Medical Policy


Subject: Gene Expression Profiling of Melanomas
Document #: GENE.00023 Publish Date:    04/24/2019
Status: Reviewed Last Review Date:    03/21/2019


This document addresses gene expression profiling to assist in the risk stratification and clinical management of cutaneous and uveal (ocular) melanoma.

Position Statement

Investigational and Not Medically Necessary:

  1. Gene expression profiling of cutaneous melanoma is considered investigational and not medically necessary.
  2. Gene expression profiling of uveal melanoma is considered investigational and not medically necessary.

Gene Expression Profiling of Cutaneous Melanoma

The DecisionDx-Melanoma test (Castle Bioscience, Inc., Friendswood, TX) is a multigene expression assay designed to predict metastasis in individuals with stage I or stage II cutaneous melanoma who have no sign of disease beyond the original tumor. The laboratory test is a signature of 31 genes, 28 discriminating genes and 3 control genes, that classifies tumors as class 1 (low risk of metastasis) or class 2 (high risk of metastasis), using reverse transcription polymerase chain reaction (RT-PCR) on formalin-fixed paraffin-embedded (FFPE) primary tumor tissue specimens obtained from either biopsy or excision of a cutaneous melanoma.

There is wide variability in metastatic rates within and across Tumor-Node-Metastasis (TNM) stage groupings in individuals with cutaneous melanoma. The DecisionDx-Melanoma test is purported to predict the risk of tumor metastasis in confirmed melanoma independent of currently used metrics of risk assessment such as Breslow thickness, ulceration status (present or absent), dermal mitotic rate, microstellitosis (present or absent), American Joint Committee on Cancer (AJCC) stage, and sentinel lymph node biopsy status. It is suggested this information would add to a comprehensive baseline evaluation and determination of the initial surveillance and treatment of an individual with high-risk stage I or II disease.

The clinical validity of the DecisionDx-Melanoma test was evaluated in a prospective, multicenter study of class 1 cutaneous melanoma tumors which analyzed microarray expression data to identify a prognostic 28-gene signature to predict risk of metastasis (Gerami, 2015b). Based on modeling analysis of cohorts of primary cutaneous melanoma tumor tissue and Kaplan-Meier analysis, the study reported the 5-year disease-free survival (DFS) rates in the development set were 100% and 38% for predicted classes 1 and 2 tumors, respectively (p<0.0001). DFS rates for the validation set were 97% and 31% for predicted classes 1 and 2 tumors, respectively (p<0.0001). The investigators suggested their preliminary analysis indicates the 28-gene signature is an independent predictor of metastasis risk in the studied cohort of cutaneous melanoma tumors.

Gerami and colleagues (2015a) assessed the prognostic accuracy of gene expression profiling for molecular staging of cutaneous melanoma in a multicenter cohort study of 217 individuals undergoing sentinel lymph node biopsy (SLNB). The prognostic accuracy of each test was determined using Kaplan-Meier and Cox regression analysis of disease-free, distant metastasis-free, and overall survival. For individuals with a negative SLNB and a class 2 gene expression profile signature (that is, a high risk outcome), Kaplan-Meier 5-year disease-free, distant metastasis-free, and overall survival rates were 35%, 49%, and 59%, respectively; however, there was no statistical difference in disease-free survival, or overall survival rates for individuals with class 2 gene expression profile signature and a negative SLNB result and individuals with a class 2 gene expression profile score and a positive SLNB result. A limitation of the study is the lack of data obtained from a randomized sample of cases, which the authors conclude as resulting “in a higher rate of distant metastasis than commonly observed or reported in the SLNB-negative group.” Additional study is needed in a randomized sample of individuals to determine how gene expression profiling combined with SLNB would contribute to the accurate staging and treatment planning of individuals with cutaneous melanoma.

Berger and colleagues (2016) performed a retrospective chart review of 156 individuals with cutaneous melanoma who were consecutively tested with the DecisionDx-Melanoma gene expression profile assay at three dermatology and three surgical oncology practices between May 2013 and December 2015. The primary purpose of the study was to evaluate clinical management plans before and after gene expression profile testing, including frequency of physical examinations (initial work-up and follow-up), frequency and modality of imaging (chest x-ray, computed tomography [CT], positron emission tomography CT [PET-CT], magnetic resonance imaging [MRI], or ultrasound), SNLB procedure recommendations and results (if performed), use and frequency of routine blood work, and referral patterns to surgical and medical oncologists. The clinical characteristics of the cohort’s tumors by AJCC staging were Stage I (n=66, 42%), Stage II (n=74, 47%), Stage III (n=13, 8%), and Unknown (n=3, 2%). Overall, 95 (61%) were classified as class 1 and 61 (39%) were class 2 by gene expression profile testing. The majority of individuals were male (62%), had a median Breslow thickness of 2.0 mm, and were 63 years old (median age). The majority of tumors were located on the extremities and had superficial spreading and nodular growth patterns. Of the 156 cases, 100 (64%) of individuals received care in surgical and oncology practices and 56 (36%) were seen in dermatology practices. Individuals categorized as class 2 by the 31-gene expression profile test were managed by surgical oncology (51% vs. 18%, p<0.001 [Fisher’s exact test]). A total of 82 (53%) individuals had a documented change in management, with the majority of class 2 (n=47 of 61, 77%) undergoing surveillance changes compared to 35 (37%) class 1 (p<0.0001 [Fisher's exact test]). When stratifying results according to low-risk and high-risk AJCC stage, gene expression profile testing confirmed a low-risk, class 1 tumor for 56% of Stage I and IIA individuals, resulting in no change in management. However, 13 of 18 early stage individuals who were identified as high-risk class 2 by gene expression profile testing had more intense management, primarily in the form of more frequent imaging requested by the surgical oncologist. Overall, the majority (77 of 82, 94%) of individuals had a change in intensity of surveillance and/or referral pattern as a result of the gene expression profile classification; these changes were concordant with the risk indicated by the test result (p<0.0001 [Fisher's exact test]), with increased management intensity for class 2 and reduced management intensity for class 1. Limitations of this study include the retrospective design and lack of follow-up data collection in the study cohort, which limits drawing conclusions on the impact of gene expression profile testing to alter clinical practice management and improve outcomes for individuals with cutaneous melanoma.

Similarly, Ferris and colleagues (2017) conducted a study that included 205 participants, retrospectively enrolled, with previously reported stage I and II melanoma in addition to sufficient clinical data to obtain 5-year survival outcomes. The primary goal of the study was the comparison of the DecisionDx-Melanoma-based classification with the AJCC online prediction tool as an independent predictor of recurrence free survival (RFS), distant metastasis-free survival (DMFS, defined as a distant metastasis detected beyond the regional basin), or overall survival (OS). The secondary aim was evaluation of the utility of gene expression profiling combined with AJCC predictions for enhancing identification of melanomas at high-risk of metastasis. The final cohort was comprised of 109 stage I and 96 stage II melanoma cases. Risk classification of DMFS and OS were determined to be significant based on a cox univariate analysis (hazard ratio range 3.2-9.4; p=0.001) for both tools. Overall, 43 (21%) cases had discordant DecisionDx and AJCC classification; 11 out of 13 (85%) deaths in that group were predicted as high risk by DecisionDx but low risk by AJCC. This study suggests that the DecisionDx test, when used in combination with AJCC, may improve identification of stage I and II cutaneous melanoma with a high-risk of recurrence or metastatic disease. The retrospective nature precludes the ability to determine if use of the combined tools would alter disease management and subsequently improve clinically meaningful outcomes.

More recently, a manufacturer-sponsored, 5-year, multi-center, prospective study of the DecisionDx-Melanoma test was published. At interim analysis (median 1.5 years), a total 322 participants had completed at least one follow-up visit and were evaluable. Individuals were enrolled in one of two on-going prospective studies, EXPAND and INTEGRATE. Eligibility for study inclusion included a diagnosis of cutaneous melanoma, age ≥16 years and a successful DecisionDx-test result. Overall 282 (88%) of 322 cases had stage I/II disease and 237 (74%) had a SLNB. A total of 248 (77%) had class-1 molecular profiles. At this interim analysis, the difference in primary outcomes between class 1 vs. class 2 profiles were as follows: 97 vs. 77% RFS, 99 vs. 89% DMFS, and 99 vs. 92% OS (p<0.0001 for each). In multi-variate analysis, Breslow thickness, mitotic rate, and DecisionDx-test result significantly predicted recurrence (p<0.01), whereas tumor thickness was the only significant predictor of DMFS and OS at 1.5 years. While these results are encouraging, the questions remains as to whether or not use of DecisionDx will improve clinically meaningful outcomes beyond current risk-stratification techniques (Hsueh, 2017).

In summary, there is insufficient evidence to evaluate the clinical validity and clinical utility of the DecisionDx-Melanoma test. Additional study is required to further validate if gene expression profile testing of cutaneous melanoma will accurately identify individuals with more aggressive disease, and how test results would alter treatment plans and improve health outcomes in the surveillance and treatment of high-risk cutaneous melanoma.

The National Comprehensive Cancer Network® (NCCN) Clinical Practice Guideline (CPG) in Oncology® recommendations for the clinical staging and workup of cutaneous melanoma (V1.2019) (principles of pathology), states:

Gene expression profiling for melanoma could be an enormously valuable contribution to understanding the biology of the disease. However, the difficulty of embracing gene expression profiling as an independent predictor or outcome is illustrated by inconsistency of results across studies aimed at defining the most predictive gene sets for melanoma (including Gerami, 2015b; Nsengimana, 2015)…While there is interest in newer prognostic molecular techniques such as gene expression profiling to differentiate melanomas at low- versus high-risk for metastasis, routine (baseline) genetic testing of primary cutaneous melanomas (before or following sentinel lymph node biopsy [SLNB]) is not recommended outside of a clinical study (trial).

Gene Expression Profiling of Uveal Melanoma

Uveal melanoma, also referred to as ocular or choroidal melanoma, while relatively rare, is the most common primary ocular malignancy in adults and has a high incidence of metastases to the liver. Even with successful treatment of the primary tumor, up to 50% of individuals will subsequently develop systemic metastases, with liver involvement in up to 90% of these individuals. Metastatic liver disease remains the most common cause of tumor-related mortality in choroidal malignant melanoma, even with aggressive systemic treatments, with a median survival rate of 2 to 7 months and a 1-year survival rate of less than 10%.

In the management of uveal melanoma, clinicopathologic features and tumor genetics are used to predict prognosis, including the risk of metastatic disease. The results of two large case series have shown that tumor size has consistently been demonstrated to be of prognostic significance, in terms of the subsequent risk of metastasis and death from uveal melanoma (Diener-West, 2005; Shields, 2009). Approximately one-half of uveal melanoma tumors will metastasize at some point prior to diagnosis of the primary eye tumor, and at the time of diagnosis of the primary eye tumor, metastatic disease (micrometastases) will only be detectable in about 3% of individuals. Recent estimates of tumor doubling time have suggested that as a result of these micrometastases, clinicians may be able to identify individuals who are at higher risk for uveal melanoma through molecular signatures unique to their specific ocular tumor or those tumors with a known tendency to metastasize (Singh, 2004).

Large-scale genetic alterations, such as the presence of only one copy of chromosome 3 (monosomy 3), have been reported in uveal melanoma and associated with metastatic disease. A type of genetic test called gene expression profiling has been proposed as a tool to identify those individuals having a high risk of developing metastasis from primary uveal melanoma. Gene expression profiling for uveal melanoma suggests that tumors can be sorted into 2 classes with different characteristics and prognosis: class 1 tumors are thought to be at lower risk for metastasis, while class 2 tumors may be at high risk.

An early study by Worley and colleagues (2007) reported that the sensitivity and specificity for a molecular classifier using two microarray gene expression profiling platforms (84.6% and 92.9%, respectively) was superior when compared to monosomy 3 detected by an array comparative genomic hybridization (aCGH) (58.3% and 85.7%, respectively) and fluorescence in situ hybridization (FISH) (50.0% and 72.7%, respectively). The investigators concluded that “molecular classification based on gene expression profiling of the primary tumor was superior to monosomy 3 and clinicopathologic prognostic factors for predicting metastasis in uveal melanoma.” This study, however, is limited in application by inconsistencies in the reported data.

The DecisionDx-UM test (Castle Biosciences Inc., Friendswood, TX) is a commercially marketed gene expression profiling test intended for use in assessing metastatic risk in individuals with uveal melanoma. It consists of a 15-gene PCR-based assay that stratifies individuals with uveal melanoma into two classes based on the molecular signature of tumor tissue. The peer-reviewed literature related to the DecisionDx-UM test consists of studies describing the derivation of the gene expression profile and analytical and clinical validation of the technology (Onken, 2010a; Onken, 2010b; Onken, 2012).

Onken and colleagues (2004) presented a derivation of class 1 and class 2 molecular signatures and explored their relationship with known prognostic factors and survival. Tumor samples were taken from 25 enucleated eyes of individuals with uveal melanoma. Gene expression was examined using high-density oligonucleotide arrays. An analysis showed that gene expression profiling of uveal melanoma tumors tends to yield two classes of molecular signatures, class 1 (14 of 25, 56%) and class 2 (11 of 25, 44%) tumors. A 3-gene set (PHLDA1, FZD6, ENPP2) that predicted tumor class with no errors (p=3.5 x 10-5) was obtained from the analysis of the top 26 discriminating genes; however, none of these three genes are included in the current DecisionDx-UM gene list (Onken, 2010a). The investigators compared tumor class with clinical and pathological features (that is, cytology rank, participant age and sex, tumor diameter and thickness, presence of local invasion, ciliary body involvement, and pigmentation rank) associated with metastasis in uveal melanoma. Advanced age, a known risk factor for metastasis, correlated significantly with tumor class, as did cytology rank. Survival analysis was performed on an additional 25 participants. Kaplan-Meier analysis showed that the 92-month survival probability for class 1 participants was 95%, compared with 31% in class 2 participants (p=0.01). Of the total individuals analyzed (n=50), 1 subject in the class 1 group died, compared with 8 subjects in the class 2 group. The investigators ranked tumors from lowest to highest proportion of epithelioid cells, an indicator of tumor severity, and found that tumor class corresponded significantly with cytology rank (p<0.0001). Cytogenetic analysis of a small subset (n=10) of samples indicated that 4 of 5 (80%) class 2 tumors exhibited monosomy 3, and no class 1 tumors exhibited monosomy 3. The investigators concluded that molecular classification may better detect high-risk individuals than chromosomal analysis (monosomy 3) testing, but stated that this finding should be confirmed using a larger data set. A limitation of this study is the use of enucleated specimens in the analysis of molecular yield. Uveal melanoma tumors that are treated with enucleation are typically larger in size and exhibit extraocular tumor extension, and currently represent a subpopulation of only 10% of all uveal melanoma due to the current use of eye-sparing treatment modalities (Onken, 2006a). According to one of the developers of the DecisionDx-UM test, it is inappropriate to generalize the results of molecular testing developed from the study of larger tumors to smaller tumors without confirming that molecular testing accurately predicts metastasis in smaller primary tumors as well.

Most tests reported in the literature to date do not provide adequate scientific and statistical validation to be used outside of an ethically supervised investigational environment…Well-controlled prospective studies are necessary to identify the most accurate, widely accessible, and affordable tests for routine clinical use (Harbour, 2009).

Onken and colleagues (2010b) conducted a prospective and technical validation study (n=609) describing the derivation of the DecisionDx-UM test utilizing the PCR-based 15-gene assay comprising 12 discriminating genes and 3 endogenous control genes from previously published data sets (Onken, 2004; Onken, 2006a; Onken, 2006b). Technical performance of the assay was assessed in tumor samples, including 553 fine needle aspiration biopsy and 56 enucleation specimens from the authors’ laboratory (n=188) and 11 collaborating sites (n=421). According to the study protocol, sample failure rate due to incorrect specimen handling was low, occurring in 32 of 609 (5.3%) of samples (p<0.0001). Preliminary data suggested the potential for increased sensitivity of gene expression profiling compared with cytologic diagnosis, as the assay failed in only 1 of 51, or 2% of samples with insufficient material for cytological diagnosis; however, point estimates of overall test accuracy (for example, sensitivity, specificity, or both) were not provided. In a subset of 172 individuals with uveal melanoma, the relationship between tumor class and metastasis was studied with available clinical data and a median follow-up time of 16 months. Within this group, the assay was reported to correctly identify individuals who went on to develop metastatic disease. Kaplan-Meier analysis showed approximately 24% class 2 individuals with uveal melanoma surviving at 48 months and close to 100% survival in the class 1 group, although more specific data was not provided. This study evaluates primarily fine-needle aspiration biopsy (FNAB) specimens (553 of 609, or 90.8%) rather than enucleation specimens (Onken, 2004); however, the data reported on the relationship between tumor class and metastasis are limited and median follow-up time was reported as a relatively short duration (16 months).

The prognostic performance of the 15-gene assay was subsequently validated by Onken and colleagues (2012) in a prospective, multicenter study involving 459 cases of posterior uveal melanoma. The Collaborative Ocular Oncology Group (COOG), comprised of 12 ocular oncology centers in North America, assigned samples obtained directly from individuals (usually at the time of treatment: FNAB, n=359; post-enucleation FNAB, n=92; and local tumor resection, n=8 cases) to prognostic subgroups: class 1 (low metastatic risk) and class 2 (high metastatic risk). After treatment of the primary tumor, participants were monitored for metastatic disease with a liver function panel every 6 months and liver imaging once per year or anytime the liver function panel was abnormal or there were symptoms suspicious for metastasis. The first 260 samples were also analyzed for chromosome 3 status (monosomy 3) using a single nucleotide polymorphism assay. Net reclassification improvement analysis was performed to compare the prognostic accuracy of the 15-gene assay with the 7th edition clinical TNM classification and chromosome 3 status. The 15-gene assay successfully classified 446 of 459 cases (97.2%). The 15-gene assay was class 1 in 276 cases (61.9%) and class 2 in 170 cases (38.1%) at the median follow-up of 17.4 months (mean, 18.0 months). Metastasis was detected in 3 class 1 cases (1.1%) and 44 class 2 cases (25.9%). Although there was an association between the 15-gene assay class 2 and monosomy 3 (p<0.0001), 54 of 260 tumors (20.8%) were discordant for the 15-gene assay and chromosome 3 status, among which the 15-gene assay demonstrated superior prognostic accuracy (p=0.0001). The 15-gene assay class had a stronger independent association with metastasis than any other prognostic factor (p<0.0001). Chromosome 3 status did not contribute additional prognostic information that was independent of the 15-gene assay (p=0.2). At 3-year follow-up, the net reclassification improvement of the 15-gene assay over TNM classification was 0.43 (p=0.001) and 0.38 (p=0.004) over chromosome 3 status. A total of 34 deaths occurred, 28 (82.4%) which were due to metastatic disease. Another 19 individuals developed metastases but were still alive at the time of the last follow-up evaluation. There was a strong association observed between the 15-gene assay class 2 and other adverse prognostic factors, including increased age of the individual, ciliary body involvement, larger tumor diameter and thickness, epithelioid cell type, and monosomy 3. The investigators stated the 15-gene assay was more strongly associated with metastasis than the other adverse prognostic factors and more accurate as a prognostic marker than monosomy 3 status for clinical validation and utility in stratifying individuals for entry into clinical trials of adjuvant therapy. The study, however, did not measure or report how classifying the tumors into subgroups altered the clinical treatment plans and improved health outcomes for these study participants with uveal melanoma.

Augsburger and colleagues (2015) performed a prospective single institution longitudinal study (in conjunction with a multicenter validation study) to determine the frequency of discordant gene expression profile classification of posterior uveal melanoma. FNAB aspirates of 80 clinically diagnosed primary choroidal and ciliochoroidal melanomas were obtained from two tumor sites prior to or at the time of initial ocular tumor treatment. Two different machine learning algorithms were calibrated against 30 uveal melanomas of known prognostic category and were used to translate results for each individual specimen into a prognostic gene expression profile class: 1) class 1: low risk of future emergence of distant metastasis; or, 2) class 2: relatively high risk of short-term emergence of distant metastases. A discordant result was defined as a “gene expression profile class assignment of different sign for the tumor cells obtained from the two sites or a failed gene expression profile test...for the tumor cells obtained from one but not both tumor sites.” The results indicated that single-site FNAB for gene expression profile testing and prognostic classification was associated with a substantial probability of misclassification of a class 2 tumor as a class 1 tumor or an inconclusive class assignment because of a low confidence result plus a small risk of a failed gene expression profile test in 9 of 80 cases (11.3%; 95% confidence interval [CI], 9.0% to 13.6%). If cases with a “low confidence” gene expression profile class assignment for one or both aspirates and the 2 cases with a failed test on one aspirate were also classified as “discordant,” as many as 13 cases (16.3%) by weighted algorithm and 15 cases (18.8%) by machine algorithm could have been classified as discordant. The authors concluded that gene expression profile testing is likely to result in the correct prognostic classification of the tumor about 85% of the time in smaller tumors. These results demonstrate the limitations of single-site FNAB sampling for obtaining a representative sample of the intraocular tumor, in particular, for prognostic gene expression profile testing. Klufas and colleagues (2015) reported on a small retrospective case series of gene expression profile testing of choroidal metastatic tumors that provided variable results with tumors receiving a class 1 or class 2 test result of nonmelanoma. A total of 3 of the 4 cases of choroidal metastasis revealed a class 1 test result. The authors stated “this further emphasizes that the test has not been validated to provide prognostic information in nonmelanoma samples and underscores the importance of obtaining a biopsy also for cytopathology for diagnostic confirmation, particularly in cases where additional molecular testing may be performed.”

Correa and Augsburger (2016) performed a prospective, single-institution interventional case series of 299 individuals to evaluate the clinical features, cytopathology, and gene expression profile of posterior uveal melanoma tumor cells sampled by FNAB at the time of or shortly prior to initial treatment. The melanoma tumor cells were classified by gene expression profile testing as class 1 in 211 cases (70.6%) and class 2 in 88 cases (29.4%). With use of a univariate prognostic model with Kaplan-Meier event rate curves and univariate Cox proportional hazard modeling, the investigators reported that gene expression profiling class was the strongest prognostic factor for metastatic death in this case series; however, it was noted that in the analysis, the largest linear basal diameter of the tumors (LBD), tumor thickness, and intraocular tumor location also proved to be significant individual prognostic factors. As with earlier studies, the authors suggested this information should be considered to identify “efficient protocol for surveillance testing and enrollment in multicenter clinical trials for adjuvant therapies.”

Two additional retrospective observational studies have attempted to validate whether any clinicopathologic factors provide independent prognostic information that may enhance the accuracy of gene expression profile classifications (Walter, 2016), and what associations exist between gene expression profile classification (class 1 or class 2), clinicopathologic features, mutation status and patient outcomes in individuals with uveal melanoma (Decatur, 2016). The study by Walter and colleagues (2016) was similar in methodology to the previously discussed Onken study (2012). The primary cohort included 339 individuals and a validation cohort of 241 individuals. Both cohorts included 132 individuals that were in the Onken (2012) study. The validation cohort was used to test a prediction model using the gene expression profile plus pretreatment largest basal diameter (LBD). Cox proportional hazards analysis was used in the primary cohort to examine gene expression profile classification and other clinicopathologic variables including age, sex, tumor thickness, LBD, ciliary body involvement, and pathologic cell type. Gene expression profile class 2 was determined to be the most significant predictor of metastases and mortality in uveal melanoma. Tumor diameter was also an independent predictor of outcomes when using a 12 mm LBD cutoff value. The authors concluded that class 2 uveal melanoma had a better prognosis when the LBD was less than 12 mm at the time of treatment. In follow-up the authors are planning “a prospective, multicenter study to validate these findings and to determine the optimal use of LBD in guiding primary tumor treatment, clinical trial inclusion criteria, and systemic adjuvant therapy.” Decatur and colleagues (2016) retrospectively studied a small group (n=81) with uveal melanoma treated with enucleation by a single ocular oncologist between 1998 and 2014. Tumor samples were used for gene expression profile testing, and were classified as class 1 in 35 (43%), class 2 in 42 (52%), and unknown in 4 (5%) individuals. Tumors with BAP1 mutations were associated with poor prognostic factors, and EIF1AX and SF3B1 mutations were associated with good prognostic factors. Gene expression profile class 2 was strongly associated with BAP1 variants (r=0.70; p<0.001). On Cox proportional hazards analysis, GEP class 2 was the strongest predictor of metastases and melanoma mortality.

Plasseraud and colleagues (2016) evaluated the clinical validity and utility of the DecisionDx-UM test in individuals (n=70) enrolled in the industry-sponsored, observational CLEAR study (NCT02376920), a cohort registry of data from four participating centers across the United States that was designed to assess information on how physicians use the DecisionDx-UM results to manage treatment plans with regards to surveillance regimens and treatment referral patterns for uveal melanoma. Surveillance regimens were not prespecified but independently determined by each participating physician utilizing the DecisionDx-UM test result and documented as part of the registry data. None of the individuals in the registry had technical failures with DecisionDx-UM testing. The intensity of surveillance was categorized based on data collection methods used in a retrospective case study and cross-sectional survey of physician practice patterns in the management of uveal melanoma in Medicare beneficiaries (Aaberg, 2014). High-intensity surveillance was defined as clinical visits every 3-6 months and liver function tests and/or liver imaging/systemic evaluation (for example, CT, ultrasound, or MRI) every 3-6 months. Low-intensity surveillance was characterized by annual imaging and/or liver function tests. A total of 70 enrolled individuals with documented class 1 (low-risk) tumors (n=37 [53%]) and class 2 (high risk) tumors (33 [47%]) were included in the analysis. Of those with class 1 tumors, 30 (81%) were class 1A and 7 (19%) were class 1B. At a median follow-up of 2.4 years, 12 (36%) class 2 individuals experienced metastasis compared to 2 (5%) class 1 (p=0.002, Fisher’s exact test). The median time to metastasis for class 2 was 1.4 years and time to death was 2.7 years. At 3 years, 100% of class 1 were metastasis-free compared to 63% (95% CI, 43%-83%) of class 2 (log rank test, p=0.003). The majority of metastases were localized in the liver (8 of 12 individuals, of which 1 person had liver/lung metastasis), but metastases were also found in the bone (n=3) and lungs/brain (n=1). Of the class 2 metastatic tumors (n=12), 9 were treated with enucleation, 2 with plaque radiotherapy, and 1 with transpupillary thermotherapy. A total of 30 of 37 class 1 were treated with low-intensity follow-up while all 33 class 2 were managed with high-intensity follow-up. Two individuals with intermediate risk class 1B results received high-intensity surveillance and 4 of 37 (11%) class 1 were referred to medical oncology. Six of 33 (18%) class 2 were referred to medical oncology and 8 (24%) class 2 were referred to adjuvant clinical trials. Four of the class 2 went on to receive systemic adjuvant therapy, of which 3 received combination chemotherapy within a clinical trial, and 1 received IVIG immunotherapy. No one in class 1 was referred to a clinical trial or had systemic adjuvant therapy. The results suggest that class 2 is managed by medical oncology (with imaging and liver function tests) and offered clinical trial participation significantly more often than class 1 (Fisher’s exact test for intensity of surveillance, p<0.0001; for medical oncology/clinical trial referral, p=0.04). The authors suggest that for class 2, with higher-intensity surveillance, the results are consistent with the “goal of potentially identifying metastases earlier, thus permitting intervention, while the patient is asymptomatic and likely amenable to treatment(s). Conversely, unnecessary surveillance can potentially be avoided for patients in whom extraocular recurrence of disease is unlikely.” And that “the decision to enroll class 2 patients in clinical trials is directly related to the level of evidence for metastatic propensity that has been reported for the [DecisionDx-UM] test.” A limitation of this study includes the high risk of bias, as it was not clear which outcome measures were prespecified or how data was collected. The CLEAR registry data collection is ongoing, with an estimated final collection date of October 2020 for the primary outcome measure (that is, individuals followed for up to 10 years with measurement for metastatic event performed at 6 month intervals).

In summary, it has been suggested that use of gene expression profile testing of primary uveal melanoma to identify individuals at high risk of metastatic disease (class 2) could select those who would benefit from adjuvant treatment to reduce the risk of metastasis and more frequent screening for the earliest development of metastatic disease. Although gene expression profiling of uveal melanoma has been shown to be an independent predictor of risk of metastasis, in the absence of effective adjuvant or systemic therapy, it is uncertain how risk stratification based upon this type of testing would improve health outcomes. There is a lack of published data from well-designed, prospective studies of sufficient sample size and follow-up that support the clinical utility of gene expression profile testing of uveal melanoma.

The current NCCN CPG for uveal melanoma (V1.2018) states:

Biopsy of the primary tumor does not impact outcome, but may provide prognostic information that can help inform frequency of follow-up and may be needed for eligibility for clinical trials. Specimen should be sent for histology, chromosome analysis, and/or gene expression profiling. The risk/benefits of biopsy for prognostic analysis should be carefully considered and discussed.

A 2015 guideline of the United Kingdom Uveal Melanoma Guideline Development Working Group (Nathan, 2015) makes two recommendations for molecular testing for uveal melanoma, stating:

The local treatment of uveal melanoma is well-established. Preservation of the eye, when attempted, is considered conservative treatment. Other conservative treatments include brachytherapy and proton beam radiotherapy. As reported in the randomized trial data from the Collaborative Ocular Melanoma Study (COMS) (Hawkins, 2011), there is no statistical difference in risk of metastasis between enucleation and plaque radiotherapy, or of brachytherapy prior to enucleation for large tumors; both strategies offer the same prognosis in terms of survival rates and risk of metastasis. Despite the established treatment protocols for primary uveal melanoma, no decrease has been observed in the mortality rate of this tumor. The 5-year survival rate has not changed over the last 3 decades (81.6%), suggesting that life expectancy is independent of successful local eye treatment (Pereira, 2013).


Cutaneous Melanoma

Cutaneous melanoma occurs in all parts of the skin, including the soles of feet, on the palms of the hand, in between toes and fingers, and underneath the finger and toe nails. The four main categories of cutaneous melanoma described by the Melanoma Research Foundation (2016) include superficial spreading melanoma (SSM), nodular melanoma, acral lentiginous melanoma (ALM) (also called subungual melanoma), and, lentigo maligna melanoma (LMM).

According to the National Cancer Institute (NCI, 2018), melanoma accounts for less than 5% of skin cancer cases but is the major cause of skin cancer deaths with the incidence rising over the past 4 decades. In 2018, it is estimated that 91,270 new cases of melanoma will be diagnosed and 9320 individuals may die of the condition in the United States. Elderly men are at highest risk; however, melanoma is the most common cancer in young adults aged 25 to 29 years.

Uveal Melanoma

According to the National Cancer Institute (NCI, 2018), melanoma of the uveal tract (iris, ciliary body, and choroid) is the most common primary intraocular (eye) malignancy in adults. The mean age-adjusted incidence of uveal melanoma in the United States is approximately 4.3 new cases per million population. The age-adjusted incidence of this cancer has remained stable since the early 1970s. Uveal melanoma is diagnosed mostly at older ages, with a progressively rising age-specific incidence rate that peaks near the age of 70. Of the three types, iris melanomas have the best prognosis, whereas those of the ciliary body have the poorest. Most uveal tract melanomas originate in the choroid. The ciliary body is less commonly a site of origin, and the iris is the least common. The comparatively low incidence of iris melanomas has been attributed to the characteristic features of these tumors, that is, they tend to be small, slow growing, and relatively dormant in comparison with their posterior counterparts. Iris melanomas rarely metastasize. Melanomas of the posterior uveal tract are cytologically more malignant, detected later, and metastasize more frequently than iris melanomas. Extrascleral extension usually confers a poor prognosis. In addition, regular screening tests for the development of liver metastases, including measurement of liver function tests, liver ultrasound, computed tomography scan, or magnetic resonance imaging, have not shown evidence of any effect on health outcomes (Augsburger, 2009).

Gene expression profiling assays are being investigated as a tool to assist in the risk stratification and clinical management of individuals with uveal (ocular) melanoma. The DecisionDx-UM test is a proprietary, multi-gene expression profiling assay intended for use in assessing metastatic risk in individuals with uveal melanoma. The DecisionDx-UM assay requires a single biopsy specimen. For individuals with the confirmed diagnosis of uveal melanoma, the tumor specimen can be obtained with a fine needle aspiration biopsy at the time of enucleation or at a later date from the FFPE slides that are made from the enucleated globe.

According to Castle Biosciences Inc., the DecisionDx-UM test results are used for the following:

Recently, the cancer-testis antigen PRAME (preferentially expressed antigen in melanoma) has been identified as a potential biomarker for increased metastatic risk in Class 1 uveal melanoma tumors. Some investigators believe this may have important implications for precision management in uveal melanoma and may aid in stratification of risk for clinical trials. It is theorized that cancers expressing PRAME may be more susceptible to immunotherapy and therefore identifying PRAME could enhance prognostic accuracy by identifying Class 1 tumors with intermediate metastatic risk (Fields, 2016a). The DecisionDx®-PRAME test (Castle Bioscience, Inc., Friendswood, TX) has been developed for this purpose. Castle Biosciences is offering PRAME testing as an optional add-on test to the DecisionDx-UM test; the newly developed DecisionDx-PRAME testing is only relevant in the context of the DecisionDx-UM results.

Regulatory Approval

Laboratories performing gene expression profiling tests are regulated under the Clinical Laboratory Improvement Amendments (CLIA) Act of 1988. Clinical laboratories may develop and validate tests in-house and market them as a laboratory service; laboratory-developed tests (LDTs) must meet the general regulatory standards of the CLIA Act. Laboratories that offer LDTs must be licensed by CLIA for high-complexity testing. To date, the U.S. Food and Drug Administration (FDA) does not require regulatory review of these tests.


Breslow thickness: Maximal thickness of a primary cutaneous melanoma measured in tissue sections from the top of the epidermal granular layer, or from the ulcer base (if the tumor is ulcerated), to the bottom of the tumor; metastatic rates correlate closely with tumor thickness.

Cytology: The study of the formation and function of cells.

Deoxyribonucleic acid (DNA): The hereditary material in humans and almost all other organisms. Genetic material contained in nearly every cell in a person’s body has the same DNA.

Gene expression profile/profiling (GEP): The individual pattern of expression of a panel of genes that is regarded as a “signature” for that tissue; a major determinant of the biology of both normal and malignant cells.

Histology: The study of the microscopic structure of tissue and cells.

Tumor node metastasis (TNM) system: One of the most widely used cancer staging systems accepted by the American Joint Committee on Cancer (AJCC). The TNM system is based on the size and/or extent (reach) of the primary tumor (T), the amount of spread to nearby lymph nodes (N), and the presence of metastasis (M) or secondary tumors formed by the spread of cancer cells to other parts of the body. A number is added to each letter to indicate the size and/or extent of the primary tumor and the degree of cancer spread.


The following codes for treatments and procedures applicable to this document are included below for informational purposes. Inclusion or exclusion of a procedure, diagnosis or device code(s) does not constitute or imply member coverage or provider reimbursement policy. Please refer to the member's contract benefits in effect at the time of service to determine coverage or non-coverage of these services as it applies to an individual member.

When services are Investigational and Not Medically Necessary:
When the code describes a procedure indicated in the Position Statement section as investigational and not medically necessary.




Molecular pathology procedure, Level 2 (eg, 2-10 SNPs, 1 methylated variant, or 1 somatic variant [typically using nonsequencing target variant analysis], or detection of a dynamic mutation disorder/triplet repeat) [when specified as the following, e.g., DecisionDx-PRAME]:

  • LINC00518 (long intergenic non-protein coding RNA 518) (eg, melanoma), expression analysis
  • PRAME (preferentially expressed antigen in melanoma) (eg, melanoma), expression analysis


Unlisted multianalyte assay with algorithmic analysis [when specified as uveal or cutaneous melanoma gene expression tests, such as DecisionDx-Melanoma]


Unlisted chemistry procedure [when specified as uveal or cutaneous melanoma gene expression tests, such as DecisionDx-Melanoma]


Oncology (uveal melanoma), mRNA, gene expression profiling by real-time RT-PCR of 15 genes (12 content and 3 housekeeping genes), utilizing fine needle aspirate or formalin-fixed paraffin-embedded tissue, algorithm reported as risk of metastasis
DecisionDx®-UM, Castle Biosciences, Inc



ICD-10 Diagnosis



Malignant melanoma of skin


Malignant neoplasm of choroid


Malignant neoplasm of ciliary body


Personal history of malignant melanoma of skin


Peer Reviewed Publications:

  1. Aaberg TM Jr, Cook RW, Oelschlager K, et al. Current clinical practice: differential management of uveal melanoma in the era of molecular tumor analyses. Clin Ophthalmol. 2014; 8:2449-2460.
  2. Augsburger JJ, Corrêa ZM, Augsburger BD. Frequency and implications of discordant gene expression profile class in posterior uveal melanomas sampled by fine needle aspiration biopsy. Am J Ophthalmol. 2015; 159(2):248-256.
  3. Augsburger JJ, Correa ZM, Shaikh AH. Effectiveness of treatments for metastatic uveal melanoma. Am J Ophthalmol. 2009; 148(1):119-127.
  4. Berger AC, Davidson RS, Poitras JK, et al. Clinical impact of a 31-gene expression profile test for cutaneous melanoma in 156 prospectively and consecutively tested patients. Curr Med Res Opin. 2016; 32(9):1599-1604.
  5. Cirenajwis H, Ekedahl H, Lauss M et al. Molecular stratification of metastatic melanoma using gene expression profiling: Prediction of survival outcome and benefit from molecular targeted therapy. Oncotarget. 2015; 6(14):12297-12309.
  6. Correa ZM, Augsburger JJ. Independent prognostic significance of gene expression profile class and largest basal diameter of posterior uveal melanomas. Am J Ophthalmol. 2016; 162:20-27.
  7. Decatur CL, Ong E, Garg N, et al. Driver mutations in uveal melanoma: associations with gene expression profile and patient outcomes. JAMA Ophthalmol. 2016; 134(7):728-733.
  8. Diener-West M, Reynolds SM, Agugliaro DJ, et al. Development of metastatic disease after enrollment in the COMS trials for treatment of choroidal melanoma: Collaborative Ocular Melanoma Study Group Report No. 26. Arch Ophthalmol. 2005; 123(12):1639-1643.
  9. Ferris LK, Farberg AS, Middlebrook B, et al. Identification of high-risk cutaneous melanoma tumors is improved when combining the online American Joint Committee on Cancer Individualized Melanoma Patient Outcome Prediction Tool with a 31-gene expression profile-based classification. J Am Acad Dermatol. 2017; 76(5):818-825.
  10. Field MG, Decatur CL, Kurtenbach S, et al. PRAME as an Independent Biomarker for Metastasis in Uveal Melanoma. Clin Cancer Res. 2016a; 22(5):1234-1242.
  11. Field MG, Durante MA, Decatur CL, et al. Epigenetic reprogramming and aberrant expression of PRAME are associated with increased metastatic risk in Class 1 and Class 2 uveal melanomas. Oncotarget. 2016b; 7(37):59209-59219.
  12. 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. 2015a; 72(5):780-785.
  13. 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. 2015b; 21(1):175-183.
  14. Harbour JW. Molecular prognostic testing and individualized patient care in uveal melanoma. Am J Ophthalmol. 2009; 148(6):823-829.
  15. Hawkins BS. Collaborative ocular melanoma study randomized trial of I-125 brachytherapy. Clin Trials. 2011; 8(5):661-673.
  16. Hsueh EC, DeBloom JR, Lee J, et al. Interim analysis of survival in a prospective, multi-center registry cohort of cutaneous melanoma tested with a prognostic 31-gene expression profile test. J Hematol Oncol. 2017. Available at: Accessed on January 15, 2019. 
  17. Klufas MA, Itty S, McCannel CA, et al. Variable results for uveal melanoma-specific gene expression profile prognostic test in choroidal metastasis. JAMA Ophthalmol. 2015; 133(9):1073-1076.
  18. McMasters KM, Noyes RD, Reintgen DS, et al. Lessons learned from the Sunbelt Melanoma Trial. J Surg Oncol. 2004; 86(4):212-223.
  19. Mruthyunjaya P, Seider MI, Stinnett S, et al. Association between tumor regression rate and gene expression profile after iodine 125 plaque radiotherapy for uveal melanoma. Ophthalmology. 2017; 124(10):1532-1539.
  20. Nsengimana J, Laye J, Filia A, et al. Independent replication of a melanoma subtype gene signature and evaluation of its prognostic value and biological correlates in a population cohort. Oncotarget. 2015; 6(13):11683-11693.
  21. Onken MD, Ehlers JP, Worley LA, et al. Functional gene expression analysis uncovers phenotypic switch in aggressive uveal melanomas. Cancer Res. 2006a; 66(9):4602-4609.
  22. Onken MD, Worley LA, Char DH, et al. Collaborative Ocular Oncology Group report number 1: prospective validation of a multi-gene prognostic assay in uveal melanoma. Ophthalmology. 2012; 119(8):1596-1603.
  23. Onken MD, Worley LA, Davila RM, et al. Prognostic testing in uveal melanoma by transcriptomic profiling of fine needle biopsy specimens. J Mol Diagn. 2006b; 8(5):567-573.
  24. Onken MD, Worley LA, Ehlers JP, Harbour JW. Gene expression profiling in uveal melanoma reveals two molecular classes and predicts metastatic death. Cancer Res. 2004; 64(20):7205-7209.
  25. Onken MD, Worley LA, Harbour JW. Association between gene expression profile, proliferation and metastasis in uveal melanoma. Curr Eye Res. 2010a; 35(9):857-863.
  26. Onken MD, Worley LA, Tuscan MD, Harbour JW. An accurate, clinically feasible multi-gene expression assay for predicting metastasis in uveal melanoma. J Mol Diagn. 2010b; 12(4):461-468.
  27. Pereira PR, Odashiro AN, Lim LA, et al. Current and emerging treatment options for uveal melanoma. Clin Ophthalmol. 2013; 7:1669-1682.
  28. Plasseraud KM, Cook RW, Tsai T, et al. Clinical performance and management outcomes with the DecisionDx-UM gene expression profile test in a prospective multicenter study. J Oncol. 2016; 2016:5325762.
  29. Shields CL, Furuta M, Thangappan A, et al. Metastasis of uveal melanoma millimeter-by-millimeter in 8033 consecutive eyes. Arch Ophthalmol. 2009; 127(8):989-998.
  30. Singh AD, Ronnie IG, Kevel T, et al. The Zimmerman-McLean-Foster hypothesis: 25 years later. Br J Ophthalmol. 2004; 88(7):962-967.
  31. Van Gils W, Ladder EM, Mensink HW, et al. Gene expression profiling in uveal melanoma: two regions on 3p related to prognosis. Invest Ophthalmol Vis Sci. 2008; 49(10):4254-4262.
  32. Walter SD, Chao DL, Feuer W, et al. Prognostic implications of tumor diameter in association with gene expression profile for uveal melanoma. JAMA Ophthalmol. 2016; 134(7):734-740.
  33. Worley LA, Onken MD, Person E, et al. Transcriptomic versus chromosomal prognostic markers and clinical outcome in uveal melanoma. Clin Cancer Res. 2007; 13(5):1466-1471.
  34. Young TA, Burgess BL, Rao NP, et al. High-density genome array is superior to fluorescence in-situ hybridization analysis of monosomy 3 in choroidal melanoma fine needle aspiration biopsy. Mol Vis. 2007; 13:2328-2333.

Government Agency, Medical Society, and Other Authoritative Publications:

  1. Nathan P, Cohen V, Coupland S, et al.; United Kingdom Uveal Melanoma Guideline Development Working Group. Uveal Melanoma UK National Guidelines. Eur J Cancer. 2015; 51(16):2404-2412.
  2. NCCN Clinical Practice Guidelines in Oncology®. © 2019 National Comprehensive Cancer Network, Inc. For additional information visit the NCCN website: Accessed on January 15, 2019.
    • Cutaneous Melanoma: V1.2019. Revised November 01, 2018.
    • Uveal Melanoma:V1.2018. Revised March 15, 2018.
  3. Sladden MJ, Balch C, Barzilai DA, et al. Surgical excision margins for primary cutaneous melanoma. Cochrane Database Syst Rev. 2009;(4):CD004835.
  4. U.S. National Institutes of Health (NIH). An ongoing, 5-year post market study to track clinical application of DecisionDx-Melanoma gene expression profile (GEP) assay results and the impact on patient outcomes and health economics. NLM Identifier: NCT02355574. Last updated February 27, 2017. Available at: Accessed on January 15, 2019.
  5. U.S. National Institutes of Health (NIH). CLEAR Registry Study: An open, 5-year registry study to track clinical application of DecisionDx-UM multigene assay results and associated patient outcomes. NLM Identifier: NCT02376920. Last updated January 25, 2018. Available at: Accessed on January 15, 2019.
  6. Wong SL, Balch CM, Hurley P, et al. Sentinel lymph node biopsy for melanoma: American Society of Clinical Oncology and Society of Surgical Oncology joint clinical practice guideline. J Clin Oncol. 2012; 30(23):2912-2918.
Websites for Additional Information
  1. National Cancer Institute (NCI). Cancer topics. Available at: Accessed on January 15, 2019.
    • Intraocular (Eye) Melanoma Treatment (PDQ®). Last modified September 20, 2018.
    • Melanoma Treatment (PDQ). Last modified September 28, 2018.


The use of specific product names is illustrative only.  It is not intended to be a recommendation of one product over another, and is not intended to represent a complete listing of all products available. 

Document History






Medical Policy & Technology Assessment Committee (MPTAC) review.



Hematology/Oncology Subcommittee review. Updated References section.



Updated Coding section with 01/01/2019 CPT changes; added 0081U.



Updated Coding section to clarify wording, removed reference to diagnostic test.



MPTAC review.



Hematology/Oncology Subcommittee review. Updated Rationale, Background, References and Websites for Additional Information.



The document header wording updated from “Current Effective Date” to “Publish Date.” Updated Coding section with 01/01/2018 CPT changes; added Tier 2 code 81401, genes LINC00518, PRAME



MPTAC review.



Hematology/Oncology Subcommittee review. Updated formatting in Position Statement section. Updated Rationale, Background, References, Websites for Additional Information, and Index sections.



MPTAC review.



Hematology/Oncology Subcommittee review. Updated Rationale, Background, References, and Websites for Additional Information sections. Removed ICD-9 codes from Coding section.



MPTAC review.



Hematology/Oncology Subcommittee review. Revised scope of document including Subject, Description, Position Statement, Rationale, Background, Coding, References, Websites for Additional Information and Index sections, adding an investigational and not medically statement for gene expression profiling of cutaneous melanoma.



MPTAC review.



Hematology/Oncology Subcommittee review. Updated Rationale, References, and Websites for Additional Information sections.



MPTAC review.



Hematology/Oncology Subcommittee review. Format changes to Coding. Updated Rationale, Background, References, Websites for Additional Information, and Index sections.



MPTAC review.



Hematology/Oncology Subcommittee review. Updated Rationale, Background, References, Websites for Additional Information and Index.



Updated Coding section with 01/01/2013 CPT changes.



MPTAC review.



Hematology/Oncology Subcommittee review. Updated Websites for Additional Information.



MPTAC review. Initial document development.