Capturing intra-tumor genetic heterogeneity by de novo mutation profiling of circulating cell-free tumor DNA: a proof-of-principle

Abstract

Background: Plasma-derived cell-free tumor DNA (ctDNA) constitutes a potential surrogate for tumor DNA obtained from tissue biopsies. We posit that massively parallel sequencing (MPS) analysis of ctDNA may help define the repertoire of mutations in breast cancer and monitor tumor somatic alterations during the course of targeted therapy.

Patient and methods: A 66-year-old patient presented with synchronous estrogen receptor-positive/HER2-negative, highly proliferative, grade 2, mixed invasive ductal-lobular carcinoma with bone and liver metastases at diagnosis. DNA extracted from archival tumor material, plasma and peripheral blood leukocytes was subjected to targeted MPS using a platform comprising 300 cancer genes known to harbor actionable mutations. Multiple plasma samples were collected during the fourth line of treatment with an AKT inhibitor.

Results: Average read depths of 287x were obtained from the archival primary tumor, 139x from the liver metastasis and between 200x and 900x from ctDNA samples. Sixteen somatic non-synonymous mutations were detected in the liver metastasis, of which 9 (CDKN2A, AKT1, TP53, JAK3, TSC1, NF1, CDH1, MML3 and CTNNB1) were also detected in >5% of the alleles found in the primary tumor sample. Not all mutations identified in the metastasis were reliably identified in the primary tumor (e.g. FLT4). Analysis of ctDNA, nevertheless, captured all mutations present in the primary tumor and/or liver metastasis. In the longitudinal monitoring of the patient, the mutant allele fractions identified in ctDNA samples varied over time and mirrored the pharmacodynamic response to the targeted therapy as assessed by positron emission tomography-computed tomography.

Conclusions: This proof-of-principle study is one of the first to demonstrate that high-depth targeted MPS of plasma-derived ctDNA constitutes a potential tool for de novo mutation identification and monitoring of somatic genetic alterations during the course of targeted therapy, and may be employed to overcome the challenges posed by intra-tumor genetic heterogeneity.

Registered clinical trial: www.clinicaltrials.gov, NCT01090960.

Keywords: breast cancer; cell-free tumor DNA; disease monitoring; intra-tumor heterogeneity; massively parallel sequencing.

Figure 1.

Patient disease presentation, treatment timeline and mutant alleles in the primary breast tumor, liver metastasis and plasma-derived DNA. Biopsies of the primary breast cancer and its synchronous liver metastasis were obtained before initiation of therapy. Following three lines of chemotherapy, the patient was treated with the AKT inhibitor Ipatasertib, and multiple plasma samples were obtained during the course of treatment. DNA samples extracted from the primary tumor, metastasis and plasma samples were subjected to targeted high-depth massively parallel sequencing. Not all mutations identified in the metastasis were reliably identified in the primary breast tumor; however, all mutations present in the primary tumor and/or liver metastasis were found in ctDNA.

Figure 2.

Representation of the mutant alleles in the primary tumor and metastasis and longitudinal monitoring of CA15.3 levels in four plasma-derived ctDNA samples. Genes whose high confidence mutations were detected at a mutant allele fraction (MAF) of ≥5% in the primary tumor are depicted in (A) and (B), whereas genes whose high confidence mutations were detected in the plasma-derived ctDNA, but either absent or present at a MAF of <5% in the primary tumor, are illustrated in (C) and (D). In (B) and (D), representative PET–CT images obtained at baseline and 2 months after initiation of Ipatasertib monotherapy; CA15.3 levels assessed during the fourth line of systemic treatment with Ipatasertib monotherapy. Arrow, initiation of Ipatasertib (AKTi) monotherapy. PD, progressive disease; *PET–CT, pharmacodynamic response.