There is increasing interest in the use of circulating tumour DNA (ctDNA) to identify targetable genomic alterations for therapy selection. However, the feasibility of next generation sequencing on ctDNA in BC patients with varying disease burden merits further investigation.
The cohort of 35 BC patients included 30 metastatic cases with paired primary and metastatic specimens in addition to plasma taken prior to commencement of a new line of palliative systemic therapy (all subtypes), + 5 patients (3 stage III, 2 stage II) about to commence neoadjuvant systemic therapy. DNA from tumour, buffy coat and plasma was sequenced on a 77-gene capture panel customised for BC. Matched tumour/normal samples were processed to discover somatic alterations using a standard GATK pipeline. Plasma samples were processed using unique molecular identifiers to identify potential alterations at low frequency.
Among the entire cohort, 74% (26/35) of patients had mutation(s) in at least 1 gene detected from ctDNA: 60% (3/5) for the neoadjuvant cases, 77% (23/30) for the metastatic cases. The most frequently mutated genes from ctDNA analysis were TP53 (17/35; 49%) and PIK3CA (8/35; 23%), with HER2, HER3, JAK2, NF1, MAGI3 and TSC2 mutations observed in 2 patients each (6%). There was no obvious correlation between detection of ctDNA mutation and disease burden, serum CA-15.3 tumour marker or circulating tumour cell levels using a microfluidic platform. Out of 35 patients, 21(60.0%) had ≥1 concordant mutation via both ctDNA and tumour genotyping. Concordance between the plasma and primary and/or metastatic lesions was observed for 59% (39/66) of the mutations detected in ctDNA. Among the metastatic cases, 14/37(38%) of the concordant mutations were shared between the plasma, primary and metastatic specimens, while 5(14%) were shared between the plasma and the primary only, and 18/37(49%) shared between the plasma and metastatic lesions only.
Detection of genomic alterations from ctDNA is feasible in BC patients, but concordance of mutations in ctDNA is better with the metastatic than the primary lesion. This is likely due to suboptimal quality of DNA from archived tissue, and spatial, temporal heterogeneity.
National Cancer Centre Singapore and Genome Institute of Singapore.
Y.S. Yap: Personal financial interests, honoraria for consultancy and talks, travel support: AstraZeneca, Eisai, Lilly, Novartis, Pfizer, Roche. All other authors have declared no conflicts of interest.