Background: Detection of BRAF mutations is an established standard of care to predict small-molecule inhibitor (vemurafenib) response in metastatic melanoma. Molecular assays should be designed to detect not only the most common p.V600E mutation, but also p.V600K and other non-p.V600E mutations.
Objective: The purpose of this study was to assess if tumor cellularity can function as a quality assurance (QA) measure in molecular diagnostics. Potential causes of discrepancy between the observed and predicted mutant allele percentage were also explored.
Methods: We correlated pathologist-generated estimates of tumor cellularity versus mutant allele percentage via pyrosequencing as a QA measure for BRAF mutation detection in formalin-fixed, paraffin-embedded melanoma specimens.
Results: BRAF mutations were seen in 27/62 (44 %) specimens, with 93 % p.V600E and 7 % non-p.V600E. Correlation between p.V600E mutant percentage and tumor cellularity was poor-moderate (r = -0.02; p = 0.8), primarily because six samples showed a low p.V600E signal despite high tumor cellularity. A QA investigation revealed that our initial pyrosequencing assay showed a false positive, weak p.V600E signal in specimens with a p.V600K mutation. A redesigned assay detected BRAF mutations in 50/131 (38 %) specimens, including 30 % non-p.V600E. This revised assay showed strong correlation between p.V600E BRAF mutant percentage and tumor cellularity (r = 0.76; p ≤ 0.01). Re-evaluation of the previously discordant samples by the revised assay confirmed a high level of p.V600K mutation in five specimens.
Conclusions: Pathologists play important roles in molecular diagnostics, beyond identification of correct cells for testing. Accurate evaluation of tumor cellularity not only ensures sufficient material for required analytic sensitivity, but also provides an independent QA measure of the molecular assays.