Background
Cutaneous melanocytic lesions represent a wide spectrum, ranging from benign to overtly malignant. This composite group includes atypical melanocytic tumors (AMTs), mostly occurring in children, adolescents, and young adults,1,2 whose biologic behavior and metastatic risk have been controversial and difficult to predict since their initial description by Spitz.1 In fact, these tumors have some attributes of malignancy, such as cytologic atypia and mitotic activity, that are considered of insufficient severity to justify a diagnosis of frank malignancy. They were later defined as “borderline melanomas”3,4 or “minimal deviation melanomas”5–7 because of the interpretative problems they presented to physicians in distinguishing them from benign melanocytic nevi. Smith et al8 and Barnhill et al,9 in different analyses, described a kind of Spitz nevus showing atypia and the capability of metastasizing, and they classified this neoplasm as Spitz tumor or atypical Spitz tumor. In subsequent studies, Barnhill10 and Elder and Xu11 introduced the term “melanocytic tumor of uncertain malignant potential” or “melanocytic proliferation with indeterminate biologic potential,” a descriptive diagnosis generally accompanied by a differential diagnosis, to underline the uncertainties in predicting the clinical behavior. Cerroni et al12 studied a series of melanocytic tumors of uncertain malignant potential and found that the presence of increased mitotic activity, mitoses near the base of the lesion, and an inflammatory reaction were statistically significant parameters unfavorably affecting outcome.
To integrate the conventional morphologic assessment methods for these difficult lesions, some authors proposed sentinel node (SN) status assessment (negative vs positive).13,14 The MSLT-1 trial demonstrated the utility of sentinel node biopsy (SNB) as a staging tool and found that SN status was the most powerful prognostic factor in patients with melanoma; this trial was the primary reference study supporting the use of SNB in cutaneous melanoma.15 The rationale behind the use of SNB also in AMT was the assumption that the detection of melanocytic cells in the draining regional lymph node basin generally indicates malignancy.14
Although some authors found that SN status did not seem to be useful in predicting outcome in atypical spitzoid tumors, a subset of AMT generally occurring in young patients and associated with a good prognosis, other studies emphasized the possible diagnostic and prognostic role of SN status in AMTs, particularly in older patients and nonspitzoid lesions.16–18 In different analyses, Gerami et al19,20 and Shen et al21 showed that the integration of molecular and histopathologic data improved the risk assessment of spitzoid tumors, and the use of fluorescence in situ hybridization (FISH) has suggested that specific genomic alterations affect the prognosis of atypical spitzoid tumors. Because the few published analyses of AMT are limited by the small series of cases and/or short follow-up,22,23 further characterization of these lesions could be of relevance. We therefore analyzed a large series of patients with a long follow-up to better define the role of SN status as a prognostic factor and to identify other potential clinicopathologic predictors of survival.
Patients and Methods
Clinicopathologic Features and FISH Analysis
A total of 238 patients eligible for the study analysis were consecutively treated from 2000 through 2010 at the Istituto Nazionale dei Tumori, Milan, University Hospitals of Brescia, Modena, Parma, and Pavia, and General Hospitals of Cremona, Macerata, Siracusa, and Trapani, all in Italy; Istituto Oncologico Svizzera Italiana, Ospedale Regionale Bellinzona e Valli, Bellinzona, Switzerland; University Hospital of Heraklion, Greece; and Queen Mary University, London, United Kingdom. Clinicopathologic data were retrieved from prospectively maintained databases. All cases were reviewed independently by a pool of dermatopathologists (B.V., M.C., M.B., A.P., S.L.P.), with disagreements resolved through discussion. Cases were subcategorized into 2 groups: spitzoid and nonspitzoid. Spitzoid tumors had a morphology resembling a Spitz tumor (large melanocytic cells, spindle cells and/or epithelioid cells, sharp lateral demarcation of the nests of intraepidermal melanocytes, maturity of cells, and rarity of individual melanocytes high above the basal cell layer).12,24 Nonspitzoid tumors included nevoid lesions (resembling a conventional or dysplastic nevus), pigmented epithelioid melanocytomas, deep penetrating nevi, and atypical blue nevi.12 Patient age and sex, histologic subtype, tumor thickness, ulceration, mitotic rate (MR)/mm2, presence of mitotic figures at the base of the lesion (marginal mitoses, within 0.25 mm of the dermal margin of the lesion),25 lymphovascular invasion (LVI), and tumor-infiltrating lymphocytes were recorded for all cases. Morphologic and FISH analysis provided support for a diagnosis of AMT according to the WHO classification criteria.26 Two hybridizations were performed: the 4-probe FISH assay targeting 6p25, 6q23, 11q13, and Cep6 was used for the first hybridization, and the 4-probe FISH assay targeting 6p25, 9p21, 11q13, and 8q24 was used for the second.19,20 The pool of dermatopathologists, having experience in molecular diagnostics and FISH testing, performed the FISH analyses. A patient was considered to have a positive FISH result based on the criteria and cutoff values used by Gerami et al.19,20,27 All patients received an initial diagnostic biopsy followed by wide (1 cm) excision; SNB was performed in 139 patients (58.4%) after the benefits and potential harms had been discussed with them. A total of 95 patients eligible for SNB declined the procedure, whereas 4 did not undergo it due to the presence of comorbidities. Pathologic assessment of SNs was performed according to the EORTC protocol.28 Patients with positive SNs were offered completion lymph node dissection (CLND) as additional therapy.
Statistical Methods
Clinicopathologic characteristics were recorded according to SNB (performed vs not performed) for the whole series of patients, according to SN status (positive vs negative) in the group undergoing SNB, and according to CLND (performed vs not performed) in the SN-positive group. The primary endpoints of the study were overall survival (OS) and progression-free survival (PFS). The OS and PFS curves were estimated using the Kaplan-Meier method29 and compared using the log-rank test. Inverse probability of treatment weighting (IPTW)30 was applied to groups affected by a selection bias. Standardized mean difference (SMD)31 was used to evaluate the balancing of clinicopathologic characteristics. The association between PFS and all investigated covariates was assessed via univariable Cox models.
We applied the elastic net penalization method in the Cox model to perform covariate selection for PFS between all investigated features in the SNB subgroup.32 More details on the statistical methods are provided in the supplemental eAppendix 1 (available with this article at JNCCN.org).
Results
Patient and Disease Characteristics
Clinicopathologic characteristics of the 238 patients with a diagnosis of AMT are summarized in Table 1. Patient age ranged from 3 to 53 years; most AMTs (52.5%) were diagnosed in young patients (age ≤20 years). The most frequent primary site was the lower limbs (38.2%). Spitzoid tumors numbered 81 (34%); nonspitzoid tumors numbered 157 (66%) and included nevoid lesions (n=89; 37.4%), pigmented epithelioid melanocytomas (n=19; 8%), deep penetrating nevi (n=43; 18.1%), and atypical blue nevi (n=6; 2.5%). Mitoses at the base of the lesion and LVI were detected in 29% and 23.5% of cases, respectively. In 28.6% of cases, the FISH test showed 9p21 loss.
Clinicopathologic Characteristics
Clinicopathologic characteristics of the 238 patients according to SNB are summarized in Table 2. The group that underwent SNB and the non-SNB group did not differ with respect to sex, tumor site, tumor thickness, ulceration, MR/mm2, and tumor-infiltrating lymphocytes. The SNB group differed significantly from the non-SNB group in age (5.8% vs 38.4% aged ≤10 years; 23.0% vs 47.5% aged 11–20 years; and 71.2% vs 14.1% aged >20 years). Histologic subtype differed in the SNB group compared with the non-SNB group: 75.5% of patients who underwent SNB had nonspitzoid tumors and 24.5% had spitzoid lesions; among those in the non-SNB group, 52.5% had nonspitzoid tumors and 47.5% spitzoid lesions. The group that underwent SNB showed significantly higher rates of mitoses at the base of the lesion (33.8% vs 22.2%), LVI (28.1% vs 17.2%), and 9p21 loss (33.8% vs 21.2%) than the non-SNB group.
Clinicopathologic Characteristics According to SNB, Before and After Propensity Score Weighting
Of the 139 patients that underwent SNB and the 99 that did not, 55 (39.6%) and 26 (26.3%), respectively, showed FISH abnormalities. SNB revealed metastatic involvement in 37 (26.6%) patients. Clinicopathologic characteristics of the 139 patients according to SN status are summarized in supplemental eTable 1. Patients who were SN-positive significantly differed from those who were SN-negative in median number of mitoses/mm2 (5 vs 2, respectively; SMD, 1.946), presence of mitoses at the base of the lesion (86.5% vs 14.7%, respectively; SMD, 2.062), and LVI (75.7% vs 10.8%; SMD, 1.733). Patients who were SN-positive showed significantly higher rates of 9p21 loss than those who were SN-negative (45.9% vs 29.4%; SMD, 1.587). Of the 37 patients who were SN-positive, 19 (51.4%) showed FISH abnormalities; 21 (56.7%) of 37 underwent CLND, among whom 1 (4.8%) showed further positive non-SNs. Patients who were SN-positive were not offered further treatment (ie, systemic therapies) other than CLND. After IPTW, SMD reached values <0.1 for all covariates included in the propensity score model.
Survival Analysis
Median follow-up was 126 months (interquartile range, 104–157 months). The 10-year OS and PFS were 97% (95% CI, 94.9%–99.2%) and 82.2% (95% CI, 77.3%–87.3%), respectively (Figure 1). The group that did not undergo SNB showed slightly better 10-year OS and PFS probabilities than the group that did: 100% versus 94.9% (95% CI, 91.3%–98.7%) and 87.8% (95% CI, 81.6%–94.5%) versus 77.9% (95% CI, 71.0%–85.4%), respectively (Figure 2A, C). Weighted comparisons showed that 10-year OS and PFS were 100% versus 96.8% (95% CI, 93.3%–100%) and 88.5% (95% CI, 82.4%–95.1%) versus 84.6% (95% CI, 77.4%–92.4%), respectively; these differences were not statistically significant (Figure 2B, D). Patients who were SN-negative and those who were SN-positive had similar 10-year OS (96.0% [95% CI, 92.3%–99.9%] vs 91.9% [95% CI, 83.5%–100%]; P=.332), whereas patients who were SN-negative showed higher 10-year PFS (84.2% [95% CI, 77.4%–91.6%] vs 63.6% [95% CI, 49.6%–81.7%]; P=.045) (supplemental eFigure 1).
Atypical melanocytic tumors. Kaplan-Meier curves for (A) OS and (B) PFS in the whole series.
Abbreviations: OS, overall survival; PFS, progression-free survival.
aNumber of patients at risk for an event.
Citation: Journal of the National Comprehensive Cancer Network J Natl Compr Canc Netw 18, 10; 10.6004/jnccn.2020.7582
Atypical melanocytic tumors. Kaplan-Meier curves for (A) OS and (B) PFS in the whole series.
Abbreviations: OS, overall survival; PFS, progression-free survival.
aNumber of patients at risk for an event.
Citation: Journal of the National Comprehensive Cancer Network J Natl Compr Canc Netw 18, 10; 10.6004/jnccn.2020.7582
Atypical melanocytic tumors. Kaplan-Meier curves for (A) OS and (B) PFS in the whole series.
Abbreviations: OS, overall survival; PFS, progression-free survival.
aNumber of patients at risk for an event.
Citation: Journal of the National Comprehensive Cancer Network J Natl Compr Canc Netw 18, 10; 10.6004/jnccn.2020.7582
Unweighted and weighted Kaplan-Meier curves for (A, B) OS and (C, D) PFS according to sentinel node biopsy (performed or not performed).
Abbreviations: OS, overall survival; PFS, progression-free survival; SNB, sentinel node biopsy.
aWeighted number of patients at risk for an event (rounded to the integer in the weighted analysis).
Citation: Journal of the National Comprehensive Cancer Network J Natl Compr Canc Netw 18, 10; 10.6004/jnccn.2020.7582
Unweighted and weighted Kaplan-Meier curves for (A, B) OS and (C, D) PFS according to sentinel node biopsy (performed or not performed).
Abbreviations: OS, overall survival; PFS, progression-free survival; SNB, sentinel node biopsy.
aWeighted number of patients at risk for an event (rounded to the integer in the weighted analysis).
Citation: Journal of the National Comprehensive Cancer Network J Natl Compr Canc Netw 18, 10; 10.6004/jnccn.2020.7582
Unweighted and weighted Kaplan-Meier curves for (A, B) OS and (C, D) PFS according to sentinel node biopsy (performed or not performed).
Abbreviations: OS, overall survival; PFS, progression-free survival; SNB, sentinel node biopsy.
aWeighted number of patients at risk for an event (rounded to the integer in the weighted analysis).
Citation: Journal of the National Comprehensive Cancer Network J Natl Compr Canc Netw 18, 10; 10.6004/jnccn.2020.7582
Supplemental eTable 2 summarizes the clinicopathologic characteristics of patients who were SN-positive according to CLND before and after propensity score weighting. Patients undergoing CLND showed observed 10-year OS and PFS probabilities similar to those of patients who did not undergo the procedure: 90.5% (95% CI, 78.8%–100%) versus 93.8% (95% CI, 82.6%–100%) and 61.0 (95% CI, 43.0%–86.3%) versus 67.0% (95% CI, 46.9%–95.8%), respectively (supplemental eFigure 2A, C). After weighting, the 10-year OS and PFS probabilities were 85.8% (95% CI, 65.6%–100%) versus 97.9% (95% CI, 89.0%–100%) and 64.9% (95% CI, 39.6%–100%) versus 61.0% (95% CI, 35.7%–100%), respectively (supplemental eFigure 2B, D).
There were 4 local recurrences, 43 regional recurrences, and 8 distant metastases as first events. Six patients died as a result of disease progression. Of the patients who developed regional recurrences, 34 (79.1%) had not undergone SNB, 6 (13.9%) underwent SNB and were SN-positive, and 3 (7%) underwent SNB and were SN-negative. Of the 51 patients who developed regional and distant metastases, 40 (78.4%) showed FISH abnormalities. Five patients who were SN-negative and 3 who were SN-positive, all belonging to the nonspitzoid group, developed distant metastases beyond the regional nodal basin; of these patients, 4 developed distant metastases after regional recurrence and 4 had distant spread in the absence of previous regional node metastases. The 8 patients who developed distant disease—6 of whom died of their disease—had an age range from 28 to 42 years; the histologic diagnosis was nevoid lesion in 6 cases, deep penetrating nevus in 1 case, and atypical blue nevus in 1 case. Three patients developed another malignancy, and 1 died of causes unrelated to AMT.
In univariable analysis of the whole series, age ≤10 versus >20 years and age 11–20 versus >20 years; MR ≥4/mm2; mitoses at the base of the lesion; LVI; and 9p21 loss (all P<.001) were strongly associated with a worse PFS, whereas SN status (P=.024) was mildly associated. All of the strongly associated variables were observed in the group in which SNB was performed (Table 3). The hazard ratio of PFS increased from 0 to 4 for MR and then reached a plateau, whereas for age, it increased up to 30 years and then decreased. Among the 55 patients with recurrences, 37 (67.3%) had MR ≥4, 35 (63.6%) had mitoses at the base of the lesion, 33 (60%) had LVI, and 34 (61.8%) showed 9p21 loss. Of the 183 patients without recurrences 21 (11.5%) had MR ≥4, 19 (10.4%) had mitoses at the base of the lesion, 24 (13.1%) had LVI, and 26 (14.2%) showed 9p21 loss. After IPTW, SMD reached values <0.1 for all covariates included in the PS model (data not shown).
Results of Univariable Cox Model Analysis for Progression-Free Survival
Predictive Covariate Selection
The elastic net penalized Cox model selection procedure led to the selection of higher MR, mitoses at the base of the lesion, LVI, and 9p21 loss as the most relevant covariates for PFS prediction. The sparsity of the events and the high correlation issue in our series prevented us from developing a reliable predictive model based on the selected variables.
Discussion
We found a positive SN rate of 26.6%, in line with data reported in other studies in which a higher proportion of positive SNs was observed in patients with AMTs compared with those with conventional melanomas.15,33 However, our study did not clearly show that SN status was a predictor of OS. Although most of the variables associated with a worse prognosis (MR/mm2 ≥4, mitoses at the base of the lesion, LVI, and 9p21 loss) were prevalent in patients who were SN-positive and could be considered to identify individuals eligible for SNB, in our analyses, SN status did not seem to provide the same important prognostic information for AMT as observed in cutaneous melanoma.15 Because our findings are not strong enough to mandate SNB for AMT currently, ultrasound imaging could be an alternative approach during follow-up.34 However, the number of positive SNs in our series was small, and further analyses of larger series of patients are needed to better evaluate the prognostic role of SN status.
Previous studies in small series with a shorter follow-up focusing on the role of SN status yielded controversial results. Busam et al35 reported that children and teenagers with spitzoid AMTs and positive SNs had a less aggressive clinical course than patients with SN-positive melanoma. Ghazi et al,16 in a study of patients with AMTs with a median age of 24 years, showed that lymph node assessment did not predict the outcome of spitzoid AMTs. Gamblin et al36 and Murali et al,37 in different analyses, reached the opposite conclusion, arguing that SNB could offer a means for assessing the metastatic potential of spitzoid AMTs and represented an adjunct tool in patient management.
In our series, the 6 deaths as a result of disease progression confirmed that AMT should be considered a provisional diagnosis pending molecular and genetic analyses, expert review, and outcome; all our data were used to predict the likelihood of melanoma in these patients. Our analysis showed that MR ≥4 and the presence of mitoses at the base of the lesion correlated with a worse prognosis. These factors seemed to be indicators of a more aggressive clinical behavior and to be associated with a higher probability of distant metastatic spread beyond the regional nodal basin. Mitosis should be taken into consideration as a prognostic factor in AMTs, and its relevance should be carefully evaluated in clinical decision-making.
Our data are in line with various previous reports in which these variables were heavily weighted for their diagnostic and prognostic significance. Cerroni et al12 identified only 3 parameters that were statistically different between 2 groups of AMTs with favorable and unfavorable behavior based on clinical follow-up: high MR, mitoses near the base of the lesion, and an inflammatory infiltrate. In a more recent study, Gerami et al38 assessed interobserver agreement in diagnosis by 13 expert dermatopathologists for 75 atypical spitzoid tumors and confirmed that frequent mitoses and deep mitoses were histomorphologic features that correlated with disease progression. In our analysis, no statistically significant association was found between an inflammatory infiltrate and patient outcome.
Our study also provides evidence that LVI was associated with shorter survival. We found that the tumor-associated lymphatic network constituted a potential criterion in the selection of high-risk patients who should be candidates for careful follow-up. Another investigation into the prognostic role of LVI in AMT conducted by Abraham et al39 in a small number of lesions related this parameter to a poorer prognosis.
Furthermore, our findings showed that the groups of younger patients (aged <10 years and 11–20 years) with AMT had a longer survival. Spatz et al40 analyzed various parameters, including age, to define a grading system for risk stratification of atypical Spitz tumors. They found that diagnosis at age >10 years carried a likelihood ratio >1.50, so this variable was used for the grading system. The low rate of local recurrences (1.7%) in our series confirmed that wide surgical excision with 1-cm margins was adequate to control local recurrence. Ludgate et al,41 in another study of 67 patients with atypical Spitz tumors, observed 1 local recurrence and recommended 1-cm excision margins to prevent local recurrences.
Finally, in our analysis, AMTs were classified into specific risk categories based on the cytogenetic changes determined by FISH. In the survival analysis, we found that 9p21 loss was associated with a higher risk of distant spread. Gerami et al20 evaluated 2 small subgroups of high-risk melanocytic neoplasms in children—including spitzoid melanomas, atypical Spitz tumors with chromosomal copy number changes, and conventional melanomas—and found that the presence of a homozygous deletion of 9p21 and a positive SNB in atypical Spitz tumors was associated with a higher risk of systemic metastasis and death. Another analysis by Gerami et al19 showed that also the presence of 6p25 and/or 11q13 gain was significantly associated with tumor progression beyond the SN when compared with FISH-negative tumors. In our series, 6p25 and 11q13 gains were not associated with a higher risk of distant spread. Additional techniques to assess molecular genetic alterations, such as comparative genomic hybridization, next-generation sequencing, microRNA and mRNA analysis, and mass spectrometric imaging, will help improve the differential diagnosis and clinical management of AMTs in the near future.42
Conclusions
Our findings showed that age >20 years, MR ≥4, mitoses at the base of the lesion, LVI, and FISH evidence of 9p21 loss were all independent predictors of more aggressive tumor behavior. SN status did not prove to be a clear prognostic factor. Our retrospective analysis may contribute to establishing a practical framework for managing patients with AMTs in which clinical decisions are supported by identifying at-risk patients who may benefit from more frequent and longer-term follow-up.
Acknowledgments
The authors thank Marije de Jager for help with the English language, and acknowledge Antonio Florita and Elena Morittu for assistance in manuscript submission.
References
- 2.↑
Kernen JA, Ackerman LV. Spindle cell nevi and epithelioid cell nevi (so-called juvenile melanomas) in children and adults: a clinicopathological study of 27 cases. Cancer 1960;13:612–625.
- 3.↑
Reed RJ, Ichinose H, Clark WH Jr, et al.. Common and uncommon melanocytic nevi and borderline melanomas. Semin Oncol 1975;2:119–147.
- 4.↑
Sagebiel RW. Histopathology of borderline and early malignant melanomas. Am J Surg Pathol 1979;3:543–552.
- 5.↑
Muhlbauer JE, Margolis RJ, Mihm MC Jr, et al.. Minimal deviation melanoma: a histologic variant of cutaneous malignant melanoma in its vertical growth phase. J Invest Dermatol 1983;80(6 Suppl):63s–65s.
- 6.↑
Phillips ME, Margolis RJ, Merot Y, et al.. The spectrum of minimal deviation melanoma: a clinicopathologic study of 21 cases. Hum Pathol 1986;17:796–806.
- 8.↑
Smith KJ, Barrett TL, Skelton HG III, et al.. Spindle cell and epithelioid cell nevi with atypia and metastasis (malignant Spitz nevus). Am J Surg Pathol 1989;13:931–939.
- 9.↑
Barnhill RL, Argenyi ZB, From L, et al.. Atypical Spitz nevi/tumors: lack of consensus for diagnosis, discrimination from melanoma, and prediction of outcome. Hum Pathol 1999;30:513–520.
- 10.↑
Barnhill RL. Malignant melanoma. In: Barnhill RL, Piepkorn M, Busam KJ, eds. Pathology of Melanocytic Nevi and Malignant Melanoma. 2nd ed. New York, NY: Springer; 2004:238–356.
- 11.↑
Elder DE, Xu X. The approach to the patient with a difficult melanocytic lesion. Pathology 2004;36:428–434.
- 12.↑
Cerroni L, Barnhill R, Elder D, et al.. Melanocytic tumors of uncertain malignant potential: results of a tutorial held at the XXIX Symposium of the International Society of Dermatopathology in Graz, October 2008. Am J Surg Pathol 2010;34:314–326.
- 13.↑
Kelley SW, Cockerell CJ. Sentinel lymph node biopsy as an adjunct to management of histologically difficult to diagnose melanocytic lesions: a proposal. J Am Acad Dermatol 2000;42:527–530.
- 14.↑
Lohmann CM, Coit DG, Brady MS, et al.. Sentinel lymph node biopsy in patients with diagnostically controversial spitzoid melanocytic tumors. Am J Surg Pathol 2002;26:47–55.
- 15.↑
Morton DL, Thompson JF, Cochran AJ, et al.. Final trial report of sentinel-node biopsy versus nodal observation in melanoma. N Engl J Med 2014;370:599–609.
- 16.↑
Ghazi B, Carlson GW, Murray DR, et al.. Utility of lymph node assessment for atypical spitzoid melanocytic neoplasms. Ann Surg Oncol 2010;17:2471–2475.
- 17.↑
Sepehr A, Chao E, Trefrey B, et al.. Long-term outcome of Spitz-type melanocytic tumors. Arch Dermatol 2011;147:1173–1179.
- 18.↑
Caracò C, Mozzillo N, Di Monta G, et al.. Sentinel lymph node biopsy in atypical Spitz nevi: is it useful? Eur J Surg Oncol 2012;38:932–935.
- 19.↑
Gerami P, Scolyer RA, Xu X, et al.. Risk assessment for atypical spitzoid melanocytic neoplasms using FISH to identify chromosomal copy number aberrations. Am J Surg Pathol 2013;37:676–684.
- 20.↑
Gerami P, Cooper C, Bajaj S, et al.. Outcomes of atypical Spitz tumors with chromosomal copy number aberrations and conventional melanomas in children. Am J Surg Pathol 2013;37:1387–1394.
- 21.↑
Shen L, Cooper C, Bajaj S, et al.. Atypical Spitz tumors with 6q23 deletions: a clinical, histological, and molecular study. Am J Dermatopathol 2013;35:804–812.
- 22.↑
Massi D, Tomasini C, Senetta R, et al.. Atypical Spitz tumors in patients younger than 18 years. J Am Acad Dermatol 2015;72:37–46.
- 23.↑
Lee CY, Sholl LM, Zhang B, et al.. Atypical spitzoid neoplasms in childhood: a molecular and outcome study. Am J Dermatopathol 2017;39:181–186.
- 24.↑
Paniago-Pereira C, Maize JC, Ackerman AB. Nevus of large spindle and/or epithelioid cells (Spitz’s nevus). Arch Dermatol 1978;114:1811–1823.
- 25.↑
Crotty KA. Spitz naevus: histological features and distinction from malignant melanoma. Australas J Dermatol 1997;38(S1):S49–53.
- 26.↑
Elder DE, Massi D, Scolyer RA, et al., eds. WHO Classification of Skin Tumours, 4th ed. Vol. 11. Lyon, France: IARC; 2018.
- 27.↑
Gerami P, Wass A, Mafee M, et al.. Fluorescence in situ hybridization for distinguishing nevoid melanomas from mitotically active nevi. Am J Surg Pathol 2009;33:1783–1788.
- 28.↑
Cook MG, Massi D, Szumera-Ciećkiewicz A, et al.. An updated European Organisation for Research and Treatment of Cancer (EORTC) protocol for pathological evaluation of sentinel lymph nodes for melanoma. Eur J Cancer 2019;114:1–7.
- 29.↑
Schemper M, Smith TL. A note on quantifying follow-up in studies of failure time. Control Clin Trials 1996;17:343–346.
- 30.↑
Austin PC, Stuart EA. Moving towards best practice when using inverse probability of treatment weighting (IPTW) using the propensity score to estimate causal treatment effects in observational studies. Stat Med 2015;34:3661–3679.
- 31.↑
Flury BK, Riedwyl H. Standard distance in univariate and multivariate analysis. Am Stat 1986;40:249–251.
- 32.↑
Zou H, Hastie T. Regularization and variable selection via the elastic net. J R Stat Soc Series B Stat Methodol 2005;67:301–320.
- 33.↑
Hung T, Piris A, Lobo A, et al.. Sentinel lymph node metastasis is not predictive of poor outcome in patients with problematic spitzoid melanocytic tumors. Hum Pathol 2013;44:87–94.
- 34.↑
Hayes AJ, Moskovic E, O’Meara K, et al.. Prospective cohort study of ultrasound surveillance of regional lymph nodes in patients with intermediate-risk cutaneous melanoma. Br J Surg 2019;106:729–734.
- 35.↑
Busam KJ, Murali R, Pulitzer M, et al.. Atypical spitzoid melanocytic tumors with positive sentinel lymph nodes in children and teenagers, and comparison with histologically unambiguous and lethal melanomas. Am J Surg Pathol 2009;33:1386–1395.
- 36.↑
Gamblin TC, Edington H, Kirkwood JM, et al.. Sentinel lymph node biopsy for atypical melanocytic lesions with spitzoid features. Ann Surg Oncol 2006;13:1664–1670.
- 37.↑
Murali R, Sharma RN, Thompson JF, et al.. Sentinel lymph node biopsy in histologically ambiguous melanocytic tumors with spitzoid features (so-called atypical spitzoid tumors). Ann Surg Oncol 2008;15:302–309.
- 38.↑
Gerami P, Busam K, Cochran A, et al.. Histomorphologic assessment and interobserver diagnostic reproducibility of atypical spitzoid melanocytic neoplasms with long-term follow-up. Am J Surg Pathol 2014;38:934–940.
- 39.↑
Abraham RM, Karakousis G, Acs G, et al.. Lymphatic invasion predicts aggressive behavior in melanocytic tumors of uncertain malignant potential (MELTUMP). Am J Surg Pathol 2013;37:669–675.
- 40.↑
Spatz A, Calonje E, Handfield-Jones S, et al.. Spitz tumors in children: a grading system for risk stratification. Arch Dermatol 1999;135:282–285.
- 41.↑
Ludgate MW, Fullen DR, Lee J, et al.. The atypical Spitz tumor of uncertain biologic potential: a series of 67 patients from a single institution. Cancer 2009;115:631–641.
- 42.↑
Hillen LM, Van den Oord J, Geybels MS, et al.. Genomic landscape of spitzoid neoplasms impacting patient management. Front Med (Lausanne) 2018;5:344.
