Circulating tumor DNA fraction predicts residual cancer burden post-neoadjuvant chemotherapy in triple negative breast cancer

Abstract

Purpose: Pathologic response after preoperative/neoadjuvant chemotherapy (NAC) is a continuum that can range from complete pathologic response (pCR) to extensive residual disease (RD). We hypothesized that post-NAC plasma circulating tumor DNA (ctDNA) fraction (TF) reflect pathologic response as continuum measured by the residual cancer burden (RCB) score.

Methods: ctDNA was assessed using the PredicineBEACON assay, that interrogates up to 50 personalized tumor variants and 500 hot-spot mutations, in 3 mL archived plasma isolated from EDTA tubes collected post-NAC but before surgery from 44 patients with stage I/III triple negative breast cancer (TNBC) who received durvalumab and weekly nab-paclitaxel followed by doxorubicin/cyclophosphamide on a clinical trial (NCT02489448). Circulating free tumor DNA methylation profiling was performed using PredicineEPIC assay in paired pre- and post-NAC plasma (N = 30). Youden's J-statistics was used to define optimal thresholds.

Results: We observed a significant positive correlation (r = 0.45, p = 0.004) between RCB scores and post-NAC TF. The median TF was significantly lower in pCR (RCB0) compared to RD patients (0.06 % vs. 0.3 %, p = 0.02). Using a TF positivity threshold of ≥0.05 %, PredicineBEACON had 58 % sensitivity at 83 % specificity for identifying RD. TF was higher in patients who experienced recurrence (n = 9) compared to those without recurrence (n = 35) (0.17 % vs. 0.05 % TF, p = 0.029). There was significant decrease in methylation signal in post-compared to pre-NAC samples, but post-treatment methylation signal was lower in cases with pCR vs RD.

Conclusions: Post-NAC plasma tumor fraction and change in tumor-derived methylation signal predict the extent of RD and recurrence in TNBC patients.

Keywords: Circulating tumor DNA (ctDNA); Minimal residual disease (MRD); Residual cancer burden (RCB).

Fig. 1

Flow diagram of sample disposition.

Fig. 2

Post-neoadjuvant chemotherapy plasma tumor fraction is associated with residual disease and recurrence. A, Plasma tumor fraction percentage in samples with pathological complete response and residual disease. B, Correlation between RCB score and tumor fraction as continuous variable. p = 0.004. C, Plasma tumor fraction in patients with no evidence of recurrent disease and local/metastatic recurrence. D, Correlation of RCB score and tumor fraction in patients with local or metastatic recurrence and no recurrence. p = 0.004 RCB = residual cancer burden, NED = no evidence of recurrent disease, RD = residual disease, pCR = pathological complete response.

Fig. 3

Tumor Fraction and ctDNA positivity are correlated with increasing RCB category. A, Tumor fraction percentages in patients based on RCB category. ∗, p < 0.05; ∗∗, p < 0.01. B, Percentage of patients based on RCB category. C, Correlation between RCB score and tumor fraction vs time to relapse. RCB = residual cancer burden.

Fig. 4

Tumor-derived cfDNA methylation signal pre- and post-NAC. A, Post-NAC methylation levels in patients who achieved pathologic complete response vs patients with residual disease. B, Methylation signal change in post-NAC samples compared to pre-NAC samples in patients who achieved pathologic complete response vs patients with residual disease. DMF = differentially methylated fragments, pCR = pathological complete response, RD = residual disease.