A number of factors are associated with an increased risk of developing BC (hereditary factors, hormonal factors, lifestyle factors, benign breast diseases, high radiographic density of the breast, environmental factors); however, most of these factors are associated with a moderate increase in individual risk, while at least half of the women who are diagnosed with BC have no identifiable risk factors, except for increasing age. Several genes have been identified that predispose individuals to hereditary BC. The majority of them are BRCA1 and BRCA2 genes. The estimated lifetime risk of BC development in carriers of BRCA1 and BRCA2 mutations ranges from 26% to 85%. Other genetic mutations associated with BC risk have a lower penetrance. These are mutations in the TP53, PTEN, CDH1, CHEK2, PALB2, and ATM genes ^[1][2].

The prognosis and the treatment of BC are generally determined by the stage (i.e., the degree of anatomical spread of the tumor) and the biological characteristics of the tumor (i.e., whether it belongs to a specific molecular biological subtype).

In essence, BC is biologically a heterogeneous disease; individual subtypes are characterized by different sensitivities to drugs and different prognoses. The key characteristic that influences the choice of treatment tactics is the expression of hormonal receptors (estrogen (ER) and progesterone (PgR)) and human epidermal growth factor receptor 2 (HER2). These markers can be used to define different functional groups of tumors: hormone receptor-positive, HER2-negative tumors (sensitive to hormone therapy); HER2-overexpressing tumors with or without hormone receptor expression (responsive to anti-HER2 therapy); hormone receptor-negative, HER2-negative tumors (“triple-negative” tumors; not sensitive to hormone therapy and anti-HER2 therapy). Studies on the molecular biology of tumors have shown that these subtypes have different well-defined genomic profiles ^[2][3][3,4], which allows for subdividing BC into luminal A (ER-positive and PgR-positive/HER2-negative with lower-grade features), luminal B (ER-positive and/or PgR-positive but with higher-grade features or HER2-positive), HER2-positive (ER-negative/PR-negative/HER2-positive), and basal-like (ER-negative/PR-negative/HER2-negative). These subgroups affect both the likelihood and timing of cancer recurrence: triple-negative/basal-like, HER2-positive, and luminal B BC are at greater risk of early recurrence relative to luminal A cancers, which have a longer latency period of possible recurrence ^[4][5]. Thus, today, BC is not considered as one disease, and when establishing a diagnosis, in addition to determining the anatomical stage, it is necessary to determine the molecular biological subtype (using genetic testing or by determining ER, PgR, and HER2 in combination with the determination of the proliferation marker Ki67). In addition to defining biologic tumor subsets, gene expression profiling (MammaPrint, Irvine, USA, Oncotype DX, Redwood City, USA, Prosigna, South San Francisco, USA) has been used to stratify tumors as having good-risk or poor-risk prognostic signatures. Retrospective analyses suggest that these gene signatures contribute independent prognostic information above and beyond that achieved with the use of traditional pathologic markers, such as stage, grade, lymphovascular invasion, and ER/PgR/HER2 status. These genomic assays have proven especially valuable in distinguishing the prognosis within the subset of luminal, ER-positive, HER2-negative breast tumors, which are the most common form of BC ^[1][2]. However, these genomic assays cannot provide all the necessary information for a precision approach for all subtypes of BC.

The discovery and implementation of prognostic and predictive tumor markers in everyday clinical practice, along with the possibility of individualization of therapy for BC, has given rise to a number of methodological problems. Firstly, BC is a heterogeneous disease, not only within the given nosological form (referencing the four molecular subtypes described previously), but also within a specific case of the disease, since one tumor node may contain foci with different expressions of ER, PgR, and HER2. Core biopsies, traditionally used for diagnostic purposes, do not guarantee the collection of representative material in the case of a heterogeneous tumor. Secondly, in the course of treatment and progression, a tumor can change its biological characteristics. For example, the sensitivity of a tumor to therapy changes due to the loss of receptors in metastatic and recurrent foci, as well as the acquisition of mutations (for example, ESR1 during hormone therapy with aromatase inhibitors). This should be taken into account when planning treatment. Reobtaining material using a core biopsy, including from a metastatic focus in order to determine the biological characteristics of the tumor, is often difficult and sometimes technically impossible. In this regard, the method of liquid biopsy with the determination of various components, such as circulating tumor DNA (ctDNA) isolated directly from plasma or from extracellular vesicles, as well as circulating tumor cells, and others, is of great value, since it allows the use of a minimally invasive procedure (for example, blood sampling from a peripheral vein) to obtain diagnostic material as often as desired.

ctDNA detection is the most widely studied, minimally invasive alternative for the molecular characterization of solid tumors. It has now been shown that ctDNA is characterized by the presence of genetic and epigenetic changes identical to those contained in the tumor tissue. The determination of specific changes in driver genes may be of diagnostic value, correlating with the response to therapy and the dynamics of the tumor process during the treatment period. Since most of the molecular abnormalities found in plasma ctDNA reflect genetic and epigenetic changes in the primary tumor, ctDNA analysis is a convenient predictive and prognostic method for monitoring the course of oncological diseases. The key advantages of ctDNA analysis include its high specificity; correlations with tumor burden, including metastasis, and response to treatment; and the representativeness of the obtained material, which excludes to a maximum extent the possible heterogeneity of the tumor.

2. Clinical Value of ctDNA in Breast Cancer Patients

The actual clinical tasks in the treatment of BC are early diagnosis, monitoring of disease progression, the evaluation of effectiveness, including in the case of NAC, the detection of minimal residual disease, and the prediction of the effectiveness of different treatment methods.

2.1. Application of ctDNA for Early Detection/Screening of Breast Cancer

Currently, the only available methods for the screening and early detection of localized BC, for which there is a possibility of radical treatment, are self-examination and objective imaging methods such as mammography, ultrasound, and magnetic resonance imaging. However, mammography has age restrictions, as it has been proved to be generally not informative in young women. In addition, mammographic screening is associated with overdiagnosis. All this requires the development of new diagnostic tests with fundamentally different capabilities, and as a result, liquid biopsies providing the determination of ctDNA may well claim the role of a screening method for BC. There are a number of studies that substantiate the possibility of using ctDNA analysis for the early detection of BC. Some works are based on the identification of mutations in ctDNA ^[5][6][122,123], while other works are based on the analysis of aberrant methylation in patients compared to healthy people ^[7][8][17,124]. Kamel A.M. et al. used parameters of DNA damage and determined that the level of DNA damage reliably identifies patients with BC compared to benign breast patients and healthy subjects (p < 0.001) ^[9][49]. The determination of mutations in genes using the CancerSEEK panel made it possible to propose a method that detects BC with a sensitivity of 33% and a specificity of 99% ^[5][122]. A use case for ctDNA was proposed by Rodriguez B.J. et al. in 2019, which consisted of the detection of mutations in the TP53 and PIK3CA genes ^[10][125]. The development of a panel-based approach including exonic regions of 33 genes involved in BC pathogenesis yielded a specificity of 86.36% and a positive predictive value of 88.46% for BC detection ^[6][123]. Li Z. et al. were the first to evaluate the methylation status of the EGFR and PPM1E promoters in the detection of BC. They observed that patients with BC had significantly higher levels of methylation than healthy individuals ^[11][126]. Fackler M.J. et al. and Visvanathan K. et al. have consistently developed assays to distinguish BC from benign breast disease and healthy normal subjects using ctDNA (the earlier cMethDNA and the more recent automated Liquid Biopsy for Breast Cancer Methylation (LBx-BCM) prototype based on the previous one). LBx-BCM achieved a ROC AUC = 0.909 (95% CI = 0.836–0.982), 83% sensitivity and 92% specificity; cMethDNA achieved a ROC AUC = 0.896 (95% CI = 0.817–0.974), 83% sensitivity and 92% specificity ^{[12][13][14][15]}[42,127,128,129]. Approaches based on whole-genome bisulfite sequencing analysis have been introduced in recent years, making it possible to obtain significant results. In one of these works, Gao Y. et al. exhibited high accuracy in early (AUC of 0.967) and advanced (AUC of 0.971) BC stages ^[7][17]. Hai L. et al. identified potential breast-cancer-specific methylation CpG site biomarkers with high specificity and sensitivity ^[16][130]. In a recent work, the genome-wide methylation method was enhanced with additional analyses of copy number alterations and four-nucleotide oligomer end motifs (4-mers), as well as multi-featured machine learning. As a result, the authors obtained a model that achieved an AUC of 0.91 (95% CI 0.87–0.95) and a sensitivity of 65% at 96% specificity ^[8][124]. Methylation changes in breast cancer-associated genes or regions can be detected in the blood in the earliest stages of the disease. This opens up the potential for non-invasive screening approaches that could complement or improve existing diagnostic methods. Thus, ctDNA can be considered as a potential biomarker in liquid biopsy samples to detect aberrant methylation and specific mutations in BC, in order to develop non-invasive methods for the early detection of BC.

2.2. ctDNA as a Predictive and Prognostic Marker in Breast Cancer Treatment

2.2.1. Two Main Approaches to Determining ctDNA in Breast Cancer

Analysis of the results of recent clinical studies in BC allowed us to identify two main methodological approaches in determining ctDNA for use in the treatment of BC. The first is the analysis of point mutations and aberrant methylation in the presence of known driver genes, which is more relevant for hormone-dependent and HER2-positive BC. The second approach is based on whole-genome or -exome analysis of cfDNA. This variant is more convenient for triple-negative BC because there are no driver genes for analysis in this subtype of BC.

2.2.2. Clinical Value of ctDNA Determination for Breast Cancer Treatment

From the point of view of clinical value, the considered studies can be divided into three main groups: studies devoted to the early detection of recurrence or progression of BC, the detection of MRD after primary treatment of BC, and usage in metastatic BC as a prognostic criterion. All of these lines of research are promising for the future development of protocols and assays for clinical use.

Early Detection of Recurrence/Progression of Breast Cancer

The significance has been shown of ctDNA in BC in the detection of preclinical metastases and the prediction of recurrence after surgery and/or NAC in patients with specific hotspot mutations ^[17][18][150,155], chromosomal rearrangements ^[19][156], and amplifications ^[20][157]. The most common driver genes in the tissues of primary and metastatic breast tumors—in particular, the TP53 gene of the p53 tumor protein, and the PIK3CA gene and others—can be analyzed using liquid biopsies. Using the panel, which includes 52 of the most common oncogenes and tumor suppressors, the detection rate of tumor-specific mutations in ctDNA was 37% and 81% for I–III stage BC and metastatic/recurrent BC, respectively. Of the 57 single nucleotide variants (SNVs) detected, the majority of the variants were found in five genes, namely TP53 (37%), PIK3CA (21%), AKT1 (7%), EGFR (5%), and KRAS (5%). The frequency of mutations’ detection in ctDNA correlated with the stage of the disease, increasing with metastatic or recurrent disease (p = 0.00026), with lymph node involvement (p = 0.00649), and with distant metastases (p = 0.0005) ^[21][61]. Significantly lower levels of promoter promA methylation of ESR1 gene at baseline were observed in patients with liver metastases (p = 0.0212) in ER-positive metastatic BC ^[22][131]. Riva F. et al. found TP53 mutations in patients’ ctDNA at baseline in 75% triple-negative BC patients. Its detection was associated with the mitotic index (p = 0.003), tumor grade (p = 0.003), and stage (p = 0.03). The patient with rising ctDNA levels experienced tumor progression during NAC ^[23][54]. As can be seen from the frequency of detection of driver mutations in ctDNA, a significant number of BC patients do not have hotspot mutations and therefore cannot be controlled using the driver gene approach. Therefore, a comprehensive solution is needed based on the use of whole-genome/exome sequencing of each patient’s tumor to search for characteristic somatic mutations. An intermediate approach was used in the c-TRAK TN trial, enrolling patients with triple-negative BC. Since these patients did not have known target genes, the analyses of tumor samples were carried out via targeted sequencing, principally with gene sequencing panels. However, for analysis in cfDNA, only one or two mutations from the sequencing were selected. ctDNA-positive patients accounted for 27.3% of all the patients over the 12 months of observation. The patients had a high rate of metastatic disease in ctDNA-positive detection (nearly 72%) ^[24][132]. Using ultra-deep sequencing to detect tumor-specific mutations in cfDNA selected from data from the whole-exome sequencing of primary tumors, the sensitivity and specificity of this parameter in relation to the incidence of recurrence in patients with BC has been revealed ^[25][151]. Tumor-specific mutations in ctDNA were found to be associated with the development of disease recurrence and progression-free survival (PFS) or disease-free survival (DFS) ^{[17][26][27][28]}[18,19,133,150]. Parsons H.A. et al. developed an ultra-sensitive, patient-specific ctDNA analysis based on exome sequencing to identify patient-specific SNVs, which can be used to design custom ctDNA tests. ctDNA detection was strongly associated with distant recurrence (HR = 20.8; 95% CI 7.3–58.9). The median lead time from the first positive sample to recurrence was 18.9 months (range = 3.4–39.2 months). It is important to note that most patients had a much lower number of mutations in the tumor for analysis in ctDNA than the test allowed (median 57, range 2–346) ^[26][18]. Aberrant methylation in the specific region EFC#93 (a pattern of five linked CpGs methylated in BC) was an independent and poor prognostic marker in pre-chemotherapy samples for death (HR = 7.689), DFS, and OS. EFC#93 positivity after chemotherapy was significantly more frequent at late stages (T2–T4) (p = 0.014) ^[29][46]. In some patients, even in the early stages of BC, a relapse of the disease develops after primary treatment. The risk of relapse weighs heavily on the minds of both patients and physicians for years after treatment of the primary tumor is complete. Regular follow-up with a medical oncologist aims to identify new symptoms or changes on physical examination. The current National Comprehensive Cancer Network (NCCN) and American Society of Clinical Oncology (ASCO) guidelines recommend the taking of a routine history and physical examination with a certain frequency. Both the NCCN and ASCO do not recommend the use of most of the examinations and blood tests (e.g., routine complete blood counts, chemistry panels, tumor markers, bone scans, computed tomography (CT) scans, magnetic resonance imaging (MRI) scans, positron emission computed tomography (PET) scans, or ultrasound examinations) in asymptomatic patients without specific clinical examination findings. Dedicated breast MRI may be considered for post-therapy surveillance in women at high risk for bilateral disease, such as carriers of the BRCA 1/2 mutations ^[30][31][158,159]. However, despite these guidelines many patients receive high-cost imaging analysis (CT, brain or body MRI, PET, and bone scans) and tumor marker blood tests during routine follow-up exams, exposing them to radiation and increasing healthcare costs. Although a number of past studies of more thorough and more financially costly dynamic follow-up (compared to the standard one) has not revealed advantages in the PFS and overall survival (OS) of patients with early detection of metastatic disease, the early detection of locoregional relapses allows for radical treatment and has a positive effect on long-term results ^[32][33][160,161]. As for metastatic disease, technologies for the local control of metastatic foci (stereotactic radiation therapy, radiofrequency, cryoablation, etc.) are currently available, which, in combination with highly effective methods of drug therapy, can significantly improve long-term treatment results. However, the scope of local methods is limited to clinical situations with a few (single) foci of small sizes (oligometastatic disease), which increases the demand for early detection of metastatic processes. Over the past decade, the development of non-invasive biomarker assays has enabled the low-cost early detection of cancer: according to the available data, the appearance of ctDNA may precede the clinical manifestation of metastatic disease by many months. Attempts are being made to develop automated assays for monitoring therapeutic response and predicting disease recurrence based on the analysis of aberrant methylation of a panel of genes ^[15][129].

ctDNA as a Surrogate Marker for Minimal Residual Disease after Primary Treatment for Breast Cancer

Neoadjuvant drug therapy is an essential treatment for early and locally advanced BC. As is known, in early and locally advanced BC, especially in triple-negative and HER2-positive variants, the achievement of a pathological complete response (pCR) has important prognostic value and is associated with an increase in OS and PFS ^{[34][35][36][37][38][39][40][41]}[162,163,164,165,166,167,168,169]. In clinical practice, to assess the effectiveness of NAC, traditional clinical and instrumental methods (examination, palpation, ultrasound of the mammary glands and regional zones, mammography) are currently used, which often do not allow for the true effect of treatment to be evaluated. For example, in the study of Ignatova E.O. et al., when analyzing the results of NAC in 41 patients with early primary operable triple-negative BC, it turned out that in the clinical and instrumental assessment of the effect of using ultrasound and mammography, the frequency of complete responses was only 30%, while the pathomorphological study of the postoperative material revealed a pCR in 65% of cases ^[42][170]. Generally, six to eight courses of chemotherapy ± targeted therapy (anti-HER2 therapy if necessary) are used in NAC. However, for some patients in the early stages with a high sensitivity to drug therapy, such a volume of treatment may be excessive, and analysis of ctDNA during treatment can help to find the optimal number of courses that each individual patient needs. To assess the degree of pathologic response in everyday clinical practice, routine pathomorphological examination with the definition of categories, such as ypTypN (evaluation of tumor T and N categories after NAC through pathological staging) and RCB (Residual Cancer Burden), is currently used. However, this method does not allow for studying 100% of the removed tissues and therefore carries the possibility of error due to the heterogeneity of the diagnostic material. Meanwhile, today, information on the degree of pathologic response of the tumor is the key to determining the tactics of post-neoadjuvant therapy in triple-negative BC and HER2-positive BC. In addition, an extremely interesting modern trend in the treatment of early BC is the study of the possibility of refusing surgical treatment in case of pCR achievement. In such studies, a biopsy of tumor tissue is used to confirm a pCR; however, this method does not allow for analyzing the entire mass of the tumor. Recent clinical studies have shown that ctDNA may play a role in detecting MRD and early detection of resistance to therapy, i.e., the molecular recurrence of BC ^{[17][18][24][43]}[132,150,152,155]. It can also be used as a method for monitoring disease progression in patients with advanced BC ^[27][44][45][19,171,172]. A relationship was found between TP53 mutations in triple-negative BC patients’ ctDNA and the effectiveness of NAC. ctDNA positivity after one cycle of NAC correlated with shorter DFS (p < 0.001) and OS (p = 0.006) ^[23][54]. Based on the whole-exome detection of mutations in tumor tissue and the formation of individual mutation panels for each patient, one of the recent studies revealed an association between the presence of tumor-specific mutations in ctDNA and response to NAC, the risk of recurrence, metastasis, and 3-year survival. It was revealed that an important time point for the prognosis of pCR is in the early (3 weeks from the start) period of NAC. At T0 (pretreatment), 61 of 84 (73%) patients were ctDNA-positive, which decreased over time (T1: 35%; T2: 14%; and T3: 9%). Patients who remained ctDNA-positive at T1 (3 weeks after the initiation of paclitaxel) were significantly more likely to have residual disease after NAC (83% non-pCR) compared with those who cleared ctDNA (52% non-pCR; OR = 4.33, p = 0.012). After NAC, all patients who achieved a pCR were ctDNA-negative (n = 17, 100%). For those who did not achieve a pCR (n = 43), ctDNA-positive patients (14%) had a significantly increased risk of metastatic recurrence (HR = 10.4). Interestingly, patients who did not achieve a pCR but were ctDNA-negative (86%) had an excellent outcome, similar to those who achieved a pCR (HR = 1.4). In this trial, the lack of ctDNA clearance (absence of tumor-specific mutations in ctDNA) was a significant predictor of poor response and metastatic recurrence (HR = 22.4; 95% CI 2.5–201, p < 0.001), while clearance was associated with improved survival even in patients who did not achieve a pCR ^[27][19]. Therefore, the personalized monitoring of ctDNA during NAC may aid in the real-time assessment of treatment response and help fine-tune pCR as a surrogate endpoint of survival ^[27][46][19,69]. ctDNA methylation analysis may also be a marker of treatment response ^[47][48][49][134,135,136]. A significant correlation was found with the extent of residual tumor burden, and response to therapy with the methylation of the RASSF1A promoter region in ctDNA after NAC ^[47][49][134,136], but not for MGMT promoter ^[50][137]. A systematic review on the prognostic potential of cfDNA methylation for hormone receptor-positive BC was conducted in 2020 and used most of the previous studies. Hypermethylation of the markers RASSF1, BRCA, PITX2, CDH1, RARB, PCDH10 and PGR, and GSTP1 showed a statistically significant correlation with poor disease outcome ^[51][173]. The authors also identified high heterogeneity in the protocols of these studies, which indicates the need for standardized assessment methods. The detection of MRD can be critical in assessing patients’ therapeutic response and subsequent treatment decisions. Several studies have evaluated the value of ctDNA in detecting MRD after NAC and surgical treatment ^[26][52][18,79]. Detection of ctDNA in serial samples using early-stage BC patients who received NAC before surgery was predictive of early relapse (HR = 12.0). Of the patients who did not relapse, 96% did not have ctDNA detected p < 0.0038) ^[17][150]. These results were confirmed by later studies of different subgroups of BC patients ^[53][54][138,153]. In the c-TRAK TN trial, the patients had a high rate of metastatic disease in ctDNA-positive detection (nearly 72%). An important conclusion was also drawn that three-month blood sampling is not frequent enough; this indicates the need for commencing ctDNA testing early, with frequent ctDNA testing regimes ^[24][132]. The presence of methylated ctDNA after the end of the chemotherapy was an indication of MRD ^[55][43].

ctDNA Mutations as Prognostic and Predictive Factors for the Effectiveness of Therapy

Several studies have shown the usefulness of analyzing point mutations in the ctDNA of HER2-positive patients ^[56][57][58][139,140,141]. TP53 mutations in ctDNA can be used as biomarkers of anti-HER2 antibody drug resistance in HER2-positive patients and HER2 tyrosine kinase inhibitor resistance in patients with HER2 mutations without amplification. Patients carrying a TP53 mutation in their ctDNA had a shorter PFS in response to anti-HER2 antibody treatment than those without a TP53 mutation (HR = 1.42, p = 0.004). TP53 mutations predicted the low efficacy of HER2 tyrosine kinase inhibitors (HR 2.83, p = 0.01) ^[56][139]. This is consistent with the data obtained by Rothé F. et al. ^[58][141]. Analysis of HER2 amplification using whole-genome sequencing data for gene copy number variation analysis showed a high concordance rate of the observed status of HER2 between tissue and ctDNA and a relationship with PFS and objective response during anticancer therapy. Among the patients that were histopathologically HER2-positive in the primary tumor, high-level amplification in HER2 copy numbers of ctDNA in the baseline treatment was significantly correlated with the best objective response during anticancer therapy (p = 0.010). Moreover, HER2 copy numbers fluctuated with the HER2-targeted therapeutic response, and the patients with a constantly positive level after 6 weeks of treatment appeared to suffer from a significantly reduced PFS (p < 0.001) ^[57][140]. In contrast to early, non-metastatic BC, ctDNA is detectable in the majority of metastatic BC: 85.71% of stage IV/M1 patients carried tumor-derived mutations in their blood, compared to only 57.81% of stage I–III/M0 patients ^[59][174]. Most of them have used a point-mutation-based approach for ctDNA analysis. Many of these studies were aimed at exploring the possibility of using the ctDNA marker to evaluate the effectiveness of different types of therapy in this group of patients. It is known that patients with advanced BC have undergone prior aromatase inhibitor therapy followed by a selection of ESR1 mutations by the time of tumor progression; therefore, attempts are being made to find new effective combinations for their treatment. Mutations of the main driver genes TP53, ESR1, PI3KCA, HER2, and AKT1 are of the greatest significance, according to the available studies. It has been observed that TP53 mutations in exons 5–8 may be an independent prognostic marker for short DFS (HR = 1.50, p = 0.009), and TP53 mutations may influence the effect of trastuzumab-based (anti-HER2) chemotherapy alone or in combination with taxanes ^[60][70]. Patients with TP53 mutations had significantly worse OS than the carriers of wild-type alleles (p = 0.0094) ^[61][142]. In the case of treatment with palbociclib (CDK4/6 inhibitor) plus fulvestrant (antiestrogen), PIK3CA or TP53 mutations were prognostic for OS (HR = 1.44 and 2.19, respectively) and associated with shorter OS (OR = 0.55 and 0.23) ^[62][143]. The presence of ESR1 mutations in ctDNA of advanced BC patients showed a worse PFS compared with those with the ESR1 wild type (HR = 1.46, p = 0.02) ^[63][20], and were highly associated with shorter OS (OR = 0.36) ^[62][143]. Dynamic monitoring of ESR1 mutations served as a predictive biomarker of acquired resistance to endocrine therapy, whereas the combination of everolimus (inhibitor mTOR) with acquired ESR1 mutations showed a longer PFS than other therapies without everolimus ^[61][142]. Patients with an ESR1 mutation in their ctDNA had a worse PFS than those with the ESR1 wild type when treated with aromatase inhibitor exemestane (2.6 months versus 8.0 months, HR = 2.12; p = 0.01), but not fulvestrant (HR 0.52; 95% CI 0.30–0.92; p = 0.02) ^[63][20]. In the case of palbociclib (CDK4/6 inhibitor) plus fulvestrant, this combination provided a PFS benefit regardless of the ESR1 mutation status in ctDNA at day 1 or at the end of treatment ^[62][143]. But an early change from an aromatase inhibitor plus palbociclib to fulvestrant plus palbociclib treatment for patients with ESR1 mutations in ctDNA ensured a gain in PFS ^[64][65][55,144]. So, this suggests the clinical efficacy of periodic monitoring of emerging or rising ESR1 mutations in ctDNA to trigger an early change from an aromatase inhibitor plus palbociclib to fulvestrant plus palbociclib treatment in patients with rising circulating ESR1 mutations detected even without tumor progression. Recently, a numerically greater improvement in median PFS for patients with ESR1 or PIK3CA mutations in their ctDNA was observed in the combination of abemaciclib (CDK4/6 inhibitor) plus fulvestrant ^[66][56]. The PIK3CA mutation status based on ctDNA also demonstrated prognostic significance for combination therapy with buparlisib (PI3K inhibitor) versus fulvestrant alone. Median PFS was significantly longer in the buparlisib versus placebo group ^[67][59]. The combination of a dual AKT and p70 ribosomal protein S6 kinase (p70S6K) inhibitor (LY2780301) with paclitaxel (taxane) in advanced BC was studied in patients with driver mutations of PI3KCA, AKT1, and TP53 in their tumors. The presence of ctDNA upon inclusion was correlated with PFS (6-month PFS was 92% for ctDNA-negative patients versus 68% for ctDNA-positive cases (HR = 3.45, p = 0.007)). Copy number alterations were associated with disease progression under paclitaxel-LY2780301. Therefore, ctDNA detection at baseline was associated with shorter PFS, while plasma-based copy number analysis helped to identify alterations involved in resistance to AKT/p70S6K inhibitor plus paclitaxel treatment ^[68][57]. The PlasmaMATCH study, which was a trial of ctDNA testing in advanced BC patients, was focused on a number of driver genes for BC. The study included 1034 patients with disease progression who were stratified based on mutations in the PIK3CA, ESR1, HER2, PTEN, and AKT1 genes identified in ctDNA. As a result, it was found that ctDNA testing enables the selection of mutation-directed therapies for patients with BC, with sufficient clinical validity for adoption in routine clinical practice. A positive response was demonstrated in carriers of HER2 mutations treated with neratinib and, if ER-positive, with fulvestrant, as well as in carriers of AKT1 mutations and ER-positive BC treated with capivasertib plus fulvestrant. In the study, neratinib for HER2-mutant BC identified via ctDNA testing had comparable activity to that observed when guided by tissue testing in a previous study ^[69][175], with durable responses. Similarly, capivasertib had high activity in patients with ctDNA-identified AKT1 mutations, both in hormone-receptor-positive BC with fulvestrant and in hormone-receptor-negative BC as a single agent, confirming the results of a previous phase I study ^[70][176]. Therefore, the results demonstrated the clinically relevant activity of targeted therapies against rare HER2 and AKT1 mutations identified via ctDNA testing ^[71][145]. The effectiveness of treatment in accordance with the mutational status of ctDNA was also shown by Lyu D. et al.; the risk of progression was lower in groups of patients who received precision therapy based on ctDNA analysis (HR = 0.55, p = 0.023) ^[72][146]. Thus, the approach using known driver genes is convenient for hormone-dependent advanced BC. A study using plasma samples from the CAMELLIA trial also confirmed the value of ctDNA analysis as a potential biomarker of therapeutic response and prognosis in patients with metastatic BC. The work was conducted using target-capture deep sequencing of 1021 genes to detect somatic variants in ctDNA. These data were used to determine the molecular tumor burden index (mTBI), which is a function based on somatic variations in ctDNA, considering the heterogeneity and dynamic evolution of the tumor. Patients with a high-level pretreatment mTBI had shorter OS than patients with a low-level pretreatment mTBI (p = 0.011). Patients with an mTBI decrease to less than 0.02% at the first tumor evaluation had longer PFS and OS (p < 0.001 and p = 0.007, respectively). The patients classified as molecular responders had longer PFS and OS than the nonmolecular responders (p < 0.001 and p = 0.036, respectively) ^[73][154]. Epigenetic changes were also studied in metastatic BC as a marker for predicting survival and disease outcome. Methylation of a number of genes separately (ESR1, SOX17) or as part of gene panels was significantly correlated to shorter OS and/or no pharmacotherapy response ^{[14][74][75][76]}[128,147,148,149]. Changes in ctDNA levels correlate with tumor volume, making ctDNA an excellent non-invasive tool for monitoring response to treatment and determining the prognosis. These circumstances allow for this method to be considered as a convenient way of obtaining diagnostic material not only at the stage of primary diagnosis, but also during the treatment process. ctDNA analysis can provide a wide range of information about the processes in tumors; in particular, it can provide a rapid diagnosis of disease recurrence in previously treated patients, and can identify existing and newly emerged gene mutations that are targets of anticancer therapy (PTEN, PIK3CA, ESR1, AKT, HER2), which can help in selecting the best therapy for each patient. Meta-analysis has confirmed the role of ctDNA as a prognostic marker for patients with BC ^[77][177]. ctDNA analysis also provides information on clonal evolution and tumor heterogeneity and can be a useful tool in studying the mechanisms of resistance to therapy in metastatic BC ^[45][172].