Cancer therapy decisions are often made according to the histopathological-molecular profile of tumor tissue obtained from surgery or biopsy. It has been shown that tumor profiles change with time and treatment, and that tumor tissue is heterogeneous. Thus, other approaches that are easily accessible and less invasive than surgery or biopsy to monitor responses to treatment and predict relapses are urgently needed. In the last few years, the term “liquid biopsy” (LB) has been introduced to represent multifunctional circulating biomarkers in the peripheral blood and other physiological fluids of patients with cancer. LB is a noninvasive alternative to tissue biopsies, but it has not been implemented in routine clinical practice for BC. In addition, liquid biopsy seems to be a promising approach for personalized medicine, which enables the prediction, monitoring, and rational selection of appropriate therapy for individual patients. Although unequivocal evidence has shown the prognostic relevance of CTCs in the peripheral blood of patients with metastatic BC, less evidence is available for the prognostic relevance of CTCs at the time of primary diagnosis. In the present study, we conducted LB through CTCs enumeration, ctcDNA and cftDNA quantification and sequencing for mutational analysis in early BC patients before and after surgery (Figure 1), providing, for the first time to our knowledge, the characterization, even though still limited to a small cohort of patients, of several prognostic tumor biomarkers, mirroring BC phenotype especially in its evolution. Specifically, we found that surgery was able to decrease the number of CTCs in a significant percentage of patients (66%). On a biological point of view, this is consistent with the assumption that cancer feed the release of malignant cells into the bloodstream, hence the reduction or reset of CTCs mirrors the removal of the tumor source. Dissemination of tumor cells from the primary tumor into the bloodstream is a critical step in tumorigenesis and is considered a precursor of distant metastases (12). CTCs have been defined “silent predictors of metastases”, this means that CTCs identifies “in progress” metastases onset, even before any other high-resolution imaging method is able to do (20). Hence, our results would suggest a potential resolution of the disease for those patients with no CTCs after surgery. Conversely, in 33% of patients the number of CTCs either remained unchanged or increased after surgical therapy. Hence, if the number of CTCs after surgery does not reset, the presence of distant micro-metastatic foci can be suspected that supply the circulatory torrent, with consequent persistence of the residual disease after surgery. In all cases, the enrolling of additional patients (which is ongoing in our laboratory) and the follow-up of those already collected are warranted, in order to prove our preliminary hypotheses from this explorative study. As above mentioned, isolation and characterization of the molecular profile of CTCs has the potential to provide a far easier liquid biopsy than tumor tissue biopsies to monitor malignant cell populations during disease progression and in response to therapies. Several CTCs capture and/or isolation technologies have been developed. The bead-based EpCAM antibody CTC capturing system is most widely used and one of them, CellSearch, the only Food and Drug Administration (FDA) approved device for CTCs analysis (enumeration), is considered the current “gold standard” (21). However, the performance and applicability of this system is receiving more and more challenge, especially regarding the capture rate (low cell recovery) and purity (low purity). Moreover, a key limitation of this technology is the ability to release and recover viable intact CTCs for downstream molecular analysis. Here, we successfully detected, counted and lysed CTCs (to get ctcDNA for subsequent NGS) from patients with early BC using the Isolation System Isoflux (Fluxion, San Francisco,CA). Although also Isoflux is an immunobead-based system, it has been projected to address CellSearch limitations, by providing high-sensitivity rare cell isolation coupled with a novel cell retrieval microfluidic-based mechanism (2). We, here, confirm that Isoflux provides a complete workflow to track oncogene mutational changes longitudinally with high success rates. Regarding the analysis of the mutation profiling of ctcDNA and matching cftDNA, our preliminary results refers to early-stage BC samples taken prior to surgery. This was a challenging issue of the study, since, as known, the fraction of ctcDNA and the total concentration of cftDNA are directly correlated with the stage of the disease and are significantly lower in earlier stages (22), which posed some technical issues in the extraction process and subsequent mutational analysis. Although we examined mutation hotspots in 59 genes by NGS, in the sequenced ctcDNA, the mutations were only detected in 9 genes (TP53, CDKN2A, FBXW7, PTPN11, KRAS, NRAS, BRAF, IDH1, ALK) and in the sequenced cftDNA, in 5 genes (PIK3CA, APC ALK, KRAS, TSC1). TP53 was the most frequently mutated gene (Table 4). On a biological point of view, this is consistent with the assumption that a non-functional TP53 has been shown to offer survival advantages to the cancer cells by eluding apoptosis, facilitating growth, conferring anoikis resistance, and the emergence of a potentially more aggressive malignancy (23). TP53 mutations are exceptionally frequent in cancer and are among the key driving factors in triple-negative breast cancer (TNBC) (24). Furthermore, TP53 mutations have been shown to predict a poor response to anthracycline-based neoadjuvant chemotherapy (25,26); others suggested that TP53 mutations confer sensitivity to taxane (27). A recent study suggested that patients with TP53 mutations are more likely to respond to anthracycline/cyclophosphamide-based neoadjuvant chemotherapy (28). Another gene that is frequently mutated in BC patients is PIK3CA. Oshiro et al. (29) reported that mutations in the PI3KCA gene could be targeted as a marker predictive of recurrence. In our study, at present, one patient (0022) had PIK3CA mutated (p. R93W) in the cftDNA. Although promising, this result is still far from being interpreted in the context of the literature due to the small number of patients (n = 6) characterized for their molecular profile by NGS. In addition, in this patient (0022) a different molecular profile of the ctcDNA compared to that of the cftDNA was found. It has been successfully demonstrated by single-cell sequencing that many breast cancers are composed of multiple distinct subclones (30). Intra-patient cellular heterogeneity is widely reported in epithelial malignancies and it is expected that CTCs will also be heterogeneous (31). Our results are consistent with previous findings which showed a heterogeneous pattern of genomic mutations on single ctcDNA or cftDNA obtained from several cancer patients, including BC ones.