second assessment, and viii) reporting of the reason for the second assessment. The criteria were adapted from the Joanna Briggs Institute (JBI) critical appraisal tool [16] to fit the context under evaluation (case series and case reports describing cases of sarcoma that were initially diagnosed as malignant melanoma). Two investigators (FC, MA) independently appraised the risk of bias of the included studies and any inconsistency was solved by consensus with all authors.

2.5. Data synthesis

The selection procedure was presented in a flow-chart. Pertinent data were extracted from included studies and summarized in tables. The inclusion of case reports and very small case series precluded the feasibility of a meaningful metanalysis, hence a narrative synthesis of included studies was conducted.

3. Results

3.1. Search results

The search identified 365 non-duplicated records. After excluding 321 records based on title/abstract, 43 potentially eligible records were retrieved for full text review. Of these, 17 were excluded due to different design (n=4), different topic (n=12) or different participants (n=1). Another two records were excluded because the full-text could not be found. Three additional records were identified via hand search. Finally, 27 records [9,11,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41] were included in the narrative synthesis (Figure 1).

3.2. Narrative synthesis of the findings

The synthesis included 23 case reports (85%) and four case series (15%). Three studies (11%) were published in 1989-2000, seven (26%) in 2001-2011 and 17 (63%) in 2012-2022. Study and patient characteristics are summarized in Table 1 and Table 2.

Overall, the studies reported on a total of 34 patients (aged 12-86 years), including 25 males (74%) and nine females (26%). The initial diagnosis was melanoma (24 patients, 71%) or suspect of melanoma (10 patients, 29%). The clinical presentation was heterogeneous (Table 1) and the involved anatomical regions included lower limbs (8/33 patients, 24%), head/neck (8/33 patients, 24%), upper limbs (7/33 patients, 21%), visceral area (4/33 patients, 12%), trunk (3/33 patients, 9%) and genital area (3/33 patients, 9%) (the information was not available in one patient).

IHC positivity was reported for S100 (22/32 patients, 69%), HMB45 (20/32 patients, 63%), MelanA (10/32 patients, 31%) and MiTF (1/32 patients, 3%).

The reason for second assessment and/or diagnostic re-evaluation included unusual presentation (12/25, 48%), uncertain diagnosis (7/25, 28%), expert opinion (2/25, 8%), no response to treatment (1/25, 4%), search for EWSR1 translocation (1/25, 4%), review after surgery (1/25, 4%) or review after death (1/25, 4%), while the information was not reported in nine patients.

EWSR1 translocation was investigated in 17/34 patients (50%), and found positive in 16 CCS and one MPNST.

The final diagnosis was clear cell sarcoma in 17 patients (50%) and soft tissue sarcoma in 17 patients (50%). The latter included sarcoma of perivascular epithelioid cells (n=4), malignant peripheral nerve sheath tumor (n=4), Kaposi's Sarcoma (n=1), chondroid syringoma (n=1), cutaneous angiosarcoma (n=1), cutaneous epithelioid angiosarcoma (n=1), epithelioid malignant Schwannoma (n=1), malignant giant cell tumor of soft tissue (n=1), malignant schwannoma (n=1), myeloid sarcoma (n=1) and pleomorphic sarcoma (n=1).

An overview of the main findings is displayed in Figure 2.

3.3. Critical appraisal of the quality of included studies

Table 3 summarizes the quality assessment of the included studies. All studies (27/27, 100%) reported clear criteria for inclusion of the patient(s). Valid methods for identification of the initial (melanoma) and final (sarcoma) conditions were described by 16/27 (59%) and 18/27 (67%) studies, respectively. All case series (4/4, 100%) included consecutive patients. Clear reporting of demographics and clinical information were found in 27/27 (100%) and 24/27 (89%) studies, respectively. Only one study (4%) specified the timing of the second assessment, and 17/27 studies (63%) clearly reported the reason for the second assessment.

4. Discussion

This systematic review evaluated the available evidence on cases of sarcoma that were initially diagnosed as malignant melanoma. Our search yielded only case reports and small case series [9,11,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41] which individually can provide sparce information to the health care provider, but summarizing data from such sources may allow a better understanding of the topic. Of note, most studies were published in the last decade, which may suggest a rising interest in differentiating sarcoma from melanoma during the diagnostic process.

Overall, the clinical presentation of such cases was heterogeneous, and some sarcomas were initially misdiagnosed because several aspects such as clinical factors, localization of the lesion, and histologic appearance suggested a malignant melanoma [9,17,18,19,20,21,22,23,24,25,26,27,28,32,33,34,35,39,40,41]. In some cases, the clinician considered other diagnoses but finally opted for a melanoma [29,30,31,36,37,38]. Furthermore, immunohistochemistry could not be helpful for differentiating sarcoma from melanoma, which was suggested by the positivity of some markers such as S100, HBMG-45 and MelanA [9,17,18,19,20,21,22,23,26,27,29,30,31,32,33,34,35,36,37,38,39]. Hence, the rarity of cases of sarcoma mimicking melanoma likely played an important role in opting for melanoma as the reasonable initial diagnosis of choice. The reader should be aware that we assume such rarity given the few cases in literature, but we do not have robust information about the real magnitude of cases of sarcoma mimicking melanoma.

In fact, the main reasons for the second assessment leading to a diagnosis of sarcoma were the unusual presentation and the uncertainty about the diagnosis, which suggested further investigations to the clinicians [9,18,19,20,21,22,23,29,30,32,33,35,38]. In few cases, the second assessment was done because the patient did not respond to the treatment [17] or during a retrospective review of cases [21,24].

This implies that the correct identification of a sarcoma mimicking melanoma relies on the health care provider being aware of the possibility of such case and being able to identify when unusual features merit further investigations.

In addition, this also means that the prevalence of such cases is unknown, because the literature does not include episodes when the health care provider did not feel the need for further investigations and no systematic investigations have been conducted in large series of melanoma patients.

We believe that misdiagnosing a sarcoma as a melanoma may have potential implications for patient care because of the different therapeutic approaches, including sentinel node biopsy, first-line therapy and surgical therapy [12,13,14]. Unfortunately, available information is insufficient to assess the prognostic effect of such misdiagnosing.

Interestingly, half of the sarcomas found at the second assessment were CCS [17,18,19,20,21,22,23,26,29,30,31,32,33]. We believe that the common features shared by CCS and melanoma [7,8] and the lower incidence of CCS were likely to be reponsible for the initial misdiagnosis. When investigated, EWSR1 translocation was found positive in almost of all cases [17,18,19,20,21,22,26,29,30,31,32,33], hence the clinician may benefit from the inclusion of the investigation of EWSR1 translocation in the initial diagnostic process.

This systematic review has some limitations that should be considered by the reader. First, the research topic was prone to be described in case reports and very small case series, hence limiting the available information and the potential for further analyses. Second, the lack of epidemiological studies prevented from any considerations about the prevalence of cases of sarcoma that were initially diagnosed as melanoma. Third, information about the timing of the second assessment would provide interesting information but it was largely missing in the literature. Fourth, the role of EWSR1 translocation in the identification of CCS could not be investigated because of selected reporting (the included studies reported some CCS with positive EWSR1 and one MPNST with negative EWSR1).

Within its limitations, our systematic review underlines an underreported problem in the diagnosis of melanoma and sarcoma, inform the physician about features that can make the differential diagnosis difficult, and highlights the importance of searching for the EWSR1 translocation in the diagnostic process.

5. Conclusions

Atypical skin lesions could be misdiagnosed as melanomas if they share many similarities. Physicians should be aware of such possibility in the diagnostic process, as it may have potential implications on the treatment strategy. In case of atypical skin lesions, it could be useful to investigate the presence of EWSR1 translocation, since CSS are the most common histology to be found in case of re-evaluation. Referral to tertiary expert centers may be recommended. Further investigations are required to better understand the epidemiology of the mislead diagnosis and to raise awareness of the issue.