SKIN November 2020 Volume 4 Issue 6 Copyright 2020 The National Society for Cutaneous Medicine 506 ORIGINAL RESEARCH Development and Validation of a Diagnostic 35-Gene Expression Profile Test for Ambiguous or Difficult-To-Diagnose Suspicious Pigmented Skin Lesions Sarah I. Estrada, MD1, Jeffrey B. Shackelton, MD2, Nathan J. Cleaver, DO3, Natalie D. Depcik- Smith, MD4, Clay J. Cockerell, MD5, Stephen N. Lencioni, BS2, Howard L. Martin, MD6, Jeffrey Wilkinson, PhD7, Lauren Meldi Sholl, MS7, Michael D. Berg, PhD7, Brooke H. Russell, PhD7, Olga Zolochevska, PhD7, Kyle R. Covington, PhD7, Aaron S. Farberg, MD8, Matthew S. Goldberg, MD7,9, Pedram Gerami, MD 10, Gregory A. Hosler, MD, PhD11 1Affiliated Dermatology, Scottsdale, AZ. 2Skin Cancer and Dermatology Institute, Reno, NV. 3Cleaver Medical Group, Cumming, GA. 4Aurora Diagnostics GPA Laboratories, Greensboro, NC. 5Cockerell Dermatopathology, Dallas, TX. 6Sagis, Houston, TX. 7Castle Biosciences, Inc., Friendswood, TX. 8Baylor University Medical Center, Dallas, TX. 9Icahn School of Medicine, Mount Sinai, NY. 10Northwestern University, Chicago, IL. 11ProPath, Dallas, TX. ABSTRACT Purpose: A clinical hurdle for dermatopathology is the accurate diagnosis of melanocytic neoplasms. While histopathologic assessment is frequently sufficient, high rates of diagnostic discordance are reported. The development and validation of a 35-gene expression profile (35-GEP) test that accurately differentiates benign and malignant pigmented lesions is described. Methods: Lesion samples were reviewed by at least three independent dermatopathologists and included in the study if 2/3 or 3/3 diagnoses were concordant. Diagnostic utility of 76 genes was assessed with quantitative RT-PCR; neural network modeling and cross-validation were utilized for diagnostic gene selection using 200 benign nevi and 216 melanomas for training. To reflect the complex biology of melanocytic neoplasia, the 35-GEP test was developed to include an intermediate- risk zone. Results: Validation of the 35-GEP was performed in an independent set of 273 benign and 230 malignant lesions. The test demonstrated 99.1% sensitivity, 94.3% specificity, 93.6% positive predictive value and 99.2% negative predictive value. 96.4% of cases received a differential result and 3.6% had intermediate-risk. Conclusions: The 35-GEP test was developed to refine diagnoses of melanocytic neoplasms by providing clinicians with an objective tool. A test with these accuracy metrics could alleviate uncertainty in difficult-to-diagnose lesions leading to decreased unnecessary procedures while appropriately identifying at-risk patients. SKIN November 2020 Volume 4 Issue 6 Copyright 2020 The National Society for Cutaneous Medicine 507 Over 5 million skin biopsies are performed annually in the US, leading to the diagnosis of over 130,000 invasive melanomas.1–5 Because melanoma is one of the most aggressive skin cancers, early detection and diagnosis are crucial.1,6 Current methods used for definitive diagnosis of melanoma are sufficient for the majority of lesions; however, histopathologic assessment can be challenging, even for experienced dermatopathologists, and high rates of diagnostic discordance have been reported.7–11 Even pigmented lesions with clear pathological features consistent with benign nevi or invasive melanoma have concordance rates of 92% and 72%, respectively7, indicating that a subset of lesions with typical histopathological presentation are subject to differential assessment. Visual assessment of hematoxylin and eosin (H&E) stained lesions is inherently subjective and relies on expert interpretation and integration of a wide spectrum of architectural and cytologic features that are weighted differently based on the presumed subtype of melanocytic neoplasm and heavily influenced by the pathologists’ personal experience and training. All melanoma subtypes, including desmoplastic, spitzoid, nevoid, lentigo maligna, and superficial spreading melanoma, can mimic benign nevus variants to varying degrees. Diagnoses require the integration of multiple factors including histopathologic and clinical features, variants of melanoma subtypes, patient age and anatomic location.12 Difficult-to- diagnose lesions are commonly sent for second opinions to expert dermatopathologists who have more experience with challenging cases; however the nature of many lesions remains ambiguous with discordant rates of lesions in this category of 25-43%.7 Studies detailing the prevalence, outcome, and misdiagnosis of these lesions indicate that improved ancillary diagnostic technologies could be greatly beneficial to the dermatopathologist and dermatologist in determining the most appropriate treatment plan.8–10,13–18 Efforts to improve melanoma diagnosis have traditionally focused on ancillary tests such as immunohistochemistry (IHC), fluorescence in situ hybridization (FISH), and comparative genomic hybridization (CGH), but each has limitations.19–22 FISH and CGH may have some limitations including less than optimal specificity and less availability than IHC. IHC is the most commonly utilized diagnostic tool for melanocytic lesions, but IHC, including Ki- 67, Melan-A/MART-1 and p16, is limited in its ability to distinguish benign from malignant melanocytic lesions.23 Similarly, a recently developed PRAME IHC assay has exhibited staining patterns in approximately 14% of nevi, some of which are above the threshold established for a diagnosis of melanoma.24 Definitive diagnosis is also complicated for a subset of lesions described as being borderline, indeterminant, of unknown malignant potential (UMP), atypical melanocytic proliferation (AMP) or in a ‘grey’ zone.25–33 Clinical management of these cases usually results in conservative treatment for the ‘most significant consideration in the differential diagnosis’.27 GEP has been employed to improve the diagnosis of these suspicious pigmented lesions. While a 2-gene pigmented lesion array (2-gene)34–39 and a 23-gene expression profile (GEP) test40,41 have been previously developed, the 2-gene utility is focused on guiding biopsy decisions by dermatologists; whereas, the 23-GEP labels a substantial number of lesions (~15% across studies) as indeterminate rather than INTRODUCTION SKIN November 2020 Volume 4 Issue 6 Copyright 2020 The National Society for Cutaneous Medicine 508 providing a result of benign or malignant.40,41 Approximately 10% of unequivocal cases and 15% of ambiguous lesions may be labeled indeterminate by the 23-GEP test.42 Although sensitivity and specificity is reported at 91.5% and 92.5%40–44, respectively, for the 23-GEP, there exists an opportunity to significantly increase the accuracy and thus optimize the management of the melanoma patient, particularly given the advances in melanoma prognosis and treatment over the past decade. In this study we describe the development and validation of a 35-GEP test to differentiate between benign and malignant pigmented lesions with greater accuracy than previously developed tests. A training cohort of samples, including subtypes considered challenging to diagnose, was established and bioinformatic and machine- learning approaches were used to select and prioritize genes associated with benign or malignant biology. The test was validated using an independent cohort of cases and demonstrates sensitivity and specificity metrics exceeding those currently reported in the melanoma diagnostic literature while maintaining a minimal indeterminate-risk zone. The novel 35-GEP test could aid in the diagnosis of suspicious pigmented lesions and improve accuracy alone or when used in combination with currently applied diagnostic tools. Sample and Clinical Data Collection Archival benign samples and associated de- identified clinical data were collected from multiple independent dermatopathology laboratories as part of this Institutional Review Board (IRB)-approved study. Formalin-fixed, paraffin-embedded (FFPE) pigmented lesion tissue was collected as 5 μm sections for subsequent diagnosis based on H&E staining and for real-time quantitative reverse transcription PCR (qRT- PCR) analysis. Additionally, archival melanoma samples and de-identified clinical data were obtained from specimens submitted to Castle Biosciences for clinical testing with the 31-GEP (DecisionDx- Melanoma). A total of 951 samples diagnosed between January 2013 and August 2020 were included in the training and validation cohorts, of which 498 were benign and 453 malignant. All laboratory personnel were blinded to clinical diagnoses for all 951 samples. Samples were excluded from the study if there was less than 10% tumor volume (cellularity of all samples was determined by a single dermatopathologist), tissue originated from melanoma metastases, lesions were not primary to the skin, tissue was derived from re-excisions (including wide local excision), diagnosis was a non- melanocytic neoplasm, or if patients had previous radiation or immunotherapy treatment. Melanoma subtypes of acral lentiginous, desmoplastic, lentiginous, lentigo maligna, nevoid, nodular, spitzoid, superficial spreading, and melanoma in situ were included. Benign subtypes of blue nevus, common nevus (compound, junctional and intradermal), deep penetrating nevus, dysplastic nevus (compound and junctional), and Spitz nevus were included. Histopathologic Examination Eight dermatopathologists participated in sample acquisition, and six dermatopathologists participated in sample review for diagnostic concordance. The majority of these dermatopathologists are affiliated with private practice and have an average of 12 years of experience reviewing skin lesions. All acquired samples were received with the original pathology report. METHODS SKIN November 2020 Volume 4 Issue 6 Copyright 2020 The National Society for Cutaneous Medicine 509 For all benign diagnoses, the contributing dermatopathologists provided a description of the lesion in a free text field, and the information was entered into the clinical research form. All benign samples then underwent H&E diagnostic review by a second and third dermatopathologist who were blinded to the original diagnosis and provided with only patient age and anatomic location of the lesion. Reviewing dermatopathologists were asked to select a diagnosis (benign, malignant, or unknown (unknown malignant potential (UMP)) as well as a subtype classification from a pre- determined list. If discordance was observed across three diagnoses, the case was reviewed by additional dermatopathologists in a blinded manner for adjudication. A total of 395 samples that were diagnosed as benign by 3 out of 3 dermatopathologists were included in the study; additionally, 78 cases diagnosed as UMP by no more than 1 dermatopathologist (i.e. 2 benign and 1 UMP) were added to the training and validation cohorts. As a result, the final training and validation cohorts consisted of benign samples with full diagnostic concordance (167/200 and 228/273 of samples, respectively) and samples with no more than one UMP classification (33/200 and 45/273, respectively). Real-Time Quantitative Reverse- Transcription PCR Pigmented lesions were processed for qRT- PCR expression analysis in a central CLIA- certified, CAP-accredited, and New York State Department of Health permitted laboratory. Tumor sections were macrodissected from unstained FFPE tissue and total RNA was extracted per manufacturer’s instructions using either the QIAsymphony SP Automated Nucleic Acid Extractor (Qiagen) or KingFisher Flex (ThermoFisher Scientific) platforms. Total RNA concentration was quantified using the NanoDrop 8000 (ThermoFisher Scientific). cDNA was obtained using the High-Capacity cDNA Reverse Transcription Kit (Applied Biosystems). cDNA pre-amplification reaction was performed utilizing the TaqMan PreAmp Master Mix (Applied Biosystems) and a 14-cycle amplification. Pre-amplified samples were diluted 1:2.5X in TE Buffer pH 7.0 (ThermoFisher Scientific) and combined with an equal volume of 2X Open Array Real-Time PCR Master Mix (Applied Biosystems). The samples were loaded onto a custom Open Array gene card using the QuantStudio 12K Flex AccuFill system (Applied Biosystems) subsequently run on the QuantStudio 12K PCR system. Expression Analysis and Diagnosis Assignment The array data was analyzed to identify genes that were best able to segregate benign and malignant lesions based on levels of gene expression.45–47 The resulting gene set was then reviewed to ensure that a wide variety of biological pathways were represented and to confirm the biological relevance of those genes. As a result, 76 candidate diagnostic genes were selected for model training. Three genes (FXR1, HNRNPL, and YKT6) were reliably and consistently expressed in the study cohort and chosen as control genes. Triplicate gene expression data were aggregated and normalized using control probes. Failure of three or more candidate genes (MGF, multiple gene failure) led to sample exclusion from training and validation cohorts, while control genes were evaluated independently, and failure of any control gene resulted in sample exclusion. Following quality control measures to assess amplification and stability of gene expression, 58 discriminant probes and 3 control probes were selected for further analysis. Deep learning techniques were applied to gene expression data for gene selection and model identification.48–50 Gene expression data analyzed with neural SKIN November 2020 Volume 4 Issue 6 Copyright 2020 The National Society for Cutaneous Medicine 510 network modeling resulted in two diagnostic algorithms.51,52 Tumors with spitzoid or melanoma in situ features had poorer initial classification accuracy; therefore, the presence of those features in diagnosis was added to the input of the algorithm to improve accuracy. Algorithm improvement continued until the mean kappa value improved by less than 0.01 for the top 25% of the assay population. Hyperparameter selection and model evaluation was performed using 4x4-fold cross validation.50,53 Kappa was determined from the average kappa value at each of the cross validation runs. The final model was trained against all training data using the optimal gene set. Two models were developed which together generated the locked algorithm for the 35-GEP test. Classification into benign (gene expression profile suggestive of benign neoplasm), intermediate-risk (gene expression profile cannot exclude malignancy) or malignant (gene expression profile suggestive of melanoma) zones was determined from the probability scores from both algorithms. Analysis was performed with R v.3.3.3. Differences in age were assessed using the Wilcoxon F test. Differences in categorical variables including sex, ulceration status and location were assessed by Pearson Chi- square test. P values <0.05 were considered statistically significant. Sample Cohorts Quantitative RT-PCR was performed on 498 benign and 453 malignant lesions accrued under an IRB-approved protocol in a multicenter cohort (Figure 1). Thirty-two samples (~3.4%) of the study cohort were excluded from further analysis due to MGF with the remaining 919 samples randomized to training or validation cohorts while conserving benign or melanoma subtype representation in each cohort. Training (200 benign nevi and 216 melanomas) and validation (273 benign nevi and 230 melanomas) cohorts’ demographic details are shown in Table 1. No statistically significant differences were observed in the training vs. validation cohorts. The median age of patients with benign lesions was 47 (range 7-85) years in the training cohort and 48 (2-90) years in the validation cohort (p=0.944), while patients with malignant tumors had a median age of 66 (range 18- 93) and 67 (25-98) years of age (p=0.203), respectively. Training and validation cohorts had 55% and 63% (p=0.071) male patients with malignant diagnosis, while 46% and 39% (p=0.17) males were included in training and validation, respectively. Ulceration was present in 29.5% (64/216) melanomas in training and 23.5% (54/230) melanomas in validation cohorts (p= 0.141). The majority of malignant lesions were biopsied from arms and legs (extremities, 40% of cases in training and validation, p=0.812), while benign lesions were mainly located on patients’ backs (36.5% in training cohort and 41% in validation, p=0.863). The distribution of different subtypes of melanoma and nevi in the training and validation sets are provided in Table 2. Development of 35-GEP Profile Artificial neural networks were selected as the model type due to their ability to recognize multiple patterns, which is critical for successfully distinguishing different subtypes of benign nevi and melanomas. Therefore, to represent biological diversity and different growth patterns and features, lesions unanimously diagnosed as benign by 3/3 reviewers and lesions with less definitive histopathology resulting in 2/3 concordance were included in the training set to ensure the resulting algorithm is RESULTS SKIN November 2020 Volume 4 Issue 6 Copyright 2020 The National Society for Cutaneous Medicine 511 Figure 1. Study Cohorts *MGF (multiple gene failure) rules: control genes were evaluated independently and failure of any control gene resulted in sample exclusion. Triplicate gene expression data were aggregated and normalized using control probes. Samples with failure to amplify ≥3 out of 73 genes were excluded from model development and validation. capable of classifying both typical and heterogenous lesions. A 35-GEP comprising 32 discriminant genes and 3 control genes was developed using neural networks for model fitting and genetic algorithms for feature selection on a diverse set of benign and malignant samples. The 35-GEP is primarily composed of genes in cytoskeletal and barrier functions, gene regulation and melanin biosynthesis (Table 3).54–78 Multiple molecular pathways have been associated with melanoma progression and the 35-GEP signature includes several genes from key signaling networks to encompass the complexity of the disease. Biological processes such as epithelial cell differentiation, tissue and epidermis development, programmed cell death, and keratinocyte differentiation were identified as top functional enrichments for this gene set. Validation of the 35-GEP Accuracy metrics within the validation cohort (all ages included) were 99.1% (95% CI: 97.9-100%) sensitivity, 94.3% (95% CI: 91.5-97.1%) specificity, 93.6% (95% CI: 90.5-96.7%) positive predictive value (PPV) and 99.2% (95% CI: 98.1-100%) negative predictive value (NPV) (Table 4), suggesting that the 35-GEP test could be a highly accurate ancillary test for diagnosis of melanocytic neoplasms. In patients ≥18 years old the 35-GEP had sensitivity of 99.1% (95% CI: 97.9-100%), specificity of 96.2% (95% CI: 93.8-98.6%), PPV of 96.1% (95% CI: 93.6-98.6%) and NPV of 99.1% (95% CI: 97.9-100%) (Table 4). Accuracy metrics were calculated without the inclusion of lesions identified as intermediate-risk (3.6% and 3.8% of the total samples in all ages and ≥18 years old, respectively). Overall, the 35-GEP was able to accurately classify different subtypes of melanoma and nevi as benign or malignant (Table 5). The 35-GEP accurately classified melanoma lesions as malignant in 14/14 desmoplastic melanomas, 25/26 lentigo maligna, 15/15 nevoid, 59/60 nodular, 72/77 superficial spreading, and 17/19 melanoma in situ. Furthermore, nevi were also appropriately classified as benign for 42/45 blue, 96/99 common nevi (including 15 compound, 40 intradermal and 10 junctional), 82/91 dysplastic nevi (including 44 compound and 38 junctional), and 26/36 Spitz nevi. Of 230 melanomas, two were identified as benign, while 15 of 273 benign lesions were classified as malignant (Table 6). One of the melanomas that was classified as benign was in situ and one was nodular melanoma with a Breslow thickness of 4.0 mm that had the low-risk prognostic Class 1B 31-GEP result. Among the 15 benign lesions that were classified as malignant by the 35-GEP, four were dysplastic (one compound with mild atypia and three junctional with mild/moderate atypia), one compound nevus, one combined blue and intradermal nevus, one blue nevus, one benign SKIN November 2020 Volume 4 Issue 6 Copyright 2020 The National Society for Cutaneous Medicine 512 Table 1. Demographic information for training and validation cohorts Training Cohort† Validation Cohort† Melanoma N=216 Benign nevi N=200 Melanoma N=230 Benign nevi N=273 Age, median (range) 66 (18-93) 47 (7-85) 67 (25-98) 48 (2-90) Sex, % male 55 46 63 39 Breslow thickness, mm (range) 1.22 (0-10) NA 1.23 (0.1-4.9) NA T stage, % (n) T1a 29 (56) - 23 (48) - T1b 13 (25) - 20 (42) - T2a 16 (31) - 16.5 (35) - T2b 14 (27) - 11 (23) - T3a 11.5 (23) - 17.5 (37) - T3b 16 (31) - 11 (23) - T4b 0.5 (1) - 1 (2) - Ulceration % (n) Present 29.5 (64) - 23.5 (54) - Absent 70.5 (152) - 76.5 (176) - Not addressed - 100 (200) - 100 (273) Location on body, % (n) Abdomen/Chest 8 (18) 11 (22) 5.5 (13) 11.5 (32) Acral 3 (6) 1 (2) 2 (5) 1 (2) Back 27 (58) 36.5 (73) 29 (67) 41 (113) Extremities 40 (86) 23 (46) 40 (91) 20 (54) Head/Neck 20 (43) 24 (48) 22 (50) 23 (63) Other 2 (5) 4.5 (9) 1.5 (4) 3.5 (9) †No statistically significant differences were observed in the training vs. validation cohorts. NA – not addressed. melanocytic nevus (not otherwise specified), and seven were Spitz nevi. Six of the seven misclassified Spitz nevi were in pediatric patients suggesting this may be a limitation of the 35-GEP (accuracy metrics for the validation cohort without lesions with spitzoid features is provided in Table 4). Spitzoid lesions are particularly difficult to diagnose as many have ambiguous histologic characteristics and may involve regional lymph nodes in the absence of increased mortality rates or malignant potential.16,30,79 35-GEP Intermediate-Risk Zone Given the potential biological transition of melanocytic lesions from a benign to a malignant state, the 35-GEP profile was developed to identify lesions with an intermediate-risk of malignancy. Inclusion of a wide variety of subtypes in the study improved the classification of lesions as benign or malignant and led to a limited intermediate-risk zone. A total of 96.4% of cases had a definitive benign or malignant test result and only 3.6% (18/503) of cases were classified as intermediate-risk, including eight melanomas and ten benign nevi. Though a definitive benign or malignant result is advantageous for implementing patient management pathways, samples with probability scores in the intermediate-risk zone may be evolving, borderline, or atypical and warrant special consideration in terms of patient management with a focus on clinicopathologic correlation. Dermatopathologists have a number of ancillary tools available to assist with the DISCUSSION SKIN November 2020 Volume 4 Issue 6 Copyright 2020 The National Society for Cutaneous Medicine 513 diagnosis of pigmented lesions, yet there is a substantial amount of diagnostic discordance that may potentially lead to overtreatment of patients with benign lesions and undertreatment of patients with melanoma.80 The 35-GEP test to distinguish benign from malignant pigmented lesions was developed to improve diagnostic accuracy and reduce diagnostic uncertainty for difficult-to-diagnose cases. Table 2. Distribution of subtypes in training and validation cohorts Training Cohort, n Validation Cohort, n P value Melanoma 216 230 0.399 Acral lentiginous 6 5 Desmoplastic 15 14 Lentiginous 3 3 Lentigo maligna 23 26 In situ 21 19 Nevoid 15 15 Nodular 47 60 Superficial spreading 66 77 Spitzoid 15 3 Not specified 5 8 Nevi 200 273 0.468 Blue 38 45 Common nevi Compound 16 16 Intradermal 20 41 Junctional 10 10 Not specified 30 32 Deep penetrating 1 2 Dysplastic Compound 40a 49b Junctional 28c 42d Spitz 17 36 P value was calculated using the Pearson Chi-square test. Dysplastic nevi had different degrees of atypia: a - mild (n=19), moderate (n=4) and severe (n=3); b - mild (n=24), moderate (n=2), and severe (n=3); c - mild (n=20) and moderate (n=6); d - mild (n=22) and moderate (n=17) atypia. Table 3. Genes Included in the 35-GEP and their Functions Gene classification Gene symbol Gene name Barrier function HAL Histidine ammonia-lyase Barrier function MGP* Matrix Gla protein Barrier function CST6* Cystatin-M Barrier function GJA1* Gap junction alpha-1 protein Barrier function CSTA Cystatin A Barrier function CLCA2* Calcium-activated chloride channel regulator 2 Cytoskeleton involved KRT17 Keratin, type I cytoskeletal 17 Cytoskeleton involved PPL* Periplakin Cytoskeleton involved KRT2 Keratin 2 Cytoskeleton involved ABLIM1 Actin binding LIM protein 1 Cytoskeleton involved DSP Desmoplakin Cytoskeleton involved NES Nestin Gene regulation KLF5 Kruppel-like factor 5 Gene regulation GATA3 GATA binding protein 3 Gene regulation BAP1* Ubiquitin carboxyl- terminal hydrolase BAP1 Gene regulation TP63 Tumor Protein P63 Gene regulation SAP130* Histone deacetylase complex subunit SAP130 Gene regulation SFN 14-3-3 protein sigma Melanin Biosynthesis GPR143 G-protein coupled receptor 143 Melanin Biosynthesis WIPI1 WD repeat domain phosphoinositide- interacting protein Melanin Biosynthesis DCT Dopachrome tautomerase Melanin Biosynthesis ATP6V0E2 ATPase H+ transporting V0 subunit E2 SKIN November 2020 Volume 4 Issue 6 Copyright 2020 The National Society for Cutaneous Medicine 514 Melanin Biosynthesis PTN Pleiotrophin Protein synthesis RPS16 40S ribosomal protein S16 Protein synthesis RPL37A 60S ribosomal protein L37a Tumorigenesis BCL2A1 Bcl-2-related protein A1 Tumorigenesis BTG1* Protein BTG1 Tumorigenesis ANXA8L1 Annexin A8-like protein 1 Tumorigenesis DUSP4 Dual specificity protein phosphatase 4 Tumorigenesis CXCL14* C-X-C motif chemokine 14 Tumorigenesis S100A8* Protein S100-A8 Tumorigenesis S100A9* Protein S100-A9 Housekeeping FXR1* RNA binding protein Housekeeping HNRNPL* mRNA function protein Housekeeping YKT6* ER membrane protein * Fourteen genes are also included in the 31-GEP test. A cross-study analysis shows that in an independent validation cohort of 503 benign lesions and melanomas the 35-GEP test demonstrated improved accuracy compared to other diagnostic tools based on their primary validation studies (Table 7). Unlike FISH, CGH or IHC, gene expression profiling captures transcriptomic events within the lesion and the surrounding tissue, allowing for a more comprehensive assessment of the biological changes that are associated with the transition to a malignant phenotype.22,81,82 IHC generally allows for evaluation of changes in the expression of a single biomarker at the protein level, which can be limited by subjective quantification systems.83 PRAME IHC has been reported as a reliable method to distinguish benign from malignant pigmented lesions; however, ~14% of nevi can have some staining for PRAME and the interpretation of positive staining (4+, ≥76% of immunoreactive tumor cells are PRAME positive) can be subjective. Thus, PRAME IHC requires further validation for widespread clinical use due to the potential for misdiagnosis of benign lesions as malignant.24,84 In the current study, PRAME expression did not improve diagnostic accuracy above the results reported for the 35-GEP (data not shown). In this study, the 35-GEP reliably diagnosed 96.4% of benign and malignant lesions. In cross-study comparison (Table 7), the 35- test out-performed a 23-GEP diagnostic test with previously reported accuracy metrics for unequivocal samples ranging from 91.5-94% for sensitivity, 90.0-92.5% for specificity, and technical failures in 14.7%. Moreover, ~15% of diagnostically concordant (i.e. 3 out of 3) cases could not be classified as benign or malignant.40,41 By comparison, the 35-GEP test demonstrated sensitivity (99.1%) and specificity (94.3%) in all ages and 99.1% sensitivity and 96.2% specificity in patients ≥18 years old, a low number of technical failures (3.4%), and no more than 3.8% of cases received an intermediate-risk result. The improved classification of lesions compared to that of the 23-GEP test is likely due to implementation of highly sophisticated modeling (neural networks) that resulted in two algorithms with 32 diagnostic and 3 control genes, the inclusion of samples with different growth patterns (total of nine melanoma subtypes and eight benign subtypes) in the training cohort as well as incorporation of lesions with 2/3 concordance. Data supporting the utility of the 23-GEP test in ambiguous or diagnostically discordant lesions is limited.44 Recently, sensitivity of 90.4% and specificity of 95.5% was reported for 125 ‘uncertain’ cases, however, the definition of uncertainty was broad and included lesions as discordant if a differing SKIN November 2020 Volume 4 Issue 6 Copyright 2020 The National Society for Cutaneous Medicine 515 Table 4. The 35-GEP Accuracy Metrics All ages N=503 >18 years old N=478 All ages* N=464 ≥18 years old* N=457 35-GEP 95% CI 35-GEP 95% CI 35-GEP 95% CI 35-GEP 95% CI Sensitivity 99.1% 97.9-100 99.1% 97.9-100 99.1% 97.8-100 99.1% 97.8-100 Specificity 94.3% 91.5-97.1 96.2% 93.8-98.6 96.5% 94.2-98.9 96.4% 94.0-98.9 PPV 93.6% 90.5-96.7 96.1% 93.6-98.6 96.5% 94.1-98.9 96.5% 94.1-98.9 NPV 99.2% 98.1-100 99.1% 97.9-100 99.1% 97.9-100 99.1% 97.8-100 Intermediate-risk result 3.6% 3.8% 3.0% 3.1% Samples that fall in intermediate-risk zone were excluded from the calculation. *Lesions with Spitzoid features were excluded. PPV – positive predictive value; NPV – negative predictive value; CI – confidence interval. diagnosis was received from just 1 of 7 dermatopathologists reviewing the cases.85 In this study, we included cases with concordance for 2 of 3 reviewing dermatopathologists in this independent 35- GEP validation. The 35-GEP was developed and validated using fully concordant lesions and a small set of ‘borderline’ cases, where no more than 1 out of 3 dermatopathologists indicated ‘unknown malignant potential’ as a diagnosis. Since the 35-GEP will be most likely used in difficult-to-diagnose lesions, inclusion of 2/3 concordant cases to capture differentially expressed genes from those histopathologically challenging cases was factored into the neural network configuration during the test development. With the improved accuracy metrics and substantially reduced intermediate-risk zone, dermatopathologists can expect a definitive result from the 35-GEP test in ≥95% of lesions submitted for testing. It is our hope that improved test characteristics for the disambiguation of pigmented lesions will help refine guidelines for when to utilize GEP in the diagnosis of challenging pigmented lesions. Although the vast majority of cases tested by the 35-GEP will have a definitive score of benign or malignant risk potential, 3.6% of cases fell into an intermediate-risk zone reflective of a molecular biology characteristic of both benign and malignant lesions. Though the prevalence is not known, evidence that there is a true ‘transition’ zone for pigmented lesions is mounting.46 Thus, interpretation of an intermediate-risk result of the 35-GEP should be considered in the context of other clinicopathological information. Specifically, in cases with an intermediate-risk score, it would be of great diagnostic importance to exclude the possibility of sampling error and ensure that the entire clinical lesion has been evaluated by routine histopathology. Unfortunately, up to 1/3 of nevi transition to melanoma, so there is a subset of lesions that may be clinically identified during this progression.12,86 In addition, there are atypical melanocytic proliferations (AMPs) that never evolve to full malignancy despite metastasis to regional lymph nodes. The spectrum of outcomes for these lesions warrants special consideration in clinical management.28 Clinical management of AMPs varies as there are no official guidelines governing their treatment, but common practice is definitive surgical treatment with removal of lesion with the margin of normal skin.28 In addition, the use of the 35-GEP can provide the dermatologist and/or patient with treatment options to cover the most severe of diagnoses, including a diagnosis of melanoma. Studies are underway in a true AMP population with SKIN November 2020 Volume 4 Issue 6 Copyright 2020 The National Society for Cutaneous Medicine 516 Table 5. Performance of the 35-GEP in Different Subtypes of Nevi and Melanoma 35-GEP result Benign, n Intermediate- risk, n Malignant, n Melanomas 2 8 220 Acral lentiginous 5 Desmoplastic 14 Lentiginous 3 Lentigo maligna 1 25 In situ 1 1 17 Nevoid 15 Nodular 1 59 Spitzoid 1 2 Superficial spreading 5 72 Not specified 8 Nevi 248 10 15 Blue 42 2 1 Common nevi Compound 15 1 Intradermal 40 1 Junctional 10 Not specified 31 1 Deep penetrating nevus 2 Dysplastic Compound 44a 4b 1c Junctional 38d 1e 3f Spitz 26 3 7 Dysplastic nevi had different degrees of atypia: a – mild (n=22), moderate (n=2) and severe (n=3); b – mild (n=1); c - mild (n=1); d - mild (n=21) and moderate (n=14); e – moderate (n=1); f – mild (n=1) and moderate (n=2) atypia. known outcomes. The 35-GEP test performed equally well in nevi and melanomas with different growth patterns. For instance, classification of lentigo maligna, nodular and superficial spreading melanomas was concordant with dermatopathologic diagnosis as were blue and common nevi (compound, intradermal and junctional), along with dysplastic nevi with varying degree of atypia with only a small percentage receiving an intermediate- risk result. Further studies to increase the number of samples in subtypes that were not represented in large enough numbers are underway. Accuracy metrics of the 35- GEP with and without spitzoid lesions and pediatric participants are presented in Table 4. The Spitz subtype is particularly challenging and thus far all available ancillary tests have had limitations in sensitivity and specificity.42,87,88 Of note, absence of spitzoid melanomas and classification of the Spitz lesions in pediatric patients was not optimal in this study and therefore further studies are being undertaken to confirm whether this is a limitation of the 35-GEP. For the dermatologist, metastatic risk assessment is critical for guiding appropriate patient management following a melanoma diagnosis. A prognostic 31-GEP test has been validated to determine individualized 5- year risk for recurrence, metastasis and melanoma-specific survival.89–91 Based on accuracy metrics and multivariate models demonstrating that the test is an independent and significant risk-prediction tool, the value of the GEP testing as an adjunct to current staging factors has been recognized by the National Comprehensive Cancer Network.92 Thus, patients diagnosed with malignant lesions have effective prognostic tools and contemporary therapies, with demonstrated improved outcomes, at their disposal. Given the availability of the prognostic 31- GEP test for cutaneous melanoma, the 35- GEP test was developed to refine the diagnosis of benign nevi and melanomas by providing dermatopathologists with an objective ancillary tool to aid in their diagnosis of difficult-to-diagnose pigmented CONCLUSION SKIN November 2020 Volume 4 Issue 6 Copyright 2020 The National Society for Cutaneous Medicine 517 Table 6. Seventeen cases (3.4% of the independent validation cohort) were misclassified by the 35-GEP Sample number 35-GEP result Sex Age Growth pattern submitted by dermatopathologist Atypia Breslow thickness, mm Ulceration Location 91 malignant Male 2 Spitz Nevus None NA NA Extremities 315 malignant Male 3 Spitz Nevus, Compound None NA NA Extremities 95 malignant Male 6 Spitz Nevus None NA NA Head/Neck 98 malignant Male 6 Spitz Nevus None NA NA Head/Neck 129 malignant Male 8 Benign Pigmented Spindle-Cell Nevus of Reed (Variant of Spitz) None NA NA Extremities 318 malignant Male 8 Pigmented Spindle Cell Variant of Spitz's Nevus None NA NA Head/Neck 297 malignant Male 33 Junctional Melanocytic Nevus Mild NA NA Back 266 malignant Female 36 Junctional Spitz Nevus None NA NA Extremities 300 malignant Male 37 Compound Dysplastic Melanocytic Nevus Mild NA NA Back 69 malignant Female 42 Combined Blue and Intradermal Nevus None NA NA Abdomen/Chest 46 malignant Female 43 Junctional Dysplastic Nevus Moderate NA NA Back 138 malignant Female 55 Benign Melanocytic Nevus None NA NA Acral 324 malignant Female 56 Compound Melanocytic Nevus None NA NA Back 323 malignant Female 58 Intradermal Melanocytic Nevus None NA NA Extremities 313 malignant Male 61 Junctional Melanocytic Nevus Moderate NA NA Back 566 benign Female 63 Melanoma in situ NA NA No Extremities 404 benign Male 83 Nodular melanoma NA 4.0 No Head/Neck NA – not addressed. lesions. Clinically implemented GEP tests for diagnostically challenging melanocytic lesions have demonstrated high impact on utility for guiding decision-making.93,94 Although not the focus of this study, assessment of 35-GEP clinical utility, as well as correlation of test results with outcomes, is underway. In the zone of significant uncertainty, the high accuracy metrics of the test might increase confidence level in diagnosis to dermatopathologists and dermatologists, while providing assurance to the patients. The test also provided a definitive result for 96.4% of the lesions in the validation study, offering an opportunity to reduce the uncertainty associated with pigmented lesions and promote more definitive management of patients by dermatologists. An ancillary test with the characteristics reported here could impact expenditure on over-diagnoses by decreasing unnecessary surgeries, imaging and follow-up while more appropriately allocating healthcare resources to those lesions where malignant risk is identified. SKIN November 2020 Volume 4 Issue 6 Copyright 2020 The National Society for Cutaneous Medicine 518 Table 7. Comparison of 35-GEP to Currently Available Ancillary Tests Study Number of cases Type of test Sensitivity Specificity Technical failure Nevi subtypes included Melanoma subtypes included Current study 503 35-GEP 99.1% 94.3% 3.4% Blue, common, deep penetrating, dysplastic, Spitz Acral, desmoplastic, lentiginous, lentigo maligna, in situ, nevoid, nodular, superficial spreading, spitzoid Clarke et al.41 437 23-GEP 94.0% 90.0% 14.7% Blue, common, dysplastic, Spitz Acral, lentigo maligna, nodular, superficial spreading Clarke et al.40 736 23-GEP 91.5% 92.5% NA Not reported Acral, lentigo maligna, nodular, superficial spreading Gerami et al.92 196 FISH# 86.7% 95.4% NA Acral, blue, common, dysplastic, Spitz Not reported Gerami et al.93 233 FISH# 83.0% 94.0% NA Blue, common, dysplastic, Spitz Acral, lentigo maligna, nodular, superficial spreading Lezcano et al.24 400 PRAME IHC 84.7%& 99.2%& NA Common, dysplastic, Spitz Acral, cutaneous paramucosal, desmoplastic, lentigo maligna, nevoid, nodular, superficial spreading Lezcano et al.87 110 PRAME IHC 75.0% 98.8% NA Blue, common, deep penetrating, dysplastic, Spitz Acral, malignant melanoma, nevoid, spitzoid # 6p25, Cep 6, 6q23, and 11q13 & Calculated from the data reported in the manuscript. NA – not addressed. Conflict of Interest Disclosures: SIE is a consultant and shareholder of Castle Biosciences, Inc. ASF, CJC, PG are consultants for Castle Biosciences, Inc. OZ, BHR, JW, LMS, LDB, and KRC are employees and shareholders of Castle Biosciences, Inc. MSG is an employee of Castle Biosciences, Inc. Funding: This study was sponsored by Castle Biosciences, Inc. Corresponding Author: Sarah I. Estrada, MD 20401 N 73rd St., #230 Scottsdale, AZ 85255 Phone: 480-563-7507 Fax: 480-563-7508 Email: sestrada@affderm.com References: 1. National Cancer Institute, National Institutes of Health. Melanoma of the Skin - Cancer Stat Facts. SEER. Published 2020. Accessed October 24, 2019. https://seer.cancer.gov/statfacts/html/melan.html 2. Wang DM, Morgan FC, Besaw RJ, Schmults CD. An ecological study of skin biopsies and skin cancer treatment procedures in the United States Medicare population, 2000 to 2015. J Am Acad Dermatol. 2018;78(1):47-53. doi:10.1016/j.jaad.2017.09.031 3. Cartee TV, Kini SP, Chen SC. Melanoma reporting to central cancer registries by US dermatologists: An analysis of the persistent knowledge and practice gap. J Am Acad Dermatol. 2011;65(5):S124.e1-S124.e9. doi:10.1016/j.jaad.2011.05.032 4. Cockburn M, Swetter SM, Peng D, Keegan THM, Deapen D, Clarke CA. Melanoma underreporting: Why does it happen, how big is the problem, and how do we fix it? J Am Acad Dermatol. 2008;59(6):1081-1085. doi:10.1016/j.jaad.2008.08.007 5. Heuring E, Chen SC. Melanoma underreporting among US dermatopathologists: A pilot study. J Cutan Pathol. 2018;45(7):550-551. doi:10.1111/cup.13149 mailto:sestrada@affderm.com SKIN November 2020 Volume 4 Issue 6 Copyright 2020 The National Society for Cutaneous Medicine 519 6. Siegel RL, Miller KD, Jemal A. Cancer statistics, 2020. CA Cancer J Clin. 2020;70(1):7-30. doi:10.3322/caac.21590 7. Elmore JG, Barnhill RL, Elder DE, et al. Pathologists’ diagnosis of invasive melanoma and melanocytic proliferations: observer accuracy and reproducibility study. BMJ. 2017;357:j2813. doi:10.1136/bmj.j2813 8. Shoo BA, Sagebiel RW, Kashani-Sabet M. Discordance in the histopathologic diagnosis of melanoma at a melanoma referral center. J Am Acad Dermatol. 2010;62(5):751-756. doi:10.1016/j.jaad.2009.09.043 9. Farmer ER, Gonin R, Hanna MP. Discordance in the histopathologic diagnosis of melanoma and melanocytic nevi between expert pathologists. Hum Pathol. 1996;27(6):528-531. doi:10.1016/s0046- 8177(96)90157-4 10. Patrawala S, Maley A, Greskovich C, et al. Discordance of histopathologic parameters in cutaneous melanoma: Clinical implications. J Am Acad Dermatol. 2016;74(1):75-80. doi:10.1016/j.jaad.2015.09.008 11. Warycha MA, Christos PJ, Mazumdar M, et al. Changes in the Presentation of Nodular and Superficial Spreading Melanomas Over 35 Years. Cancer. 2008;113(12):3341-3348. doi:10.1002/cncr.23955 12. Damsky WE, Bosenberg M. Melanocytic nevi and melanoma: unraveling a complex relationship. Oncogene. 2017;36(42):5771-5792. doi:10.1038/onc.2017.189 13. Diwan AH, Lazar AJ. Nevoid Melanoma. Clin Lab Med. 2011;31(2):243-253. doi:10.1016/j.cll.2011.03.002 14. DeWane ME, Kelsey A, Oliviero M, Rabinovitz H, Grant-Kels JM. Melanoma on chronically sun- damaged skin: Lentigo maligna and desmoplastic melanoma. J Am Acad Dermatol. 2019;81(3):823- 833. doi:10.1016/j.jaad.2019.03.066 15. King R. Lentiginous Melanoma. Arch Pathol Lab Med. 2011;135(3):337-341. doi:10.1043/2009-0538- RA.1 16. Harms KL, Lowe L, Fullen DR, Harms PW. Atypical Spitz Tumors: A Diagnostic Challenge. Arch Pathol Lab Med. 2015;139(10):1263-1270. doi:10.5858/arpa.2015-0207-RA 17. Yang GB. Risk and Survival of Cutaneous Melanoma Diagnosed Subsequent to a Previous Cancer. Arch Dermatol. 2011;147(12):1395. doi:10.1001/archdermatol.2011.1133 18. Magro CM, Crowson AN, Mihm MC. Unusual variants of malignant melanoma. Mod Pathol. 2006;19(2):S41-S70. doi:10.1038/modpathol.3800516 19. Troxel DB. Pitfalls in the Diagnosis of Malignant Melanoma: Findings of a Risk Management Panel Study. Am J Surg Pathol. 2003;27(9):1278–1283. 20. High WA. Malpractice in Dermatopathology— Principles, Risk Mitigation, and Opportunities for Improved Care for the Histologic Diagnosis of Melanoma and Pigmented Lesions. Clin Lab Med. 2008;28(2):261-284. doi:10.1016/j.cll.2007.12.006 21. Fryback DG, Thornbury JR. The Efficacy of Diagnostic Imaging. Med Decis Making. 1991;11(2):88-94. doi:10.1177/0272989X9101100203 22. Miedema J, Andea AA. Through the looking glass and what you find there: making sense of comparative genomic hybridization and fluorescence in situ hybridization for melanoma diagnosis. Mod Pathol. Published online February 17, 2020. doi:10.1038/s41379-020-0490-7 23. Ohsie SJ, Sarantopoulos GP, Cochran AJ, Binder SW. Immunohistochemical characteristics of melanoma. J Cutan Pathol. 2008;35(5):433-444. doi:10.1111/j.1600-0560.2007.00891.x 24. Lezcano C, Jungbluth AA, Nehal KS, Hollmann TJ, Busam KJ. PRAME Expression in Melanocytic Tumors: Am J Surg Pathol. 2018;42(11):1456-1465. doi:10.1097/PAS.0000000000001134 25. Berk DR, LaBuz E, Dadras SS, Johnson DL, Swetter SM. Melanoma and Melanocytic Tumors of Uncertain Malignant Potential in Children, Adolescents and Young Adults-The Stanford Experience 1995-2008: Melanoma and Melanocytic Tumors of Uncertain Malignant Potential. Pediatr Dermatol. 2010;27(3):244-254. doi:10.1111/j.1525- 1470.2009.01078.x 26. Cockerell CJ. Commentary on Atypical Melanocytic Proliferations: Dermatol Surg. 2018;44(2):175-176. doi:10.1097/DSS.0000000000001365 27. Elder DE, Xu X. The approach to the patient with a difficult melanocytic lesion. Pathology (Phila). 2004;36(5):428-434. doi:10.1080/00313020412331283905 28. Ensslin CJ, Hibler BP, Lee EH, Nehal KS, Busam KJ, Rossi AM. Atypical Melanocytic Proliferations: A Review of the Literature. Dermatol Surg Off Publ Am Soc Dermatol Surg Al. 2018;44(2):159-174. doi:10.1097/DSS.0000000000001367 29. Hillen LM, Van den Oord J, Geybels MS, Becker JC, zur Hausen A, Winnepenninckx V. Genomic Landscape of Spitzoid Neoplasms Impacting Patient Management. Front Med. 2018;5. doi:10.3389/fmed.2018.00344 30. Raghavan SS, Peternel S, Mully TW, et al. Spitz melanoma is a distinct subset of spitzoid melanoma. Mod Pathol. 2020;33(6):1122-1134. doi:10.1038/s41379-019-0445-z 31. Sepehr A, Tahan SR. “Nevus/Melanocytoma/Melanoma”: That Which We Call a Rose by Any Other Name Would Smell as Sweet. Arch Pathol Lab Med. 2012;136(2):135-135. doi:10.5858/arpa.2011-0148-LE 32. Urso C. Tertium Non Datur ? Legitimacy of a Third Diagnostic Category in Melanocytic Lesions. Arch SKIN November 2020 Volume 4 Issue 6 Copyright 2020 The National Society for Cutaneous Medicine 520 Pathol Lab Med. 2012;136(10):1181-1183. doi:10.5858/arpa.2012-0071-LE 33. Zembowicz A, Scolyer RA. Nevus/Melanocytoma/Melanoma: An Emerging Paradigm for Classification of Melanocytic Neoplasms? Arch Pathol Lab Med. 2011;135:7. 34. Ferris LK, Moy RL, Gerami P, et al. Noninvasive Analysis of High-Risk Driver Mutations and Gene Expression Profiles in Primary Cutaneous Melanoma. J Invest Dermatol. 2019;139(5):1127- 1134. doi:10.1016/j.jid.2018.10.041 35. Ferris LK, Gerami P, Skelsey MK, et al. Real-world performance and utility of a noninvasive gene expression assay to evaluate melanoma risk in pigmented lesions: Melanoma Res. Published online July 2018:1. doi:10.1097/CMR.0000000000000478 36. Ferris LK, Jansen B, Ho J, et al. Utility of a Noninvasive 2-Gene Molecular Assay for Cutaneous Melanoma and Effect on the Decision to Biopsy. JAMA Dermatol. 2017;153(7):675. doi:10.1001/jamadermatol.2017.0473 37. Gerami P, Yao Z, Polsky D, et al. Development and validation of a noninvasive 2-gene molecular assay for cutaneous melanoma. J Am Acad Dermatol. 2017;76(1):114-120.e2. doi:10.1016/j.jaad.2016.07.038 38. Lee JJ, Lian CG. Molecular Testing for Cutaneous Melanoma: An Update and Review. Arch Pathol Lab Med. 2019;143(7):811-820. doi:10.5858/arpa.2018-0038-RA 39. Shah A, Hyngstrom J, Florell SR, Grossman D. Use of the Pigmented Lesion Assay to rapidly screen a patient with numerous clinically atypical pigmented lesions. JAAD Case Rep. 2019;5(12):1048-1050. doi:10.1016/j.jdcr.2019.10.004 40. Clarke LE, Flake DD, Busam K, et al. An independent validation of a gene expression signature to differentiate malignant melanoma from benign melanocytic nevi. Cancer. 2017;123(4):617- 628. doi:10.1002/cncr.30385 41. Clarke LE, Warf MB, Flake DD, et al. Clinical validation of a gene expression signature that differentiates benign nevi from malignant melanoma. J Cutan Pathol. 2015;42(4):244-252. doi:10.1111/cup.12475 42. Minca EC, Al-Rohil RN, Wang M, et al. Comparison between melanoma gene expression score and fluorescence in situ hybridization for the classification of melanocytic lesions. Mod Pathol. 2016;29(8):832-843. doi:10.1038/modpathol.2016.84 43. Clarke LE, Pimentel JD, Zalaznick H, Wang L, Busam KJ. Gene expression signature as an ancillary method in the diagnosis of desmoplastic melanoma. Hum Pathol. 2017;70:113-120. doi:10.1016/j.humpath.2017.10.005 44. Ko JS, Matharoo-Ball B, Billings SD, et al. Diagnostic Distinction of Malignant Melanoma and Benign Nevi by a Gene Expression Signature and Correlation to Clinical Outcomes. Cancer Epidemiol Biomarkers Prev. 2017;26(7):1107-1113. doi:10.1158/1055-9965.EPI-16-0958 45. Kabbarah O, Nogueira C, Feng B, et al. Integrative Genome Comparison of Primary and Metastatic Melanomas. PLOS ONE. 2010;5(5):e10770. doi:10.1371/journal.pone.0010770 46. Shain AH, Joseph NM, Yu R, et al. Genomic and Transcriptomic Analysis Reveals Incremental Disruption of Key Signaling Pathways during Melanoma Evolution. Cancer Cell. 2018;34(1):45- 55.e4. doi:10.1016/j.ccell.2018.06.005 47. Scatolini M, Grand MM, Grosso E, et al. Altered molecular pathways in melanocytic lesions. Int J Cancer. 2010;126(8):1869-1881. doi:10.1002/ijc.24899 48. Goldberg DE. Genetic Algorithms in Search, Optimization, and Machine Learning. 1st Edition. Addison-Wesley Professional; 1989. 49. Holland JH. Adaptation in Natural and Artificial Systems. The MIT Press; 1975. Accessed September 9, 2020. https://mitpress.mit.edu/books/adaptation-natural- and-artificial-systems 50. Wysong A, Newman JG, Covington KR, et al. Validation of a 40-Gene Expression Profile Test to Predict Metastatic Risk in Localized High-Risk Cutaneous Squamous Cell Carcinoma. J Am Acad Dermatol. Published online April 25, 2020. doi:10.1016/j.jaad.2020.04.088 51. Ripley BD. Pattern Recognition and Neural Networks. Cambridge University Press; 1996. Accessed September 9, 2020. http://www.stats.ox.ac.uk/~ripley/PRbook/ 52. Leshno M, Lin VYa, Pinkus A, Schocken S. Multilayer feedforward networks with a nonpolynomial activation function can approximate any function. Neural Netw. 1993;6(6):861-867. doi:10.1016/S0893-6080(05)80131-5 53. Sayed S, Nassef M, Badr A, Farag I. A Nested Genetic Algorithm for feature selection in high- dimensional cancer Microarray datasets. Expert Syst Appl. 2019;121:233-243. doi:10.1016/j.eswa.2018.12.022 54. Ableser MJ, Penuela S, Lee J, Shao Q, Laird DW. Connexin43 Reduces Melanoma Growth within a Keratinocyte Microenvironment and during Tumorigenesis in Vivo. J Biol Chem. 2014;289(3):1592-1603. doi:10.1074/jbc.M113.507228 55. Ancans J, Thody AJ. Activation of melanogenesis by vacuolar type H+-ATPase inhibitors in amelanotic, tyrosinase positive human and mouse melanoma cells. FEBS Lett. 2000;478(1-2):57-60. doi:10.1016/S0014-5793(00)01795-6 56. Brand M, Moggs JG, Oulad-Abdelghani M, et al. UV-damaged DNA-binding protein in the TFTC complex links DNA damage recognition to SKIN November 2020 Volume 4 Issue 6 Copyright 2020 The National Society for Cutaneous Medicine 521 nucleosome acetylation. EMBO J. 2001;20(12):3187-3196. doi:10.1093/emboj/20.12.3187 57. Bresnick AR, Weber DJ, Zimmer DB. S100 proteins in cancer. Nat Rev Cancer. 2015;15(2):96-109. doi:10.1038/nrc3893 58. Choi WJ, Kim M, Park J-Y, Park TJ, Kang HY. Pleiotrophin inhibits melanogenesis via Erk1/2-MITF signaling in normal human melanocytes. Pigment Cell Melanoma Res. 2015;28(1):51-60. doi:10.1111/pcmr.12309 59. Davis DG, Siddiqui MT, Oprea-Ilies G, et al. GATA- 3 and FOXA1 expression is useful to differentiate breast carcinoma from other carcinomas. Hum Pathol. 2016;47(1):26-31. doi:10.1016/j.humpath.2015.09.015 60. De Filippo E, Manga P, Schiedel AC. Identification of Novel G Protein–Coupled Receptor 143 Ligands as Pharmacologic Tools for Investigating X-Linked Ocular Albinism. Invest Ophthalmol Vis Sci. 2017;58(7):3118-3126. doi:10.1167/iovs.16-21128 61. Eckhart L, Schmidt M, Mildner M, et al. Histidase expression in human epidermal keratinocytes: Regulation by differentiation status and all-trans retinoic acid. J Dermatol Sci. 2008;50(3):209-215. doi:10.1016/j.jdermsci.2007.12.009 62. Gou R, Zhu L, Zheng M, et al. Annexin A8 can serve as potential prognostic biomarker and therapeutic target for ovarian cancer: based on the comprehensive analysis of Annexins. J Transl Med. 2019;17(1):275. doi:10.1186/s12967-019-2023-z 63. Guyonneau L, Murisier F, Rossier A, Moulin A, Beermann F. Melanocytes and Pigmentation Are Affected in Dopachrome Tautomerase Knockout Mice. Mol Cell Biol. 2004;24(8):3396-3403. doi:10.1128/MCB.24.8.3396-3403.2004 64. Ho H, Kapadia R, Al-Tahan S, Ahmad S, Ganesan AK. WIPI1 Coordinates Melanogenic Gene Transcription and Melanosome Formation via TORC1 Inhibition. J Biol Chem. 2011;286(14):12509-12523. doi:10.1074/jbc.M110.200543 65. Hu D, Ansari D, Bauden M, Zhou Q, Andersson R. The Emerging Role of Calcium-activated Chloride Channel Regulator 1 in Cancer. Anticancer Res. 2019;39(4):1661-1666. doi:10.21873/anticanres.13271 66. Jacob JT, Coulombe PA, Kwan R, Omary MB. Types I and II Keratin Intermediate Filaments. Cold Spring Harb Perspect Biol. 2018;10(4). doi:10.1101/cshperspect.a018275 67. Murali R, Wiesner T, Scolyer RA. Tumours associated with BAP1 mutations. Pathology (Phila). 2013;45(2):116-126. doi:10.1097/PAT.0b013e32835d0efb 68. Nishimoto SK, Nishimoto M. Matrix Gla Protein Binds to Fibronectin and Enhances Cell Attachment and Spreading on Fibronectin. International Journal of Cell Biology. doi:https://doi.org/10.1155/2014/807013 69. Pelletier J, Thomas G, Volarević S. Ribosome biogenesis in cancer: new players and therapeutic avenues. Nat Rev Cancer. 2018;18(1):51-63. doi:10.1038/nrc.2017.104 70. Qendro V, Lundgren DH, Rezaul K, et al. Large- Scale Proteomic Characterization of Melanoma Expressed Proteins Reveals Nestin and Vimentin as Biomarkers That Can Potentially Distinguish Melanoma Subtypes. J Proteome Res. 2014;13(11):5031-5040. doi:10.1021/pr5006789 71. Schmidt A, Bekeschus S, von Woedtke T, Hasse S. Cell migration and adhesion of a human melanoma cell line is decreased by cold plasma treatment. Clin Plasma Med. 2015;3(1):24-31. doi:10.1016/j.cpme.2015.05.003 72. Schultz J, Ibrahim SM, Vera J, Kunz M. 14-3-3σ gene silencing during melanoma progression and its role in cell cycle control and cellular senescence. Mol Cancer. 2009;8:53. doi:10.1186/1476-4598-8- 53 73. Xia T, Lau K-M, Cheng CK, Chan NC, Ng MHL. Abstract 2498: Over-expression of dual-specificity phosphatase 4 (DUSP4) in multiple myeloma. Cancer Res. 2018;78(13 Supplement):2498. doi:10.1158/1538-7445.AM2018-2498 74. Yang X-Y, Ozawa S, Kato Y, et al. C-X-C Motif Chemokine Ligand 14 is a Unique Multifunctional Regulator of Tumor Progression. Int J Mol Sci. 2019;20(8). doi:10.3390/ijms20081872 75. Yang Y, Tetreault M-P, Yermolina YA, Goldstein BG, Katz JP. Krüppel-like Factor 5 Controls Keratinocyte Migration via the Integrin-linked Kinase. J Biol Chem. 2008;283(27):18812-18820. doi:10.1074/jbc.M801384200 76. Zeeuwen PLJM, Cheng T, Schalkwijk J. The Biology of Cystatin M/E and its Cognate Target Proteases. J Invest Dermatol. 2009;129(6):1327- 1338. doi:10.1038/jid.2009.40 77. Zhu R, Li W, Xu Y, Wan J, Zhang Z. Upregulation of BTG1 enhances the radiation sensitivity of human breast cancer in vitro and in vivo. Oncol Rep. 2015;34(6):3017-3024. doi:10.3892/or.2015.4311 78. Matin RN, Chikh A, Law Pak Chong S, et al. p63 is an alternative p53 repressor in melanoma that confers chemoresistance and a poor prognosis. J Exp Med. 2013;210(3):581-603. doi:10.1084/jem.20121439 79. Luo S, Sepehr A, Tsao H. Spitz nevi and other Spitzoid lesions: Part II. Natural History and Management. J Am Acad Dermatol. 2011;65(6):1087-1092. doi:10.1016/j.jaad.2011.06.045 80. Jafry M, Peacock S, Radick A, et al. Pathologists’ Agreement on Treatment Suggestions for Melanocytic Skin Lesions. J Am Acad Dermatol. Published online 2019. doi:10.1016/j.jaad.2019.12.020 SKIN November 2020 Volume 4 Issue 6 Copyright 2020 The National Society for Cutaneous Medicine 522 81. Di Blasio S, van Wigcheren GF, Becker A, et al. The tumour microenvironment shapes dendritic cell plasticity in a human organotypic melanoma culture. Nat Commun. 2020;11(1):2749. doi:10.1038/s41467-020-16583-0 82. Pruessmann W, Rytlewski J, Wilmott J, et al. Molecular analysis of primary melanoma T cells identifies patients at risk for metastatic recurrence. Nat Cancer. 2020;1(2):197-209. doi:10.1038/s43018-019-0019-5 83. Lilyquist J, White KAM, Lee RJ, Philips GK, Hughes CR, Torres SM. Quantitative Analysis of Immunohistochemistry in Melanoma Tumors. Medicine (Baltimore). 2017;96(15). doi:10.1097/MD.0000000000006432 84. Field MG, Decatur CL, Kurtenbach S, et al. PRAME as an independent biomarker for metastasis in uveal melanoma. Clin Cancer Res. 2016;22(5):1234-1242. doi:10.1158/1078- 0432.CCR-15-2071 85. Clarke LE, Mabey B, Flake II DD, et al. Clinical validity of a gene expression signature in diagnostically uncertain neoplasms. Pers Med. Published online June 17, 2020:pme-2020-0048. doi:10.2217/pme-2020-0048 86. Bevona C, Goggins W, Quinn T, Fullerton J, Tsao H. Cutaneous Melanomas Associated With Nevi. Arch Dermatol. 2003;139(12):1620-1624. doi:10.1001/archderm.139.12.1620 87. Lezcano C, Jungbluth AA, Busam KJ. Comparison of Immunohistochemistry for PRAME With Cytogenetic Test Results in the Evaluation of Challenging Melanocytic Tumors: Am J Surg Pathol. Published online April 2020:1. doi:10.1097/PAS.0000000000001492 88. Reimann JDR, Salim S, Velazquez EF, et al. Comparison of melanoma gene expression score with histopathology, fluorescence in situ hybridization, and SNP array for the classification of melanocytic neoplasms. Mod Pathol Off J U S Can Acad Pathol Inc. 2018;31(11):1733-1743. doi:10.1038/s41379-018-0087-6 89. Gerami P, Cook RW, Russell MC, et al. Gene expression profiling for molecular staging of cutaneous melanoma in patients undergoing sentinel lymph node biopsy. J Am Acad Dermatol. 2015;72(5):780-785.e3. doi:10.1016/j.jaad.2015.01.009 90. Gerami P, Cook RW, Wilkinson J, et al. Development of a prognostic genetic signature to predict the metastatic risk associated with cutaneous melanoma. Clin Cancer Res Off J Am Assoc Cancer Res. 2015;21(1):175-183. doi:10.1158/1078-0432.CCR-13-3316 91. Vetto JT, Monzon FA, Cook RW, Johnson C, Covington KR, Leachman S. Clinical utility of a 31- gene expression profile test to determine eligibility for sentinel lymph node biopsy in melanoma patients >65 years of age. In: ; 2018. 92. Swetter SM, Thompson JA, Coit DG, et al. NCCN Clinical Practice Guidelines in Oncology. Cutaneous Melanoma. Version 3.2020. Published online May 18, 2020. 93. Cockerell C, Tschen J, Billings SD, et al. The influence of a gene-expression signature on the treatment of diagnostically challenging melanocytic lesions. Pers Med. 2017;14(2):123-130. doi:10.2217/pme-2016-0097 94. Cockerell CJ, Tschen J, Evans B, et al. The influence of a gene expression signature on the diagnosis and recommended treatment of melanocytic tumors by dermatopathologists: Medicine (Baltimore). 2016;95(40):e4887. doi:10.1097/MD.0000000000004887