Variable Genomic Landscapes of Advanced Melanomas with Heavy Pigmentation

Abstract Background In the current study, we examined the real-world prevalence of highly pigmented advanced melanomas (HPMel) and the clinicopathologic, genomic, and ICPI biomarker signatures of this class of tumors. Materials and Methods Our case archive of clinical melanoma samples for which the ordering physician requested testing for both PD-L1 immunohistochemistry (IHC) and comprehensive genomic profiling (CGP) was screened for HPMel cases, as well as for non-pigmented or lightly pigmented advanced melanoma cases (LPMel). Results Of the 1268 consecutive melanoma biopsies in our archive that had been submitted for PD-L1 IHC, 13.0% (165/1268) were HPMel and 87.0% (1103/1268) were LPMel. In the HPMel cohort, we saw a significantly lower tumor mutational burden (TMB, median 8.8 mutations/Mb) than in the LPMel group (11.4 mut/Mb), although there was substantial overlap. In examining characteristic secondary genomic alterations (GA), we found that the frequencies of GA in TERTp, CDKN2A, TP53, and PTEN were significantly lower in the HPMel cases than in LPMel. A higher rate of GA in CTNNB1, APC, PRKAR1A, and KIT was identified in the HPMel cohort compared with LPMel. Conclusions In this study, we quantified the failure rates of melanoma samples for PD-L1 testing due to high melanin pigmentation and showed that CGP can be used in these patients to identify biomarkers that can guide treatment decisions for HPMel patients. Using this practical clinical definition for tumor pigmentation, our results indicate that HPMel are frequent at 13% of melanoma samples, and in general appear molecularly less developed, with a lower TMB and less frequent secondary GA of melanoma progression.


Introduction
Melanoma remains a deadly disease worldwide, with 106 110 new diagnoses and 7180 deaths estimated in the US in 2021. 1 Surgical excision remains the main treatment for patients with localized melanoma. In clinically advanced disease, targeted therapies such as BRAF and MEK inhibitors are given for tumors that carry activating mutations in BRAF. More recently, monoclonal antibodies against cytotoxic T lymphocyte-associated-4 (CTLA-4), programmed death ligand-1 (PD-L1), and programmed death-1 (PD-1) have become widely used. 2 The 2 PD-1 inhibitors, nivolumab and pembrolizumab, that have an indication for advanced melanoma from the US Food and Drug Administration (FDA) were both approved without companion diagnostics (CDx) to assess likely tumor responsiveness before initiation of ICPI. While the key clinical trials of these 2 agents showed tumor objective response rates (ORRs, 40% and 52%, respectively) and 5-year overall survival (OS, both 34%) that are superior to the benefits from standard systemic therapies for patients with melanoma, new biomarkers are needed to stratify patients with advanced melanoma into likely responders vs non-responders to ICPI. [3][4][5][6] Currently, PD-L1 immunohistochemistry (IHC) is FDAapproved as a CDx for several ICPIs in specific non-melanoma tumor types. 7,8 In 2 different pan-solid tumor approvals that include melanoma, tumor mutational burden (TMB) ≥10 mutations/Mb and microsatellite instability-high (MSI-High) were each approved as a separate CDx for pembrolizumab in patients that have exhausted standard of care options. 9,10 In addition, evidence exists in the literature that CD274 (encodes for PD-L1) amplifications and losses can predict responses to ICPI. [11][12][13][14] Finally, GA in PBRM1, STK11, or KEAP1, amplifications of MDM2/4, and APOBEC mutational signature have also been suggested as predictive biomarkers for ICPI response in other tumor types. [15][16][17][18][19][20][21][22] Despite the importance of ICPI in the treatment of advanced melanoma, current methods for PD-L1 IHC do not allow interpretation when the tumor contains abundant melanin that obscures the IHC stain. In the current study, we examined the real-world prevalence of highly pigmented advanced melanomas (HPMel), defined here by the presence of melanin pigmentation that rendered IHC uninterpretable. In addition, we examined the clinicopathologic, genomic, and ICPI biomarker signatures of this class of tumors. Non-pigmented or lightly pigmented advanced melanoma cases (LPMel), defined here by interpretable IHC for PD-L1, served as our comparator group.

Patients and Tumor Samples
Our overall archive of 367 651 tumor samples, each from a different patient, were sent from medical care facilities across North America for comprehensive genomic profiling (CGP) for detection of targetable GA during routine clinical care. We screened our archive for melanoma cases between January 2018 and May 2021 for which the ordering physician had also requested PD-L1 IHC assay (n = 1268 consecutive cases). Age and sex of patient and specimen site were extracted from accompanying pathology reports. Approval for this study was obtained from the Western Institutional Review Board Protocol No. 20152817, including issuing an informed consent waiver and a HIPAA waiver of authorization.

PD-L1 IHC and Determination of Heavily Pigmented Melanomas vs Non/Lightly Pigmented Melanomas
All PD-L1 IHC testing were performed using the DAKO 22C3 PharmDx assay (Agilent, Santa Clara, CA) following the manufacturer's instructions in a Clinical Laboratory Improvement Amendments (CLIA)-certified and College of American Pathologists (CAP)-accredited reference laboratory (Foundation Medicine, Morrisville, NC). A positive PD-L1 IHC produces a brown stain that can be obscured by heavy melanin pigmentation, rendering it impossible to distinguish positive from negative results. Whenever this technical problem resulted in uninterpretable IHC, it was documented, allowing us to identify HPMel cases. As noted above, cases with interpretable IHC were designated LPMel.

Comprehensive Genomic Profiling
Comprehensive genomic profiling was performed using the FDA approved FoundationOne CDx assay (Foundation Medicine, Cambridge, MA) in a CLIA-certified, CAPaccredited laboratory, as previously described. 23 FoundationOne CDx uses a hybrid capture methodology combined with bioinformatics to detect base substitutions, insertions/deletions, and copy number alterations in 324 genes and select gene rearrangements in 36 genes, as well as TMB and MSI. A board-certified pathologist reviewed a hematoxylin and eosin (H&E) slide from each sample under light microscopy to determine tumor adequacy (at least 20% tumor nuclei present) and review diagnosis. After CGP, a board-certified pathologist reviewed and approved the report. Tumor mutational burden was determined on a standard 0.79 Mb of sequenced DNA and assessment of MSI was performed by analysis of DNA sequencing across 114 loci as previously described. 24,25 High TMB (TMB-H) was defined as ≥10 mutations/Mb. CD274 gene copy number (CN) gain was defined as a CD274 CN of at least one above the overall ploidy of the tumor specimen, copy number loss as a CD274 CN at least one below the overall ploidy, and amplification as a CD274 CN at least 4 above overall tumor ploidy. Predominant genetic ancestry of each patient was determined using principal component analysis of single-nucleotide polymorphisms based on their known variation amongst 5 ancestral superpopulations in the 1000 Genomes Project: African, Central and South American, East Asian, European, and South Asian. [26][27][28] Ultraviolet (UV) mutational signatures were defined as described by Zehir et al 29

Statistical Analyses
We compared characteristics between our HPMel and LPMel cohorts, using Fisher's exact test for categorical data, owing to the sizes of the cohorts, and the non-parametric Wilcoxon test for continuous parameters. A 2-tailed P-value of <.05 was considered to be statistically significant.
Here, we saw a significantly reduced TMB (median 8.8 vs 11.4 mut/Mb, P = .007) and  Table S2). Otherwise, there was no significant difference between the HPMel and LPMel cohorts in the ICPI biomarkers examined in this study (P > .05; Table 2).

Genomic Differences Between HPMel and LPMel
In the overall melanoma cohort (n = 1268), GA were present in  Fig. 2A Table S2).

Discussion
In this study, we have shown that less pigmented melanomas harbor more frequent GAs in TERTp, CDKN2A, TP53, and PTEN, mutations that are classic in the progression of melanomas. These results suggest that LPMel have progressed further along this genomic pathway than HPMel. 30 This, along with the lower TMB and TMB-H status in the HPMel, suggests that pigmented melanomas in general are molecularly less developed and may have a reduced response to ICPI. However, there is marked overlap in TMB between the 2 groups, suggesting that HPMel would nevertheless benefit from ICPI.
Melanomas in sun damaged skin are divided in the current World Health Organization Classification into low cumulative sun damage (CSD) and high CSD, based on the degree of solar elastosis in background dermis, with resulting clinicopathologic and genomic differences. 31 Crucial genomic differences include the identification of BRAF V600E alterations in low CSD melanoma and by contrast BRAF non-V600E and NF1 alterations in high CSD melanoma. A limitation of this study was that stratification into these 2 CSD types was not possible as background solar elastosis was not evaluable for the majority of cases. However, the frequency of GAs in BRAF V600E, BRAF non-V600E, and NF1 did not show significant differences between the HPMel and LPMel groups ( Table 3), indicating that it is likely that both groups contain similar proportions of low and high CSD tumors.
An enrichment of GA in CTNNB1/APC and PRKAR1A was identified in the HPMel cohort. This is consistent with  the literature that GA in these genes are associated with heavily pigmented melanocytic tumors. Specifically, it is well-established that GAs in CTNNB1/APC and PRKAR1A are associated with deep penetrating nevus and pigmented epithelioid melanocytoma, respectively, both of which have abundant melanin pigment as a histolopathologic feature. [32][33][34][35] These alterations have also been described in specific subsets of low CSD melanoma. 31 The increased percentage of specifically KIT alterations in the HPMel group is likely secondary to enrichment for anogenital mucosal primary site cases (8/18 cases), which often contain increased pigmentation. 36,37 From a practical clinical standpoint, this is the first study in the literature that presents real-world failure rates of PD-L1 IHC in HPMel. These cases account for 13.0% (165/1268) of the total population of melanomas submitted for testing at our institution. Almost half were TMB-H, a predictive biomarker that these patients with TMB-H HPMel are likely to respond to ICPI. In addition to the presence of other positive predictive biomarkers such as CD274 amplification and CN gains, negative predictive biomarkers such as CD274 CN loss, STK11 GA, and MDM2 amplification were also present. These findings are further exemplified by the 2 patient cases presented in this study where we saw one patient with HPMel with high TMB and one patient with HPMel with CD274 amplification. Given that a high percentage of these melanoma cases that fail PD-L1 IHC show other biomarkers predictive of response to ICPI, it appears important to test HPMel with CGP in order to guide treatment decisions.
With respect to limitations, it is unclear whether the HPMel cases were sent because the samples were the best available or if the sending physicians were unaware of the complications with samples with extensive pigmentation. Further education in best samples to submit for both PD-L1 IHC and CGP is warranted and may lower real-world failure rates of PD-L1 IHC. Additional studies on the potential utility of red chromogen, such as Fast Red, for PD-L1 IHC may also be warranted. 38 Another limitation of this study is that the cases included were collected from patients with advanced malignancies, submitted for detection of targetable biomarkers and GA. Thus, these cases likely exemplify the aggressive end of the biologic spectrum without representing early, thin, and/or indolent melanomas. Finally, follow-up data were not available in this study, but will be important to obtain for future studies of HPMel to correlate with therapeutic outcomes.

Conclusions
In this study, we quantified the failure rates of melanoma samples for PD-L1 testing due to high melanin pigmentation and showed that CGP can be used in these patients to identify biomarkers that can guide treatment decisions for patients with HPMel. We show that pigmented melanomas in general are molecularly less developed, with a significantly lower but highly overlapping TMB. Additional studies are needed to evaluate if extensive pigmentation in melanoma may be an independent negative predictor for response to immunotherapy.

Data Availability
The data underlying this article are available in the article and in its online supplementary material.