Association of Cell-Free DNA Tumor Fraction and Somatic Copy Number Alterations With Survival in Metastatic Triple-Negative Breast Cancer

Abstract

Purpose Cell-free DNA (cfDNA) offers the potential for minimally invasive genome-wide profiling of tumor alterations without tumor biopsy and may be associated with patient prognosis. Triple-negative breast cancer (TNBC) is characterized by few mutations but extensive somatic copy number alterations (SCNAs), yet little is known regarding SCNAs in metastatic TNBC. We sought to evaluate SCNAs in metastatic TNBC exclusively via cfDNA and determine if cfDNA tumor fraction is associated with overall survival in metastatic TNBC. Patients and Methods In this retrospective cohort study, we identified 164 patients with biopsy-proven metastatic TNBC at a single tertiary care institution who received prior chemotherapy in the (neo)adjuvant or metastatic setting. We performed low-coverage genome-wide sequencing of cfDNA from plasma. Results Without prior knowledge of tumor mutations, we determined tumor fraction of cfDNA for 96.3% of patients and SCNAs for 63.9% of patients. Copy number profiles and percent genome altered were remarkably similar between metastatic and primary TNBCs. Certain SCNAs were more frequent in metastatic TNBCs relative to paired primary tumors and primary TNBCs in publicly available data sets The Cancer Genome Atlas and METABRIC, including chromosomal gains in drivers NOTCH2, AKT2, and AKT3. Prespecified cfDNA tumor fraction threshold of ≥ 10% was associated with significantly worse metastatic survival (median, 6.4 v 15.9 months) and remained significant independent of clinicopathologic factors (hazard ratio, 2.14; 95% CI, 1.4 to 3.8; P < .001). Conclusion We present the largest genomic characterization of metastatic TNBC to our knowledge, exclusively from cfDNA. Evaluation of cfDNA tumor fraction was feasible for nearly all patients, and tumor fraction ≥ 10% is associated with significantly worse survival in this large metastatic TNBC cohort. Specific SCNAs are enriched and prognostic in metastatic TNBC, with implications for metastasis, resistance, and novel therapeutic approaches.

Fig 1.

Genome-wide copy number profiles in cell-free DNA (cfDNA) are highly concordant with metastatic biopsy specimens. (A) REporting recommendations for tumor MARKer prognostic studies (REMARK) diagram. (B) Copy number plots of four representative pairs of metastatic biopsy (left panels) and cfDNA (right panels) with copy number (log2 ratio) indicated on the y-axis and chromosome on the x-axis. Sensitivity and specificity of tumor biopsy somatic copy number alterations detected in cfDNA (n = 10 pairs) are indicated for overall, gain, or loss. Examples of private somatic copy number alterations present in cfDNA but not metastatic biopsy (top panels, white arrow) and conversely metastatic biopsy but not cfDNA (top panels, gray arrow) are indicated. TFx, tumor fraction; TNBC, triple-negative breast cancer.

Fig 2.

Metastatic triple-negative breast cancers (TNBCs) demonstrate enrichment of driver and targetable copy number alterations. (A) Gene-level copy number alterations in primary TNBCs from METABRIC and The Cancer Genome Atlas (TCGA), and (B) from metastatic TNBCs from cell-free DNA (cfDNA). The frequency of gene-level copy number gains (red) or losses (blue) across the genome (top panel) and per-sample copy number alteration for 25 breast cancer–related genes (bottom panel). (C) Percentage of samples with gain (top panel) or loss (bottom panel) for 25 breast cancer–related genes in primary (gold) versus chemoresistant metastatic TNBCs (blue). Genes with significant alteration in metastatic TNBC (Fisher’s exact false discovery rate adjusted [FDR] P < .05) indicated by asterisk. (D) Percent of genes altered in primary TNBCs versus chemoresistant metastatic TNBCs. ULP-WGS, ultra-low-pass whole-genome sequencing.

Fig 3.

Prognostic copy number alterations in metastatic triple-negative breast cancer (mTNBC). (A) Significance (Fisher exact P value) of gene-level alteration of gain versus no gain (left panel) or loss versus no loss (right panel) across the genome. Associations with false discovery rate adjusted (FDR)-adjusted P value < .05 indicated in green (gain in light blue, loss in pink) above solid line. Gold line indicates nominal P value < .01, blue line indicates Bonferroni corrected P < .05 (nominal P value < .001). (B) Volcano plot of negative log₁₀ hazard ratio for overall metastatic survival from highest tumor fraction blood draw by chromosomal cytoband versus hazard ratio significance. Only cytobands significantly gained (light blue) or lost (pink) in mTNBC relative to primary TNBC are plotted (Fisher exact FDR P < .05; corresponding segments in A). Size of individual point indicates the frequency altered among patients with mTNBC. Genes listed are those with expression likely altered by chromosomal alteration.

Fig 4.

Cell-free DNA tumor fraction is an independent prognostic biomarker in metastatic triple-negative breast cancer. (A) Representative copy number plots and tumor fraction (TFx) demonstrating dynamic range of TFx within an individual patient whose chest wall disease initially responded then recurred on clinical trial of cabozantinib with associated drop then rebound in TFx. (B) Scatter plot of individual sample TFx measurements for all samples (center), first blood sample collected per patient (left), and maximum TFx per patient (right). Boxplots indicate 25th to 75th percentiles, with median indicated by central line and whiskers representing 1.5 times interquartile range. (C) Kaplan-Meier curve of overall metastatic survival from first blood draw for patients with metastatic triple-negative breast cancer stratified by TFx of first blood draw. (D) Multivariable Cox proportional hazards model of overall metastatic survival from first blood draw. HR, hormone receptor; HER2, human epidermal growth factor receptor 2.

Fig A1.

Metastatic triple-negative breast cancer cohort survival by baseline clinicopathologic characteristics. (A-C) Kaplan-Meier curves of survival from metastatic diagnosis stratified by stage at diagnosis (A), age at primary diagnosis by decade over 40 (B), and germline BRCA1/2 mutation status (C). (D) Kaplan-Meier curve of survival from first blood draw stratified by line of therapy in the metastatic setting. P-value indicates log-rank test.

Fig A2.

Genome-wide copy number profiles across patients reveal high within-patient correlation and two distinct patterns of copy number alterations. (A) Unsupervised hierarchical clustering of correlation matrix of spearman rho for ichorCNA copy number call in 1 million base pair bins across the genome for all plasma samples with TFx 10%. Colors bars indicate unique patients. (B) Association of factors with metastatic TNBC Cluster. (C) Proportion of samples by primary receptor status of two largest clusters from. Chi-square test of independence on proportions indicated. (D) Gene-level copy number frequency plots of primary TNBCs from METABRIC, stratified by study-reported IntClust10 (bottom) versus other IntClust groups (top). (E) Gene-level copy number frequency plots of metastatic TNBCs from cfDNA, stratified by metastatic TNBC Cluster from Figure 1C. Gain and loss frequencies are shown in red and blue, respectively. Deletion frequencies are negated for comparison. (F) Principal component analysis (PCA) of gene-level copy number calls for METABRIC1 triple-negative breast cancers (n = 123) with projection of gene-level copy number calls for mTNBC from plasma (n = 101) onto the PCA coordinate basis. Samples were designated as basal-like IntClust10 versus other (non-basal) IntClust for METABRIC or Cluster1 versus Cluster2 for mTNBC.

Fig A3.

Copy number profiles of metastatic relative to primary TNBCs. (A-B) Regions of significant loss (A) and gain (B) from GISTIC2.0 analysis of metastatic TNBC copy number profiles derived from cfDNA. (C) Absolute difference in gene-level copy number alteration frequency across the genome in metastatic TNBC samples from cfDNA (n = 101) versus primary TNBCs (TCGA + METABRIC; n = 433). (D-E) Frequency of copy number gain or loss (D) and per-sample gene-level comparison (E) for 25 breast cancer-related genes in 20 patients with paired primary tumor and metastatic cfDNA copy number data. Genes with significant alteration (Fisher exact P < .05) in metastatic samples indicated by dot.

Fig A4.

Chromosomal gains of 18q11 and 19p13 are associated with poor survival in metastatic TNBC. (A) Proportion of patients with primary TNBC (TCGA + METABRIC, total n = 433) or metastatic TNBC (n = 101) with gain or amplification of 18q11, 19p13, or both. (B) Kaplan-Meier curve of overall metastatic survival from highest TFx blood draw for metastatic triple-negative breast cancer patients stratified by gain or amplification of 18q11, 19p13, or both. (C-D) Univariate (C) and multivariable (D) Cox proportional hazards model of overall metastatic survival from highest TFx blood draw. (E) Kaplan-Meier curve of overall survival for primary triple-negative breast cancer patients in METABRIC dataset stratified by gain or amplification of 18q11, 19p13, or both.

Fig A5.

Tumor fraction and copy number profiles are reproducible from independent blood draws. (A) Schematic of independently processed same-day blood draws. Two separate blood tubes from a single venipuncture had plasma separated and were frozen in independent laboratories. Equivalent volumes of plasma then underwent DNA extraction, library construction, low-coverage sequencing, and TFx calculation via ichorCNA. (B) Total cell-free DNA yield (ng per mL plasma; left panel) and TFx (right panel) per patient. (C-D) Representative ichorCNA copy number plots from same-day blood draws for the two patients with the highest detected TFx.

Fig A6.

Tumor fraction is associated with liver metastasis and metastatic survival in patients with primary TNBC. (A) Multiple linear regression of TFx with covariates. (B) TFx by presence or absence of liver metastasis in first blood sample collected per patient (left) or maximum TFx per patient (right). Boxplots indicate 25th-75th percentiles with median indicated by central line and whiskers representing 1.5 times interquartile range. (C) Kaplan-Meier curve of survival from first blood draw stratified by TFx above versus below 10% for only those patients with primary TNBC. (D) Multivariable Cox proportional hazards model of TFx above versus below 10% for only those patients with primary TNBC. (E) Multivariable Cox proportional hazards model of TFx as a continuous variable. Hazard ratio for TFx reported based on increments of 10%.