Next-Generation Molecular Imaging of Thyroid Cancer

Simple Summary Molecular imaging utilizes radionuclides or artificially modified molecules to image particular targets or pathways which are important in the pathogenesis of a certain disease. Transporter-based probes like radioiodine and [18F]fluoro-D-glucose ([18F]FDG) are widely used for diagnosing thyroid cancer (TC) and predicting the prognosis thereafter. However, newly developed probes (peptide, antibody, nanoparticle probes, and aptamer) image the fine molecular changes involved in the pathogenesis of TC and enable target-specific diagnosis and treatment of TC. Furthermore, novel molecular probes have high specificity and sensitivity, imparting a high level of objectivity to the research areas of TC. Abstract An essential aspect of thyroid cancer (TC) management is personalized and precision medicine. Functional imaging of TC with radioiodine and [18F]FDG has been frequently used in disease evaluation for several decades now. Recently, advances in molecular imaging have led to the development of novel tracers based on aptamer, peptide, antibody, nanobody, antibody fragment, and nanoparticle platforms. The emerging targets—including HER2, CD54, SHP2, CD33, and more—are promising targets for clinical translation soon. The significance of these tracers may be realized by outlining the way they support the management of TC. The provided examples focus on where preclinical investigations can be translated. Furthermore, advances in the molecular imaging of TC may inspire the development of novel therapeutic or theranostic tracers. In this review, we summarize TC-targeting probes which include transporter-based and immuno-based imaging moieties. We summarize the most recent evidence in this field and outline how these emerging strategies may potentially optimize clinical practice.


Introduction
Thyroid cancer (TC) is one of the most common cancer types and its occurrence has been rapidly increasing over the last several years [1]. TC represents around 2-2.3% of new cancer cases and 0.2-0.4% of deaths from all cancer types [2,3]. In 2021, the USA may have approximately 44,280 new cases of TC and about 2200 deaths [2]. Around 90,000 new cases along with 6800 deaths were estimated in 2015 in China [2]. By 2030, TC is anticipated

Transporter-Targeting Probes
Most transporter-targeting probes are small-size molecules, carried into the intracellular space by transporters on the cell surface. Some transporter-associated probes may take part in cell metabolism [20,21]. Many transporter-based isotopes, including radioiodine, are routinely to image the recurrence and metastases of TC. Several alternatives to radioiodine can identify RR-DTC metastases lacking radioiodine uptake. Other [5,22]. In particular, [ 18 F]FDG has been widely applied in the management of TC [5].

Sodium Iodine Symporter (NIS)-Targeting Probes
NIS is the protein mainly locating at the cell plasma membrane, which carries Na + /Iions from the extracellular matrix into the intracellular fluid [23]. The transported iodine, as an element, helps produce thyroid hormone (iodide organification) [24]. Unlike the expression pattern of other tumor targets (low expression in normal tissues and high expression in tumor tissues), NIS is usually present at high levels in normal thyroid tissues and DTC cells, enabling radioiodine collection in normal thyroid and TC cells [25]. Nevertheless, NIS downregulation happens in RR-DTCs, poorly differentiated TCs (PDTCs), and ATCs, causing these TC cells hard to benefit much from radioiodine treatment [26].

Radioiodine
Radioiodine, a widely used radioisotope, has a crucial role in the diagnosis and treatment of DTC. There are several medically useful radioisotopes of iodine ( 125 I, 131 I, and 124 I, etc.). However, only 131 I and 124 I are commonly applied in clinical settings due to their clinically acceptable radiation half-life, diagnostic or therapeutic performance, economic cost, and safety [27]. [ 131 I]NaI can track thyroid and TC cells with γ radiation on SPECT, and damage those cells by emitting βradiation [28]. [ 131 I]NaI allows ablation of thyroid remnant, adjuvant therapy of TC, and therapy of TC, which vastly improves the prognosis of patients with TC [5]. 124 I is another isotope of iodine emitting positron, which can be exploited for PET imaging. [ 124 I]NaI-PET/CT has superior spatial resolution and quantification ability over [ 131 I]NaI-SPECT [29].
Recently, numerous reports have focused on radioiodine for improving the diagnosis performance, and efficacy of treatment [30][31][32]. Tg tests coupled with iodine uptake assay [32], or [ 124 I]NaI PET/CT only [30,31], are used for 131 I dosimetry. Apart from performing dosimetry before 131 I treatment, much attention should be given to increase the membranous expression of NIS, induce the concentration of 131 I, and improve the therapeutic efficacy of 131 I treatment. For RR-DTC, PDTC, and ATC, it is essential to explore agents that could increase NIS expression and augment the migration of NIS to the cell membrane. These agents mainly include but are not limited to, retinoic acid [33], mechanistic target of rapamycin kinase (mTOR) inhibitors [34], and very recently, V-Raf murine sarcoma viral oncogene homolog B (BRAF) and mitogen-activated protein kinase kinase (MAP2K1/2, MEK1/2) inhibitors, which inhibit the extracellular signal-regulated kinase (ERK) pathway responsible for tumor progression and radioiodine uptake [35,36] (Figure 1). RR-DTCs would be stabilized, or shrinkage after treatment with kinase inhibitors, owing to the suppressed signaling pathway and enhanced 131 I treatment efficacy [26,37].
Lately, estrogen-related receptor gamma (ERRγ), one of the estrogen-related receptors, has gained more traction as a potential target to enhance or enable radioiodine uptake. ERRγ, a member of NR3B nuclear receptor superfamily, is a biomarker for multiple cancers, including breast cancer and prostate cancer [38]. Previous reports have shown that the ERRγ inverse agonist GSK5182 increased NIS expression and NIS-mediated iodine uptake in Kirsten rat sarcoma viral oncogene homolog (KRAS) or BRAF mutated ATC cells in vitro [39]. In addition, another ERRγ inverse agonist, DN200434, was recently shown to increase the uptake of radioiodine in ATC tumors, identifying ERRγ as a target to enhance 131 I therapy responsiveness [40] (Figure 2). It remains to be determined if DN200434 has a re-differentiative effect in patients with either RR-DTC or ATC.   Detecting local recurrence and metastases of DTC in radioiodine imaging is particularly important for the arrangement of local treatments, e.g., surgery or radiotherapy [5]. Whereas negative radioiodine imaging with increased serum thyroglobulin is a barrier for finding malignant lesions, the so-called "Thyroglobulin Elevated and Negative Iodine Scintigraphy" (TENIS) needs radically diverse diagnostic and therapeutic methods [26]. TENIS could be caused by poor NIS expression, iodide organification defect, or radioiodine stunning [41,42]. A failure to find NIS-expressing DTC lesions might delay the diagnosis and also the timely onset of the treatment [43]. Despite [ 18 F]FDG-PET/CT could be applied for finding TENIS metastases, but its uptake might be partially caused by tumor-infiltrating immune cells [44] [45][46][47][48].
[ 18 F]TFB is an analog to radioiodine, having similar NIS affinities, same charge, and similar ionic radius to iodide. Therefore, [ 18 F]TFB can be transported by NIS [46,49] differed from radioiodine, [ 18 F]TFB can be readily synthesized at medical cyclotrons, and it provides a satisfactory half-life, dosing, biodistribution, and PET imaging quality [46,47] ( Figure 3). [ 18 F]TFB-PET can exclusively reveal NIS expression in tumor cells, therefore reclassifying TENIS metastases into partial or complete dedifferentiation, and helping metastasis localization and prognosis evaluation [47]. [ 18 F]FS and [ 18 F]HFP are two other newly discovered NIS-targeting tracers having favorable targeting efficiency, image contrasts, and biodistribution features [45,48]. Although it has been reported that FS and HFP had higher NIS affinity than TFB [45]
[ 18 F]FDG-PET/CT has shown great sensitivity in patients who otherwise do not benefit from 131 I treatment. This is because metastases without radioiodine uptake tend to have high glycolytic rates, causing enhanced [ 18 F]FDG uptake [54]. Furthermore, [ 18 F]FDG can help detect TC recurrence or metastases and predict radioiodine uptake [54,55]. More specifically, [ 18 F]FDG maximum standard unit value (SUVmax) higher than 4.0 would predict poor radioiodine uptake [54]. Besides, most PDTC, ATC, and MTC cells do not concentrate 131 [66][67][68]. Published data for [ 11 C]MET in TC is limited. Only one case with hyperparathyroidism has been reported, which showed intense focal [ 11 C]MET uptake in a cold nodule with highly increased sestamibi uptake. The nodule was finally diagnosed as FTC, indicating the incremental value of [ 11 C]MET in imaging DTCs [63]. [  F]FGln, an analog of natural glutamine regulated by several glutamine (Gln) transporters (solute carrier family 1 member 5, SLC1A5; solute carrier family 38 member 1, SLC38A1; and SLC7A5; etc.), has been tested and subsequently considered as a promising probe for assessing glutamine metabolism in tumors [61]. Its use is justified by the understanding that tumor cells need extra nutrition and energy for rapid growth and proliferation, while glutamine metabolism is occasionally used by the cell as an alternative to glucose [74]. [ 18 F]FGln can further complement the diagnostic capacity of [ 18 F]FDG by detecting Gln metabolic changes in PTCs [62]. In [ 18 F]FGln imaging, excellent contrast images can be made only 10 min after injection, while late-phase imaging (60 min) would cause a high background to some extent [62] (Figure 4).  [21,75]. However, data for those probes remain inadequate now. Furthermore, unlike other amino acid tracers transported by multiple unspecific amino acid transporters, [ 18 F]FAMT has an α-methyl moiety that allows it to be exclusively specific to SLC7A5, making it highly tumor-specific [76,77] [20]. Future studies are warranted to investigate the amino acid metabolism in TCs and the diagnostic value of amino acid tracers in large cohorts.

Nucleoside Transporter-Targeting Probes
Radiolabeled or fluorescent nucleobase analogs are currently used to diagnose solid tumors, including cancers of the bladder, breast, lung, ovary, and pancreas. Regarding diagnosis of TC specifically, only [ 18 F]fluorothymidine ([ 18 F]FLT) has been tested to date. [ 18 F]FLT, which can be taken up by equilibrative nucleoside transporter 1 (ENT1), is a marker of cell proliferation [78]. In one study, 20 [79]. So far, [ 18 F]FLT PET/CT has not progressed very far in diagnosing TCs.

Peptide-Based Probes
Peptide tracers have played vital roles in MI due to their unique advantages, notably their low molecular weight and ability to bind tumor biomarkers specifically, with low toxicity to surrounding non-cancer cells.

Somatostatin Receptor (SSTR)-Targeting Probes
Somatostatin receptors have become typical therapeutic targets in neuroendocrine tumors (NETs) because they are often overexpressed on the surface of tumor cells. This has led to the development of several 68 Ga-labelled somatostatin analogs as PET imaging probes [81], which could be used for the diagnosis of MTC [80]. 68 [82]. Nevertheless, studies reporting the diagnostic value of SSTR-targeted PET in recurrent MTC are limited. A meta-analysis involving nine studies reported that the tumor detection rate on SSTR-based PET or PET/CT is only 63.5% in recurrent MTC, which is lower than that in other NETs [83].

αvβ3 Integrin-Targeting Probes
The integrin αvβ3 expression on epithelial cells and mature endothelial cells is relatively low, however, it is commonly and highly expressed in solid tumors. RGD and RGD 2 are peptides that bind integrin αvβ3 [84]. Recently, the dimeric [ 68 Ga]Ga-DOTA-RGD 2 has been successfully applied for PET imaging of RR-DTCs in clinical settings [85], showing sensitivity, specificity, and accuracy of 82.3%, 100%, and 86.4%, respectively, which exceeds the same measurements in [ 18 F]FDG of 82.3%, 50%, and 75%, respectively. For RR-DTCs, the advantage provided by [ 68 Ga]Ga-DOTA-RGD2 is the ability to detect lesions not detected by [ 18 F]FDG [85] ( Figure 5). Furthermore, diagnosis of RR-DTCs using [ 68 Ga]Ga-DOTA-RGD 2 is better accompanied by [ 177 Lu]Lu-DOTA-RGD 2 , a potential treatment option for RR-DTCs [86]. Considering that [ 68 Ga]Ga-DOTA-RGD 2 and [ 177 Lu]Lu-DOTA-RGD 2 are a useful theranostic pair for RR-DTCs, the potential to improve the theranostic landscape of RR-DTCs by sequentially using these agents is high. Nuclear medicine approaches have revolutionized the theranostic arsenal for DTCs, and we are confident that there is room to optimize the management of RR-DTCs with these novel agents.

PSMA-Targeting Probes
PSMA is overexpressed on the prostate cancer cell membrane. Recently, several studies found unexpected PSMA-targeted radiotracer uptake by TCs, including RR-DTCs [87][88][89][90][91] ( Figure 6). In addition,~50% of TC microvessels showed high expression of PSMA related to tumor size and vascular invasion [89]. Thus, it is reasonable that high-grade TCs can be targeted by PSMA-specific radioligands like [ 177 Lu]Lu-PSMA and [ 225 Ac]Ac-PSMA [92], establishing a novel theranostic platform for TCs that are refractory to radioiodine treatment. Currently, the clinical interest and focus of PSMA-targeted theranostics remain primarily oriented towards prostate cancers. It is worth exploring the performance of PSMA-targeted agents in RR-DTCs. The authors wonder if PSMA-targeted agents will open a new horizon for RR-DTCs in the future.

Antibody-Based Probes
Antibodies are high-affinity molecules with strict targeting abilities that are used for highly specific binding [97]. The development and translational use of antibody therapeutics have shaped the model of molecular targeted therapy and immunotherapy. The high affinity of monoclonal antibodies for their targets promotes the rational and efficacious use of antibody therapeutics [98]. We have advocated that PET imaging with radiolabeled antibodies or antibody fragments (i.e., immunoPET) provides a powerful platform for visualizing the tumor targets, selecting suitable patients for targeted therapies or immunotherapies, and assessing the therapeutic responses thereafter [19]. The first-generation monoclonal antibodies (mAbs) were of murine origin, making them immunogenic, limited for their clinical use. Consequently, chimeric mAbs, humanized mAbs, and complete human mAbs were produced to solve this issue [98]. One limitation of the full-size antibody probes is their considerable size (~150 kDa), which leads to a long circulatory half-life and reduced tissue penetration [99]. To ameliorate the imaging quality and efficiency and accelerate clinical translation, some smaller molecule substitute probes have been investigated, including antigen-binding fragments (Fabs) and engineered Fab variants, single-chain variable fragments (scFv), diabodies, minibodies (~25-100 kDa), and other types of therapeutic proteins, such as affibodies and nanobodies [19]. Facilitated by these developments, multiple antibodies, and antibody derivatives have been designed as either imaging probes or therapeutic agents to induce cancer cell death and elicit host immune effector responses in TC [19].

Single Target Immunoglobulin G (IgG) Probes
Full-size IgG antibody probes have been applied to tumor detection, staging, guidance of local treatment, identification or validation of tumor targets, and assessment of therapeutic response or tumor prognosis [100]. Once the first-rank antigen has been selected, the corresponding IgG can be labeled with a radionuclide or fluorescent tag [19]. The radionuclide labeled IgG can be visualized via immunoPET imaging, and the fluorescent tags can be visualized through the fluorescence system during thyroidectomy or metastasectomy [101].
By labeling the HER2-targeting mAb pertuzumab with 89 Zr, we have developed the [ 89 Zr]Zr-DFO-pertuzumab and evaluated its diagnostic efficacy in subcutaneous and orthotopic ATC models [114] (Figure 8). ImmunoPET and fluorescence imaging indicated that radiolabeled or fluorescence-labeled HER2 probes are promising for the management of ATCs, which may become helpful tools for image-guided tumor removal or identifying HER2-positive ATCs for HER2-targeted therapies. However, clinical studies are needed for further translation. To facilitate clinical translation and broad clinical use, we have developed a series of novel nanobody-based tracers to delineate HER2 expression. We will test the performance of the tracers in TC models very soon.

Intercellular Adhesion Molecule-1 (ICAM-1, CD54)-Targeting Probes
ICAM-1, belonging to the immunoglobulin superfamily of cell adhesion molecules, consists of five extracellular IgG-like domains and one cytoplasmic tail [115]. ICAM-1 is found to be expressed at low levels in normal tissue, but at high levels in multiple types of cancer, including TCs [116,117]. One of its important features is that it can initiate tumor transmigration and invasion [116,118]. Furthermore, ICAM-1-targeted chimeric antigen receptor T (CAR-T) cells can robustly kill TC cells [119,120]. Research thus far has suggested ICAM-1 as an ideal target for TC diagnosis and treatments. For this purpose, Wei et al. created an immunoPET probe [ 64 Cu]Cu-NOTA-ICAM-1, which targets ICAM-1. [ 64 Cu]Cu-NOTA-ICAM-1 immunoPET imaging showed high contrast in diagnosing the subcutaneous and orthotopic ATCs in preclinical settings [101] (Figure 9). With the published data and unpublished data in hand, we believe that ICAM-1 may serve as a viable biomarker for certain types of TCs. However, it remains to see the diagnostic utility of ICAM-1-targeted tracers in patients with TCs.

Lectin Galactoside-Binding Soluble 3 (LGALS3, Galectin-3, or Gal3)-Targeting Probes
Gal-3 is a protein that is undetectable in normal and benign thyroid tissues but highly expressed in DTC cytosol, cell membranes, and intercellular substance [121]. The expression of galectin-3 as a biomarker for TCs has been validated in two multicenter studies [122,123]. The sensitivity and specificity of Gal-3 immunodetection reached 94% and 98% in distinguishing benign from TC lesions, with positive and negative predictive values of 98% and 94%, respectively, and diagnostic accuracy of 96% [122]. [ 89 Zr]Zr-labeled Gal3 mAb ([ 89 Zr]Zr-DFO-Gal3) or Gal-3 mAb with F(ab') 2 conjunction ([ 89 Zr]Zr-Gal3-F(ab') 2 ) has shown good binding to TC in vivo, allowing it to be potentially used for the detection of recurrence and metastases [124,125] (Figure 10). The particular design of [ 89 Zr]Zr-DFO-Gal3-F(ab') 2 , a protein formed of two F(ab') fragments, results in faster blood clearance and lower liver uptake than traditional mAb-based tracers [125]. The high uptake of [ 89 Zr]Zr-DFO-Gal3-F(ab') 2 in kidneys is due to the urinary excretion [125], which should not be problematic because metastatic TC to the kidney is very rare [126]. For a diagnostic purpose, the dose is usually quite low and so is the nephrotoxicity. Thus [ 89 Zr]Zr-DFO-Gal3-F(ab') 2 might be an excellent candidate for translation into the preoperative evaluation and postoperative follow-up.

Bispecific IgG Probes
Bispecific antibody (BsAb) probes have filled the vacancy of single target IgG probes in theranostics by providing higher antigen-binding capacity in tumor tissues than the monomeric counterparts [19]. Additionally, the pharmacokinetics of BsAbs could be improved by protein modification. The ability of BsAbs to bind to two targets allows these bispecific IgG probes to display an enhanced role for targeting two antigens on a tumor cell surface, linking the tumor cells and immune cells, for instance [127,128]. However, until recently, only one BsAb targeting CEA and HSG has been thoroughly investigated in the diagnosis of MTC.
As stated previously, the intense expression of CEA is a biomarker of MTC. Prior clinical studies have shown the high sensitivity of the combination of anti-CEA BsAbs and 111 In or 131 I labeled haptens-peptides [129][130][131]. IMP288, an HSG hapten, was reported to have the ability to bind multiple radionuclides [132]. Meanwhile, a trivalent BsAb (called TF2), was engineered composing one HSG glycine Fab fragment and two anti-CEA Fab fragments [133]. The combination of 68 Ga labeled IMP288 and TF2 in PET imaging yields high sensitivity and specificity; Nevertheless, the pretargeting conditions may still need to be modified to reduce or avoid IMP288-induced adverse effects (malaise, bronchospasm, tachycardia, and hypertension) [132,134]. The delivery method of the tracer may challenge patients' acceptability because the combination of IMP288 and TF2 requires two injections: the first injection for TF2 BsAb, and a second injection for [ 68 Ga]Ga-IMP288, with a time lag (one or two days) between the two injections [134] (Figure 11). The pretargeting strategy was used for the diagnostic purpose in the study. Replacement of 68 Ga with beta-emitter (e.g., 177 Lu) or alpha-emitter (e.g., 225 Ac) will further develop pretargeting therapeutic strategies, which will hopefully maximize the therapeutic index and minimize the adverse effects.

Fab-Based Probes
Fab is characterized by a light chain and a heavy chain of an immunoglobulin, containing variable regions, constant domain of the light chain (CL), and first constant domain of the heavy chain (CH1) [135]. The Fab, therefore, takes the specificity of the immunoglobulin. Unlike the traditional antibodies (produced from mammalian cells), Fabs could be generally and easily produced from bacteria cells, like E. coli [136]. One drawback of Fab is the limited retention on the antigen and rapid clearance [137]. Some Fabs have been discovered for the treatment of TC (targeting cluster of differentiation 276 [CD276] [138], etc.), and some publications reported the potential value of Fab as diagnostic probes targeting Galectin-3 [125,139,140].
[ 89 (Figure 12). Unlike the uptake of full-size 89 Zr-labeled Gal-3 mAb which lasts over five days after injection, the [ 89 Zr]Zr-DFO-αGal3-Fab-PAS 200 was supposed to have a shorter lasting time, but the exact time was undetermined [125]. Research concerning a head-to-head comparison between the anti-Gal3 IgG probe and the corresponding fragment probe is lacking.

Nanobody-Based Probes
A single-domain antibody (sdAb, nanobody) is an engineered antibody fragment containing a single monomeric variable antibody domain. Compared to the large size of full-size antibodies (~150 kDa), nanobodies (~15 kDa) can be delivered to tumors with comparatively less obstruction [141]. Nanobodies can be reconstructed to Fc-domains or conjugated to molecular inhibitors, radioisotopes, fluorescent dye, and nanoparticles, making them suitable for targeting tumors with many applications [142]. For example, Jailkhani et al. established nanobody libraries against extracellular matrix (ECM) proteins, which are hallmarks of many diseases, including cancers. PET/CT imaging showed that 64 Cu-labeled NJB2 nanobody probes targeted ECM and detected breast cancer and melanoma for primary and metastatic foci (including thyroid) with excellent contrast [143]. Thus, nanobody probes may open up a promising opportunity for application in TCs. So far, nanobody probes remain absent in TC research [144]. Our team has developed a series of nanobodies targeting various targets (e.g., tumor-associated calcium signal transducer 2 [TACSTD2, TROP-2], ICAM-1, integrin associated protein [CD47], and melanoma cell adhesion molecule [MCAM, CD146]) and are fully exploring the theranostic potential of the nanobodies in TCs.

Aptamer-Based Probes
Aptamers are nucleic acids with antigen selectivity rivaling that of antibodies [145]. They bind to their target through electrostatic interactions, hydrophobic interactions, and induced fitting. Aptamers also offer target recognition that is comparable to traditional antibodies. Unlike antibodies, however, aptamers can be produced more feasibly. Its additional advantages include favorable storage properties and limited immunogenicity in vivo [146]. The major drawback of aptamers is the lack of stability in vivo. Regarding their application in TC, only a few aptamer probes have been reported [15].

Prominin 1 (PROM1, CD133)-Targeting Probes
CD133 is a kind of glycoprotein mainly expressed in hematopoietic stem and progenitor cells [147]. As a marker of cancer stem cells of brain tumor, colon cancer, melanoma, and ATCs [148][149][150][151], it is known to be responsible for the rapid growth of ATC and PTC cells [152,153]. Ge et al. synthesized and characterized an aptamer AP-1-M targeting CD133 in an ATC xenograft model. The synthesized AP-1-M-doxorubicin conjugates can effectively bind CD133-expressing tumor cells, and an intense signal may reflect the tumor proliferation at a fast pace [15] (Figure 13).

PTC Tissue-Targeting Probes
Zhong et al. generated a PTC tissue-specific aptamer (TC-6) via tissue-based systematic evolution of ligands by exponential enrichment (SELEX), with clinical PTC tissues (positive control) and non-tumor thyroid tissues (negative control). The TC-6 can specifically distinguish PTC from other non-tumor tissues (Figure 14), and suppress the migration and invasion of PTC cells [16]. However, the exact molecular target remains unknown.

Nanoparticles-Based Probes
Nanoparticles have been emerging with widespread attention in MI, drug delivery, and disease treatment. Nanoparticles have brought their potential as MI agents to TC, primarily through their applicability in fluorescence imaging, ultrasound, and MRI [154,155]. These modalities enable nanoparticles to accumulate in cells by activation through US, light, temperature, and pH change, depending on the nanoparticle structures and their surface molecules. The ligand options for targeted nanoparticles are somewhat limited. Antibodies and peptides are the primary ligand choices due to their specific affinity to targets in TC fields. Although applications of targeted nanoparticles in TC have so far been limited, there have been publications investigating nanoparticles conjugated to antibodies targeting epidermal growth factor receptor (EGFR) or Src homology 2 (SH2) domain-containing phosphatase 2 (SHP2) [154,155].

EGFR-Targeting Probes
EGFR is a receptor binding the extracellular epidermal growth factor family (EGF family) [156]. In many tumor types, including TC, increased EGFR expression or activity initiates the tumor cell progression [112]. Recently, EGFR has been the target of the newly created nanoparticle (called C-HPNs) based on a core-shell system loaded with EGFRtargeted cetuximab and 10-hydroxycamptothecin (10-HCPT, chemotherapy agent). The EGFR antibody ligands enable nanoparticles to attach to cells which overexpress EGFR. With low-intensity focused ultrasound (LIFU) assistance, the liquid perfluoropentane (PFP) core in the nanoparticles would become vaporized and transformed into microbubbles, enhancing ultrasound contrast for tumor diagnosing. The core explosion induced by PFP boiling causes the release of 10-HCPT, providing more targeted delivery of the chemotherapeutic drug [154] (Figure 15).

Protein Tyrosine Phosphatase Non-Receptor Type 11 (PTPN11, SHP2)-Targeting Probes
Another example is the SHP2, which is a tumor biomarker, acting as a signal of cell proliferation and immortality [157]. Hu et al. created an SHP2-targeted core-shell nanoparticle chelated with the contrast agent Gd 3+ on the surface (NPs-SHP2). Similar to the EGFR-targeted nanoparticles mentioned previously, PFP-based LIFU can facilitate the probe carrying contrast agent been accumulated in the thyroid tumor area for enhanced MRI [155] (Figure 16).

Conclusions and Future Perspectives
In summary, MI plays a vital role in evaluating and managing TCs, especially in accurately finding occult foci that are undetected by traditional ultrasound, CT, and MRI, thereby helping TC patients get the precise therapeutics (Table 1). Transporter-based probes tend to have high sensitivity, and immune-based probes generally have high specificity. So far, there is no single probe to reveal all the lesions both specifically and sensitively. Moreover, although TCs can be classified in a single pathological category, the biomarker expression in TC can vary dramatically. The apparent heterogeneity of TC requires the availability of more than one therapeutic method. By fully elucidating the biological characteristics of TCs and thoroughly exploring the biomarkers enriched in TCs [26,158,159], we believe that we can discover helpful targets for developing diagnostic probes and companion therapeutic agents for different molecular types of TC, not limited to pathological typing and phenotyping. With the development of biological techniques and imaging tools, more valuable imaging methods are emerging. Integration of multiple imaging modalities and anatomical features would help physicians diagnose and treat TCs in a timely target-specific manner. Nanoparticles may enable anti-TC drug delivery and multimodality imaging, which may further improve the management of TC. However, the authors are cautious because of the limited clinical evidence in the field of TC. Bispecific MI probes have gained more traction in the past decade. The synthesis of bispecific antibody or antibody fragment tracers has been thoroughly elucidated elsewhere [170]. The importance of bispecific probes is that they enable enhanced affinity and high image quality, providing the ability targeting more occult foci than traditional single-target probes, and inducing more comprehensive application across TC patients with complex biological characteristics.
For antibody probe design, the traditional antibody-based radioactive or fluorescent probes for TCs have simply been prepared by non-selective conjugation on lysine/cysteine residues with or without chelators [171,172]. This approach may reduce target binding affinity, especially at a high conjugate/protein ratio, and in any case, leads to a mixture of products with different numbers of tags per protein molecule [173]. The method may also cause undesirable biodistribution (e.g., high kidney uptake and poor tumor targeting due to in vivo cleavage of the S-S linkage) and pharmacokinetics (fast antibody clearance for modification of interchain disulfide cysteine) [174]. Protein engineering techniques are progressing very rapidly. In this setting, the site-specific and homogeneous introduction of the tags into the targeting moieties would be more advantageous. The emerging techniques mainly include chelator conjunction to the antibody glycan region, enzymeassisted chelator attachment, and incorporating and chelating radioisotope into amino acid sequences [171]. This may help us design molecular imaging agents and companion therapeutic agents with increased possibility for clinical translation.
For the diagnosis of TC, the superiority of MI over conventional anatomical imaging is clear, with advantages of favorable spatial and temporal resolution and functional imaging [175]. The emergence of MI has fundamentally changed the management of TC. For instance, [ 18 F]FDG PET/CT plays its role in optimizing initial therapy, which is mandatory for improving DTC outcome. [ 18 F]FDG PET/CT could be conducted if some foci are radioiodine non-avid before treatment planning. Rosenbaum-Krumme et al. found that the TNM staging and management were changed from standard therapy (surgery plus 131 I therapy) to individual therapy (standard therapy plus external beam therapy or targeted therapy etc.) because of the [ 18 F]FDG PET/CT results in 21% of the high-risk DTCs [176]. We suppose that MI in TCs would become a helpful modality for tumor staging, prognosis evaluation, which may lead to immense changes in what treatments the TC patients are given, maximizing the benefits of individual therapy.
For TC therapy, despite the rapid progress in molecular imaging, we firmly believe that the innovation of therapeutic agents should accompany the development of diagnostic agents. For example, radionuclides such as 177 Lu,225 Ac 188 Re, 67 Cu, 47 Sc, 166 Ho, 90 Y, 161 Tb, 149 Tb, 212 Pb, and 213 Bi emitting α-particles, β-particles, and Auger electrons can be feasibly chelated with DOTA, NOTA, etc., which is similar to the diagnostic isotopes mentioned in this review, 64 Cu or 68 Ga chelated with NOTA or DOTA [177]. For theranostic application, diagnostic probes with a chelating agent and a radionuclide suitable for imaging (e.g., NOTA and 64 Cu) can be used to map the target and assess the therapeutic potential of the same probe labelled with the same chelating agent and a therapeutic radionuclide (e.g., NOTA and 67 Cu) [19]. 149 Tb is another theranostic radioisotope that has not been investigated in thyroid cancer research to date. 149 Tb can simultaneously emit positrons (β + particles), α-particles, and γ-radiation, allowing PET and α particle-based therapy to go on at the same time [177]. The current review focuses on emerging probes for imaging with examples. Further introduction and discussion about the concept of TC therapy or theranostics will be updated and illustrated in an upcoming review. We hope the ever-developing diagnostic and therapeutic probes and theranostic applications can substantially improve the management of TCs, especially the aggressive RR-DTCs, MTCs, and ATCs.
In the last two decades, there has been a general increase in the prevalence of TC. The phenomenon is partially due to environmental factors (e.g., chemical pollution, anthropogenic or natural radiation), but is also due to overdiagnosis by increased screening with more sensitive methods (e.g., high-resolution ultrasound), especially in developed countries (e.g., South Korea and the United States), leading to unnecessary treatment. As mentioned previously, most incident TCs are low-risk DTCs that tend to retain their stability over the years until death due to aging. Therefore, it is unnecessary to treat indolent TCs because of the low cost-effectiveness and the practically unchanged mortality. Something to note is that the ever-developing MI techniques might more sensitively detect TCs, potentially causing overdiagnosis for indolent TC and overtreatment via invasive methods (e.g., thyroidectomy or metastasectomy) or noninvasive therapeutics (e.g., chemotherapy agents or multikinase inhibitors). To avoid these pitfalls, future work could focus on discovering prognosis-or progress-related targets or probes, thereby classifying TCs into indolent and active disease. In other words, future MI research should not be limited to the field of finding latent TC lesions.

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

Conflicts of Interest:
Weibo Cai is a scientific advisor, stockholder, and grantee of Focus-X Therapeutics, Inc. All other authors declare no conflict of interest.