Diagnostic utility of the 23-gene expression profile test for an atypical intradermal melanocytic proliferation

Histopathologic evaluation of suspicious pigmented lesions is the gold standard for accurate diagnostic determination. However, some melanocytic neoplasms cannot be definitively diagnosed by traditional work-up; these lesions include atypical intradermal melanocytic proliferations, as well as others, which may pose ambiguity with regard to their malignant potential [1–4]. Various immunohistochemistry (IHC) stains and other ancillary diagnostic tests are available to alleviate uncertainty regarding malignant potential. IHC stains for proteins such as SOX10/HMB-45/Melan-A (to aid in identification and localization of melanocytes) and Ki-67 (for estimating the proliferation rate) are often informative for diagnostic distinction [5]. PRAME IHC interpretation can also aid in evaluating the malignant potential of melanocytic lesions on a grading scale of 0 (no staining) to 4+ (positive staining in ≥75% of melanocytes), but even a 4+ score has shown a broad range of sensitivity between various studies from 60–93.5% in ambiguous lesions [6–9]. Patchy, weak, and/or focal staining often has limited utility for resolving the ambiguity of the lesion in question [10].

The 23-gene expression profile (23-GEP) test is one of several ancillary diagnostic testing modalities recommended by multiple guidelines and appropriate use criteria committees that may help facilitate a more definitive diagnosis in cases that are diagnostically uncertain by histopathology [4,11,12]. The 23-GEP (MyPath^® Melanoma, Castle Biosciences, Inc., Friendswood, TX, USA) uses a gene expression signature based on 14 discriminant genes in three categories (PRAME, immune and S100A-related), along with nine control genes for normalization [13,14]. Clinicians submit biopsy specimens for testing as formalin-fixed, paraffin-embedded (FFPE) tissue, either as nine unstained slides and one H&E-stained slide (5 μm sections) or as an FFPE block. A dermatopathologist reviews every order submission and marks the appropriate area on the slide confirmed to contain a melanocytic lesion with at least 10% tumor content. The tissue is manually macrodissected and pooled for RNA extraction, quantification and reverse transcription. After a final pre-amplification step, the sample is loaded onto a custom OpenArray microfluidics chip for real-time, quantitative PCR on a QuantStudio 12K Flex instrument (ThermoFisher Scientific) for measurement of discriminant gene expression, normalization to control genes, and assignment of the final numerical score from -16.7 to +11.1 which corresponds to one of three final test results of: suggestive of a benign neoplasm (-16.7 to -2.1); intermediate, cannot rule out malignancy (-2.0 to -0.1); or suggestive of a malignant neoplasm (0 to +11.1) [13–15].

The 23-GEP test has demonstrated 90.4–94.9% sensitivity and 88.7–96.2% specificity in a variety of studies, including some cohorts with known long-term outcomes or lesions lacking full diagnostic concordance from histopathological review [13,15–18]. The evidentiary support for the 23-GEP has strengthened since it was first made clinically available in 2015, with several clinical advantages over other diagnostic ancillary tests such as a quick turnaround time, modest tissue utilization, analytical validation (laboratory repeatability) [14], and an objective, easily interpretable/clinically actionable test result output [13,17]. Like all ancillary testing, GEP test results are designed to be interpreted in conjunction with all available patient information such as histopathological assessment, clinical features and other available laboratory testing before rendering a final diagnosis.

The timely and accurate diagnosis of malignant melanoma is especially important for those with a personal or family history melanoma, as they are at a higher overall risk for developing malignant melanoma. Establishing a definitive diagnosis in these patients leads first to the appropriate treatment of their melanocytic neoplasm, and subsequently to the appropriate level of follow-up surveillance to promote identification of additional lesions of concern at an earlier timepoint, thereby reducing the likelihood of delayed identification of a malignant melanoma with concomitant risk of recurrence and metastasis [19,20].

Case report

A 34-year-old male patient was seen by his dermatologist for a large, heterogenous, cutaneous melanocytic neoplasm on the right lateral malar cheek (Figure 1). The patient reported a family history of melanoma (paternal aunt), which, along with the concerning size and coloring of the lesion, increased the suspicion for melanoma. Two punch biopsies were performed (Figure 1, superior [A] and inferior [B]) and each portion of the lesion was assessed.

Upon histopathological review of hematoxylin and eosin (H&E)-stained sections, an atypical intradermal melanocytic proliferation was observed, with some unusual features more prominent in the upper half of the lesion (Figure 2A&D). Large pleomorphic melanocytes with large inclusions/pseudoinclusions, dusky, abundant cytoplasm and some spitzoid features were noted in both the superior (Figure 2B&C) and inferior biopsy specimens (Figure 2E&F). The abnormal melanocytes extended into the reticular dermis, and a single mitotic figure was identified in the superficial papillary dermis of the superior biopsy specimen. At the base of the lesion, there appeared to be some maturation of melanocytes with less atypia. Some tumor-infiltrating lymphocytes were appreciated, but no ulceration or features or regression were present.

IHC staining of HMB45, PRAME, SOX10 and Ki-67 was concomitantly performed. HMB45 demonstrated patchy staining throughout the section but was mostly positive in the larger, more pleomorphic melanocytes (Figure 3A). PRAME staining showed patchy, weak and diffuse positivity in 60–75% of cells that was stronger in the more superficial melanocytes with abundant pleomorphism in comparison to weaker staining in the deeper melanocytes (Figure 3C). SOX10 IHC highlighted the melanocyte populations (not shown). Lastly and most worrisome, Ki-67 staining showed a mitotic rate of 5–7% in the more superficial melanocytes, while the deeper melanocytes were mostly negative (Figure 3B&D).

Based on these inconclusive IHC results, the preferred diagnosis was a mostly intradermal melanocytic proliferation with unusual features. A preliminary pathology report was issued, and the lesion was sent for 23-GEP testing to alleviate the histopathological ambiguity due to high concern for a differential diagnosis of melanoma arising within an intradermal nevus, a concern shared by multiple other dermatopathologists who were consulted on the case. Due to the size and the histopathologically heterogenous presentation, 23-GEP testing was performed simultaneously on both biopsy specimens, which both returned results of ‘gene expression profile suggestive of a malignant neoplasm’ (+2.1, superior biopsy; +6.4, inferior biopsy), which was communicated as an amendment to the initial report.

Considering the 23-GEP results together with the extensive (but inconclusive) IHC testing, the final diagnosis rendered was malignant melanoma, pT3a (AJCC, 8th Edition [21]) with 2.5 mm Breslow depth, Clark's level IV. The patient was treated with wide-local excision and sent to a cancer center for continued medical oncology follow-up. Notably, an MRI was performed which showed no evidence of brain metastasis. As of 9 months after the initial biopsies, the patient is alive and well and has had no evidence of recurrence or metastatic disease, and no additional worrisome lesions have been detected.

Discussion

Accurate diagnosis of suspicious pigmented lesions is paramount for optimizing patient management decisions aligning with malignant potential. Difficult-to-diagnose cases often require additional time, effort and expense in the form of extra IHC stains, second opinion, or other ancillary testing. PRAME IHC staining in this case was inconclusively positive, which is difficult to interpret histopathologically and to incorporate into risk-aligned management decisions. Indeed, a recent report designed and analyzed using the stringent requirements of the Standards for Reporting Diagnostic accuracy studies (STARD) recommended that any PRAME scores of 2+ or 3+ (26–75% positively stained melanocytes) be considered noncontributory to diagnostic evaluation [10]. The 23-GEP test provided malignant results for both biopsies, a finding that reinforces the value of a test method and algorithm that considers the expression levels of multiple, objective biomarkers simultaneously alongside PRAME RNA expression to calculate the score [13].

Diagnostic GEP testing has been evaluated by the American Society of Dermatopathology Appropriate Use Criteria, and GEP testing (qRT-PCR) is considered ‘majority usually appropriate’ for lesions similar to the one described in this case as ‘adult patient with pathology suggestive or suspicious for melanoma: severely atypical compound melanocytic proliferation versus melanoma on cosmetically sensitive areas and special sites, including digits, acral, genital, ears and scalp’ [4]. GEP offers advantages such as an objective test result, a modest 10% tumor volume requirement, limited use of tissue (requires nine unstained slides and an H&E slide), and fast turnaround. The clinical utility of the 23-GEP has been demonstrated by studies showing a 42.7% decrease in indeterminate diagnoses when the test result was utilized during the diagnostic process [22], and in a few case series and case study reports [23–26]. An 80.5% decrease in excision recommendations was also noted in the context of a benign 23-GEP result [22]. Benign results with the 23-GEP have also been demonstrated to reduce the actual rates of re-excision by 72.7% in patients with diagnostically ambiguous lesions [27].

Conclusion

Overall, 23-GEP testing in this case was advantageous for facilitating the final diagnosis of a pT3a melanoma in the context of the inconclusive features of malignancy identified during extensive histopathological review. This in turn led to appropriate medical oncology follow-up. Additionally, due to the size and depth of the lesion, the added confidence in a malignant diagnosis provided by 23-GEP testing can further verify the need for a wide-local excision in sensitive or surgically complicated sites such as the face.