Liquid Biopsies: Applications for Cancer Diagnosis and Monitoring

The minimally—or non-invasive detection of circulating tumor-derived components in biofluids, such as blood, liquid biopsy is a revolutionary approach with significant potential for the management of cancer. Genomic and transcriptomic alterations can be accurately detected through liquid biopsies, which provide a more comprehensive characterization of the heterogeneous tumor profile than tissue biopsies alone. Liquid biopsies could assist diagnosis, prognosis, and treatment selection, and hold great potential to complement current surveilling strategies to monitor disease evolution and treatment response in real-time. In particular, these are able to detect minimal residual disease, to predict progression, and to identify mechanisms of resistance, allowing to re-orient treatment strategies in a timelier manner. In this review we gathered current knowledge regarding the role and potential of liquid biopsies for the diagnosis and follow-up of cancer patients. The presented findings emphasize the strengths of liquid biopsies, revealing their chance of improving the diagnosis and monitoring of several tumor types in the near future. However, despite growing evidence supporting their value as a management tool in oncology, some limitations still need to be overcome for their implementation in the routine clinical setting.


Introduction
As the cancer burden keeps on growing globally, continuous efforts are made in order to improve the diagnosis and the management of this disease. Early diagnosis remains one of the main challenges in cancer care [1]. Thus, the development of screening and early detection tests, along with the establishment of efficient monitoring methods, are of great importance to improve the efficacy of therapies and to reduce cancer mortality [2]. In this sense, precision medicine has gained particular attention in the oncology field [3]. Molecular profiling can be applied to gain insight into the alterations underlying tumorigenesis [4], contributing to the identification of diagnostic and prognostic biomarkers, and to the selection of treatments considering individual variability.
duplicates, a total of 539 studies were screened by their title and abstract, and 123 remained. The full text of these studies was examined to determine their eligibility, and 52 articles were excluded. Finally, a total of 71 studies were included in this review.

Discussion
The review of the selected articles revealed that liquid biopsies are a topic of growing interest in cancer research, given their particularly enticing characteristics and large number of potential clinical applications. This type of biopsies hold promise to improve several aspects of cancer management, including tumor profiling, with implications for diagnosis and treatment selection, prognosis, and long-term monitoring, paving the way towards precision oncology and better patient outcomes.
The main findings of the retrieved studies were summarized in Table 1.

Discussion
The review of the selected articles revealed that liquid biopsies are a topic of growing interest in cancer research, given their particularly enticing characteristics and large number of potential clinical applications. This type of biopsies hold promise to improve several aspects of cancer management, including tumor profiling, with implications for diagnosis and treatment selection, prognosis, and long-term monitoring, paving the way towards precision oncology and better patient outcomes.
The main findings of the retrieved studies were summarized in Table 1. The studied 24 plasma miRNA signature seems to have utility as a prognostic and monitoring tool in lung cancer.

Blood circulating tumor cells (CTCs) and plasma ctDNA
Longitudinal ctDNA analysis can be used to monitor treatment response and to identify mechanisms of resistance in melanoma patients.
CTC-derived xenografts can be used as a complement to improve personalized treatment selection.
Ogle, L.F. et al., 2016 [67] 69 hepatocellular carcinoma (HCC) patients Blood CTCs Multiparametric analysis, using flow cytometry, size, morphology and the investigation of DNA improved the detection of CTCs, which have predictive potential in HCC.
Wang, X. et al., 20167 [28] 200 non-small cell lung cancer (NSCLC) patients Urine ctDNA The KRAS mutational profile is highly concordant in urine and in corresponding tumor tissues. The longitudinal monitoring of mutations in transrenal DNA is helpful to stratify NSCLC patients according to predicted outcomes. It is possible to detect programmed death ligand 1 (PD-L1) expression in CTCs of NSCLC patients. CTCs were more frequently PD-L1(+) than tumor tissues and PD-L1(+) CTCs were found in all patients at progression. Methylation and copy number analysis of urinary cfDNA allows to detect bladder cancer, which could be valuable for diagnosis and monitoring of tumor burden. ctDNA is useful to identify therapeutic targets, to monitor therapy and to find mechanisms of resistance in NSCLC.

Liquid Biopsies for Diagnosis and Tumor Profiling
Currently, cell-free DNA (cfDNA) is one of the most studied analytes in liquid biopsies. Both the quantity [17,21,22,26,39,47,52] and the integrity [21,22,26] of cfDNA in circulation have shown the ability to distinguish cancer patients from healthy individuals. The total levels of cfDNA tend to be higher in cancer patients than in healthy subjects [17,21,22,26,39,47,52], and seem to increase with stage [17] and metastasis [55]. The increased concentration of cfDNA in these patients is thought to reflect the additional release of genetic material from tumor cells, but it could also be a result of the defective clearance of circulating DNA by phagocites [79]. However, high cfDNA levels are not specific of cancer and have been identified in other pathological and non-pathological conditions, including exercise, trauma and surgery [22], which might hamper their direct application for cancer diagnosis. Regarding integrity, cancer patients seem to display higher cfDNA fragmentation (<100 bp) than healthy controls [22,26], although a study in thyroid carcinoma revealed the contrary [21], once more disclosing that this type of analyses still lack the sensitivity and specificity needed for diagnosis.
Nevertheless, through the analysis of tumor-specific alterations, including single nucleotide variants (SNVs), insertions, deletions, copy number variations (CNVs) [40], and methylation alterations [16,57,59], one can identify tumor-derived DNA-ctDNA, among the total pool of cfDNA, providing a much more accurate form of cancer genotyping and, consequently, of diagnosis. Importantly, these (epi)genetic alterations seem to be highly concordant in blood ctDNA and in corresponding tumor tissues in a variety of cancers, including lung [17,29,42], breast [35], colorectal [17,44,50,52], pancreatic [32], liver [57], esophageal [17], gastric [6,43], and ovarian [17] cancers. In this sense, these minimallyinvasive and less risky liquid biopsies [57] could be used as an alternative to tissue biopsies in cases in which the latter cannot be performed [24,39,49] or when these do not gather enough high-quality DNA [17,39]. In fact, since 2016, non-small cell lung cancer (NSCLC) patients who are unable to provide tumor specimens can be tested for EGFR mutations in plasma, using the U.S. Food and Drug Administration (FDA)-approved cobas EGFR Mutation Test v2 [80]. Since then, several other liquid biopsy-based tests have been approved by this agency: plasma samples are now being used to detect specific gene mutations and rearrangements in patients with ovarian, lung, breast, and metastatic castration resistance prostate cancers. These tests are mainly used as companion diagnostic tests to identify patients who are eligible for targeted treatments [81].
However, the concordance of alterations found in ctDNA and tumor tissues varies not only according to the type of cancer [19], since different tumors have different probabilities of shedding DNA into the bloodstream [20], but also depending on the stage of the disease [17,36]. In fact, cfDNA levels tend to be lower in earlier stages [17,18], suggesting a limited role of liquid biopsies for the early detection of cancer [17,48], although increased technological sensitivity might help overcome these problems [17]. Even though the use of liquid biopsies for cancer screening is still largely undeveloped, a liquid biopsy test based on the detection of abnormal methylation of the SEPTIN9 gene in blood has already been approved to screen for colorectal cancer (CRC) [82]. In addition, other blood tests that use ctDNA to screen for several cancer types, including ovarian, liver, stomach, pancreatic and esophageal, which, to date, do not have screening tests available, are currently being studied [83].
In addition to accurately representing the tumor profile, ctDNA also captures tumor heterogeneity [40,44,53]. This is particularly relevant in metastatic disease, in which multiple tissue biopsies might not be viable and are associated with increased cost and risks to the patients [40]. Correctly profiling all tumor sites is especially important for the detection of actionable mutations and, as mentioned above, for the selection of patients that might benefit from targeted therapies [19,44]. Still, ctDNA analysis might be less sensitive to detect some alterations that are found in tumor tissues [19,41], since these are diluted in a background of germline DNA [12,20] and some might be present in small subsets of cells and, thus, exist in lower quantities in circulation [6,60]. Therefore, currently, liquid biopsies seem to have a more complementary rather than alternative role to tissue biopsies for diagnostic and profiling purposes.
Besides blood, other biofluids, such as urine [7,23,28,37,39,47,52], cerebrospinal fluid (CSF) [60], and gastric washes [54], have been shown to harbor ctDNA. Depending on the type of cancer, tumors might more closely contact with different fluids that, consequently, might contain higher ctDNA concentrations than blood [23]. For example, urinary ctDNA has been detected in bladder [23,37,58] and other urothelial cancers [23]. Moreover, transrenal DNA, resulting from the clearance of bloodstream cfDNA by the kidneys, has also been detected in non-urological tumors, such as NSCLC [28] and CRC [39,52]. Notably, urinary ctDNA also presents cancer-specific mutations [23], CNVs and methylation al-terations [58] that are highly concordant with the ones found in tumor tissues [39,47,52]. Given this, the use of urine for liquid biopsies is particularly promising considering that its collection is totally non-invasive, resulting in improved patient compliance for serial sampling for diagnosis or follow-up [39]. Nevertheless, although there seems to be a positive correlation between plasma and urine cfDNA levels [37,52], mutation detection in urine is usually limited in comparison to blood [39]. Regarding CSF, this is a specially relevant source of information in brain tumors, such as gliomas [19,60], as an alternative to surgical tissue biopsies. Importantly, CSF-derived ctDNA has also been shown to display tumor-concordant mutations, CNVs, and structural rearrangements [60].
In addition to ctDNA, other components with diagnostic potential can be investigated in liquid biopsies, such as messenger RNA (mRNA) [4,62] and micro-RNAs (miRNAs) [61,64]. For instance, Malczewska et al. evaluated gene expression in the blood of patients with bronchopulmonary carcinoid (BPC) tumors and revealed that the levels of target transcripts were significantly increased in comparison to healthy controls and enabled to distinguish metastatic and localized disease. Importantly, gene expression was highly correlated in tumor tissue and blood [4]. Similarly, the levels of several miRNAs are often altered in cancer patients, allowing to determine miRNA signatures with diagnostic and prognostic potential [61,65]. Tumors and their microenvironment release miRNAs that exist in the bloodstream in ribonucleoprotein complexes or incorporated into EVs [65]. In particular, circulating miRNA profiles seem to be concordant with the ones of tumor tissues [64]. However, EV-incorporated miRNAs seem to represent only a small fraction of the miRNAs present in circulation and to have distinct diagnostic performance [61].
The DNA in exosomes (exoDNA) can also be a valuable source of information, given that it is less prone to degradation than cfDNA and released from living cells, thus probably reflecting tumor-driving alterations more accurately than ctDNA that is derived from apoptotic and necrotic cells [32]. Contrarily to exosomes, that are derived from intracellular compartments, tumor-associated microparticles (taMPs) are membrane-derived vesicles that display cell surface markers from their source. As so, the expression of the cancer markers epithelial cell adhesion molecule (EpCAM) and CD147 in taMPs was shown to be specific of cancer patients, which could potentially be used for diagnostic purposes. In addition, these double-positive taMPs significantly correlated with tumor burden in CRC [66].
Blood platelets are local and systemic responders that act during carcinogenesis and metastasis [78]. Platelets are able to directly ingest circulating mRNA and tumor-associated proteins released by tumor cells as well as undergo splice events in response to signals released by cancer cells and the tumor microenvironment [84]. Platelets exposed to tumorinduced education present an altered platelet behavior [77,78]. These platelets (TEPs) have been shown to be highly specific and are able to predict and facilitate early screening of NSCLC [77], predict and discriminate patients with primary glioblastoma from patients with brain metastases and with neuro-inflammatory conditions [78] and predict treatment outcome after therapy [76].
Lastly, the number of CTCs cells that detach from primary and metastatic tumor sites [5], seems to increase with stage and disseminated metastasis [15], revealing its potential as a diagnostic and prognostic marker. Although most CTC isolation systems, including the CellSearch platform, rely on the detection of cell surface markers, such as EpCAM and cytokeratins, important subpopulations of CTCs that do not express these markers or that have undergone epithelial-to-mesenchymal transition (EMT) might be missed using these methods [5,8,67]. In this sense, efforts are currently being made to integrate the isolation of such subsets of cells, for example, through multiparametric analysis [67] or detection of additional relevant (e.g., mesenchymal [8]) markers [5,15]. In addition to the identification and quantification of CTCs, their molecular characterization provides relevant information regarding mechanisms of CTC release and tumor biology [8]. Particularly, methylation status has been shown to be highly concordant in CTCs and in paired plasma ctDNA, suggesting a common source of this DNA [16]. Furthermore, CTCs' (epi)genetic profiles mirror the ones of the tumors [12,16] and could also be used for targeted treatment selection [11,15].
Besides having potential to aid in diagnosis and tumor profiling, liquid biopsies seem to have prognostic value. Firstly, studies have shown that ctDNA detection positively correlates with tumor size [6,18,20,32], informing about disease burden. Secondly, high ctDNA levels at baseline have been correlated with lower survival [17,24,32,45,46] and poorer outcomes [24,60]. Similarly, the number of CTCs also seems to increase with tumor size and to be a predictor of poor survival [9,67].
In addition to ctDNA, exoDNA has also been shown to have prognostic value, with KRAS mutant allele fraction (MAF) being associated with progression-free survival (PFS) and overall-survival (OS) in pancreatic cancer [32]. Further, tumor-derived platelets biomarkers, as KLK3, FOLH1, and NPY, enable prediction after abiraterone therapy in castration resistant prostate cancer (CRPC) and KLK2, KLK3, and FOLH1 were associated with short OS [67].

Liquid Biopsies for the Follow-up of Cancer Patients
A particularly promising application of liquid biopsies concerns its use to monitor disease evolution and treatment response in cancer patients. In fact, the non-or minimallyinvasive nature of these biopsies make them a favorable alternative for long-term followup [52]. Contrarily, multiple invasive tissue biopsies are not always feasible and, more importantly, these often miss alterations found in sites other than the primary tumor that might influence therapy response and efficacy [46]. In addition, imaging techniques that are often used to monitor cancer patients undergoing treatment present many limitations. For example, computed tomography (CT) scans expose the patients to radiation, are costly [48], have low sensitivity to detect small lesions [3,48] and do not provide information regarding genetic changes induced by treatment [31].
On the other hand, the longitudinal collection of biofluids allows for the monitoring of the disease, assessment of treatment response, and identification of mechanisms of resistance [48,53]. Importantly, given that cfDNA has a short half-life [85][86][87], liquid biopsies allow for the follow-up of cancer patients in real-time. For instance, ctDNA analysis has been successfully used to detect resistant mutations in genes such as EGFR [29,47], ERBB2 [43], PIK3CA, and RAS [28,50], in various cancers including NSCLC [28,29,47], CRC [50], and gastric cancer [43]. Moreover, monitoring the dynamics of ctDNA alterations is useful to predict response to treatment and clinical outcome [3,6,29,37,39,41,42,45,46,49], and allows to reorient treatment regimens in a more timely manner [25,59]. As a matter of fact, a ctDNA decrease after treatment has been associated with lower risk of progression [3,41,42] and longer survival [42], while persistent or increased ctDNA levels have been associated with progression [3,28,37,42,45,49,55], relapse [33,35,37], and decreased survival [35]. In particular, a promising study has shown that it might be possible to quantitatively predict the time needed for disease progression by combining longitudinal profiling of cfDNA and mathematical modeling [44].
In addition to ctDNA, both the count and characterization of CTCs have been shown to have predictive value [9] and to be useful to monitor therapy response and the dynamic progression of tumors [88]. Given that these cells are closely involved in the metastatic process, monitoring them might help predict the patients' post-treatment outcomes and risk of metastasis [88,89]. A decline in the number of CTCs after treatment could be related with a better prognosis [89,90]. As a matter of fact, some studies have shown that the number of CTCs tends to decrease in response to therapy [15,90], whereas a postoperative increase in CTC number and in the percentage of mesenchymal CTCs has been shown to precede recurrence [8]. However, a better understanding of the patterns of CTC shedding and clearance is imperative for a correct interpretation of CTC fluctuations [91].
Additionally, similarly to ctDNA, CTCs also seem to predict therapy response more efficiently than imaging techniques. In fact, a study using a gene expression panel to detect CTCs was able to identify metastatic CRC (mCRC) patients that were not responding to therapy, which was unnoticed by CT scans. Besides demonstrating a higher sensitivity than CT, CTC analysis also identified non-responders presenting lower PFS and OS earlier than this imaging technique [9].
Furthermore, the characterization of CTCs during treatment also allows to detect genetic alterations or gene expression changes related to resistance [5,12]. Remarkably, a study has revealed that the detection of the EGFR T790M resistant mutation in patients undergoing treatment with tyrosine kinase inhibitors (TKIs) occurred earlier in CTC than in ctDNA analysis [12]. In addition, as proposed by Girotti et al., xenografts established from patients' CTCs could be used to select second-line personalized therapeutic options [75].
Transcriptomic data with monitoring potential can also be investigated in EVs or in circulating RNA. Interestingly, EVs seem to present different mRNA profiles than CTCs and, although the expression of various transcripts might be associated with progression or response to therapy, depending on the source, EVs or CTCs, the same transcript might be associated with different outcomes, revealing the need of exploring which is the most relevant monitoring marker [10]. In addition, Malczewska et al. have shown that blood mRNA levels allowed for the identification of patients with residual disease and for the differentiation of progressive from stable disease in lung neuroendocrine tumors [4]. Assessing the levels of circulating miRNAs might also contribute to monitor disease, as revealed by the decrease of certain miRNAs after management of prostate cancer (PCa) [63] or hepatocellular carcinoma (HCC) [64]. Importantly, significantly higher levels of miRNA-224 were detected in HCC patients with residual tumors, whereas conventional markers and magnetic resonance imaging (MRI) could not distinguish residual disease [64]. Moreover, a study in lung cancer demonstrated that a circulating miRNA signature was able to stratify patients according to their risk of relapse, which adds prognostic value and eventually allows to identify patients that might benefit from adjuvant therapies, for example [65].
The levels of the previously mentioned EpCAM(+) taMPs might also serve as a monitoring tool, as revealed by their decrease after surgical removal of CRC [66]. Another way of assessing successful surgical resection, relying on the evaluation of postoperative Apo10-and transketolase like 1 (TKTL1)-expressing monocytes, has been tested in oral squamous cell carcinoma (OSCC) [74].
Lastly, liquid biopsies have also been used to detect and monitor circulating cellfree Epstein-Barr virus (cfEBV) DNA, for example, in EBV-associated gastric carcinoma (EBVaGC) [51] and nasopharyngeal carcinoma [38]. Monitoring the dynamics of cfEBV DNA in EBVaGC revealed that the levels decrease significantly after treatment, but increased in a patient that developed recurrence [51]. Distinct patterns of cfEBV DNA clearance associated to treatment sensitivity have also been identified in nasopharyngeal carcinoma, demonstrating its prognostic value and usefulness for treatment adaptation. Similarly to what has been depicted in ctDNA studies, the persistence of cfEBV DNA after treatment was associated with worse survival. In addition, the levels of cfEBV at baseline were also associated with the likelihood of treatment response or resistance [38].
Overall, this data reveals that liquid biopsies are useful and advantageous to monitor cancer patients, providing early information regarding disease evolution and treatment efficacy, and allowing to reorient treatment strategies in time. In particular, several studies indicate that the use of liquid biopsies as a complement to current surveilling strategies, such as tissue biopsies and imaging, could substantialy improve the detection of resistance, residual disease and relapse.
Recently, the use of liquid biopsies to predict and to assess response to cancer immunotherapy has emerged as a new area of investigation, which is explored in more detail below.

Liquid Biopsies in the Immuno-Oncology Field
Liquid biopsies have gained particular attention in the immuno-oncology field lately, with several ongoing clinical trials and studies focusing on this subject. The use of programmed death 1 (PD-1)/programmed death ligand 1 (PD-L1) or cytotoxic T-lymphocyteassociated antigen 4 (CTLA-4) immune checkpoint inhibitors (ICIs) has become a standard in the treatment of various cancers, including NSCLC [13,14,56,73], with significant clinical benefits.
Currently, one of the most recognized predictors of response to PD-1/PDL-1 ICIs is PD-L1 expression in tumor tissues [73]. However, not all patients respond to this type of treatment, including some patients with PD-L1(+) tumors. In addition, a subset of PD-L1(−) patients seem to benefit from ICI treatment. This could be, at least in part, due to the heterogeneous PD-L1 expression that is not always captured by tissue biopsies [13]. Thus, new biomarkers to predict and monitor ICI treatment response and outcomes are needed [56]. In this regard, liquid biopsies could be a more appropriate way to identify and track these markers. Given this, studies have investigated the prognostic potential of PD-L1 expression in CTCs [13] and in populations of nucleated blood cells [73]. A study in NSCLC has shown that CTCs were most frequently PD-L1(+) than tumor tissues, which could be a result of tumor heterogeneity or insufficient tissue sampling. A higher number of CTCs at baseline was associated with worse outcomes in patients treated with ICIs, probably on account of underlying extended disease. In addition, although PD-L1(+) CTCs did not correlate with outcomes, all patients with progressive disease presented PD-L1(+) cells and these were observed in higher number at baseline in non-responders, suggesting a possible mechanism of resistance involving this cell subpopulation [13]. Another study has revealed that a higher quantity of PD-L1-expressing peripheral circulating cells was associated with worse survival [73]. Overall, these studies suggest that the detection and monitoring of PD-L1(+) cells might be useful for prognosis and early assessment of ICI treatment response.
Besides PD-L1 expression, serial monitoring of CTC-derived transcripts has also been shown to be predictive of clinical outcome in melanoma patients undergoing ICI treatment [14]. In addition, the evaluation of dynamic CNVs in cfDNA could serve as an early indicator of immunotherapy response or progression [56].
Further studies should evaluate the relationship between PD-L1 expressing CTCs and efficacy of ICI therapies and explore the potential of other liquid biomarkers to predict response and to identify patients eligible for immunotherapy.

Advantages and Disadvantages of Liquid Biopsies
The described evidence revealed that liquid biopsies provide a minimally-invasive way of representing the heterogenous tumor profile [19] at baseline and/or during followup [12,19,39], improving patient care without the limitations and risks associated with tissue biopsies. The main clinical applications and limitations of each analyte found in liquid biopsies are summarized in Table 2. Rare in circulation [5,12,16] Difficult and costly isolation [5,12,16,30] Circulating RNAs Techniques in early development [42,78] The use of liquid biopsies for follow-up of cancer patients is one of its most investigated applications and seems to be the one that is closer to a broad implementation in the clinical setting. On the other hand, the use of liquid biopsies for screening and early cancer detection remains challenging [40], mainly because analytes, such as CTCs [2] and ctDNA [17,48], are often below optimal levels for analysis, particularly in early-stage patients. A potential solution to improve the sensitivity of cancer detection would be to evaluate multiple circulating analytes (e.g., ctDNA, RNA, proteins [94], and metabolites [95]), however, the specificity and cost-effectiveness of such tests would probably be compromised [2]. In fact, the low specificity of liquid biopsy tests is a concern, given that their use for cancer screening would result in the detection of high numbers of false positives, making healthy individuals undergo unnecessary invasive procedures and distress. For instance, alterations resulting from clonal hematopoiesis in peripheral blood cells can also be found in cfDNA of individuals that do not have cancer [23,53], which, if not carefully analysed, may lead to less correct diagnosis. Despite these challenges, the non-invasive screening and early detection of cancer is still one of the most attractive and awaited applications of liquid biopsies, since it could substantially improve treatment efficacy and patient survival [2,96], especially for cancer types lacking screening tools and that are often diagnosed in advanced stages [96]. Hopefully, the technological advances and growing interest in this topic will result in more evidence regarding the value of liquid biopsies as a tool for cancer screening and diagnosis in the coming years.
Ensuring that cfDNA samples are of sufficient quantity and quality is crucial for the success of downstream applications. For that purpose, contamination of samples with genomic DNA must be prevented, for example, using white blood cell stabilizers. The isolation of cfDNA from plasma samples instead of serum is also preferable, given that it avoids the release of cellular DNA from lysing cells during the clotting process. In addition, due to low concentrations, extraction methods must guarantee high yields of cfDNA [92]. The low quantities and small fraction of ctDNA in circulation require the use of highly sensitive detection techniques, such as droplet digital Polymerase Chain Reaction (ddPCR), Next Generation Sequencing (NGS) [41] or BEAMing (beads, emulsion, amplification and magnetics) [12]. However, due to temporal heterogeneity, targeted sequencing limits treatment response monitoring and detection of resistance mutations [36], evidencing the need to use broader panels to analyze ctDNA during follow-up [41], which might compromise sensitivity of detection [44]. ctDNA profiling is still complex and expensive to apply in clinical routine [30,40]. Additionally, the isolation of CTCs, which are extremely rare in circulation, is also difficult [5,12,16] and costly [30]. Instead, the quantification of total cfDNA could be used as a simple and cheap alternative to predict disease response and outcomes, as revealed by a study in metastatic breast cancer (mBC) [30], although its diagnostic value is limited, as explained previously.
In conclusion, despite still presenting some limitations, liquid biopsies hold great potential to improve clinical care in oncology. In particular, this type of biopsy offers opportunities to improve the surveillance of cancer patients during treatment and might be helpful to supplement current diagnosis and tumor profiling strategies in the near future.

Data Availability Statement:
No new data were created or analyzed in this study. Data sharing is not applicable to this article.

Conflicts of Interest:
The authors declare no conflict of interest.