Histopathologic abundance of pigmentation correlates with disease‐specific survival in malignant melanoma but is not independent of current AJCC pT stage

The increasing number of melanoma patients makes it necessary to develop best possible strategies for prognosis assessment in order to recommend appropriate therapy and follow‐up. The prognostic significance of tumor cell pigmentation has not been fully elucidated. Hematoxylin and eosin (H&E)‐stained sections of 775 melanomas diagnosed between 2012 and 2015 were independently assessed for melanin pigment abundance by two investigators, and the impact on melanoma‐specific survival was calculated. Unpigmented melanomas (n = 99) had a melanoma‐specific survival of 67.7%, melanomas with moderate pigmentation (n = 384) had a melanoma‐specific survival of 85.9%, and strongly pigmented melanomas (n = 292) had a melanoma‐specific survival of 91.4% (p < .001). In an analysis of melanoma‐specific survival adjusted for pT stage and pigmentation, we found a nonsignificant impact of pigmentation abundance with a hazard ratio of 1.277 (p = .74). The study presented here provides evidence in a German cohort that patients with pigmented melanomas have a more favorable prognosis than those diagnosed with nonpigmented melanomas. Moreover, the abundance of pigmentation already seems to provide a first prognostic estimate. However, it does not appear to provide significant additional value for prognostic assessment according to the AJCC 2017 pT classification.

At the moment several gene expression-based tests are being developed to distinguish patients at high risk from those at low risk of metastasis (Amaral et al., 2020;Bellomo et al., 2020;Gambichler et al., 2021;Gerami et al., 2015).Nevertheless, these tests are costly and cannot be broadly applied yet.In addition, artificial intelligencebased models that can estimate classification, prognosis, and, in some cases, treatment response in melanoma are being developed (Brinker et al., 2021;Forchhammer et al., 2022;Grant et al., 2022;Johannet et al., 2021;Kulkarni et al., 2020).
The gold standard of daily practice is still based on histologic diagnosis and the 8th AJCC classification of 2017 which incorporates Breslow's tumor thickness, the presence of ulceration of the primary tumor, and metastasis to lymph nodes and/or other organs (Gershenwald et al., 2017).Depending on the AJCC Stage, the prognosis of the patients varies considerably.For example, 5-year survival in stage I is 99.4%, while in stage IV it drops to 23% (Buja et al., 2021).
Furthermore, different subtypes of melanoma are characterized.
The WHO distinguishes melanomas associated with chronic sun exposure from those with low chronic sun exposure.Low chronic sun exposure is associated with superficial spreading melanoma, while lentigo maligna melanomas occur in severely and chronically sundamaged skin.Desmoplastic melanomas are also found to be associated with sun-damaged skin.Nodular melanomas, on the contrary, are found in both sun-damaged skin and non-sun-damaged skin.The underlying genetic mechanisms are different in these two cases.In addition, there are melanoma subtypes that occur in skin that is not chronically sun-damaged.These include acral lentiginous, mucosal, and choroidal melanomas, as well as Spitz melanomas and melanoma originating in blue or congenital nevi (Elder et al., 2020).
In addition to the melanoma subtype, the tumor thickness and, if present, an ulceration of the tumor is specified in the initial histological report.Using this information, the tumor stage (by AJCC 2017) is calculated in an attempt to estimate the prognosis of the patient.Also, for other histologic features, such as regression and mitotic rate, an influence on the further prognosis is ensured although they are not directly included in the AJCC classification (Evans et al., 2018;Hale et al., 2013;Ribero et al., 2016).
There is no established grading system for malignant melanoma.By contrast, grading for other tumor entities like squamous cell carcinoma is based on the degree of dedifferentiation.In keratinocytes, for example, the ability to form horn is associated with a well-differentiated state, since the primary task of keratinocytes corresponds to the formation of the stratum corneum.Therefore, the grading of cutaneous squamous cell carcinomas aligned to Broders is based on the ability to form horn centers (Que et al., 2018).A similar argument could be made for malignant melanoma.The basic function of melanocytes is the formation of melanin pigment.Therefore, preserved pigmentation could be attributed to a highly differentiated form, while lost pigmentation could be attributed to a more dedifferentiated form.
Besides tumor diameter and ulceration, one of the most obvious morphologic features at first sight is the pigmentation of melanocytic lesions.The darker a pigmented lesion, the more alarmed patients and initial investigators might be.
Pigmented melanomas are excised at earlier stages and diagnosed with smaller tumor thickness.In one study that examined various tumor parameters in melanotic and amelanotic melanomas, amelanotic melanomas were found to have significantly worse survival.They were also more proliferative (had more mitotic figures) and were diagnosed at later tumor stages (Thomas et al., 2014).
On the contrary, epidemiological data from Nigeria show that dark-skinned Africans are significantly more likely to develop melanoma than albinos and thus hypopigmented individuals of the same population (Kiprono et al., 2014;Okafor & Onyishi, 2021).Also, melanogenesis seems to be able to increase the resilience of tumor cells.At least experimentally, the efficacy of lymphocytes, chemotherapy, and radiotherapy can be increased when inhibitors of melanogenesis are added (Brożyna et al., 2016;Slominski et al., 2009).
There is therefore conflicting evidence for the prognostic relevance of pigmentation.The predictive value of pigmentation assessment at the histologic level remains to be satisfactorily elucidated.
The aim of this extensive retrospective study was to use real-world data to expand our knowledge of the prognostic significance of tumor cell pigmentation in melanoma tumors.We examined a total of 775 melanomas to determine their level of pigmentation and compared melanoma-specific survival between patients with histopathologically nonpigmented/very low pigmented, moderately pigmented, and strongly pigmented melanomas.

| Study population
All 2223 patients that were diagnosed with primary melanoma at the University Department of Dermatology in Tuebingen between 01/01/2012 and 12/31/2015 who gave written informed consent to the nationwide melanoma registry were identified.All 775 patients with follow-up data of at least 2 years and histological

Significance
Malignant melanoma of the skin is one of the most common cancers and shows high incidence gains.Adjuvant therapy is available at early stages but comes with non-negligible side effects.Therefore, an accurate prognosis assessment is already necessary at the time of initial diagnosis.
The gold standard is still based on histologic diagnosis.In contrast to most solid tumors there is no established histologic grading of melanoma despite enormous diversity in morphology and pigmentation.The results show that the degree of pigmentation alone can be used to estimate melanoma prognosis.sections in our archive were included in further analysis.Patients that died of melanoma within 24 months were also included.
Patients with a follow-up of less than 2 years were excluded from the study.Melanoma diagnosis was made by at least two experienced board-certified dermatopathologists (SF and GM).
The vast majority of cases (773 melanomas; 99.7%) included in the study were cutaneous melanomas, while two cases were identified as mucosal melanomas.

| Evaluation of pigmentation
All H&E sections of primary melanomas were reviewed by an experienced board-certified dermatopathologist (SF) and an aspiring dermatopathologist in training (VA) in nonidentical order.Pigmentation was stratified as zero for eosinophilic cytoplasm and visible pigmentation in less than 5% of tumor cells, one for pale pigmentation of more than 5% and less than 90% of tumor cells, and two for strong pigmentation in more than 90% of tumor cells.
All cases that received unequal grading of pigmentation from VA and SF were reviewed together.Without knowledge of the previous assessment, pigmentation was then jointly graded using the criteria mentioned above.

| Statistics
Statistical calculations were performed using IBM SPSS Statistics version 28.0 (IBM SPSS, Chicago, IL, USA).Numeric variables were described by mean and standard deviation or median values and interquartile range (IQR).p values in Kaplan-Meier curves were calculated using the log-rank test (Mantel-Cox).Cohens-Kappa calculations were performed using the cross-tabulation tool in SPSS.
Receiver operating characteristic (ROC) curve analyses were conducted using the ROC analysis tool in SPSS.Linear regressions to create a predictor were obtained using the linear regression tool in SPSS.Cox analyses were performed using the Cox analysis tool in SPSS.Throughout the analysis, p values of <.05 were considered statistically significant.

| Licensing by the ethics committee
The retrospective study was approved by the Ethics Committee of the Medical Faculty of the University of Tubingen as part of the project "Prediction of recurrence-free survival in patients with malignant melanoma using image analysis of histological sections by artificial neural networks" (No. 874/2019/BO2).

| RE SULTS
Patients had a mean age of 59.64 (±15.3)years.In our cohort, 55.4% of patients were male and 44.6% were female.The most common subtype was superficial spreading melanoma with 462 cases (59.6%), followed by nodular melanoma with 128 cases (16.5%), lentigo maligna melanoma with 65 cases (8.4%), and acrolentiginous melanoma with 48 cases (6.2%).Mucosal melanoma and nevus-associated melanoma accounted for two cases each (0.3%), and the remaining 68 cases (8.8%) were either characterized as other subtypes or were not classified at all.
Regarding location, melanomas of the trunk (40.0%) and melanomas of the lower extremity (27.9%) were the most frequent.Lastly most of the tumors were operated at stage pT1a (39.1%) whereas only 7.9% were stage pT4b.
All 775 melanomas were examined and a grading of 0, 1, or 2 was given as stated under methods.(Figure 1).In 199 cases, different scores were assigned by the two investigators.Discordant assessment was predominantly between 1 and 2, while 0 was usually clearly assigned.There was substantial correlation between the two investigators.Cohen's kappa was 0.674 (p < .001).
The confidence intervals of the mean ages all overlap.We did not see significant differences in the sex distribution of nonpigmented, moderately pigmented, and strongly pigmented melanomas either (Table 1).Of the unpigmented melanomas 26.3% were located on the head/neck, 24.2% on the trunk, 18.2% on the upper extremity, and 30.3% on the lower extremity.Respectively, 19.8% of melanomas located on the head/neck, 7.7% of the melanomas on the trunk, 16.2% of those on the upper, and 13.9% of those on the lower extremity were not pigmented.The most common nonpigmented subtype was nodular melanoma.34.4% of nodular melanoma were nonpigmented.By contrast, most moderately pigmented and most strongly pigmented melanomas were in the subgroup of superficial spreading melanoma (62.5% and 69.1%, respectively).The most common nonpigmented pT Stage (AJCC 2017) was pT4b at 45.9%.By contrast, only 3% of pT1a melanomas were nonpigmented.
First, we investigated whether the presence of pigment has an impact on melanoma-specific overall survival.Of 775 melanomas, 676 cases (87.2%) showed pigmentation of tumor cells, 99 melanomas were unpigmented.The patients with pigmented melanomas showed significantly better survival (n = 676 with 79 events; melanoma-specific survival of 88.3%) than patients with TA B L E 1 Demographics of the cohort, tumor subtype, location, and pT stage (AJCC 2017).nonpigmented melanomas (n = 99 with 32 events; melanomaspecific survival of 67.7%) (p < .001)(Figure 2).
There were significant survival advantages (p < .001)for patients with pigmented versus less pigmented as well as nonpigmented tumor cells.Interestingly moderately pigmented melanoma patients had better survival than nonpigmented melanoma patients (Figure 3).
In 24 cases (3.1%), an accumulation of a nonpigmented tumor cell clone was found within an otherwise pigmented tumor (see Figure S1).Those patients had a melanoma-specific survival of 79.2% (n = 24 with five events).Their prognosis was therefore between the unpigmented and moderately pigmented.F I G U R E 3 Melanoma-specific survival of patients with unpigmented, moderately pigmented, and strongly pigmented melanomas.Kaplan-Meier curves of melanoma-specific survival for histologically unpigmented (0-blue), moderately pigmented (1-green), and strongly pigmented primary melanomas (2-pink) (n = 775).Median follow-up 6.31 years (p < .001,log-rank test).
this purpose, linear regression was used to calculate a predictor from pigmentation abundance and pT stage.Abundance of pigmentation alone has an area under the curve (AUC) of AUC = 0.636 (95% confidence interval (95% CI): 0.578-0.693).The pT classification (tumor thickness and presence of ulceration) has an AUC of 0.787 (95% CI: 0.0739-0.835)and the combined score of pT classification and pigmentation has an AUC of 0.802 (95% CI: 0.757-0.846)(Figure 4).
In an analysis of melanoma-specific survival adjusted for pT stage and pigmentation, there was a nonsignificant impact of pigmentation abundance on melanoma-specific survival with a hazard ratio of 1.277 (p value = .74).In a regression adjusted for age, sex, pT stage, presence of regression and pigmentation, pT stage showed the largest and more significant impact on melanoma-specific survival of patients (HR 1.426, p value <.001) (Table 2).Therefore, adding pigmentation to the established pT classification does not yield a statistically significant prognostic prediction.

| Results
In our retrospective study, nonpigmented melanomas were less common than pigmented melanomas.In higher pT tumor stages, melanomas were more likely to be nonpigmented than in lower pT stages, and patients with nonpigmented melanomas had a worse survival prognosis than patients with moderately and highly pigmented melanomas.However, the assessment of prognosis based on pT classification was not significantly improved by adding pigmentation to the analysis, and there was no significant impact on prognosis when pT stage was included in the analysis.
The frequency of 12.8% histopathologically nonpigmented melanomas in our study is within the range of approximately 2%-20% reported elsewhere in the literature (Balch et al., 1978;Barnhill et al., 1996;Huvos et al., 1972;Larsen & Grude, 1978;Moreau et al., 2013;Thomas et al., 2014).In accordance with other studies, nodular melanomas were more frequently nonpigmented, or described as amelanotic melanomas (Thomas et al., 2014).In contrast to other studies, we saw no particular association of the occurrence of nonpigmented melanomas with the occurrence of the tumor on the head and neck (Moreau et al., 2013;Thomas et al., 2014).Though we did see, that melanomas located on the trunk were more often pigmented than melanomas of other locations.
Our data confirms and extends the multicenter and international data of Thomas et al. (2014) from Australia, the USA, Italy, and Canada.Additionally, in our German cohort we can demonstrate poorer survival rates in histopathologically amelanotic melanomas, as well as, a worse prognosis for patients with less pigmentation present in the tumor cells.
Although Larsen and Grude (1978) described difficulty in grading pigmentation, we were able to grade with a high interrater correlation using a simple division into non or only weakly pigmented, moderately pigmented, and strongly pigmented melanomas.However, we were similarly unable to provide a further reliable breakdown of pigmentation into more precise expression groups.Our data also provide support for the hypothesis that the least pigmented clones contribute substantially to a progressive course, although this could not be proven in the small subcollective.
The background of the influence of pigmentation on melanoma development and tumor progression and dissemination is a matter of discussion.For example, intravital imaging studies revealed a subpopulation of low to nonpigmented cells mobile in the primary tumor.
These cells exhibited high Brn-2 expression and invaded the vasculature during the experiments (Du et al., 2004).Therefore, rather less pigmented clones could be responsible for forming metastases.
Some evidence also points in the opposite direction.Reactive intermediates are released within melanin synthesis.These intermediates can generate a genotoxic and mutagenic milieu in melanocytes and locally inhibit the immune system (Slominski et al., 1998(Slominski et al., , 2004;;Slominski & Goodman-Snitkoff, 1992).Quite astonishingly, Nigerian studies have shown that dark-skinned patients are significantly more likely to develop melanoma than albinos.Hence hypopigmented individuals of the Nigerian population are less likely to develop melanoma (Kiprono et al., 2014;Okafor & Onyishi, 2021).
Melanogenesis also appears to be capable of increasing tumor cell resilience.At least experimentally, the efficacy of lymphocytes, of chemotherapy and of radiotherapy can be increased when inhibitors of melanogenesis are added (Brożyna et al., 2016;Slominski et al., 2009).The role of melanin as a double-edged sword in melanoma formation and later tumor progression is further elucidated, for example, in (Cabaço et al., 2022).
Based on our research, the pT stage according to AJCC 2017 still provides the best possible prognosis estimation for survival at the time of initial histologic diagnosis, nevertheless, there might be other clinical implications.Given the invasive nature of sentinel lymph node biopsies and the occurrence of unnecessary procedures, it is vital to enhance the identification of patients who would derive the greatest benefit (Quaglino et al., 2011).In addition to pT stage, patient age, and mitotic rate, tumor cell pigmentation could potentially serve as a relevant risk factor in the evaluation process.
However, further studies are required to investigate this aspect.
Additionally, we reaffirm the rationale that the less pigmented a melanoma is, the later it is diagnosed.Therefore, the designation as black skin cancer, which is commonly used in German-speaking countries, may in fact be misleading.The less black the malignant melanoma is, the worse it seems to behave.

| Limitations
The study presented here has several limitations.First, it is a retrospective evaluation at a single center in Germany.Second, clinical information on pigmentation is lacking.It was only examined by light microscopy.
All evaluated sections were from the same pathological institute.
It is possible that the results can only be valid for other institutes to a limited extent, since a slightly different staining pattern could be seen in the H&E staining and possibly therefore the assessment of pigmentation could vary.In addition, an ethnic bias of our data cannot be excluded.Data regarding the ethnicity of our cohort is lacking.
The treatment of the patients in our study was not investigated.
It is possible that changes in treatment regimens during the study period may limit the predictive accuracy of the graded pigmentation.Mutations of the mitogen-activated protein kinase/extracellular signal-regulated kinase (MAPK/ERK) signaling pathway has been described to be associated with histologic pigmentation and is found in approximately half of the melanomas (Cabaço et al., 2022;Viros et al., 2008).Since our patient population could already be offered targeted BRAF inhibition and, in the context of early-access programs, also possibly combined BRAF and MEK inhibition, an influence on prognosis at least in higher tumor stages is quite conceivable.This problem is already mentioned by Thomas et al. (2014).
To obtain more conclusive results, a larger, specially designed, multicenter study with many independent investigators would be required.

| CON CLUS IONS
The study presented here provides evidence within a German cohort that patients with pigmented melanoma have a more favorable prognosis than those with nonpigmented melanoma.Moreover, the abundance of pigmentation already seems to provide some prognostic estimation.However, it does not seem to add significant additional value to the prognosis estimation medium pT classification according to AJCC 2017.

ACK N OWLED G M ENTS
We thank the medical laboratory assistants of the Department of

1
Representative images of H&E sections of malignant melanomas with differing degrees of pigmentation.(a) Graded as 0. Eosinophilic cytoplasm and visible pigmentation in less than 5% of the tumor cells.Scale bar is 50 μm (b) Graded as 1.Moderate pigmentation in more than 5% and less than 90% of the tumor cells.Scale bar 50 μm (c) Graded as 2. Strong pigmentation in more than 90% of tumor cells.Scale bar is 50 μm.
Later, we wanted to evaluate whether pigmentation alone can be considered an independent prognostic factor or whether it correlates with established predictive factors.The prognosis estimation by pigmentation was compared with the prediction quality of pT stage by ROC analysis.Furthermore, it was investigated whether pigmentation permitted an additional added value in prognosis estimation.For F I G U R E 2 Melanoma-specific survival of patients with unpigmented and pigmented melanomas.Kaplan-Meier curves of melanoma-specific survival for histologically unpigmented (0-blue) and pigmented (1-pink) primary melanomas (n = 775).Median follow-up of 6.31 years (p < .001,log-rank test).
Dermatology, Eberhard Karls Universität Tübingen for excellent technical support and Anneli Vollert, who was a great help with the linguistic revision of the manuscript.Open Access funding enabled and organized by Projekt DEAL.FU N D I N G I N FO R M ATI O NThis research received no external funding.CO N FLI C T O F I NTE R E S T S TATE M E NTV.A,A.A-G., G.M., and L.R. declares no conflict of interest.C.G. reports personal fees from Amgen and MSD, grants and personal fees from Novartis, NeraCare, BMS, Philogen, Roche, and Sanofi, outside the submitted work.L.F. reports grants from the Swiss National Science Foundation, Swiss Cancer League, Hookipa Pharma, Krebsliga Schweiz, and Novartis Foundation as well as an advisory role for Novartis, Sanofi and Bristol-Myers Squibb, outside the submitted work T.E.reports payment or honoraria for lectures, presentations, speakers bureaus, manuscript writing or educational events, and participation in advisory board from Bristol Myers Squibb, Merck Sharp & Dohme, Novartis, Pierre Fabre, Sanofi, Immunocore, CureVac and Almiral Hermal; and unpaid leadership as board member of the DeCOG.S:F: received personal fees from Kyowa Kirin and Takeda Pharamceuticals (speaker's honoraria), as