Colorectal carcinoma 5-year survival ranges from 90% in patients with localized disease to 15% for those with distant metastasis. Recent research has suggested that hypermutation may predict response to immunotherapy in these tumors, but routine tests for microsatellite (MS) instability miss tumours which acquire hypermutation through other mechanisms. Tumour mutational burden (TMB) measured by next-generation sequencing has been proposed to overcome this limitation.
A commercially available next-generation sequencing panel targeting 409 cancer-relevant genes was validated for TMB measurement using 12 MS stable and 14 MS instable metastatic colorectal carcinomas, defined by routine testing of MS loci. The same panel was applied to 53 untested colorectal carcinomas with matched synchronous metastases, collected across 10 years to determine both their TMB and presence of KRAS/BRAF mutations.
All samples could be sequenced at a mean coverage depth of 766x and uniformity of 97%. Mean TMB (mutations/megabase) was 8.84 for MS stable vs. 33.36 for MS instable cases in the assay validation cohort. A cut-off value of 15.56 reached 100% sensitivity (95% CI 77-100%) and 100% specificity (95% CI 73-100%). Analysis of the 10 years cohort showed KRAS mutation in 25 cases and BRAF mutation in 4; automated TMB analysis was feasible for samples collected within 7 years (n = 16), while in older specimens DNA deamination caused artefactual calls. 14 cases showed a low TMB in both primary and metastasis, one MS instable case showed high TMB in both primary and metastasis, and one MS stable case showed low TMB (11.61) in the primary and a higher TMB (21.37) in the metastasis. The mutational signature of the metastatic sample showed C>A transversions (11%), missing in the primary tumour, suggesting an additional mutational mechanism.
Gene panel-based TMB analysis can be performed on routine histology samples to detect both hypermutation and cancer relevant somatic mutations. Analysis of older samples may lead to deamination artifacts, which can however be revealed by mutational signature analysis.
Aldo Scarpa (ARC-NET Cancer Research Centre).
Associazione Italiana Ricerca Cancro [AIRC grant n. 12182].
All authors have declared no conflicts of interest.