**6. Discussion**

Starting in the early 1990s, CDs became of large interest for developing contrast agents and molecular probes for medical imaging. Although the primary application of CDs in medical imaging was the basis for developing novel Gd-based MRI contrast agents, CD-based contrast agents commence to be utilized by other imaging modalities such as PET, CT, US, and PAI. This growing interest is caused by excellent biocompatibility, low toxicity, and relative ease of modifying of CD molecules. Currently, the most frequently used CD molecule for contrast agent development is β-CD. Despite the extensive development of β-CD-based contrast agents, some of the medical imaging areas started utilizing α- and γ- CD-based contrast agents and molecular probes as well.

The working mechanism of the vast majority of CD-based contrast agents for conventional MRI relies on the Freed's theory [57] and the decreasing of the relaxation times of the solvent protons once either paramagnetic ion or metal–organic complex containing the paramagnetic ion binds to a supramolecular cage. Initially, this binding was done through hydrophobic interaction between CDs cavity and metal–organic complexes. Although contrast agents synthesized using this approach demonstrated significantly better performance compared to the clinically available analogs, the main disadvantage of this technique is blocking the supramolecular cage cavity, which makes the developed contrast agents and molecular probes uncapable to image some drug transport, biodistribution, and treatment monitoring. In addition, the number of paramagnetic ions per molecule is also limited, since usually only one metal–organic complex enters the cavity. This issue was overcome by creating different CD-based polymers and dendrimers, and the number of macromolecules determined the maximum number of paramagnetic ions per molecule. Starting from the early 2000s, the direct labeling of the CD macrocycles with chelates containing paramagnetic ions has been developed intensively. This method allows keeping the CD cavity free for interactions with different molecules. Therefore, the contrast agents based on the direct labeling of the CD macromolecules have a significantly larger field of applications compared to agents developed using non-covalent interactions.

Currently, CD-based MRI contrast agents contain mostly Gd(III) ions and are designed as T<sup>1</sup> contrast agents. On the other hand, the development of CD-based T<sup>2</sup> and T<sup>2</sup> \* contrast agents have been demonstrated in vitro and in vivo [23,71,75]. Although some of the developed CD-based contrast agents for proton MRI demonstrated an overall low level of cytotoxicity, dedicated toxicology studies are needed prior to the further translation of this imaging agents into clinics. In addition, the vast majority of the relaxivity measurements were conducted at low magnetic fields, which are not for clinical imaging purposes, such as 0.47 and 0.5 T. The studies that focused on animal imaging were mostly conducted at high magnetic fields, such as 7 and 9.4 T. In order to facilitate the potential clinical transition of the developed CD-based contrast agents, the accurate measurement of relaxivities and comparisons to clinical analogs should be performed at 1.5 and 3 T, which are currently used for clinical MRI. During the last eight years, CDs drew attention from researchers working with hyperpolarized MRI. Although only a few studies were conducted, the obtained results demonstrated the high potential of CD macromolecules to become a basis for further development of hyperpolarized probes. Recently, Hane et al. demonstrated that γ-CD-based pseudorotaxanes can be a valuable platform for developing the molecular imaging probes that utilizes the HyperCEST effect [95]. The alkyne chains have a high affinity to the γ-CD cavity and can be easily designed to serve as a high-affinity probe that has a selective binding to the disease site. Functionalizing the alkyne chain should be done carefully such that there is no effect on the interaction with hyperpolarized <sup>129</sup>Xe. The HyperCEST depletion observed was only around 50% indicating partial depolarization of hyperpolarized <sup>129</sup>Xe encapsulated by γ-CD-based pseudorotaxanes. An accurate study for radiofrequency saturation pulses is required to maximize the HyperCEST depletion and to translate γ-CD-based pseudorotaxanes to in vivo imaging applications. Although the demonstrated results of hyperpolarized MRI molecular imaging probes look promising, further in vivo imaging studies are needed, as well as biodistribution and toxicity evaluations of the proposed biosensors.

Following the success of the CD-based MRI contrast agents, several attempts have been made to use CDs as a contrast agent for US imaging [44,96]. The developed US CD-based contrast agents caused significant acoustic impedance difference between the tissue and the contrast agents. Despite successful in vitro demonstration of this proof-of-principle [44], there was no further development of the CD-based contrast agent for US imaging purposes during the last five years. On the contrary, CD-coated NPs became of interest for the PAI. It was found that coating of UCNPs with α-CD cause photoacoustic signal enhancement [49] and, therefore, can serve as efficient PAI contrast. The most recent achievement in the field demonstrated PAI detection of prosthetic joint infection using β-CD-conjugated indocyanine green in mice [48]. All studies with CD-based PAI tracers were conducted in small animals. Since the PAI tracers required irradiation with infrared light to emit ultrasound, further studies should move to the larger animals in order to evaluate penetration depth of the excitation light and to estimate the performance of PAI contrast agents prior to clinical translation. Furthermore, the development of larger pallet of PAI contrast dedicated for specific diseases would be beneficial to facilitate the future translation into the clinic.

The vast majority of the radiolabeled CDs derivatives were used for PET and SPECT imaging of the drug delivery and biodistribution in vivo. To be suitable for PET, multiple CD-based tracers containing <sup>64</sup>Cu,68Ga, and <sup>18</sup>F were developed. For SPECT imaging purposes, CD macrocycles were radiolabeled with either 99mTc or <sup>125</sup>I. The further widening of the radiolabeled CD derivatives might become a useful tool for pharmaceutics and drug development.

The most recent CD-based NPs were studied as contrast agents for CT. To be able to produce a sufficient contrast, the CD-based NPs must contain atoms with a high atomic number. The heavy atoms absorb X-rays with higher efficiency, increasing the X-ray attenuation coefficient of the tissue in which the contrast agent is present [107–109]. With modern advances in X-ray detection allowing lower dose image acquisition [110–113], the utilization of the novel CD-based NPs containing heavy elements will be highly beneficial for accurate anatomical imaging purposes. In addition, the novel dual energy CT approach [114,115] will benefit even more from the implementation of the novel CD-based NPs. Implementation and further development of the demonstrated NPs with substantially higher X-ray attenuation compared to currently available iodine and barium contrasts could potentially allow a superior improvement of CNR of the dual energy CT image. CT contrast of CD-based NPs was initially observed from contrast agents developed for multimodal imaging [50,53]. Only during the last three years has the research in synthesis of dedicated CT CD-based contrast agents been conducted extensively. Currently, the vast majority of the CD-based CT contrast agents were developed specifically for cancer imaging, whereas only one agent for angiography was developed. Therefore, it might be potentially useful to develop CD-based contrast agents suitable for different types of clinical CT imaging. In addition, the potential CD-based CT contrast agents must undergo an accurate toxicity study prior to further translation to in vivo imaging. Furthermore, special attention should be given to the investigation of extraction pathways of the contrast agents from the living organism during in vivo studies.

One of the potential approaches for further advances in the field of CD-based contrast agents is the development of a contrast agent that can serve for dual imaging modalities such as PET/MRI and PET/CT. Usually, the PET-active component of the tracer is small, and therefore, the development of the PET/MRI and PET/CT contrast agents could be built around radiolabeling of the existing CD-based agents for MRI and CT imaging modalities, respectively.

Overall, CDs are of high interest in the medical imaging field and are currently a very promising basis for developing various contrast agents. The successful clinical translation of CD-based contrast agents for proton MRI can help significantly improve the quality of clinical MRI scans. Further development of functionalized CD-based imaging agents for MRI has the potential to make molecular imaging using clinical proton MRI possible. Despite the enormous development level of CD-based contrast agents for conventional proton MRI, other imaging modalities started utilizing CD-based contrast agents recently and further developments and investigations are needed prior to successful clinical translation. Nevertheless, CD-based contrast agents demonstrated exceptional performance in the areas of CT, PET, and PAI.

**Author Contributions:** Y.S. performed literature search, wrote the manuscript, and revised and edited the paper. C.J.N. performed literature search and wrote the manuscript. V.G. wrote the manuscript and revised and edited the paper. L.E.S. wrote the manuscript and edited the paper. B.D. wrote the manuscript and revised and edited the paper. F.T.H. revised and edited the paper. M.S.A. revised and edited the paper. All authors have read and agreed to the published version of the manuscript.

**Funding:** This work was partially supported by the Natural Science and Engineering Research Council of Canada (NSERC) Discovery grant. Y.S. was supported by an Ontario Graduate Fellowship. C.J.N. was supported by the NSERC Undergraduate Student Research Award (USRA). V.G. was supported by an Ontario Trillium Scholarship.

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