Characterization of PRAME immunohistochemistry reveals lower expression in pediatric melanoma compared to adult melanoma

Pediatric melanomas are rare tumors that have clinical and histological differences from adult melanomas. In adult melanoma, the immunohistochemical marker PRAME is increasingly employed as a diagnostic adjunct. PRAME is also under investigation as a target structure for next‐generation immunotherapies including T‐cell engagers. Little is known about the characteristics of PRAME expression in pediatric melanoma. In this retrospective study, samples from 25 pediatric melanomas were compared with control groups of melanomas in young adults (18–30 years; n = 32), adult melanoma (>30 years, n = 30), and benign melanocytic nevi in children (0–18 years; n = 30) with regard to the immunohistochemical expression of PRAME (diffuse PRAME expression >75%/absolute expression). Pediatric melanomas show lower diffuse PRAME expression (4%) and lower absolute PRAME expression (25%) compared to young adult melanomas (15.6%/46.8%) and adult melanomas (50%/70%). A significant age‐dependent expression could be observed. An analysis of event‐free survival shows no prognostic role for PRAME in pediatric melanoma and young adult melanoma, but a significant association with diffuse PRAME expression in adulthood. The age dependency of PRAME expression poses a potential pitfall in the diagnostic application of melanocytic tumors in young patients and may limit therapeutic options within this age group. The immunohistochemical expression of the tumor‐associated antigen PRAME is an increasingly important diagnostic marker for melanocytic tumors and is gaining attention as a possible immunotherapeutic target in melanoma. As the available data primarily stem from adult melanoma, and given the clinical and histological distinctions in pediatric melanomas, our understanding of PRAME expression in this specific patient group remains limited. The age‐dependent low PRAME expression shown here constrains the use of this marker in pediatric melanoma and may also limit the use of immunotherapeutic strategies against PRAME in young patients.


| INTRODUC TI ON
With an incidence of 1.3-1.6/millionpediatric melanoma is a tumor rarely occurring in the age group under 15 years (Ferrari et al., 2019).
However, the incidence increases significantly in adolescents and reaches about 10 times the incidence of children in the age group of 15-19 years (Ferrari et al., 2019;Strouse et al., 2005).Most pediatric melanomas have a rather benign course, but there are also fatal outcomes in this age group (Pampena et al., 2023).The prognosis differs significantly between the subtypes of pediatric melanoma.
While spitzoid melanomas (SM) and in particular the subgroup of spitz melanomas have a predominantly favorable prognosis, the subgroup of conventional type pediatric melanomas (including superficial spreading melanoma (SSM), nodular melanoma (NM), acrolentiginous melanoma (ALM) and lentigo-malignant melanoma (LLM)) as well as melanomas arising on congenital melanocytic nevi (CMN) show more aggressive courses (Merkel et al., 2019).The histologic classification sometimes poses a challenge.Morphologically, there are substantial overlaps between the subtypes but also between borderline lesions such as the group of atypical Spitz tumors.An exact diagnosis is often only possible by combining morphology, immunohistochemistry, and an additional molecular workup (Church et al., 2022;Raghavan, Peternel, et al., 2020).
PReferentially expressed Antigen in MElanoma (PRAME) is a tumor-associated antigen (TAA) and belongs to the group of socalled "cancer testis" antigens.PRAME is an immunohistochemical marker that has become increasingly important in the routine diagnosis of melanoma in recent years.The use of this antibody in the diagnosis of melanocytic tumors was described in a study by Lezcano et al. (2018).Here, it was shown that a diffuse PRAME expression of over 75% has high sensitivity and specificity for nonspindle cell primary melanomas and lymph node metastases and can be found in around 90% of conventional melanomas, but only in 0.7% of benign nevi (Lezcano et al., 2018).Numerous studies on the diagnostic use of PRAME in adult melanomas have followed in recent years, which have confirmed these results but mostly found lower expression levels in melanoma (Bello et al., 2023;Gassenmaier et al., 2021;Gradecki, Slingluff et al., 2021;Gradecki, Valdes-Rodriguez, et al., 2021;Tio et al., 2020;Turner et al., 2023).
PRAME has a prognostic role in uveal melanoma (Field et al., 2016), but its prognostic significance in cutaneous melanoma is still unclear.In contrast, little is known about the expression of PRAME in childhood melanomas (Umano et al., 2021).Studies suggest that the use of PRAME in the setting of spitzoid tumors is of limited value (Raghavan, Wang, et al., 2020).In contrast, a recent publication found a possible role of PRAME in the differentiation of benign proliferating nodules of melanomas at the base of congenital giant nevi (Boutko et al., 2023).
In addition to its diagnostic use, PRAME has also recently gained importance as a target for T-cell enhancement in the context of immunotherapy due to its role as a tumor-associated antigen (Bunk et al., 2019;Gezgin et al., 2017;Gutzmer et al., 2016;Sailer et al., 2022).
It is therefore of great interest to characterize PRAME expression in pediatric melanomas both in the diagnosis of childhood melanocytic tumors and with regard to future targeting of immunotherapy in melanomas.The aim of our study is therefore to analyze the immunohistochemical PRAME expression in a cohort of pediatric melanomas in comparison to melanomas in young adults, adult patients, and childhood nevi.In addition, it will be investigated whether the expression of PRAME can be used as a prognostic marker in this setting.

| Study cohort
The retrospective study was performed on formalin-fixed and paraffin-embedded (FFPE) tissue of pediatric melanomas, melanomas of young adults, adult melanomas, and samples from benign melanocytic nevi of childhood.Written informed consent was given by all patients or their legal guardians.Consent was waived for deceased patients or samples older than 5 years.
The study cohort included 30 patients in childhood and adolescence (0-18 years) who were diagnosed with pediatric melanoma as a primary tumor (n = 28) or metastasis (n = 2) at various tumor stages.

Significance
The immunohistochemical expression of the tumorassociated antigen PRAME is an increasingly important diagnostic marker for melanocytic tumors and is gaining attention as a possible immunotherapeutic target in melanoma.As the available data primarily stem from adult melanoma, and given the clinical and histological distinctions in pediatric melanomas, our understanding of PRAME expression in this specific patient group remains limited.The age-dependent low PRAME expression shown here constrains the use of this marker in pediatric melanoma and may also limit the use of immunotherapeutic strategies against PRAME in young patients.
the cohort of pediatric melanomas by Breslow tumor thickness (range of 0.1 mm) for primary melanomas and location (lymph nodes, organs, skin) for metastases.Due to the rarity of the diagnosis in the young adult cohort, tumor parameters could not be matched to the study cohort of pediatric melanomas.Five pediatric melanomas were excluded during the study (4 samples had no detectable tumor tissue left and 1 tumor was re-classified as nevus).Benign melanocytic nevi of childhood (under 18 years) were selected as a control  9,11,13,14,17,18; GNAQ Exon 5; GNA11 Exon 5; CTNNB1 Exon 3; PDGFRA Exon 12,13,14,18;KRAS Exon 2,3,4;MAP2K1 Exon 2,3,6,7,11;TERTp).Histologic diagnoses of the control cohorts were performed by at least two experienced board-certified dermatopathologists.The classification of chronic sun damage (CSD) in adult primary melanomas was semi-quantitative (no-CSD, low-CSD, high-CSD) according to the solar elastosis in the H&E-section.
PRAME expression was independently quantified by two authors (S.F. and M.H.) in the epidermal and dermal melanocytic component without knowledge of the histological diagnosis and the clinical course.In case of discrepant findings, consensus was achieved together with a third reviewer (V.A.).

| Statistics
Statistical calculations were performed using IBM SPSS version 26; p values <.05 were considered statistically significant.Numerical results are reported as means and standard deviations or medians and interquartile ranges (IQR).Event-free survival rates of PRAME-positive versus PRAME-negative samples were assessed in primary melanomas with Kaplan-Meier curves and compared using the log-rank test.An event was defined as the first occurrence of metastasis or death by melanoma.

| Ethics statement
The study was approved by the institutional review board of the University of Tübingen (Project ID: 786/2018BO2) and was conducted in accordance with the Declaration of Helsinki.

| Expression profile of PRAME in pediatric melanoma, melanoma in young adults, adult melanoma, and benign nevi of childhood
PRAME expression in all samples was analyzed by immunohistochemical staining and diffuse PRAME staining was defined by a PRAME score of 4+ (>75% of cells) in the in situ or invasive component.Tables S1-S4 list PRAME expression for each cohort in relation to clinical data, follow-up, and outcome.

PRAME expression has been detected in pediatric melanomas
as well as in adult melanomas (Figure 1).In total, 5 of the 25 pediatric melanomas (20%) were found to have PRAME expression, with only 1 of 25 samples (4%) showing diffuse PRAME expression (Table S1).Of the metastases, one of two samples showed PRAME expression which, however, was not classified as diffuse.Of the primary cutaneous melanomas, PRAME expression was found in 1 of 12 spitzoid melanomas (no diffuse PRAME expression), in 1 of the 6 melanomas of the conventional type (no diffuse PRAME expression), in 2 of 3 melanomas on congenital nevus (one sample with diffuse PRAME expression) and none of the two samples with other melanoma subtypes.

| Expression of PRAME in relation to patient age
To determine whether the variations in PRAME expression in the different cohorts were related to patient age, we created a histogram of the frequency of diffuse PRAME expression versus age (20year bins) (Figure 2a,c) and performed a logistic regression model (Figure 2b,d  with positivity in 7.4% of cases in the 0-20 years age group compared to 80% positivity in the 80+ years age group (Figure 2a).The regression analysis showed that the prevalence of diffuse PRAME expression increased significantly with advancing patient age (p < .0001; Figure 2b).Similar results were obtained by analyzing the absolute PRAME expression.Again, PRAME expression increased from 25.9% positivity in the 0-20 age group to 80% positivity in both the 61-80 and 81+ age groups (Figure 2c).A regression analysis revealed a significant increase in absolute PRAME expression in relation to age (p = .0002;Figure 2d).
Since chronic UV damage is an important difference between pediatric and adult melanomas, the influence of chronic sun damage was examined in our cohort of adult melanomas.Comparing diffuse PRAME expression in the group without, with low, and with high UV damage, there is no significant difference in the subgroup of melanomas with particularly high UV damage (3 of 6 melanomas with diffuse PRAME expression) (see Figure S1, Table S3).Thus, in our cohort, chronic light damage does not seem to explain the decisive difference between PRAME expression in the age groups.These results must be interpreted with caution as the corresponding subgroups have rather small numbers of cases.

| Expression of PRAME in relation to patient survival
To investigate whether the expression of PRAME may have prognostic value for pediatric melanoma patients, we performed Kaplan-Meier analyses of event-free survival in a cohort of 20 patients with primary melanoma samples and available follow-up data, four of whom died during the analyzed period (Figure 3).No prognostic significance was found for diffuse PRAME expression (p = .591;Figure 3a).When survival analysis is based on mere PRAME expression, there is a relationship between PRAME and melanoma-related events which however does not reach statistical significance (p = .149;Figure 3b).A similar picture was observed for melanoma-associated overall survival.
Three of the four patients who died in our pediatric melanoma cohort showed PRAME expression, with diffuse PRAME expression detectable in one of these patients (see Table S1).
For comparison, we performed the same survival analysis in the control groups of young adult melanomas (n = 30) and adult melanomas (n = 28) with primary melanoma samples and available follow-up data.The group of young adults also demonstrated no significant association between diffuse PRAME expression and event-free survival (p = .835;Figure 3c).Looking at the non-diffuse PRAME expression in this group, there is also a trend toward melanoma-related events in association with PRAME expression which does not reach statistical significance (p = .169;Figure 3d).In this age group, there was only one deceased patient in our cohort.This patient exhibited diffuse PRAME expression (see Table S2).In adult melanoma patients over 30 years of age, there is a clear significant association between diffuse PRAME expression and melanoma-related events (p = .026;Figure 3e).The 14 diffuse PRAME-positive patients had a median event-free survival of 821 days following metastasis or death in 7 of the 14 patients.A median event-free survival of 2000 days was achieved in the 14 diffuse PRAME-negative patients with metastasis or death in only two cases.This association is weakened if non-diffuse PRAME expression is used for the analysis of eventfree survival in this age group (p = .112;Figure 3f).Looking at the deceased patients in this age group, five of the six deceased patients showed PRAME expression, four of whom had a diffuse expression pattern (see Table S3).Analyzing PRAME expression in pediatric melanoma samples and control groups yields crucial results that could influence the utility of PRAME as a diagnostic marker and as a target for immunotherapy.

F I G U R E 3
We have confirmed that PRAME is also expressed in pediatric melanoma to a diminished degree compared to adult melanoma.
Only 20% of the melanomas in the age group under 18 years of age showed detectable PRAME positivity and only one of those 25 cases reached the level of diffuse expression which is established as a cut-off marker in diagnostic use (Lezcano et al., 2018).
Similar data have already been described in a series of pediatric melanoma cases by Umano et al. (2021).This study also shows that PRAME expression in pediatric melanomas is lower than in numerous publications that mainly refer to an adult population.
The ideal diagnostic cut-off for diffuse PRAME expression (>75%) as a marker for melanoma must be questioned, especially in this setting.In adult melanoma, there is a controversial discussion as to whether a lower PRAME cut-off might provide more accurate information (Gassenmaier et al., 2021;Gradecki, Valdes-Rodriguez, et al., 2021;Raghavan, Wang, et al., 2020).A cross-sectional study by O'Connor et al. (2022) confirms that there is a higher joint sensitivity/specificity for a PRAME cut-off of 2+ (>25%) and 3+ (>50%), however a lower accuracy compared to the cut-off at 4+ (>75%).If the mere expression in melanocytes is used as a differentiation criterion in line with our data, the number of positive samples in pediatric melanoma increases significantly.There is also a higher correlation with melanoma-related events, which indicates increased PRAME expression in pediatric melanomas with an unfavorable course.Nevertheless, we would have found two unequivocally benign pediatric nevi to be positive in our cohort with a lower cut-off using only the positivity of PRAME as a criterion (see Figure S2).The accuracy of differentiation between benign and malignant tumors thus suffers from such an adjustment.
In summary, it can therefore be stated that PRAME expression is limited in its diagnostic value in pediatric melanomas and that PRAME negativity in histology cannot be used as an argument for the benignity of a pediatric melanocytic lesion.Melanomas arising from congenital nevi may be an exception to this rule.In a recent case report and a recently published study, it was shown in the differentiation of benign proliferating nodules and melanomas arising from congenital nevi that the latter express an increased amount of PRAME (Boutko et al., 2023;Gill et al., 2021).Our data also support this hypothesis.Of the three melanomas arising on congenital nevi in our cohort, two demonstrated PRAME expression, one of them with a diffuse pattern above 75%.PRAME positivity in the setting of a proliferative tumor arising from a congenital nevus should therefore be critically checked for the diagnosis of melanoma and further molecular work-up should follow if necessary.
An unexpected finding is seen in the assessment of PRAME expression in young adult melanomas.Compared to our cohort of adult melanomas (over 30 years of age), but also compared to published data (Gassenmaier et al., 2021;Gradecki, Valdes-Rodriguez, et al., 2021;Lezcano et al., 2018;Tio et al., 2020;Turner et al., 2023), a considerably lower rate of diffuse PRAME expression was found.
One possible reason for this may be the fact that rather thin primary melanomas were investigated in our study (median tumor thickness of 0.75 mm).However, in a previous study by Gassenmaier et al. (2021) (Turner et al., 2023).These results therefore support our conclusion that there is an age-dependent increase in PRAME expression in melanocytic tumors.Knowledge of such age dependence of PRAME expression is crucial not only for the diagnostic but also for the therapeutic use of PRAME, as this could limit the potential use of such therapies in young melanoma patients.
Lastly, we further investigated PRAME expression in our cohort of adult melanoma patients.An absolute PRAME expression rate of 70% and a diffuse PRAME expression rate of 50% were found, which is consistent with published data (Gassenmaier et al., 2021;Gradecki, Valdes-Rodriguez, et al., 2021;Lezcano et al., 2018;Tio et al., 2020).Although Kaplan-Meier survival analysis in pediatric melanoma and young adult melanoma does not find a significant association between diffuse and absolute PRAME expression and melanoma-specific events, a significant association between diffuse PRAME expression and event-free survival was found in our cohort of adult melanoma patients over 30 years of age.In a previous study, we were unable to demonstrate a prognostic benefit for PRAME in a cohort of thin (Breslow tumor thickness <1 mm) non-metastatic versus metastatic melanomas (Gassenmaier et al., 2021).In contrast, the prognostic role of PRAME in uveal melanoma is established (Cai et al., 2018;Field et al., 2016).Our data suggest that at least diffuse PRAME expression may have a prognostic role in the setting of adult melanoma.However, as our analysis is based on a small control cohort with few melanoma-specific events, this should be further investigated in a larger cohort.
The limitations of our study are that the selection of the melanoma sample for children and young adults was limited by the rarity of the disease in these age groups.As a result, the total cohort size was small and the young adult patient group could not be matched with the pediatric melanoma cohort in terms of tumor thickness and therefore included thinner melanomas.This group also had a lower rate of ulceration and melanoma-related events compared to the adult melanoma group.In addition, a proportion of the follow-up data were censored due to the early follow-up dates, and follow-up data were missing for a few cases because the samples were collected at many different sites and some of the sampling dates were many years ago.
In summary, our study characterizes the expression of PRAME in pediatric melanomas, adult control cohorts, and benign nevi through immunohistochemistry.The findings reveal comparatively low expression levels in pediatric samples, indicating age-dependent expression.This should be considered when interpreting PRAME as a diagnostic marker but also in the context of PRAME as a target for future immunotherapies.
cohort to compare age-matched non-melanoma patient samples with tumor samples from the test cohort.Patients were identified from either the German Registry for Rare Pediatric Tumors (STEP Registry), the Melanoma Registry of the German Dermatological Society in Tübingen, the Archives of Dermatopathology of the University Hospital of Tübingen, or the Archives of Pathology of the University Clinical Center/Medical University of Gdansk.The histopathological diagnosis of pediatric melanoma was made by the respective examining dermatopathologist or pathologist from German and European centers.All samples were further evaluated morphologically and in conjunction with available molecular workups (Comparative Genomic Hybridization (CGH) and Next Generation Sequencing (NGS)) by a reference dermatopathologist from the STEP registry and classified into groups of spitzoid melanoma, conventional melanoma (SSM, NM, ALM, LMM), melanoma on congenital nevus (CNM), other melanoma and melanoma metastasis.The diagnosis was based primarily on the external immunohistochemical stainings and the molecular pathology reports already carried out.In individual cases, previously incomplete immunohistochemical staining was supplemented or repeated.In the absence of tumor sequencing, an NGS panel was added which contains the most important driver mutations of malignant melanoma (Panel containing BRAF Exon 11, 15; NRAS Exon 2, 3, 4; KIT Exon 8, ).The analysis was performed over all melanoma samples including metastatic samples.Samples were combined in age groups of 20 years.The first group (0-20 years) contains all pediatric samples (n = 25) and young adults until 20 years (n = 2, n total = 27).The age group 21-40 years contains samples of young adults over 20 years of age (n = 30) and adult melanoma patients until 40 years (n = 3, n total = 33).The other age groups include adult melanoma patients and are represented by n = 7 (age group 41-60), n = 15 (age group 61-80), and n = 5 (age group over 80) patients.This analysis shows an increase in diffuse PRAME expression in every age group, TA B L E 1 Epidemiological data and tumor characteristics.