Keywords

1 Introduction

Histological examination of tumor tissues is still regarded as the gold standard method for assessing cancer biomarkers. However, it entails invasive collection procedures, which are not only costly and time-consuming but also associated with a variety of potentially serious adverse effects [1]. It is therefore poorly suited for serial use in evaluating the evolution of tumors and their response to treatment. Biomarkers that can be detected and measured in blood samples offer a number of potential practical advantages in this setting, as well as potential shortcomings involving their diagnostic/prognostic performance.

The principal serum biomarkers used to detect and monitor thyroid cancer are thyroglobulin and calcitonin [2, 3]. Thyroglobulin, the protein precursor of thyroid hormone, is produced exclusively by follicular thyrocytes (normal or neoplastic). Assays of its levels in the serum are of little use in the preoperative work-up of patients with thyroid nodules since elevations can be associated with most types of thyroid disorders [4]. However, in patients with differentiated thyroid cancer (DTC) who have undergone total thyroidectomy and radioiodine remnant ablation, detectable thyroglobulin levels in the serum have long been considered a sensitive and specific indicator of persistent or recurrent cancer [5]. In terms of sensitivity, the major shortcoming of this approach is the possible interference by circulating autoantibodies directed against thyroglobulin itself, which are present in approximately 20–25% of patients with DTC [6]. Specificity issues are becoming more and more relevant as the management of DTC evolves to meet the challenges of the rising incidence and earlier diagnosis of these tumors. The specificity of serum thyroglobulin assays declines substantially in the presence of residual normal thyroid tissue [7,8,9,10], and this situation is being encountered with increasing frequency as treatment choices shift toward the use of less extensive surgery, more selective use of radioactive iodine remnant ablation, and in some cases even active surveillance alone [5, 11,12,13,14,15]. The use of serum calcitonin levels as a marker of MTC recurrence is also limited by problems of specificity. This polypeptide hormone is produced mainly (but not exclusively) by the C cells of the thyroid gland, and immunoassays of calcitonin levels in the serum thus provide information on the activity of the cells that give rise to MTC and to the related but benign condition known as C-cell hyperplasia [16]. However, hypercalcitoninemia can also be caused by a number of other diseases, including chronic autoimmune thyroiditis and several non-thyroid-related conditions (e.g., neuroendocrine tumors, hypergastrinemia, hypercalcemia, chronic kidney disease) [16]. To further complicate matters, serum calcitonin levels are influenced by a number of analytical factors, drug use, and physiological variables, such as age, sex, diet, and lifestyle. As a result, reliable, standardized cutoffs for interpreting serum calcitonin assays are lacking [16].

2 Liquid Biopsy: Concept and Potentialities

One of the solutions being developed to meet clinicians’ increasing need for cancer biomarker assays that combine ease of use and low cost with high-level diagnostic accuracy is the concept known as “liquid biopsy.” The term refers to all diagnostic procedures performed on cancer-derived materials that can be isolated from a peripheral blood sample rather than from a tissue biopsy or resected tumor [17]. This new and powerful approach for the study of cancer has been made possible by recent advances in technologies, such as next-generation sequencing (NGS) and digital PCR [18]. They allow detection and quantitative analysis of materials in the bloodstream (including those present at very low levels), which originate specifically from solid tumors [18].

Aside from its practical benefits, the liquid biopsy approach offers several other potential advantages over tissue-based analyses. For example, compared with tissue biopsy or aspirates for cytological analysis, which depict very restricted areas of the tumor, examination of various tumor derivatives that find their way into the bloodstream are likely to furnish a more comprehensive picture of intra-tumoral heterogeneity. Indeed, used serially, liquid biopsies can potentially provide ongoing documentation of the evolving genetic panorama of the entire tumoral landscape, including metastatic lesions [19]. This knowledge may be prognostically informative, as shown in studies of breast cancer and other types of cancer, where certain genetic alterations display correlation with disease stage, vascular invasion, lymph node metastasis, as well as overall and disease-free survival [1]. It might also improve decisions regarding appropriate targeted therapy and other aspects of clinical management [20].

A growing body of evidence suggests that liquid biopsies are potentially valuable tools for identifying early-stage malignancy and for monitoring tumor burden, including minimal residual disease [18]. Use of this approach could conceivably eliminate some of the frequent drawbacks of imaging studies, which are currently used for this purpose, including high cost, exposure to ionizing radiation, and limited sensitivity for the detection of micrometastases and/or for monitoring minimal residual disease [1]. Uninformative imaging findings are often bolstered with supplementary information on circulating levels of protein biomarkers, such as prostate-specific antigen (PSA), carcinoembryonic antigen (CEA), and cancer antigens (CA) 19-9 and 125. Here, too, liquid biopsy offers potential advantages, particularly when reliable, disease-specific protein markers for the patient’s cancer are simply not available. Moreover, protein biomarkers persist in the circulation for weeks, whereas the turnover of circulating cancer cells or circulating free DNA is much more rapid. As a result, liquid biopsy should be able to detect tumor changes long before they are revealed by imaging findings or protein biomarker studies [21].

The most important areas in which liquid biopsies are expected to make major contributions are the prediction and monitoring of responses to treatment. The presence of a single genetic alteration in the tumor can decisively influence the selection of targeted drug therapies, particularly for advanced-stage cancers. Today, these decisions are made on the basis of genetic analysis of archived tumor tissues collected weeks or months earlier, which are probably poorly representative of the current genetic status of the disease. In contrast, a liquid biopsy can provide a “real-time” picture of the current molecular status of the tumor [1, 22]. Moreover, recent studies have shown that serial biopsies collected during the course of treatment can facilitate earlier detection of multiple drug-resistant clones [21], allowing prompt discontinuation of expensive, potentially toxic drug therapy that is unlikely to be effective and rapid initiation of more suitable treatment [1].

The expected benefits of liquid biopsy appear to be supported by growing bodies of evidence in several types of cancers. However, before this approach can be used in the clinic, numerous hurdles must be overcome, including the lack of standardized techniques, and the preliminary findings must be substantially strengthened by the addition of data from larger studies and prospective trials [1, 18, 22].

3 The Cancer-Derived Materials Found in the Bloodstream

As noted above, whole tumor cells as well as cell-free nucleic acids (cfDNA, cfRNA, and circulating miRNAs) can all be found within the peripheral blood [18]. Whole tumor cells can enter the bloodstream after breaking free from primary or metastatic tumors. It is unclear whether this represents an active invasion of the vascular tree or is simply the result of passive shedding of tumor cells [23]. One of the most plausible hypotheses is that individual CTCs or clumps of CTCs detach from the tumor and penetrate the bloodstream via an active process that probably involves the epithelial-to-mesenchymal transition (EMT) [23]. Cell-free nucleic acids (DNA and RNA) can be found in the bloodstream as freely circulating (cfNAs) species or encapsulated in extracellular vesicles (EVs). EVs are membranous lipid structures produced by healthy and non-healthy cells for specific purposes, such as intercellular communication and immunoregulation [24]. The mechanisms by which cfNAs are introduced into the blood vessels are unclear. They probably include passive release during apoptotic or necrotic events occurring in the tumor microenvironment, as well as the active secretion of cfNA fragments, alone or incorporated into protein or lipid complexes [1, 24,25,26]. It is also conceivable that cfNAs can be released by CTCs after they enter the bloodstream, although conclusive evidence of this mechanism has yet to be presented [1].

The following section provides a closer look at circulating RNAs as a potential class of biomarkers for thyroid cancer.

4 Circulating RNAs

Messenger RNAs (mRNAs), ribosomal RNAs (rRNAs), transfer RNAs (tRNAs), small nuclear RNAs (snRNAs), microRNAs (miRNAs), and long noncoding RNAs (lncRNAs) can all be found in the human bloodstream. Most of the research conducted thus far on circulating RNAs has focused on the analysis of mRNAs and more recently miRNAs. The presence and stability of RNAs in body fluids was surprising, given the high levels of RNase in the extracellular environment. The circulating RNAs were found to be protected from these enzymes by their incorporation in lipoprotein complexes or sequestration within lipid vesicles [27]. Notably, miRNAs were found to be the most abundant RNA species in lipid vesicles [28]. As compared to miRNAs, mRNAs were reported to be more susceptible to degradation [29]. For this reason, the past decade has witnessed a shift in the focus of research toward the study of circulating miRNAs, which are tissue-specific and highly stable, even after exposure to high temperatures, low or high pH, prolonged storage at room temperature, and multiple freeze-thaw cycles [26, 30].

4.1 Isolation and Detection

Research on circulating mRNAs and miRNAs as potential biomarkers has produced highly variable results, clearly reflecting the absence of methodological standardization in this field [31]. This variability can be caused by multiple factors, including sample-related factors (i.e., patient’s sex, sample collection time, diet, exercise), pre-analytical factors (i.e., sample type, processing, and/or storage conditions), and experiment-related factors (i.e., RNA isolation protocol, quantification methods) (Table 6.1).

Table 6.1 Summary of factors that affect reproducibility of circulating RNA analysis
Full size table

Different biological fluids can be expected to have different circulating RNA profiles, and yet assays continue to be performed on whole peripheral blood samples, as well as their serum, plasma, and mononuclear cell fractions. In more recent studies, serum and plasma samples tend to be preferred since they eliminate the interference caused by the presence of normal blood cells. Cell-free mRNA in plasma or serum undergoes rapid degradation and is thus difficult to identify, even with highly sensitive techniques [29]. For this reason, circulating mRNA is almost always isolated from the mononuclear cell layer of peripheral blood, and transcripts for thyroid-specific genes are assumed to be derived from circulating tumor cells. Rare attempts have also made to extract mRNA from whole-blood samples for this purpose.

Plasma and serum generally have similar miRNA expression patterns [30], but in certain instances, significant differences between these biological fluids are evident [32]. Indiscriminate use of plasma and serum specimens within a given study is not recommended, and comparing results from studies in which different sample collection protocols were used is also unadvisable. If plasma is used, care should be given to the choice of anticoagulants and EDTA, since citrate and heparin can inhibit the activities of reverse transcriptase and DNA/RNA polymerases [33], which are commonly used in circulating RNA detection methods (qPCR, NGS, microarray).

Other factors affecting circulating RNA analysis are experiment-related: these include the method used for total RNA isolation as well as the choice of a measurement platform. Older RNA extraction methods use a phenol/chloroform-based technique that is often facilitated by the addition of guanidinium thiocyanate. The newer methods, which are faster and in some cases automated, use phenol/chloroform extraction with mini-columns, which allows the inclusion of multiple wash steps and produces RNA samples with higher purity [34]. Farina and coworkers tested two of the most widely used extraction kits, mirVana PARIS and miRNeasy Serum/Plasma Kits, and found the overall performance of the latter to be superior [35].

Several technologies are currently available for quantifying circulating RNAs, including microarrays, quantitative PCR (qPCR), NGS, and digital PCR (dPCR), and each has advantages and limitations (Table 6.1). Microarrays are the least sensitive and specific, whereas quantitative real-time PCR is probably the most popular. It can be used to quantify single RNAs or hundreds of RNAs; it is also relatively cheap, easy to carry out, and sensitive. As for miRNAs, several companies offer quantitative PCR-based assays for the detection of specific miRNAs, including some based on the stem-loop real-time PCR technique used for the relative quantification of low-abundance circulating miRNAs [36]. However, characterizing circulating miRNA expression using relative quantification is limited by the absence of a universally invariant endogenous control in body fluids [37]. This limitation can be overcome with innovative methodologies like NGS and digital PCR, which allow quantitative analysis of circulating miRNA expression without internal controls. NGS is ideal for profiling circulating RNAs, because it provides comprehensive, definitive information on low-abundance species, including sequence data, which can be used to distinguish different isoforms of mRNA and miRNA and to identify new miRNAs and changes related to RNA editing. It cannot, however, be used for absolute quantification of circulating RNAs. Digital PCR is currently the only technique that can directly quantify the absolute number of circulating RNA copies. For this reason, dPCR should be the method of choice for validation studies and for setting diagnostic and prognostic tests. It also eliminates the need for the standard curves and endogenous controls required for qPCR, and it is also superior to the latter in terms of sensitivity, precision, and susceptibility to PCR inhibitors [38, 39].

4.2 Progress Toward the Analysis of Circulating mRNAs in Thyroid Cancer

Circulating thyroid-specific mRNAs have been evaluated as potential diagnostic and prognostic biomarkers in thyroid cancer (see Table 6.2). Several studies have shown that assays of circulating TSHR and TG mRNA levels can improve the diagnosis of most (78–85%) thyroid nodules with indeterminate FNA [55, 56, 59, 60]. The findings of one study suggested that these markers could be particularly useful for distinguishing follicular adenomas from carcinomas, which is difficult with FNA cytology, although this conclusion needs to be confirmed in an extended cohort of patients [55]. Combined assessment of circulating TSHR mRNA levels and neck ultrasound findings in patients with cytologically indeterminate thyroid nodules resulted in more sensitive detection of all types of DTCs, but it also reduced specificity [56]. Rates of assay positivity for circulating TSHR mRNAs reportedly vary with the thyroid cancer histotype. Detectable levels are found mainly in PTC patients (principally in tall cell variants). Assay negativity is more common in patients with micro-PTCs, follicular variant of PTCs, FTC, and dedifferentiated or anaplastic thyroid cancers [57, 59].

Table 6.2 Circulating messenger RNA as biomarkers of thyroid cancer
Full size table

The possible utility of circulating thyroid-specific mRNA assays to detect thyroid cancer recurrence after total thyroidectomy has been explored as an alternative to serum thyroglobulin assays, which are unreliable in several situations, such as the presence of Tg autoantibodies. Historically, TG mRNA is the marker candidate that has been most widely studied for this purpose. Studies conducted to evaluate the value of circulating TG mRNA assays for detecting local or distant metastasis in DTC patients showed that the most reliable results were obtained only with primers carefully designed to detect all TG splice variants [60, 64]. The specificity of TG mRNA in this setting is more controversial, since ectopic TG expression was also detected in lymphocytes [64]. In one study, TG mRNA assay identified distant or local DTC recurrence with high sensitivity and specificity (100% and 94%, respectively). The results obtained with TG mRNA assay showed a high concordance with TSHR mRNA data [60]. Higher TG mRNA levels have also been found in metastatic vs. nonmetastatic DTC, with more pronounced increases in those with distal vs. proximal metastasis [63]. Other investigators found that circulating TG mRNA levels were of no value in predicting post-thyroidectomy recurrence of disease [65]. Moreover, it is not yet clear if it could represent a more sensitive marker than serum thyroglobulin assays [61, 63, 66, 67].

Higher presurgery TSHR mRNA levels appear to predict a higher risk for thyroid cancer recurrence [56, 59, 60], but they displayed no association with lymph node metastasis, multifocality, or extrathyroidal extension [56]. Assays of TSHR mRNA levels (beginning on the day after surgery) were a powerful posttreatment surveillance tool for identifying patients with residual/metastatic disease [56, 59]. Notably, TSHR mRNA assay positivity is reportedly associated with a higher responsiveness to radioactive iodine treatment [57].

Comparative analysis of the performance of circulating TG, NIS, TSHR, TPO, and PDS (Pendred syndrome) mRNA assays in the detection of residual or recurrent thyroid cancer found that, under thyroid hormone suppressive therapy, TPO and TSHR mRNA offered good specificity but low sensitivity. In contrast, circulating TG, NIS, and PDS mRNA assays were unaffected by TSH status (suppressed or stimulated) and displayed good sensitivity but limited specificity [61]. The high specificity in this setting of TPO mRNA levels was also confirmed in a more recent study, where assay positivity displayed correlation with early stage PTC [62].

4.3 Progress Toward the Analysis of Circulating miRNAs in Thyroid Cancer

A growing body of evidence points to circulating miRNAs as an ideal class of biomarkers for many cancer types, owing mainly to their tissue-specific expression patterns and impressive stability in biological fluids [30, 68,69,70]. However, their widespread use in clinical practice has been delayed for several reasons. First of all, it is becoming increasingly clear that circulating miRNA levels can be affected by a number of physiological factors and pathological conditions. Second, studies conducted on circulating miRNAs thus far are characterized by marked methodological diversity, and consequently the results vary widely and are often inconclusive.

As far as thyroid cancer is concerned, little or no attention has been given to the expression of circulating miRNAs and their clinical significance in less common forms, such as MTC, PDTC, and ATC. The majority of studies published thus far have been conducted on patients with PTC, and in this setting, circulating miRNAs show undeniable promise as novel diagnostic and prognostic biomarkers. However, as shown in Table 6.3, the protocols used and the results obtained from these studies are highly variable. Importantly, only few groups have validated their findings in a larger cohort of patients [71, 73, 74, 77, 80] or specified the isoforms of the miRNAs identified [74, 75, 77, 80, 81].

Table 6.3 Circulating microRNAs as biomarkers of thyroid cancer
Full size table

Circulating levels of miR-146b-5p, miR-221-3p, and miR-222-3p in PTC patients have been found to be higher than that in healthy controls [71, 72, 80], while miR-222 and miR-146b levels also reportedly discriminate between PTCs and benign nodules [71, 76, 79]. Plasma levels of miR-21 in FTC patients are reportedly higher than those found in patients with benign nodules or PTC, whereas miR-181a is more highly expressed in PTC patients than in those with FTC [78].

In PTC patients, circulating levels of miR-146b-5p, miR-221-3p, miR-222-3p, and miR-146a-5p have been shown to decline after tumor excision [71, 72, 78,79,80]. Notably, miR-221-3p and miR-146a-5p levels in PTC patients have been shown to predict clinical responses, with significantly increased levels observed at the 2-year follow-up in patients with structural evidence of the disease, including some in which serum thyroglobulin assays remained persistently negative [80].

The association of circulating miR-146b-5p, miR-221-3p, and miR-222-3p with the presence of thyroid cancer is strengthened by the evidence of their upregulated expression in PTC [82, 83], FTC [84, 85], and ATC [84] tissue. Upregulated tumor tissue expression of miR-21 has been found in these tumors as well as in MTC [86].

Conclusions

The liquid biopsy technique can potentially change clinical practice by providing us with more specific and reliable biomarkers for the early detection of cancer. The rapid, low-cost, noninvasive nature of sample collection makes this new technique ideal for ongoing real-time assessment of neoplastic disease, including changes in tumor burden, intra- and interlesional genetic heterogeneity, and response to therapy. Recent studies have shown that assays of circulating RNAs can be used for the early diagnosis of thyroid cancer and for monitoring treatment responses. Compared with circulating mRNAs, circulating miRNAs are emerging as the more promising candidates, due to their higher stability and tissue-specific origin. Realization of this enormous potential will depend largely on our ability to successfully address the outstanding issue of low reproducibility for the biomarkers that have been discovered. A major goal of future research in this field should therefore be the development of standardized methods for evaluating circulating miRNAs and validation of their performance in clinical settings.