Discussion:
We report a case series of pediatric melanoma adding to the growing data on the diagnosis and management of this rare malignancy. Sixty-five percent of the patients in this cohort were diagnosed with melanoma in adolescence (age ≥ 11 years). Moreover, both patients that progressed were older than 11 years old at diagnosis, supporting the existing literature stating that adolescent melanoma has a more aggressive disease course compared to pre-pubertal melanoma5.
Outcomes were favorable, with event-free and overall survival of 91% and 100%, respectively. Our data suggest that including nivolumab in the initial treatment of melanomas diagnosed at Stage III may help avoid recurrence or progression of melanoma. Four out of the 9 patients diagnosed at Stage III did have Spitzoid melanoma and 3 of them were diagnosed in younger childhood (age < 11 years). Multiple other studies similarly found that Spitzoid melanomas occur more commonly in pre-pubertal age 5-8. It is generally thought that patients with Spitzoid melanoma have higher rates of positive SLN despite an excellent overall prognosis8. Additionally, we observed no progression or recurrence of melanoma in the remaining 3 patients diagnosed at Stage III with either melanoma of unspecified subtype (n=2) or severely atypical spindle cell melanocytic tumor (n=1). We also observed that among the 19 patients who had an SLNB, prepubertal patients (< 10 years of age) had a lower percentage of SLN positivity (57%) when compared to adolescent patients (75%). This differs from the findings by Moore-Olufemi et al.,Pol-Rodriquez et al. and El Sharouni et al. , but our small sample size is much smaller7,9,10.
Nivolumab therapy was initiated in 7 patients, and was well tolerated in most (5/7, 71%). Patients experienced either no adverse events (n=3) or Grade 1 or 2 toxicities including mild dermatitis (n=1) and chills (n=1) that improved with steroid administration. Nivolumab treatment was discontinued in one patient who had a grade 3 lichenoid drug eruption and withheld in 1 patient who developed hyper-eosinophilia. Although the safety of nivolumab has not been extensively studied in the pediatric population, a study by Davis et al. showed that nivolumab is safe and well-tolerated in children and young adults11. A small case series (n=10) that included 5 patients with skin melanoma further supports the safety and efficacy of nivolumab in pediatric patients. Complete remission was achieved in 3 patients with melanoma (60%)1. In our cohort, we observed complete remission in all the patients (100%) treated with nivolumab.
Melanoma in pediatric patients can present differently than in adults: 39% of the pediatric population in this study were diagnosed with Stage III melanoma which aligns with previous studies suggesting children are diagnosed at later stages than adults8. This is most likely due to the relative rarity of melanoma and paucity of research on the clinical and histopathological features indicative of melanoma in the pediatric population. Our case series also suggests that ulceration is less of an adverse prognostic finding in children than it is in adults as only 3 patients had ulcerated lesions. Of those 3 patients, 2 received nivolumab and all three achieved CR after initial therapy. Current literature supports these findings with similar reports of lower frequencies of ulceration histologically12.
While Sharouni et al ., observed that a Breslow thickness > 4 mm predicted worse survival, in our case series, none of the patients with a reported Breslow thickness of > 4 mm had recurrence or progression7. Of the 2 patients that had a recurrence, 1 was diagnosed with a Breslow thickness of 2 mm, while the other did not have a reported Breslow thickness. Sharouniet al ., also noted that all 3 children who recurred had melanomas located on the head or neck. Only 3 of our patients were diagnosed with head/neck melanomas and of those 1 (33.3%) developed a recurrence.
Another factor influencing the prognosis of pediatric melanoma is its genomic subtype. Non-spitzoid melanomas typically have evidence of UV exposure with a large number of C>T transitions at dipyrimidine sites, BRAF and TERT promoter mutations and inactivation of PTEN 13. Of the 4 patients diagnosed with a non-spitzoid type melanoma who underwent genetic testing, none had these mutations. Instead, 2 patients had negativeBRAF mutations, 1 only underwent a chromosomal microarray andp53 sequencing and deletion/duplication testing and 1 underwent a diagnostics gene expression profile that did not include BRAF . These findings are limited by the size of our case series. Additionally, most of our patients with non-spitzoid melanoma did not undergo genetic testing (12/18, 67%). All patients diagnosed with melanoma should undergo genetic testing as taking a combined genomic and clinicopathologic approach has been shown to optimize diagnosis and treatment13. Pappo et al. , demonstrated that the genomic differences between spitzoid melanoma, conventional melanoma and melanoma arising from a congenital nevus can influence treatment and thus clinical outcomes13.
Lastly, our findings differ from existing literature suggesting that the majority of melanomas arise de novo rather than in conjunction with a pre-existing nevus5,14. While there is no universally agreed upon histological definition of nevus-associated melanoma, pathology-based studies have found that 20% to 30% of melanomas contain nevus cells in histologic continuity with melanoma15. In our study, only two of the melanomas had evidence of normal nested melanocytes to support an origin in an acquired nevus. This is similar to adult melanoma where research indicates most melanomas develop as new growths and supports the findings that the lifetime risk for an individual nevus to transform into melanoma is quite low, particularly in younger individuals6. However, while nevus-associated melanomas are not common, they should not be considered incidental phenomena as incidence is shown to be higher than what would be expected by chance alone5.