CtDNA agrees (mostly) with tissue analysis in mCRC

In 329 patients who had detectable ctDNA, defined as at least one mutation or one methylated biomarker, the kappa coefficient rose to 89% and the accuracy to 94.8%.

Also as noted, in the 293 patients with liver metastases, the accuracy improved to close to 100% with both NGS and methylated biomarker detection.

“Based solely on the plasma analysis using our prespecified analysis, 14 patients would have received ineffective and potentially deleterious anti-EGFR therapy. All were considered as truly RAS wild-type in their plasma because of the presence of other mutations detected by NGS and/or a positive methylation assay,” the researchers wrote.

The lack of sensitivity of ctDNA in these patients may have been due to the fact that the fraction of mutated alleles in these patients was lower than in patients for whom both plasma and tissue analysis were negative for RAS mutations. All of these patients had early metastases, which could explain why the sensitivity of the plasma assay was lower than in patients with advanced-stage disease, they said.

“On the contrary detection of RAS mutations in plasma but not in tumor cells could exclude some patients from potentially effective anti-EGFR therapy,” they added. They identified eight patients with plasma-positive, tissue-negative discordance, which could be due to errors in tissue sampling due to heterogeneity of mutations within tissues, or to a lack of sensitivity of RAS mutation detection in tissues.

The trial was sponsored by AGEO and supported by a grant from Merck Serono. Dr. Laurent-Puig disclosed grants from AGEO and personal fees from Merck Serono and other companies. Multiple coauthors reported fees from Merck Serono and others.

SOURCE: Laurent-Puig P et al. Ann Oncol. 2018 Feb 9. doi: 10.1093/annonc/mdy061/4846852.