Background: Malignant melanoma is a genetically diverse disease. The most frequent mutation is BRAF, followed by NRAS and cKIT mutations. While BRAF, NRAS and cKIT mutations represent the largest fraction of patients, there is also an important group of triple wild type patients (3xWT) in dire need for potential targeted therapies. We compared mutational and PD-L1 profiles of 3xWT tumors using the database of the Caris Molecular Intelligence.

Methods: We analyzed 541 patient samples for immunohistochemistry (IHC) and next generation sequencing (NGS) data available from the Caris database. Out of the 541 samples 89 samples also had PD-L1 expression data available. Samples were grouped in 4 subtypes: BRAF^mut (n = 169), NRAS^mut (n = 151), cKIT^mut(n = 25) and 3xWT (n = 197).

Results: The database is skewed with an underrepresentation of BRAF^mut patients and enrichment of 3xWT tumors. BRAF^mut, NRAS^mut, cKIT^mutand 3xWT patient were 30%, 26%, 3.7% and 40% of the total population, respectively. BRAF^mut and 3xWT subgroups have more frequent cMET expression (p = 0.002). While NRAS^mut tumors show lack of ERCC1 expression, cKIT^mut tumors lack PGP expression. All mutation subtypes present equally frequent expression of MGMT, SPARC, TOP2A, TOPO1, TS, TUBB3 and RRM1 suggestive of multiple chemoresistance pathways. TP53 mutations are recurrent (~15%) in all mutation subgroups. In 3xWT tumors a different spectrum of mutations arise with low frequency ( < 10%). Mutant KRAS, JAK3, cMET, GNA11, GNAQ, APC, KDR, BRCA1, ERBB4, represent actionable alterations in up to 40% of that subgroup. None of the 541 tumors presented mutation in Akt, BRCA2, IDH1, CSF1R, GNAS, Notch1, Smo, STK11, VHL, MLH1, MPL, MPM1 A PD-L1 subanalyzes showed an overall 75% positivity. 3xWT tumors were more frequently PD-L1 negative as compared to BRAF^mut(p = 0.077)

Conclusions: To our knowledge,the Caris database provides one of the largest profiling of 3xWT. In contrast to BRAF^mut, NRAS^mut and cKIT^mut tumors, 3xWT melanoma harbor a more complex mutational landscape. Low frequency mutations can identify targets in up to 40% of these patients. 3xWT are less frequently PD-L1 positive. Overall 3xWT patients should be tested for multiple markers in order to identify low frequency mutations