American Melanoma Institute · 4 min read
The stains, chromosome tests, and gene expression tests used on difficult moles and melanomas, and what guidelines say about each.
Most moles and melanomas are diagnosed with a standard stain and a trained eye. When a lesion falls in a gray zone, the dermatopathologist can order additional tests on the same tissue block. The original sample is enough, so you won't need another biopsy.
These tests fall into 3 groups that answer different questions. Diagnostic tests ask whether a lesion is a mole or a melanoma. Prognostic tests ask how a confirmed melanoma is likely to behave. Mutation tests ask which drugs could work. Mixing up the groups causes a lot of confusion, so this article takes them one at a time.
Immunohistochemistry uses antibodies that attach to a specific protein and mark it with color. PRAME is one of the newer stains. In a study of 400 pigmented tumors, most melanomas showed PRAME in nearly all of their cell nuclei, while most moles showed none or only scattered positive cells (Lezcano and colleagues, reviewed in Surgical Pathology Clinics, 2021).
Other stains add different clues. The p16 protein acts as a brake on cell division, and its complete loss leans toward melanoma. HMB-45 normally fades as mole cells descend deeper into the skin, so strong staining at the bottom of a lesion raises concern. Ki-67 marks cells preparing to divide, and doctors often read it together with HMB-45.
Exceptions exist for every stain. Some melanomas lack PRAME, and some benign lesions show it. Pathologists treat each result as one piece of evidence alongside what they see on the routine slide.
Melanomas tend to gain or lose large pieces of chromosomes. Most moles keep a normal set. Fluorescence in situ hybridization (FISH) uses glowing probes aimed at a handful of chromosome regions that are often abnormal in melanoma, and the pathologist counts the signals in each cell.
Comparative genomic hybridization (CGH) and the related SNP array survey every chromosome at once. They need more tissue and more time than FISH. Many experts favor them for Spitz tumors and other unusual lesions, where the pattern of gains and losses can point toward a specific diagnosis.
Gene expression profiling measures how active a set of genes is. Labs offer 2 commercial diagnostic tests, one using 23 genes and one using 35, and each generates a score that falls into a likely benign, likely malignant, or intermediate range. Pathologists use them on lesions of uncertain malignant potential and interpret the score in the context of the microscope findings. An intermediate score leaves the question open.
A separate category of test is run after melanoma is already diagnosed. The most widely marketed in the United States is a 31-gene test (31-GEP) that sorts tumors into classes by estimated risk of recurrence or spread. Another, CP-GEP, combines 8 genes with the patient's age and the tumor's thickness to estimate the chance that a sentinel lymph node contains melanoma.
Guidelines have been cautious. The 2019 American Academy of Dermatology guidelines of care said these tests needed further validation in prospective studies before becoming part of routine care. For years the National Comprehensive Cancer Network (NCCN) guidelines said prognostic GEP tests were not recommended outside of a clinical trial.
In February 2026, NCCN made a limited change. According to published summaries of Version 1.2026, the guidelines now say CP-GEP may be used in selected patients with T1b melanoma, and possibly T2a, to support shared decision-making about sentinel lymph node biopsy. The update followed a prospective study in which patients with a high-risk result were about 3 times as likely to have a positive sentinel node. NCCN continues to state that other GEP tests for predicting sentinel node results are not recommended outside a clinical trial.
More data is needed on the central question. Several studies show that these tests can sort patients into higher-risk and lower-risk groups. What hasn't been shown in a randomized trial is that changing care based on the result (more scans, fewer scans, skipping a node biopsy, starting a drug) leads to better outcomes than decisions based on thickness, ulceration, and sentinel node status. Trials are under way, and guidance may shift again as results come in.
If a prognostic test is offered to you, useful questions include what decision the result would change, whether your insurer covers it, and what the doctor would recommend if the test didn't exist.
The third group of tests looks for DNA changes that drugs can target. About half of skin melanomas carry a change in the BRAF gene. For those tumors, a BRAF inhibitor paired with a MEK inhibitor (dabrafenib with trametinib, vemurafenib with cobimetinib, or encorafenib with binimetinib) can shrink advanced disease. Dabrafenib with trametinib is also used after surgery for stage III melanoma.
Doctors usually order BRAF testing for stage III and stage IV melanoma, and sometimes for high-risk stage II. It isn't part of the workup for a thin melanoma treated with surgery alone, because the result wouldn't change anything. Melanomas on the palms, soles, nail beds, and mucosal surfaces more often carry changes in a different gene, C-KIT, which has its own targeted drugs.
The lab can run mutation tests on the original biopsy block, even months or years later. Labs store these tissue blocks for years, so the material is normally available if the need arises later.
