New Test to Aid Melanoma Diagnosis: Ready for Clinical Use?

Alexander M. Castellino, PhD

March 10, 2015

Differentiating between malignant melanoma and benign skin moles can be difficult in about 15% of cases, where histopathologic analysis is not straightforward because of ambiguous findings.

A new 23-gene signature test (myPath Melanoma, Myriad Genetics) could be helpful in these instances, researchers working with the company suggest.

The test was developed with a training set and then validated in an independent cohort. The results were published online March 2 in the Journal of Cutaneous Pathology.

"myPath Melanoma is a powerful new molecular diagnostic test that analyzes genetic information inside skin cells to help us understand the biology of a patient's skin lesion and objectively differentiate benign moles from potentially lethal melanomas," researcher Loren Clarke, MD, medical director for dermatology at Myriad, said in a company press release.

"It is a tool for a pathologist to be used as an adjunct to microscopy," he told Medscape Medical News.

However, it would be premature to base a diagnosis and clinical management on this test on the basis of the results reported so far, said Klaus J. Busam, MD, professor of pathology and laboratory medicine at the Weill Medical College of Cornell University and director of dermatopathology in the Department of Pathology at the Memorial Sloan Kettering Cancer Center in New York City.

"The test strongly correlates with the current gold standard of diagnosis — histopathologic examination," he told Medscape Medical News after reviewing the study.

Developing the Test

To develop the gene signature for the test, Dr Clarke's team identified 79 candidate biomarker genes from the literature that had expression patterns expected to differentiate benign nevi from malignant melanoma.

Archival formalin-fixed, paraffin-embedded tissue samples of melanocytic lesions were used to determine gene-expression levels.

From a dataset of 83 melanocytic lesions, the 79 candidate genes were narrowed to the 40 considered to be most promising as biomarkers.

From the training set of 595 melanoma and nevus samples, 51 samples were excluded because of re-excision, lack of sufficient tissue, or because benign or malignant classification was not possible; 42 were excluded because they did provide an expression of housekeeping genes; and 38 were excluded because the expression of one or more genes could not be detected. The remaining 464 samples were used in the initial development of the signature.

Of the 40 candidate genes, 27 differentiated between nevi and melanoma in all 464 samples. The researchers then used a statistical approach to refine and define the 23-gene signature.

To validate this 23-gene signature, they used a "predetermined score calculation" from the training set on another 571 samples obtained from four different institutions.

Of the 571 samples, 20 samples were excluded because they were re-excision specimens, 24 were excluded because they did not have sufficient tissue for processing, and 90 were excluded because they had gene-expression levels below the level of detection and did not generate a score.

Of the remaining samples, 211 were from melanoma subtypes, 67 were from dysplastic nevi, and 149 were from conventional nevi.

When these validation samples were tested with the 23-gene signature, scores ranged from –16.7 to 11.1, with scores below 0 indicating benign lesions and scores above 0 indicating malignant lesions.

The test correctly identified samples that were melanoma and that were benign skin moles with a sensitivity and specificity of 90%.

Does the Test Fulfill an Unmet Need?

This test holds promise, but "all we know so far is that the test gets the diagnosis right in up to 90% of cases of obvious nevi and melanomas," Dr Busam told Medscape Medical News.

"For cases classified by consensus as being unequivocally benign or malignant, no additional testing is needed. Routine histopathologic examination is, in most cases, accurate, faster, and cheaper than gene-expression profiling," he said.

However, the test could have clinical value as an ancillary tool for problematic lesions when a pathologist cannot determine whether a lesion is benign or malignant after routine histopathologic review, Dr Busam explained.

In the validation sample, 17 samples were histopathologically ambiguous. Nine of the samples adjudicated to be malignant were confirmed as such with the 23-gene signature. Of the eight samples adjudicated to be benign, four were classified as benign with the 23-gene signature and four were classified as malignant.

However, the test has not yet been validated for problem cases. To judge the clinical value of the test, outcomes-based studies are needed that compare the test with routine histopathologic review and other ancillary tests, such as cytogenetic analysis, Dr Busam told Medscape Medical News.

The researchers agree that "additional outcomes-based and prospective studies are needed to further assess the performance of the test."

For now, the test is being offered to pathologists in the United States, through an early-access launch, to diagnose their difficult lesions at no charge, Dr Clarke told Medscape Medical News.

Myriad continues to gather data in real-world cases, he added. The test requires the submission of five microscopic slides, which are processed within 5 days.

Several authors are employees of Myriad Genetic Laboratories. Dr Busam has disclosed no relevant financial relationships.

J Cutan Pathol. Published online March 2, 2015. Abstract


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.
Post as: