Several population studies have shown that SM is more frequent in elderly patients with a mean age of 65 years, which is nearly 10 years older than the age of patients with melanoma located in the trunk or limbs 31. The mean age of patients with SM in the present study was slightly lower than that described in the literature, which is at 61.85 years. These results could be explained by the demographic characteristics of Brazil, which has a population with a lower life expectancy than Europe and North America 32. Additionally, the mean age was higher of patients with androgenetic alopecia (mean 73.18) than of those without androgenetic alopecia (mean 52.26).
In line with previous studies, SM occurred six times more frequently in men than in women 4,33,34. Furthermore, SM was associated with chronic sun damage but not with skin phototype or personal and family history of melanoma. Previous studies proposed that the hair shaft provides physical scalp protection against UV radiation 35 and we observed that androgenetic alopecia was significantly associated with elastosis and a dermoscopic pattern of photodamage. Androgenetic alopecia is typically more frequent in men; however, in the present study, we detected an equal frequency of the disease in SM patients of both sexes, reinforcing its importance as a risk factor for this type of tumor.
Because most SM studies are retrospective, a precise diagnosis of androgenetic alopecia is not always feasible. Additionally, in some cases, patients may have severe androgenetic alopecia yet develop melanoma in areas covered by hair that are not exposed to chronic sun damage (Fig. 3). Therefore, we evaluated histopathological criteria suggestive of androgenetic alopecia and elastosis at the site of the tumor, aiming to assess whether it is an area of chronic solar damage. According to this evaluation, 45.8% of patients had moderate to severe androgenetic alopecia and 54.1% had moderate to severe elastosis at the tumor site (Fig. 4).
SM is frequently diagnosed with high Breslow thickness 6, likely owing to late diagnosis because the tumor is hidden under the hair. Nevertheless, androgenetic alopecia is highly prevalent in SM patients 4,35 with a rate of nearly 50% in the present study. Additionally, we observed slightly thicker SMs in patients without androgenetic alopecia, strengthening the hypothesis that high Breslow thickness is due to the tumor being concealed under the hair. Another possible cause of the higher Breslow thickness in SM is the higher proportion of aggressive subtypes 36,37. Of the 48 cases we observed, 4.1% were nodular melanoma (Fig. 5) and 2% were desmoplastic melanoma (Fig. 6), which are consistent with the rates of these subtypes of melanoma in other sites 38–41.
BRAF-mutated melanoma represents approximately 50% of all melanomas and is associated with younger age, lack of elastosis at the tumor site, and truncal location 19,42. Generally, SM comprises head and neck melanomas and is associated with solar elastosis and less frequently with BRAF mutations 43,44. Despite this tendency, 41.5% of the SMs in our cohort were BRAF-mutated. In general, the BRAF V600 gain-of-function mutation is most common from V to E (70%), V to K (10–30%), or V to R (5%) 45. In contrast, the present study showed different frequencies of these mutations: 50% was V to K, 31.2% was V to E, and 18.8% was other BRAF mutation locations. Patients with BRAF V600K mutations tend to be older and have more chronic sun damage 46. Some evidence suggests that these patients have a significantly shorter disease-free survival and overall survival 19. The high frequency of BRAF V600K mutations observed in SM can partially explain the unique behavior of this subgroup of melanomas.
KIT mutations are reported in 3% of melanomas and are more common in acral and mucosal sites as well as sun-exposed skin melanomas. Moreover, NRAS mutations are associated with congenital melanocytic nevi 42. Both, KIT and NRAS mutations occur in approximately 4.8% of SM cases.
NF1 is a tumor suppressor gene, and its loss-of-function has been described in up to 10–15% of melanomas. This mutation is more common in chronically sun-exposed skin of older male patients 47. NF1 mutation is associated with poor survival outcomes in patients with melanoma, as well as those with other cancers, such as breast cancer and head and neck squamous cell carcinoma 20. Although identification of NF1 mutations is challenging owing to the size and complexity of the gene, we observed these mutations in 20% of SMs, which is a possible explanation for the worse prognosis of patients with SM.
Regardless of location, early diagnosis with dermoscopy increases the chances of a favorable prognosis for patients with melanoma 21. The scalp is a frequently neglected region during dermatological examination and is difficult for patients to self-examine. Additionally, identifying melanomas hidden under the hair or interspersed with other benign, pigmented lesions in a chronically sun-damaged bald scalp is challenging 26,33.
Until 2012, only two case reports described dermoscopic patterns of SM. The first report described a pattern similar to that of trunk and limb melanomas 22, while the second referred to a pattern similar to that of facial melanomas 23. In 2012, Stanganelli et al. described the dermoscopic patterns of 71 SMs. They found that regression and atypical pseudo-networks were more common in thin melanomas, whereas irregular pigmented blotches and unspecific patterns were more common in thick lesions. The authors did not find a correlation between dermoscopic pattern and androgenetic alopecia 24. In our study, the most frequent dermoscopic finding was regression structures (53%), followed by an atypical network, diffuse or multifocal hypopigmentation, and perifollicular granularity (38.7%, each). However, we observed that androgenetic alopecia was associated only with the photodamaged dermoscopic pattern.
The high dermoscopic variability reported in SM is likely owing to the variety of histopathological subtypes of melanoma in the scalp area: LM especially in patients with alopecia and chronic sun-damaged skin; superficial spreading melanoma in patients with a hairy scalp; and nodular and desmoplastic melanomas. The last two subtypes are rare but most commonly found on the scalp 31.
In the Stanganelli et al. publication, the authors did not find a correlation between androgenetic alopecia and Breslow thickness, while Benati et al. found thicker melanomas in bald patients than in those with hair 33. Garbarino et al. analyzed 97 flat SMs and found that superficial spreading melanoma was most often invasive while the LM subtype was most often in situ. The authors hypothesized that LM demonstrates slower growth and easier diagnosis because it is visible to clinical examination 48. In our study, melanomas in patients with androgenetic alopecia and elastosis were associated with lower Breslow thickness. These findings suggest that a higher Breslow thickness in SM is related to late diagnosis and the tumor’s concealment under the hair.
In conclusion, SM has distinct clinical, dermoscopic, histopathological, and molecular profiles in areas with or without androgenetic alopecia. SM in areas with androgenetic alopecia and elastosis occurs more frequently in elderly patients with chronic sun damage, presenting as the LM subtype, and with BRAF V600E mutation than SM without androgenetic alopecia and elastosis. A simple, generalized linear model showed that SM in areas with androgenetic alopecia and elastosis is more likely to present with photodamaged dermoscopic patterns and have lower Breslow thickness and absence of mitosis. The lower Breslow thickness in SM patients with androgenetic alopecia could be owing to easier tumor visualization and earlier diagnosis, in contrast to tumors hidden under the hair. Thus, proactive screening of scalp areas should be encouraged.
Finally, SM has a distinct molecular profile with a high frequency of BRAF V600K and NF1 mutations. These adverse prognostic markers can explain the unique behavior of this subgroup of melanomas. Further studies are necessary to validate these findings.