NEW YORK (GenomeWeb) – In a study appearing online today in Cell, members of the Cancer Genome Atlas outlined a genomics-based classification scheme for cutaneous melanoma.
The researchers used genome, exome, and microRNA sequencing, as well as array-based copy number, methylation, and proteomic profiling, to assess tumor samples from more than 300 individuals with cutaneous melanoma. Their integrated analysis of these data pointed to four genomics-based subtypes, which typically converged on the same signaling, cell cycle, and apoptosis pathways.
The team's transcriptomic analysis also uncovered three expression-based subclasses for cutaneous melanoma, including a group of tumors with elevated expression of genes related to T cell activity and tumor infiltration by immune cells — features that might boost tumor vulnerability to immunotherapy-based treatment strategies.
"This comprehensive classification of melanomas allows us to create a framework that could be used to further personalize therapeutic decision-making in both the targeted and immunotherapy arena, as well as to develop more impactful prognostic and predictive models to inform patient care," co-corresponding author Jeffrey Gershenwald, a surgical oncology researcher at the University of Texas MD Anderson Cancer Center, said in a statement.
Targeted treatments developed in recent years are showing promise against metastatic melanomas harboring specific types of mutations, the team noted. Nevertheless, a clearer picture of the disease is expected to help in designing treatments against subtypes that don't respond to existing treatments and to ward of recurrence in those that do.
With this in mind the TCGA performed a comprehensive analysis of DNA, RNA, methylation, miRNA, copy number and protein patterns in up to 333 tumors from 331 individuals with cutaneous melanoma — a collection that included 266 metastatic tumor samples and 67 primary tumors.
The team did whole-exome sequencing on 58 primary tumor samples, 262 metastatic tumors samples, and matched normal tissue from 318 of these individuals, for example. Nearly 200 samples were assessed using all six molecular profiling methods.
As described in past skin cancer studies, the tumors contained mutation signatures attributed to ultraviolet radiation. The researchers also saw the highest mutation rate detected by TCGA so far, with nearly 17 single nucleotide changes or small insertions and deletions turning up per million bases of protein-coding sequence, on average.
Despite the rampant mutations, though, the team was able to classify the tumors into four main genomic subtypes based on the presence or absence of BRAF hotspot mutations, RAS hotspot mutations, and NF1 mutations.
In addition to the BRAF-mutant, RAS-mutant, and NF1-mutant subtypes, the team introduced a "triple-wild type" subtype — a group of tumors characterized by complicated structural rearrangements, focal amplifications, and frequent KIT gene mutations.
Though the researchers did not detect any clear prognostic correlations for the genomics-based subtypes, they did see survival differences in cutaneous melanoma subclasses formed with the help of tumor transcriptomic profiles.
That analysis pointed to three expression-based tumor groups: one with enhanced expression of keratin-related genes, another with lower-than-usual pigmentation and epithelial gene expression, and a third expressing immune signatures related to lymphocyte-led tumor invasion and a protein marker called LCK that's linked to T cell immune signaling.
Based on the data available, the team found that individuals with tumors from the latter group — dubbed the "immune" subclass — tended to fare better in terms of overall survival.
Together, the results point to new possibilities for treating melanoma and predicting patient outcomes, the study's authors noted. In particular, they argued that the NF1-mutant subtype may be susceptible to treatments based on MEK or ERK inhibition, while immune subclass tumors could potentially be tackled by immunotherapy-based approaches.