In Mexico, there have been few studies on primary oral and sinonasal melanoma, an aggressive neoplasm with a low survival rate and few therapeutic alternatives. Further, there is limited information about its clinical and histopathological characteristics. The aim of this retrospective study was to describe the clinicopathological profile of these tumours in patients attending a major oncology reference centre in Mexico City over a 12-year period. Demographic and clinical data were obtained from the clinical charts, and histopathological features were evaluated. χ 2 , Fisher’s exact, and Mann–Whitney U -tests were used for analysis; significance was set at P < 0.05. Thirty-three cases were studied (73% sinonasal melanoma (SNM) and 27% oral melanoma (OM)); 58% were female and the median age was 66 ( Q 1 – Q 3 55.5–75) years. Compared with OM patients, SNM patients had a shorter time to diagnosis (16.7 vs. 11.7 months, P = 0.022), were identified at earlier stages (33.3% vs. 58.3%, P = 0.010), and all presented symptoms (66.7% vs. 100%, P = 0.015). All samples showed vertical growth and 96.9% exhibited pleomorphism. A higher proportion of cases with pleomorphism developed metastases at follow-up than those without (60% vs. 12.5%, P = 0.026). The present study provides valuable information that could form the basis of future studies in the search for advanced therapy modalities.
Primary oral and sinonasal melanomas (POSNMs) are rare and aggressive neoplasms that comprise less than 1% of all melanomas and 4% of all head and neck cancer. The incidence of POSNM is apparently steady, in contrast to melanoma of the skin, which has been increasing in countries such as the USA and Australia (from 18/100,000 in 1999 to 25.5/100,000 in 2008 ). There have been few studies on POSNMs in Latin America, and information on their clinicopathological profile is missing. In Mexico, few studies have defined the characteristics of head and neck melanomas and most have described only oral melanoma (OM) characteristics; none have compared sinonasal melanomas (SNMs) to OMs.
The histopathological diagnosis of POSNMs represents a challenge because of its rarity and variable appearance. The poor clinical outcome is related to patient difficulties in recognizing the early signs and symptoms, resulting in late diagnosis.
Recent studies have associated certain tumour suppressors and oncogenes (NRAS, CKIT, and BRAF) with melanoma tumourigenesis, related to the regulation of proliferation and apoptosis pathways. The characterization of these genes has allowed the development of molecular therapeutic targets including BRAF and KIT kinase inhibitors (vemurafenib and sunitinib, respectively), which are now being applied in metastatic cutaneous melanoma, modifying the conventional treatment.
The aims of this study were to provide information related to the epidemiological, clinical, and histopathological characteristics of 33 cases of POSNM seen at a major oncology centre in Mexico City over a 12-year period (January 2000 to September 2012) and to evaluate the differences between SNM and OM.
Materials and methods
This was a retrospective study that included data from POSNM patients aged >18 years, attending a major referral oncology centre in Mexico City over a 12-year period (January 2000 to September 2012). The hospital received 129,095 histopathological samples during the study period; 7.7% (9890) of these were head and neck tumours, of which 43 (0.4%) were POSNM. This study was approved by the ethics and research board of the hospital.
The diagnosis of the primary tumour was based on the information registered in the clinical charts. Cases with a history of melanoma at another site or evidence of metastatic melanoma from an extramucosal site were excluded.
Demographic characteristics and details of tobacco and alcohol consumption, anatomical site (International Classification of Diseases (ICD) 10th revision, C00–05 and C30–31), clinical appearance (number, shape, surface, colour, and size), time to diagnosis (defined as the period from the onset of the lesion to the definitive diagnosis), symptoms, follow-up period, recurrence, metastases, and treatment were obtained from the clinical charts. Clinical stages were defined according to the report by Kanetaka et al. , as follows: stage Ia: local, limited to one anatomical region; stage Ib: local, extending to adjacent regions; stage II: regional, cervical nodal metastasis only; and stage III: distant metastasis. Taking into consideration the retrospective nature of this study, which comprised cases from 2000 to 2012, the AJCC staging sytem published in 2010 was not applied to clinically classify the patients included.
Histological haematoxylin and eosin (HE) stained slides of the cases were assessed in accordance with the modified criteria of the Western Society of Teachers of Oral Pathology (WESTOP). The parameters evaluated included cellular invasion (considered as non-invasive (in situ) or invasive (presence of cell invasion in the superficial or deep lamina propria, or invasion into skeletal muscle, bone, or cartilage)) and tumour growth pattern (horizontal or vertical). Cell morphology was classified as monomorphous (epithelioid or spindle cells) or polymorphous, defined as cytologically heterogeneous and characterized by mixed cells varying in shape and size (clear, spindle, plasmacytoid, rhabdoid, and/or multinucleated tumour cells). Also, the type of arrangement (alveolar, organoid, solid, pagetoid, papillary, or mixed), the mitotic rate (number of mitoses per 4 high-power fields, mm 2 ), the presence of melanin, necrosis, perineural, neural, perivascular, and vascular invasion, and the inflammatory content, were evaluated. The diagnosis of lesions without melanin-pigmented cells, amelanotic epithelioid melanomas, and spindle and undifferentiated tumours were confirmed by immunohistochemical analysis of S-100, vimentin, HMB-45, or melan-A. Immunohistochemistry techniques were not performed routinely in all cases, particularly in those with a clear melanoma morphology and melanin-pigmented cells; these techniques are often too costly to be applied on a regular basis.
The statistical analysis was performed using IBM SPSS Statistics for Windows software, version 19.0 (IBM Corp., Armonk, NY, USA). Medians and interquartile intervals ( Q 1 – Q 3 ) were calculated, and χ 2 , Fisher’s exact, and Mann–Whitney U -tests were applied as necessary. The level of significance was set at P < 0.05.
Thirty-three oral and sinonasal melanoma cases were included, 24 SNM (73%) and nine OM (27%); 19 (58%) were females and the median patient age was 66 ( Q 1 – Q 3 55.5–75) years. A higher proportion of women was found in the SNM group (62.5%) in comparison to the OM group (44.4%). As shown in Table 1 , tobacco and alcohol consumption was more frequent in OM patients (55.6% and 66.7%, respectively) than in SNM patients (45.8% and 37.5%). Five patients (15.2%) reported a family history of cancer (liver, prostate, stomach, and ovarian cancer, and skin melanoma; one each).
|Characteristics||OM ( n = 9)||SNM ( n = 24)||P -value|
|Sex, n (%)|
|Female||4 (44.4)||15 (62.5)||0.442 a|
|Male||5 (55.6)||9 (37.5)|
|Age, years, median ( Q 1 – Q 3 )||60.5 (48.5–76.5)||65.7 (57.7–75)||0.486 b|
|Occupation, n (%)|
|Housewife||4 (44.4)||13 (54.2)||0.354 c|
|Farm worker||1 (11.1)||5 (20.8)|
|Employee||2 (22.2)||1 (4.2)|
|Other||2 (22.2)||5 (20.8)|
|Tobacco consumption, n (%)|
|Yes||5 (55.6)||11 (45.8)||0.645 c|
|No||3 (33.3)||9 (37.5)|
|NA||1 (11.1)||4 (16.7)|
|Alcohol consumption, n (%)|
|Yes||6 (66.7)||9 (37.5)||0.227 c|
|No||3 (33.3)||11 (45.8)|
|NA||0 (0.0)||4 (16.7)|
|Family history of cancer, n (%)|
|Yes||1 (11.1)||4 (16.7)||0.111 a|
|No||8 (88.9)||20 (83.3)|
All 33 lesions were nodular and hyperpigmented, and described as coffee-brown (7, 21.2%), black (3, 9.1%), or grey-brown (4, 12.1%). A higher proportion of ulcerated lesions was identified in OM cases (33.3%) than in SNM cases (16.7%). All 24 SNM were solitary lesions, whereas five of the nine OM (55.6%) showed satellite lesions at the time of oral examination.
The nasal fossa (23/24, 95.8%) and hard palate (6/9, 66.7%) were the most affected locations. Eighteen (54.5%) of the 33 POSNMs involved the adjacent mucosa (nine OMs). The time to diagnosis was significantly shorter for the SNM group than for the OM group (median 11.7 vs. 16.7 months, P = 0.022). All SNM cases showed signs and symptoms such as pain, nasal obstruction, epistaxis, rhinorrhoea, and progressive enlargement of the affected area. In contrast, only two thirds of OM patients (66.7%) reported symptoms such as bleeding and swallowing difficulty ( P = 0.015). Pain was an important symptom in 20 of the 33 POSNM patients (60.6%), of whom 14 (70%) had SNM. Earlier stages of the disease (Ia and Ib) were identified in 58.3% (14/24) of SNM cases compared to 33.3% (3/9) of OM cases ( P = 0.010) ( Table 2 ).
|Features||OM ( n = 9)||SNM ( n = 24)||P -value|
|Tumour size, mm, median ( Q 1 – Q 3 )||50 (30–62.5)||48.5 (29–60)||0.852 a|
|( n = 8)||( n = 19)|
|Time to diagnosis, months, median ( Q 1 – Q 3 )||16.7 (7–24)||11.7 (3.2–11.2)||0.022 a|
|Signs and symptoms, n (%)|
|Yes||6 (66.7)||24 (100.0)||0.015 b|
|No||3 (33.3)||0 (0.0)|
|Clinical stage, n (%) d|
|Ia||1 (11.1)||7 (29.2)||0.010 b|
|Ib||2 (22.2)||7 (29.2)|
|II||6 (66.7)||4 (16.7)|
|III||0 (0.0)||6 (25.0)|
|Recurrence at the 6-month follow-up ( n = 25), n (%)|
|Yes||3 (37.5)||11 (64.7)||0.344 b|
|No||5 (62.5)||6 (35.3)|
|( n = 8)||( n = 17)|
|Total metastasis, n (%)|
|Yes||8 (88.9)||18 (75.0)||0.642 c|
|No||1 (11.1)||6 (25.0)|
|Metastasis at diagnosis ( n = 26), n (%)|
|Yes||7 (87.5)||9 (50.0)||0.057|
|No||1 (12.5)||9 (50.0)|
|( n = 8)||( n = 18)|
All cases were invasive ( Fig. 1 A) and showed a vertical growth pattern; 96.9% exhibited pleomorphism. As shown in Table 3 , the solid arrangement was the most frequent presentation (48.5%), followed by mixed and organoid arrangement ( Fig. 1 B). With regard to cellular morphology, epithelioid was the most common histological feature (45.5%) ( Fig. 1 C), followed by the plasmacytoid and the clear cell morphology. It is interesting to note that while all lesions were clinically hyperpigmented, only 26 (78.8%) cases showed melanin deposits in the tissue analyzed. Mitosis was present in 63.6% of cases and was more frequent in the SNM group (75%) than in the OM group (33.3%) ( P = 0.044). Most of the samples analyzed showed less than five mitoses per 40× field ( Fig. 1 D). Necrosis was present in 63.6% of POSNM. The most frequent invasion patterns identified were perivascular (90.9%) and vascular (45.5%), and only two cases showed perineural and neural invasion (6%). Thirty-two tumours (97.0%) had focal and mixed inflammatory infiltrates.
|OM ( n = 9), n (%)||SNM ( n = 24), n (%)||P -value|
|Alveolar||0 (0.0)||2 (8.3)||0.791 a|
|Organoid||2 (22.2)||4 (16.7)|
|Solid||4 (44.4)||12 (50.0)|
|Mixed||3 (33.3)||6 (25.0)|
|Epithelioid c||4 (44.4)||11 (45.8)||0.614 b|
|Spindle d||2 (22.2)||2 (8.3)|
|Plasmacytoid e||1 (11.1)||6 (25.0)|
|Clear cells f||1 (11.1)||5 (20.8)|
|Rhabdoid||1 (11.1)||0 (0.0)|
|Present||3 (33.3)||18 (75.0)||0.044 b|
|Absent||6 (66.7)||6 (25.0)|
|0||4 (44.4)||8 (33.3)||0.690 b|
|1–30%||5 (55.6)||9 (37.5)|
|31–60%||0 (0.0)||4 (16.7)|
|>61%||0 (0.0)||3 (12.5)|