Introduction

Melanoma is the most common skin cancer in children and young adults1,2. The difference in the cutoff value of age to define “pediatric melanoma” (PM) among various studies is one hurdle in studying PM3. Another challenge is the rarity of PM which accounts for approximately only 2% of all melanomas4. Clinical characteristics of PM resembling benign lesions and showing amelanotic features also pose a diagnostic challenge5,6. A prior report has suggested that the majority of PM cases are Caucasians7. The analysis of the Surveillance, Epidemiology, and End Results (SEER) database reported that 85% of PM patients were Caucasians7. Data on melanomas in Asians are rare with Asian/Pacific Islander patients altogether comprise only 2%8,9,10.

With an age-adjusted incidence of 7.8 per 100,000 adolescents and young adults (AYAs, 15–39 years old), a recent review has addressed the need of attention to this age group11. Prior reports have specifically addressed melanoma in AYAs11,12,13. Although one study has noted a lower incidence of melanoma in Asian AYAs than in Western countries13, clinical and histopathologic data of melanoma on Asians are lacking. Moreover, a racial difference in characteristics of melanoma between Caucasians and non-Caucasians (mostly Hispanics) even in young patients has been reported14,15. Thus, the aim of this study was to evaluate clinical and histopathological features of melanoma Asian under age 40.

Methods

This study was approved by the Institutional Review Board of Asan Medical Center (IRB No. 2020 − 1102). The study was conducted in compliance with the relevant guidelines and regulations. Patients and/or their legal guardians signed informed consent regarding publishing their data and photographs. A total of 45 patients under age 40 were diagnosed with melanoma in situ or invasive cutaneous melanoma in Asan Medical Center between January 2001 and December 2020. All melanomas were confirmed with skin biopsies. Patients diagnosed with non-cutaneous melanoma or melanoma of unknown origin were excluded.

Clinical variables including age at diagnosis, sex, anatomic site of lesion, nodularity, ulceration, amelanosis, TNM stage at diagnosis, and types of treatment were collected from medical records and clinical photographs. Based on the nodularity, lesions were classified as flat patch-like lesions and bumps. Hematoxylin and eosin (H&E) stained slides were evaluated for Breslow thickness, Clark’s level, ulcer, mitosis, lymphovascular invasion, perineural invasion, greatest dimension, and histologic types, which were classified as superficial spreading, acral lentiginous, nodular, lentigo maligna, desmoplastic, spitzoid, and unknown. Two dermatopathologists diagnosed spitzoid melanomas for cases that showed more than two of the following features: no maturation with HMB45 staining, loss of p16, and Ki-67 proliferation index 10% or higher.

Overall survival (OS) was defined as the duration between the date of initial diagnosis of melanoma and the date of patient’s death or the latest follow-up. Progression free survival (PFS) was defined as the duration between the date of initial diagnosis and the date of local recurrence or metastasis. Local recurrence and metastasis to lymph nodes were included when they were proved histologically. Metastasis to distant organs was determined when imaging studies were consistent with metastasis. Survival analysis was conducted using the Kaplan-Meier method. Tumor staging was assessed according to the American Joint Committee on Cancer (AJCC) staging system, 8 th edition. To identify prognostic factors, univariate Cox regression analysis was performed using the survival package in R language (version 3.6.3). The variables with p-value < 0.05 in univariate Cox regression model were further investigated in multivariate Cox regression analysis.

Subgroup analysis according to age was conducted. Patients were divided into three groups (group A for children aged 14 or under, group B for adolescents aged between 15 and 21, and group C for young adults aged 22 to 39) as described in a previous study16.

Racial differences in characteristics of pediatric melanoma were explored. Patients in this present study who were all Asians were compared with Caucasian and Hispanic patients whose SEER data were previously reported. Continuous variables were compared using the Kruskal-Wallis test. Categorical variables were compared using Fisher’s exact test. All analyses were performed using Windows SPSS 21.0 (SPSS Inc., Chicago, IL, USA) and R studio (version 3.6.3). P-values of < 0.05 were considered statistically significant.

Results

Clinical and histopathological characteristic of melanomas in Asians under age 40

Clinical and histopathological characteristics of patients are summarized in Table 1. A total of 45 patients were analyzed. The median age was 35 years old (range: 4–38). Anatomical sites of melanomas were classified as head and neck, upper extremities, trunk, and lower extremities. The lower extremities were the most prevalent locations for melanomas in the study population with 22 (45.3%) cases, followed by 11 (24.4%) cases on upper extremities, 8 (17.8%) on trunk, and 4 (8.9%) on head and neck. Eleven (30.6%) cases showed nodular lesions. Ulceration was clinically shown in 4 (10.6%) patients and 8 (20.5%) patients who had amelanotic lesions.

Table 1 Summary of clinical and histologic characteristics of patients enrolled in the study.
Full size table

Eight (17.8%) in situ melanoma lesions and thirty-seven invasive melanoma cases were included. They showed a mean Breslow thickness of 4.98 ± 4.73 mm. Ulceration was noted in 5 (12.8%) of 39 specimens. Mitosis of 1/mm2 or more was noted in 11 specimens. The most common histologic type was acral lentiginous (n = 18, 40%) of which twelve (26.7%) patients were subungual, followed by nodular (n = 6, 13.3%) and spitzoid (n = 5, 11.1%) types. Superficial spreading type was found in only one (2.2%) patient.3.1.3. Genetic characteristics of melanomas in Asians under age 40.

The patients’ status was also investigated prior to targeted treatment. Five of 17 patients tested positive for BRAF-V600E mutation. None of the patients was investigated for PD-L1 protein expression, although 1 patient used nivolumab for a year.

Progression and survival outcomes and prognostic variables of patients

Of 45 patients enrolled for this study, 13 (28.9%) died during the follow-up. The mean overall survival (OS) was 48.4 ± 33.5 months. Eleven (24.4%) patients showed recurrence. The mean progression free survival period for the entire cohort was 28.4 ± 14.7 months (range, 9 to 60 months). Table S1 provides a summary of the patients with disease progression or metastasis. The most common site of primary lesion in the patients who progressed (6/11, 54.5%) and died (6/13, 46.2%) involved lower extremities, followed by trunk (3/11, 27.3% and 5/13, 38.5%). Histological subtypes of melanoma were available for 5 patients who progressed and 7 patients who died. Among the 5 progressed patients, one patient had nodular type of melanoma and the other 4 manifested acral lentiginous melanoma. Three of the 7 patients who died were diagnosed with acral lentiginous melanoma and the other 4 were nodular type. None of the cases was initially diagnosed as benign or other melanoma subtypes including spitzoid. The anatomical sites of progression were classified into histologically confirmed local recurrence and lymph node metastasis, and radiologically reported distant metastasis. Local recurrence was found in one patient whose primary lesion involved the 1 st toe. 7 patients developed lymph node metastasis, including 3 involving inguinal areas, 1 iliac, 1 cervical, and 2 axillary cases. 5 patients showed distant metastasis, involving bone, lung, and liver in one patient, bone alone in the second patient, brain in the third patient, stomach in the fourth case, and brain and skin metastasis in the fifth patient. The last patient had an acral melanoma on thumb, which metastasized to neck.

Figure 1 represents a collection of Kaplan-Meier curves according to statistically significant prognostic variables (p < 0.05). Log-rank sum test was statistically significant for anatomical site (extremities vs. non-extremities, p = 0.039), histological type (nodular vs. acral vs. others, p = 0.018), tissue ulceration (p = 0.023), perineural invasion (p = 0.047), lymph node involvement at diagnosis (p < 0.0001), and recurrence (p = 0.0035). Table S2 provides a summary of Cox-regression analyses. Histological subtypes, lymph node involvement at diagnosis, metastasis at diagnosis, and recurrence were statistically significant in univariate analysis. P-values < 0.05 were used for multivariate analysis. Metastasis at diagnosis was excluded as the p-value of multivariate Cox analysis was not available. Lymph node involvement at diagnosis (p < 0.01) and nodular histology (p < 0.01) were statistically significant in multivariate analysis.

Fig. 1
figure 1

Kaplan-Meier graphs for statistically significant prognostic variables in log-rank sum test (p < 0.05). (A) Anatomical site: Extremities (33) vs. Non-extremities (12) (p = 0.039), (B) Tissue ulceration: Presence (6) vs. Absence (32) (p = 0.023), (C) Histological subtypes: Acral lentiginous (18) vs. Nodular (6) vs. Others (8) (p = 0.018), (D) Perineural invasion: Presence(1) vs. Absence (19)(p = 0.047), (E) Lymph node involvement at diagnosis, presence (8) vs. absence (35) (p < 0.001), (F) Progression (11) vs. No progression (34) (p = 0.0035).

Full size image

Age subgroup analysis

Results of subgroup analysis are shown in Table 2. Clinically, lesions with an amelanotic feature were more prevalent in children and adolescents (groups A (age: 0–14) and B (age: 15–21)) than in young adults (group C (age: 22–39)) (p < 0.01). There was no significant difference in anatomical ___location between the groups. Histologically, the spitzoid type of melanomas was identified more than other types of melanomas in children and adolescents (groups A and B) than in young adults (group C) (p < 0.001). Clinical photographs of spitzoid melanoma characterized by an amelanotic nodule in children and adolescents are shown in Fig. 2A and B, respectively. Clinical features of acral lentiginous type of melanomas from each group are presented in Fig. 2C and D, and 2E.

Table 2 Subgroup analysis according to age group.
Full size table
Fig. 2
figure 2

Representative clinical photos. (A) A spitzoid melanoma presenting as an amelanotic nodule on the right knee of a 4-year-old girl. (B) A spitzoid melanoma presenting as an amelanotic nodule on a 15-year-old male. The feature is similar to pediatric spitzoid melanoma as shown in (A). (C) Clinical picture of a nail unit melanoma, a subtype of acral lentiginous melanoma on the 4 th finger of a 10-year-old boy. (D) An acral lentiginous melanoma on the 5 th toe of a 19-year-old male. Histopathologic examination revealed mitoses with a Breslow depth of 2.5 mm. (E) A nail unit melanoma of a 38-year-old female presented as a longitudinal dark brown melanonychia in a single digit.

Full size image

Pediatric melanoma (age < 20 years old) in the study and comparison with other races

Among 45 patients included in the study, nine were under age 20. We classified their melanomas as PM. The median age of PM was 10 years old (range: 4–19). Clinical and histopathologic characteristics of PMs are summarized in Table 3. The most common anatomical sites for melanomas in the pediatric population were lower extremities (55.6%), followed by upper extremities (22.2%), head and neck (11.1%), and trunk (11.1%). Four (44.4%) cases had nodular configuration. Four (44.4%) cases were amelanotic.

Table 3 Summary of pediatric melanomas (age < 20 years old) in the present study and comparison with other races using SEER database.
Full size table

Three (17.8%) in situ melanoma lesions and six invasive melanoma cases were evaluated. They showed a mean Breslow thickness of 5.5 ± 4.13 mm. Mitosis of 1/mm2 or more was found in two specimens. The most common histologic type was the spitzoid type (n = 4, 44.4%), followed by the acral lentiginous type (33.3%).

A summary of PM characteristics according to race is presented in Table 3. In previous studies that reported racial difference comparing non-Caucasians’ PM with Caucasians’ PM, more female patients were reported15. In this study, however, more male patients were included (p < 0.01). There was a statistically significant difference in the ___location of lesions according to race (p < 0.001), with extremities being the most common sites involved in non-Caucasians’15. According to national data (SEER database), the most common anatomical site was the trunk (35.9%) for melanomas in Whites under age 20 and the lower extremities (40.1%) for Hispanics15. Similar to Hispanics, Asian PMs in the present study most commonly occurred on the extremities (77.8%), especially the lower parts (55.6%) with high percentages. A statistically significant difference in histologic type was also found according to race (p < 0.001). The most common subtype of melanoma was spitzoid (36.5%) in Hispanics under age 20 according to SEER database, followed by the superficial spreading type (17.3%), while superficial spreading type (32.1%) was the most common one in Whites, followed by the spitzoid type (22.5%)15. In the present study, the most common subtype in Asians under the age of 20 was spitzoid (44.4%), followed by acral lentiginous type (33.3%). Significant difference in lymph node involvement at diagnosis was found (p < 0.047).

Discussion

Diagnosis and management of melanoma in young individuals is clinically challenging as there is a paucity of studies on melanoma in young ae, especially Asians. This study sought to investigate the characteristics, risk factors, and outcomes in melanoma in young Asian patients. Importantly, understanding of melanomas in this demographic may improve risk stratification and help clinicians decide on management for such patients.

First, this study found that characteristics of melanoma differ with specific age groups. Among melanoma patients below 40 excluding pediatric and adolescents, they exhibit features of melanoma in keeping with the classic adult-type melanoma. Pediatric and adolescents display a more atypical phenotype with statistically more frequent amelanosis as clinical feature and spitzoid type as histologic feature. As such, we demonstrate that the age of the patient at presentation may be an important consideration on the management of melanoma17.

Recently, melanoma in AYA has drawn attention. According to our results, melanomas in adolescents (Group B) and young adults (Group C) showed different characteristics. Adolescents (Group B) showed characteristics similar to those of children under age 15 (Group A) who had high frequencies of clinically amelanotic tumors and histologically spitzoid melanomas, while melanomas in young adults (Group C) were mostly clinically melanotic and histologically acral lentiginous. Thus, it may be worthy to distinguish the group of AYA into two distinct groups of adolescents and young adults, although further studies with larger sample sizes are needed.

While the cutoff value of age to define PM differs in various studies, PM is most commonly defined as a melanoma in patients aged 20 or under at diagnosis3. Clinically, as PMs are reported to not fulfill conventional ABCDE criteria with atypical features such as amelanosis and uniform color, Cordoro et al. have proposed a modified ABCD criteria for children, in which ABCDE stands for amelanotic, bleeding or bump, color uniformity, de novo of any diameter, and evolution, respectively5. Although previous study reported amelanosis of PM to be up to 50%18, a recent multi-centered Italian study reported more than 90% of PMs to be pigmented and non-amelanotic19. Clinical features in the present study showed that PM lesions had higher frequencies of nodularity and amelanosis. As shown in Table 2, not just children, but also adolescents (Group B) showed high nodularity (66.7%) and amelanosis (50%). This might be due to a high percentage of acral and spitzoid melanomas in this study in both children and adolescents. Unlike Caucasians, acral melanomas are common in Asians and more commonly present as amelanotic20 and spitzoid melanomas often show nodular and amelanotic lesions21.

PM can be classified into three subtypes: spitzoid melanoma, melanoma arising in a congenital melanocytic nevus (CMN), and conventional adult-type melanoma22. A previous study has reviewed PMs and found that spitzoid melanomas frequently occur on limbs21. These results were consistent with results of the current study showing that spitzoid melanomas occurred on extremities including knees, legs, and elbows. While the prognosis of spitzoid melanoma remains controversial, several recent reports have suggested favorable outcomes23,24,25,26. No patient with a spitzoid melanoma in the present study died or recurred during at least 24 months of follow-up. It is worth noting, however, that the cases in the present study, previously reported as spitzoid melanomas in pathological reports, may now be reclassified as atypical spitzoid tumors, as the classification has changed over time27. The favorable prognostic outcome, even in cases of positive sentinel lymph node suggests atypical spitzoid tumors than spitzoid melanomas28. Although the gold standard for diagnosing atypical spitzoid tumors involves identifying genetic alterations alongside histopathological examination, the retrospective design of the study limited the ability to perform such testing.

Second, this study shows that acral lentiginous melanomas and melanomas on lower extremities are common in pediatric Asians as shown in Table 3, alike melanomas in adults. While, the most common conventional adult-type of melanoma in adults is the acral lentiginous type in Asians, the superficial spreading type is most common in Hispanics and Whites29. Even at an early age (< 20 years), a high frequency of superficial spreading PM was seen in Caucasians and Hispanics and the acral lentiginous type was detected in Asians in the present study. Along with histological subtype, which differed significantly according to race, sexual predisposition, anatomical site, and presence of lymph node involvement at diagnosis also showed statistically significant difference. The finding that acral lentiginous melanoma accounts for 58% of all cutaneous melanomas in Asians may explain the higher incidence of melanomas located on extremities in our study29,30. Extremities were the most prevalent locations of lesions in all age groups as seen in Table 2. Although spitzoid melanomas accounted for lesions arising on extremities in younger groups (Groups A and B), none in the young adult group (Group C) had spitzoid melanomas.

Third, while one third of melanomas are reported to develop from nevi, nevus-associated melanoma was not observed in the present study. This suggests that nevi in young Asian patients are less likely to progress to melanoma at least in young age below40. Melanoma may develop on CMN. Its risk increases with size, multiple satellite lesions, and abnormalities in magnetic resonance imaging31,32. Although the risk of melanoma arising on CMN remains low with an overall risk of 0.7%, the risk is higher (10–15%) for giant CMNs where multiple sporadic evolutions occur at different sites33,34,35. This type of melanoma is known for its fatality36. No patient in the present study had melanoma arising on CMN, which might have contributed to an excellent prognosis of PM in this study.

Survival analysis showed that lymph node involvement at diagnosis and nodular histology represented statistically significant prognostic factors. Several previous studies have reported that patients with lymph node involvement at diagnosis show poor prognosis37,38,39, consistent with results of this study.

Finally, in our study, 26.7% of observed melanomas in patients below 40 were the subungual type in Table 1. Due to the high frequency of subungual melanoma in patients below 40, subungual nevi or pigmentary changes in fingernails and toenails should be investigated and carefully monitored in the clinical setting. Of note, a recent report in Korea noted that early onset cutaneous malignant melanomas in Asian patients aged 10 to 39 more frequently presented as subungual melanomas (27.3%) compared with their counterparts aged 50 to 79 (17.2%), whereas acral lentiginous melanoma excluding subungual types were less prevalent (age 10–39 (18.2%), age 50–79 (54.0%))40.

This study has several limitations. First, it was designed retrospectively with a small sample size, limiting the genetic testing of spitzoid tumors. Second, it was carried out in a single tertiary hospital. Further multi-centered studies with large samples and comparative genetic studies are warranted. However, the present study is meaningful in that we discovered that melanomas in young Asians under age 40 frequently arose on extremities with acral lentiginous, nodular, and spitzoid types being common. Lymph node involvement at diagnosis and nodular histology are significant prognostic factors. Statistically significant racial differences were found in PM based on sex, anatomical site, histological types, and lymph node involvement at diagnosis. In addition, subungual type of melanomas in young Asian patients suggest careful monitoring of pigmentary lesions in nail units.