The first two authors contributed equally to this article.
Melanoma in the young: Differences and similarities with adult melanoma
A case-matched controlled analysis
Article first published online: 18 JUN 2007
Copyright © 2007 American Cancer Society
Volume 110, Issue 3, pages 614–624, 1 August 2007
How to Cite
Livestro, D. P., Kaine, E. M., Michaelson, J. S., Mihm, M. C., Haluska, F. G., Muzikansky, A., Sober, A. J. and Tanabe, K. K. (2007), Melanoma in the young: Differences and similarities with adult melanoma. Cancer, 110: 614–624. doi: 10.1002/cncr.22818
- Issue published online: 18 JUL 2007
- Article first published online: 18 JUN 2007
- Manuscript Accepted: 19 MAR 2007
- Manuscript Revised: 16 MAR 2007
- Manuscript Received: 27 JUN 2006
- sentinel lymph node;
Melanoma in the first 2 decades of life is rare and is less well characterized than melanoma in adults. Previously published comparisons of melanoma of the young (age ≤20 years) and in adults have been biased by the observation that primary melanomas of the young are significantly thicker on average than adult melanomas. In this study, the use of an adult control group that was matched for tumor thickness adjusted for this bias and allowed for a comparison of the biology of pediatric melanoma and adult melanoma.
The authors identified young patients with melanoma who were treated at their institution between 1971 and 2002, and matched each patient by thickness and year of diagnosis to 2 adult control patients with melanoma. The study group of 73 young patients with melanoma (aged ≤20 years) was compared with to 146 adult case-matched control patients for clinical presentation, treatment, and clinically relevant outcome parameters, including disease-free and cause-specific survival. A subset of pediatric patients aged ≤13 years also was analyzed against case-matched adult controls.
The demographics and clinical presentation of the 2 groups were similar. Of the young patients with melanoma who underwent pathologic staging of clinically negative lymph nodes by either elective lymph node dissection or lymphatic mapping/sentinel lymph node biopsy, 11 of 25 patients (44%) had positive lymph nodes compared with 11 of 46 patients (23.9%) among the adults. The overall incidence of positive lymph nodes was 17.8% in young melanoma patients and 9.6% in thickness-matched adult control patients. Nonetheless, 10-year cause-specific survival was similar between young melanoma patients and adult melanoma patients (89.4% and 79.3%, respectively). No significant differences were observed comparing young patients with melanoma aged <13 years versus age >13 years.
Melanoma of the young had some important differences and similarities relative to adult melanoma. Lymph node metastases were more prevalent in young patients with melanoma compared with adult (thickness-matched) control patients; however, the 5- and 10-year survival rates were similar. Cancer 2007. © 2007 American Cancer Society.
Melanoma in the pediatric and adolescent population is rare. Patients age <20 years account for between 1.3% and 2% of all cutaneous melanoma diagnoses,1 whereas prepubertal children account for only 0.3% to 0.4% of newly diagnosed cutaneous melanomas.2 Consequently, less is known regarding melanoma of the young relative to adult melanoma, and most of our knowledge regarding melanoma of the young is derived from studies of adult patient populations.
It is a logical supposition that the natural history of melanoma of the young indeed may differ from that of adult melanoma. For example, at least 33% of prepubertal melanomas arise from large congenital nevi,3 and 50% of the cases of malignant melanoma arising from large congenital nevi occur in the first decade of life.4 This illustrates that the pathogenesis of malignant melanoma in children often is different from that in adults.
Furthermore, lesions such as atypical Spitzoid proliferations and melanocytic tumors of uncertain malignant potential (MELTUMPs) may obscure the diagnosis of malignant melanoma in children.5 This is particularly a problem when analyzing large, multicenter databases, such as the Surveillance, Epidemiology, and End Results database. In a recent study of that database on young melanoma patients, between 44% and 65% of lesions in the various age groups were categorized histologically as not otherwise specified,6 and it is likely that a significant portion of those actually were MELTUMPs.
In general, most studies of young patients with melanoma report that primary tumor thickness7 and stage at presentation8 are important prognostic features for melanoma of the young. Some have suggested that young patients with melanoma may experience a worse outcome9 with earlier recurrences compared with adults,10, 11 whereas others report a more favorable outcome relative to that observed in adults.12 However, comparisons between melanoma of the young and adult melanoma have been influenced by differences in tumor thickness, which may obscure age-related differences in tumor biology.
In an effort to compare tumor biology accurately between young patients and adult patients with melanoma, we matched adult melanoma patients by thickness to the population of young melanoma patients who were seen at the Massachusetts General Hospital (MGH). To accommodate for the evolving insights into diagnosis and therapy of melanoma over the study period, we also matched for year of diagnosis.
MATERIALS AND METHODS
We queried the MGH melanoma database of 3202 patients as well as institutional pathology records to identify all young individuals who were diagnosed with invasive cutaneous melanoma between the years 1971 and 2002. The cut-off age that is used to define melanoma of the young varies widely in published reports, ranging from the 10th birthday to the 21st birthday (Table 1). Most studies use the term pediatric melanoma instead of melanoma of the young. We believed that pediatric melanoma may be different from melanoma in young adults, but we also recognized that both are very rare and may differ in their own from adult melanoma. Therefore, we preferred the term melanoma of the young, and we selected the 21st birthday as the cut-off age, thereby capturing all of the groups described in previous studies. We also performed subgroup analyses on pediatric patients before their 13th birthday versus adolescents/young adults past their 13th birthday in an effort to investigate whether there were differences between pediatric melanoma patients and adolescents/young adults with melanoma.
|Institution or database||Cut-off age, y*||No. of patients||Reference|
|United States National Cancer Data Base||18||2585||Lange et al., 200522|
|University of Colorado Health Sciences Center, Denver||21||15||Roaten et al., 200626|
|SEER database||20||1255||Strouse et al., 20056|
|Sydney Melanoma Unit and Newcastle Melanoma Unit, New South Wales||13||32||Milton et al., 197727|
|University of Colorado Health Sciences Center, Denver||16||27||Lange et al., 200528|
|Memorial Sloan-Kettering Cancer Center, New York||18||40||Saenz et al., 199910|
|Swedish National Cancer Registry||20||177||Sander et al., 19999|
|Massachusetts General Hospital, Boston†||14||22||Scalzo et al., 199729|
|EORTC-MCG||16||60||Spatz et al., 199630|
|Children's Hospital, Boston||15||23||Barnhill et al., 199517|
|St. Jude Children's Research Hospital, Memphis||20||33||Rao et al., 199019|
|Duke University Melanoma Clinic||20||78||Reintgen et al., 198911|
|Frenchay Hospital, Bristol||21||29||Moss and Briggs, 198631|
|Christie Hospital and Holt Radium Institute, Manchester||21||31||Pratt et al., 198116|
Patients who received no treatment or clinical follow-up at the MGH were excluded from the analysis. Eighty-six young melanoma patients were identified, and data were collected for these patients from clinical records, the Social Security Death Index, and the Massachusetts Department of Public Health Registry of Vital Records and Statistics death certificate database. If these sources showed conflicting results, then chart review was considered to be the most accurate source. Medical charts were available for 68 of 86 young melanoma patients (79.0%) and for 134 of 146 adult melanoma patients (91.2%).
Although the diagnosis of dysplastic nevus syndrome was noted in medical records, the foundation for such a diagnosis (eg, the number of nevi in these patients) often was not recorded. Therefore, dysplastic nevus syndrome was considered present when it was recorded as a diagnosis by treating physicians. Information on the presence and type of precursor lesions, such as congenital nevi, also was collected.
An initial comparison between the young patients and the adult patients with melanoma in the database (3202 patients) revealed that young patients had significantly thicker primary tumors. The median tumor thickness in young melanoma patients was 1.3 mm versus 0.9 mm in adult melanoma patients (P = .05). Because thickness is a major prognostic factor for melanoma, we selected an adult control group that was matched to the young melanoma patients by tumor thickness. In addition, controls were matched by year of diagnosis to ensure that each group was treated according to similar standards of care, because melanoma patient care evolved over the 32 years of the study. Elective lymph node dissections generally were performed on patients who had melanomas that measured from 1 mm to 4 mm in thickness prior to 1993. This technique of lymph node staging changed in 1993, when the technique of lymphatic mapping and sentinel lymph node biopsy was introduced. All patients with positive sentinel lymph nodes underwent completion lymphadenectomy. The size of the database precluded matching for additional parameters, such as ulceration; however, it is noteworthy that subsequent analysis demonstrated no difference in the frequency of ulceration between young melanoma patients and case-matched adult controls.
We matched adult melanoma patients in a ratio of 2:1 to obtain greater statistical power. Although it was attempted, the size of the database did not allow for higher matching ratios. If >2 matches for a young melanoma patient were found in the initial query, then 2 controls were selected using computerized randomization. If the initial query did not yield enough matches, then the range of years was broadened to a maximum of 5 years. For thickness, even though an exact match was not always possible, all matches were within the same American Joint Committee on Cancer (AJCC) tumor (T) category, according to the 2002 AJCC staging system.13 Subsequent analysis confirmed that there was no difference in thickness or year of diagnosis between the control group and the pediatric group. Although we attempted to rereview all slides of young melanoma patients, we were unsuccessful in locating slides for a majority of patients. Fifteen cases were rereviewed by our senior dermatopathologist (M.C.M.), and it was determined that the original diagnosis was correct in 100% of these rereviewed patients without any change in measurement of tumor thickness. Thirteen young melanoma patients had primary tumors in which the thickness was unknown. Thus, they could not be included in the matching process and were not taken into account in further comparisons between young and adult melanoma patients. The lack of information on primary tumor thickness was mostly because of the lack of pathology reports/slides. Ten of these 13 patients were diagnosed before 1990.
The young melanoma patients also were analyzed in 2 groups: 1 group of children before their 13th birthday and 1 group of adolescents/young adults past their 13th birthday. Thus, because each young patient with melanoma had their own 2 corresponding control patients, 4 groups were created: a pediatric group of children before their 13th birthday and their controls, and a group of adolescents/young adults past their 13th birthday and their controls. The pediatric group aged <13 years consisted of 8 children; thus, their control group consisted of 16 patients. The group of adolescents/young adults aged ≥13 years consisted of 65 patients; thus, the corresponding control group consisted of 130 patients. In the accompanying tables, these subgroups are presented separately. Statistical analyses always were performed between each subgroup and their controls as well as between all young melanoma patients and all adult controls.
To compensate for the age difference between young melanoma patients and the adult control group, cause-specific survival was compared instead of overall survival. Survival was calculated from the date of initial diagnosis to the date the patient was last known to be alive or to the date of death. The date of first recurrence or of last reported contact during which disease status was known was used as the endpoint for disease-free survival determination. Recurrences were categorized as follows: local, regional, and distant.
All continuous variables were compared using the Mann-Whitney test. Categorical variables were compared using the Pearson chi-square test. Disease-free, cause-specific, and overall survival were determined by the Kaplan-Meier method.14 Patients for whom cause of death was unknown were censored in cause-specific survival analyses. Survival distributions were compared using the log-rank test. Factors that affected cause-specific survival were analyzed using a Cox proportional-hazards model. SAS statistical software (SAS Institute Inc., Cary, NC) was used for statistical analyses. This study was approved by the MGH Institutional Review Board.
General Patient Characteristics
Eight pediatric patients were aged <13 years, and the mean age at diagnosis for the whole group of young patients with melanoma patients was 17.2 years (Table 2). In the adult control group, the mean age was 53.8 years, and the distribution according to sex was similar in each group. The median follow-up for the young and adult populations were 5.4 years and 4.6 years, respectively.
|No. of patients (%)||P||No. of patients (%)||P||P|
|Prepubescent: aged <13 y, n = 8||Controls, n = 16||Adolescent/Young adults: aged ≥13 y, n = 65||Controls, n = 130||Ages ≪13 y vs ≥13 y||Young vs adult patients|
|Male gender||3 (37.5)||6 (37.5)||NS||28 (43.1)||56 (43.1)||NS||NS||NS|
|White||8 (100)||14 (87.5)||59 (90.8)||121 (93.1)|
|Hispanic||0||1 (6.3)||NS||2 (3.1)||0||NS||NS|
|Unknown||0||1 (6.3)||4 (6.2)||8 (6.2)|
|Positive||2 (25)0||2 (12.5)||NS||15 (23.1)||16 (12.3)||NS*||NS||NS|
|Missing||0||2 (12.5)||16 (24.6)||39 (30)|
|Dysplastic nevus syndrome||1 (12.5)||2 (12.5)||NS||17 (26.2)||10 (7.8)||<.001||NS||<.001|
|Multiple primaries†||0||0||NS||2 (3.1)||6 (4.6)||NS||NS||NS|
|T1||1 (12.5)||2 (12.5)||31 (47.7)||62 (47.7)|
|T2||4 (50)||8 (50)||13 (20)||26 (20)|
|T3||2 (25)||4 (25)||NS||15 (23.1)||30 (23.1)||NS||NS||NS|
|T4||1 (12.5)||2 (12.5)||6 (9.2)||12 (9.2)|
|2||1 (12.5)||2 (12.5)||18 (27.7)||31 (23.8)|
|3||1 (12.5)||2 (12.5)||20 (30.7)||36 (27.7)|
|4||4 (50)||11 (68.8)||NS||22 (33.8)||53 (40.8)||NS||NS||MS|
|5||1 (12.5)||1 (6.3)||2 (3.1)||8 (6.2)|
|Missing||1 (12.5)||0||3 (4.6)||2 (1.5)|
|Positive||0||2 (12.5)||NS||8 (12.3)||19 (14.6)||NS||NS||NS|
|Missing||5 (62.5)||8 (50)||38 (58.5)||67 (51.5)|
|SSM||4 (50)||8 (50)||41 (63.1)||81 (62.3)|
|NM||1 (12.5)||3 (18.8)||8 (12.3)||23 (17.7)|
|ALM||0||2 (12.5)||1 (15.4)||3 (2.3)|
|LMM||0||1 (6.3)||NS||0||6 (4.6)||.003||NS||<.001|
|Minimal deviation‡||2 (25)||0||7 (10.8)||0|
|MM NOS||1 (12.5)||2 (12.5)||7 (10.8)||14 (10.8)|
|Unknown||0||0||1 (1.5)||3 (2.3)|
|Mitotic rate per 10 HPF|
|Positive||4 (50)||2 (12.5)||NS||25 (38.5)||33 (25.4)||NS||NS||NS|
|Missing||3 (37.5)||10 (87.5)||24 (36.9)||61 (46.9)|
|Head/neck||1 (12.5)||3 (18.8)||10 (15.4)||22 (16.9)|
|Trunk||4 (50)||6 (37.5)||NS||33 (50.8)||52 (40)||NS||NS||NS|
|Extremity||3 (37.5)||7 (43.8)||22 (33.8)||56 (43.1)|
|I or II||6 (75)||14 (87.5)||NS||54 (83.1)||118 (90.8)||NS||NS||NS|
|III or IV||2 (25)||2 (12.5)||11 (16.9)||12 (9.2)|
There were no young or adult melanoma patients with a history of xeroderma pigmentosum, human immunodeficiency virus, or a history of immune-modulating medication use, although 1 young patient with melanoma was known to have a history of ulcerative colitis, which was in remission at time of diagnosis. Clinically atypical nevi were noted more commonly in pediatric patients compared with adults; and, overall, there was a statistically greater instance of atypical nevi or clinical dysplastic nevus syndrome (as diagnosed by their treating physicians) in the young patients compared with the adult controls.
Of 8 patients with young melanoma aged <13 years, 3 patients (37.5%) had congenital nevi as precursor lesions, 1 patient (12.5%) had a nevus that was not otherwise specified, 1 patient (12.5%) had no precursor lesion, and, in 3 patients (37.5%), it was unknown whether they had precursor lesions. Of 65 young patients with melanoma aged >13 years, 8 patients (12.3%) had congenital nevi as precursor lesions, 17 patients (26.2%) had a nevus that was not otherwise specified, 16 patients (24.6%) had no precursor lesion, and, in 24 patients (36.9%), it was unknown whether they had precursor lesions. Of 146 adult melanoma patients, 4 patients (2.7%) had congenital nevi as precursor lesions, 29 patients (19.8%) had a nevus that was not otherwise specified, 2 patients (1.4%) had lentigo maligna, 40 patients (27.4%) had no precursor lesion, and, in 71 patients (48.6%), it was unknown whether they had precursor lesions.
A family history of malignant melanoma was more common among the group of young patients compared with the matched adult control group, although the difference did not reach statistical significance; 25.6% of pediatric patients had a positive family history compared with only 17.3% of adults. Aside from age, there were no statistically significant differences in general patient characteristics between children (before their 13th birthday) and adolescents/young adults (after their 13th birthday).
Primary Tumor Characteristics
There was no difference between the young melanoma patients and adult controls in the distribution according to Clark level (Table 2). Unfortunately, ulceration data were unavailable in a large number of patients, as the series included patients who were diagnosed with melanoma as early as 1971, and it was not possible to obtain slides for rereview. Among those patients for whom ulceration status was recorded (ie, 34.2% of all young patients with melanoma and 48.6% of adult patients with melanoma), there was no statistically significant difference between the 2 populations.
Among the young melanoma population, the most common histologic classification was superficial spreading melanoma (n = 45, 61.6%), and there were no diagnoses of lentigo maligna melanoma. There were no adult patients with minimal deviation melanomas, and 9 patients in the pediatric group who had a mean tumor thickness of 3.8 mm (range, 0.4–8.8 mm). There was no significant difference between adult controls and the young melanoma population in the distribution of conventional histologic subtypes. There were no statistically significant differences in primary tumor characteristics between prepubescent children and adolescents/young adults. Overall, young melanoma patients were significantly more likely to have histologically identifiable precursor nevi than adults. There was no significant difference with respect to tumor localization between young melanoma patients and adult control patients, although this analysis may have been influenced by the process of matching for tumor thickness.
Stage at Presentation
Stage at presentation was determined after local therapy and any lymph node evaluation procedures were performed. Because of the small numbers of patients in each group, we only analyzed whether there was a difference in localized (stage I or II) versus disseminated (stage III or IV) disease. In this analysis, there was no significant difference between young melanoma patients and their thickness-matched controls, although it should be noted that this comparison was influenced by the process of matching the 2 populations for T-stage. Similar proportions of pediatric and control patients presented with AJCC stage III melanoma.
Lymph Node Evaluation
The frequency and types of initial lymph node surgery were similar between young melanoma patients and adult control patients (Table 3). There were no differences in the results of initial lymph node evaluation between children aged <13 years and adolescents/young adults aged ≥13 years.
|Procedure*||No. of patients (%)||P||No. of patients (%)||Ages <13 y vs ≥13 y||P|
|Prepubescent: aged <13 y, n = 8||Controls, n = 16||Adolescent/Young adults: aged > 13 y, n = 65||Controls, n = 130||P||Young vs adult patients|
|Performed||2 (25)||4 (25)||NS||14 (21.5)||18 (13.8)||NS||NS||NS|
|Positive||2 (100)||2 (50)||NS||5 (35.7)||6 (33.3)||NS||NS||NS|
|Performed||2 (25)||3 (18.8)||NS||7 (10.7)||21 (16.1)||NS||NS||NS|
|Positive||0 (0)||0 (0)||NS||4 (57.1)||4 (19)||.05||NS||NS|
|TLND||0 (0)||0 (0)||NS||2 (3.1)||3 (2.3)||NS||NS||NS|
|Patients with positive lymph nodes||2 (25)||2 (12.5)||NS||11 (16.9)||13 (10)||NS||NS||NS|
Similar proportions of patients in both groups underwent elective lymph node dissection or sentinel lymph node biopsy. Sentinel lymph nodes were analyzed by step-sectioning.15 All young and adult patients with melanoma who had positive sentinel lymph node biopsies subsequently underwent completion lymphadenectomies, only 1 of which revealed additional positive lymph nodes. Of the young patients with melanoma who underwent pathologic staging of clinically negative lymph nodes by either elective lymph node dissection or lymphatic mapping, 11 of 25 patients (44%) had positive lymph nodes compared with 11 of 46 adult patients (23.9%). Young patients with melanoma were significantly more likely to have metastatic melanoma identified in elective lymph node dissection specimens. Although the overall 17.8% incidence of positive lymph nodes at initial presentation in young patients with melanoma was higher than the 9.6% incidence in adult patients, with a relatively small sample size of lymph node-positive patients, this difference only approached statistical significance (P = .08). Of 49 young melanoma patients with T1 or T2 tumors, 8 patients (16%) underwent initial lymph node surgery, and none had positive lymph nodes. Of 24 young patients with melanoma who had T3 or T4 primary tumors, 19 patients (79%) underwent initial lymph node surgery. Thirteen of those patients (68.4%) had positive lymph nodes, including 2 patients who presented with clinically enlarged lymph nodes, and underwent therapeutic lymphadenectomy. Two patients with positive lymph nodes were aged 10 years at time of diagnosis, all others were aged ≥15 years.
Detailed characteristics of the 16 young melanoma patients who underwent sentinel lymph node biopsy are shown in Table 4. The relatively high risk for positive sentinel lymph nodes observed in this cohort was associated with a median tumor thickness of 2.8 mm in patients who underwent lymphatic mapping. All young melanoma patients who had positive sentinel lymph nodes had either AJCC T3 or T4 primary lesions. Three of 5 patients (60%) who had primary tumors that were classified as minimal deviation melanomas had a positive sentinel lymph node biopsy.
|Variable||Prepubescent: aged <13 y||Adolescents/Young adults: aged ≥ 13 y|
|Negative SNB, n = 0||Positive SNB, n = 2||Negative SNB, n = 8||Positive SNB, n = 5|
|Median thickness (range), mm||—||2.9 (2.8–3)||2.3 (1.1–5.2)||2.7 (2.2–6.8)|
|AJCC tumor classification|
|Primary tumor location|
Of the 13 young patients with melanoma for whom information on primary tumor thickness was lacking and who could not be included in the matching process, 2 patients had positive sentinel lymph node biopsies, 1 patient had positive lymph nodes detected during elective lymph node dissection, and 1 patient underwent therapeutic lymph node dissection. Thus, 4 of these 13 patients (30.8%) had positive lymph nodes after initial lymph node evaluation.
Twelve young patients with melanoma and 33 adult control patients experienced at least 1 recurrence (Table 5). No patients had first recurrences at >1 site. The distribution of locations of first recurrence was similar between young and adult melanoma patients, with approximately equal proportions of patients recurring locally (including in transit), in regional lymph nodes, or at distant sites. The median time to first recurrence was 2.1 years for young melanoma patients and 1.4 years for the control group.
|Variable||Prepubescent: aged <13 y, n = 8||Controls, n = 16||P||Adolescent or young adults: aged ≥13 y, n = 65||Controls, n = 130||P||P|
|Ages <13 y vs ≥13 y||Young vs adult patients|
|Median follow-up, y||14.4||4.3||NS||7.3||6.5||NS||NS||NS|
|Local and in-transit||2||2||1||8|
|Regional lymph nodes||1||3||—||3||—||—||—|
|Patients with ≥1 recurrence||3||8||NS||9||25||NS||NS||NS|
|Sites of distant recurrence*|
|Mean time to first recurrence, y|
Among the young melanoma patients with positive sentinel lymph nodes, 1 patient developed a recurrence and ultimately died as a direct result of disease. To date, none of the 9 patients who had negative sentinel lymph node biopsies have recurred.
Because there were no statistically significant differences between prepubescent children (aged <13 years) versus adolescents/young adults (aged ≥13 years) with respect to general patient characteristics, primary tumor characteristics, or results from initial lymph node evaluation, comparisons for survival analyses were only made between the group of young patients with melanoma as a whole and the control group to obtain greater statistical power. There was no difference in outcome between the young melanoma patients and adult control patients with respect to cause-specific or disease-free survival (Figs. 1, 2). The 5-year cause-specific survival rate was 91.3% for young melanoma patients and 86.2% for adults; the 5-year disease-free survival rate was 83.2% in young melanoma patients and 75.7% in adults; and the 10-year cause-specific survival rate was 89.4% and 79.3% for pediatric and adult patients, respectively. There was no difference in survival between young and adult melanoma patients after stratification by stage of disease.
In univariate analysis, significant predictors of cause-specific survival in the group of young melanoma patients were thickness, lymph node status, AJCC stage, and Clark level (Table 6). Because of the small sample size and the low number of total events, multivariate analysis was not undertaken. Not surprisingly, cause-specific survival was significantly worse among young melanoma patients with stage III or IV disease at presentation compared with paients who presented with localized disease (P < .01); at 10 years, the survival rate among young patients with melanoma who presented with localized disease (stages I and II) was 94.4%, whereas the survival rate for young patients with stage III melanoma was 60.1%. No young melanoma patients with tumors that measured ≤1 mm in thickness died from disease.
|Positive lymph nodes||6.092||.014||4.488||1.36–14.78|
|AJCC stage III or IV||2.346||.005||2.346||1.29–4.25|
The incidence of melanoma of the young is increasing, although the increase is a bit more rapidly for adolescents than for young children (aged <10 years).6 The incidence of melanoma in prepubescent children is extremely low; thus, individual institutions have access to clinical information on only a small group from which to draw meaningful conclusions. To our knowledge, the current series of young melanoma patients is the largest single-institutional cohort series published since 1994. Our relatively large number of young melanoma patients allowed us to match them to adult melanoma patients, so that the clinical course of melanoma in young patients could be compared with that in adult patients. In addition, this more modern series included a large number of young patients who underwent surgical and pathologic staging of their lymph nodes.
We also present subgroup analyses of prepubescent melanoma patients (before their 13th birthday). We chose age 13 years as the cut-off point, because this separated prepubescent children from pubescent children. Although Table 1 shows that there is much controversy regarding the cut-off point for defining pediatric melanoma, we believed that, if there were a biologic basis for any difference between very young children and older children with melanoma, then it would be most logical to choose this as the cut-off age. We did not note significant differences between prepubescent and pubescent children in any of the variables that we investigated. However, our groups were small, and we provided subgroup analyses mainly to facilitate compilation and comparison with published and future studies of pediatric melanoma that include only prepubescent subjects.
Melanoma of the young is distinct from adult melanoma with respect to several general demographic and histologic characteristics that, together, suggest that a different set of factors may contribute to the development of melanoma in young individuals. Our findings that a higher percentage of young melanoma patients have positive family histories and have atypical nevi suggest that a stronger predisposing genetic component may be operant in this group. The early age of disease occurrence as well as the absence of the histologic subtype of lentigo maligna melanoma support the logical supposition that sun exposure plays a less important role in the development of melanoma in young individuals. The histologic types of melanoma varied significantly between young patients and adults, with young melanoma patients having a greater frequency of minimal deviation melanoma.
Although the correlation between stage and outcome is well accepted for melanoma of the young,6, 8, 10, 16 there is controversy concerning the aggressiveness of adult melanoma versus melanoma of the young.17 Our results indicate that young melanoma patients have slightly thicker tumors at presentation compared with their adult counterparts, a finding that is consistent with data from previous studies.7, 18, 19 The reasons why young melanoma patients present with thicker tumors are not clear from our study. Because of the relative rarity of melanoma in young individuals, physicians likely have a lower index of clinical suspicion with respect to skin lesions in young patients. Thus, it is reasonable to speculate that young melanoma patients may experience a greater delay in diagnosis relative to adults.
Although it appears from Figures 1 and 2 it that the prognosis for young melanoma patients in fact may be better than for adult patients, we detected no statistically significant difference in recurrence-free or cause-specific survival between young and adult patients. Because tumor thickness is the most important prognostic variable in patients with localized melanoma, prior comparisons of tumor biology between young and adult melanoma patients have been biased by the greater tumor thickness observed in young melanoma patients. The use of thickness-matched controls is 1 strategy to eliminate this bias. To our knowledge, this study is the first to make such a comparison. Thus, our observation that recurrence-free and cause-specific survival in young melanoma patients is similar to that in adult melanoma patients who are matched for thickness is noteworthy and important.
Previously published series of young melanoma patients reported 5-year survival rates of 70% to 80%,12, 20, 21 which were lower than the 91.3% 5-year survival rate observed in our study. In the current study, we analyzed melanoma-specific mortality, a technique that is both logical and necessary when comparing survival rates between adult and pediatric patients.
Another factor that complicates the comparison of survival between published studies of young melanoma patients is that there is no consistency in these reports concerning the age used to define melanoma of the young (Table 1). It is a logical hypothesis that melanoma biology may differ between young melanoma patient subgroups: neonates, prepubescent children, adolescents, and young adults. Although our study was not large enough to draw meaningful conclusions with respect to this hypothesis, we did provide subgroup analyses between prepubescent melanoma patients before their 13th birthday and adolescents/young adults past their 13th birthday to facilitate comparisons and data pooling with other studies.
It has been suggested that young patients with melanoma may experience earlier recurrences.10, 11 However, our data indicate no differences compared with thickness-matched adults in either the frequency or nature of recurrences.
A relatively high frequency of regional lymph node metastases in young individuals with melanoma has was reported in a study of childhood melanomas collected in the National Cancer Data Base, in which a 26.2% rate of regional metastases was observed in children ages 10 years to 18 years.22 Because of the rarity of melanoma of the young as well as the relatively recent development of lymphatic mapping, little is known about the utility of sentinel lymph node biopsy in young melanoma patients. Our data strongly support a role for sentinel lymph node biopsy in the identification of metastatic melanoma in young melanoma patients. Of the 18 patients aged <21 years who underwent sentinel lymph node biopsy at the MGH, 50% were positive, which was substantially higher than the 25% rate of positive lymph nodes in a comparably sized series recently published by Toro et al.23 This large difference probably is because the mean tumor thickness in our group of patients who underwent sentinel lymph node biopsy was substantially greater–2.9 mm compared with 1.7 mm in the series by Toro et al. Nevertheless, these data support the applicability of sentinel lymph node biopsy to the population of young patients with melanoma.
It is worth noting that 60% of patients with minimal deviation melanoma (also referred to as borderline lesions or MELTUMPs)24 in our series had positive lymph nodes. These results are similar to those recently published by Lohmann et al., who reported that 50% of patients with cutaneous lesions of diagnostic uncertainty had positive sentinel lymph nodes.25 Thus, although these datasets do not allow for definitive conclusions, they suggest that sentinel lymph node biopsy also may be of utility in young individuals with minimal deviation melanoma.
Several limitations of this study deserve reiteration: 1) the study was retrospective; 2) the study population represents patients who were treated in a single tertiary-care institution, which may have been a basis for selection bias; 3) despite the institution's recognized, long-standing expertise in melanoma dermatopathology, this study ranged over 3 decades, and not all slides were available for rereview; 4) young patients with melanoma who had lesions to which a tumor thickness was not assigned were excluded from the study by virtue of the methodology of matching for tumor thickness; and 5) prospectively defined criteria for lymph node evaluation in children with melanoma were not in place over the years of this study. Young melanoma patients without thickness information were not matched to adults without thickness information, because the reasons for the lack of information on primary tumor thickness were not known. Thus, matching these patients could result in matched young and adult patients with 2 very different primary tumors, which would obscure comparisons between the groups. Nevertheless, the value of case matching in this study is the ability to compare more accurately melanoma in young patients with melanoma in adults.
This study highlights important similarities and differences in melanoma biology between adults and young patients with melanoma. The prognostic factors that have been identified for adult melanoma are similar to those observed for melanoma in younger patients. Our study demonstrated that the survival of young patients with melanoma was similar to the survival of adult thickness-matched controls. Finally, our data suggest there is a clear role for sentinel lymph node biopsy in young patients with melanoma, although further study is warranted to help establish the exact indications for sentinel lymph node biopsy in this group of young patients.
- 22Childhood melanoma: demographics and clinical presentation (P-127). Presented in part at: 6th World Congress on Melanoma, September 6–10, 2005, Vancouver, British Columbia, Canada., , , .
- 24Benign pigmented lesions and malignant melanoma. In: ElderDE, ElenitsasR, JohnsonBL, MurphyGF, eds. Lever's Histopathology of the Skin, 9th ed. Lippincott, Williams and Wilkins; 2004: 715–787., , , .
- 27Cutaneous melanoma in childhood: incidence and prognosis. Australas J Dermatol. 1997; 38( suppl 1): S44–S48., , , .