Theranostic Imaging of Yttrium-90

This paper overviews Yttrium-90 (90Y) as a theranostic and nuclear medicine imaging of 90Y radioactivity with bremsstrahlung imaging and positron emission tomography. In addition, detection and optical imaging of 90Y radioactivity using Cerenkov luminescence will also be reviewed. Methods and approaches for qualitative and quantitative 90Y imaging will be briefly discussed. Although challenges remain for 90Y imaging, continued clinical demand for predictive imaging response assessment and target/nontarget dosimetry will drive research and technical innovation to provide greater clinical utility of 90Y as a theranostic agent.


Yttrium-90 and Its Role in Targeted Radiotherapy
In general, theranostics are agents that possess diagnostic and therapeutic attributes for personalized patient treatment for various diseases [1]. A commonly used theranostic agent is radioactive iodine (e.g., Iodine-131 or 131 I) for the evaluation of thyroid physiology and pathophysiology, treatment of hyperthyroidism, treatment of thyroid cancer, and posttreatment assessment of radioactive iodine distribution in the body. The rare-earth lanthanide, Yttrium-90 ( 90 Y), is almost exclusively a high-energy beta-particle (i.e., electron or − ) emitting radionuclide used for radiotherapy with a maximum particle energy of 2.28 MeV (average energy of 0.94 MeV) that allows for high dose deposition with an average and maximum soft tissue penetration of 2.5 mm and 11 mm, respectively [2,3]. 90 Y has a physical half-life of 64.1 h [4] which makes it amenable for a variety of targeted radiotherapy applications including 90 Y-labeled colloid [5,6], somatostatin-receptor targeting peptides [7,8], tumortargeting antibodies [9,10], and resin/glass microspheres for catheter-directed embolization of hepatic malignancy and metastases [3,[11][12][13]. Regardless of the targeted delivery agent used, the selection of 90 Y and its use for radiotherapy are complex and necessitate close collaboration among various medical specialties including nuclear medicine, interventional radiology, medical oncology, and radiation medicine [14]. 90 Y can be administered via direct injection into a space or cavity (e.g., radiosynovectomy), intravenously for peptide receptor radionuclide therapy (PRRT) and radioimmunotherapy (RIT), and intra-arterially for radioembolization (RE) therapy. Other therapeutic − emitting radioisotopes (e.g., 131 I for thyroid cancer [15] and Samarium-153 ( 153 Sm) for osseous metastases [16]) also produce discrete gamma photons which can be imaged after therapy but contribute to additional absorbed radiation dose. One advantage of 90 Y is that it is an almost pure − emitting radioisotope which lacks such gamma photons [6]. On the other hand, because of the lack of gamma photons from 90 Y, conventional scintigraphic imaging and assessment of the posttherapy distribution of its radioactivity are challenging. This lack of gamma photons led to the development and use of surrogate gamma-emitting radioisotopes (e.g., Indium-111-( 111 In-) labeled peptides and antibodies) with analogous chemical properties as a tracer for 90 Y dosimetric assessment and pharmacokinetics [2,17]. Likewise, Technetium-99m-( 99m Tc-) labeled macroaggregated albumin (MAA) is currently used as a surrogate radiotracer for planning 90 Y microsphere RE therapy [18][19][20]. It is important to note that use of such surrogate tracers may not always accurately predict 90 Y radiotherapy effects in vivo and such discrepancies may result in unanticipated and unintended toxicities [17,[21][22][23]. Given that surrogate tracer agents may not always predict the precise posttherapeutic distribution of 90 Y, subsequent imaging assessment of 90 Y radioactivity is an important adjunctive step to assess and verify delivery and dosimetric distribution of the 90 Y agent to the target(s) and exclude any nontargeted delivery [24]. Likewise, accurate quantification of 90 Y radioactivity in both targeted lesions and nontargeted tissues would allow for improved comparisons of radiotherapy outcomes in patients. This review will subsequently discuss the different diagnostic imaging approaches used for therapeutic 90 Y radioactivity assessment ( Figure 1).

Bremsstrahlung Radiation
Conventional scintigraphic imaging and quantification of monoenergetic gamma-emitting medical radioisotopes (e.g., 99m Tc) have driven the evolution of current planar gamma cameras with optimized collimators and detector crystals for detecting and counting primary (i.e., unscattered) photons in discrete energy windows. − particle emission from 90 Y produces bremsstrahlung photons which can also be imaged scintigraphically [6,25]. The 90 Y bremsstrahlung photons are generated when the high-energy − particle (i.e., electron) is emitted from the 90 Y nucleus and then slows (i.e., it loses its kinetic energy) while interacting with adjacent atoms. As the electron slows down, its kinetic energy is converted into the continuous energy spectrum of both primary and scattered photons with no dominant energy photopeak for conventional scintigraphic imaging (i.e., bremsstrahlung radiation).
In 1967, Simon and Feitelberg described posttherapy bremsstrahlung imaging assessment of intra-arterially administered 90 Y-labeled plastic microspheres in oncology patients [25]. Furthermore, they described an early clinical case of nontargeted deposition of 90 Y-labeled microspheres within the lungs of a patient with a radioembolized left renal mass. The radioembolized left renal mass and bilateral lungs demonstrated 90 Y radioactivity on posttherapy bremsstrahlung imaging and the bilateral lung radioactivity was presumed arteriovenous shunting of microspheres through the tumor and then trapped in the lungs. Subsequently, others have described posttherapy planar bremsstrahlung imaging for patients following direct injection of 90 Y (e.g., radiosynovectomy) [5,6], intravenous administration of 90 Y-labeled RIT [26], and intra-arterial administration of 90 Y-labeled microspheres [27][28][29][30][31]. In addition, one study demonstrated the capability for planar bremsstrahlung imaging to detect focal 90 Y radioactivity using a phantom model simulating soft tissue extravasation of an intravenous 90 Y dose [32].
Although technically feasible, image quality for 90 Y bremsstrahlung is limited by overlying tissue attenuation, internal photon scattering, variable count rates of emitted bremsstrahlung photons, a wide range of photon energies produced, low spatial resolution (which worsens with is a high-energy − emitting radioisotope used clinically for targeted radiotherapy (upper left). The targeted radiotherapy applications include direct injection of 90 Y into a body space or cavity, conjugation of 90 Y to a peptide for peptide receptor radionuclide therapy (PRRT), or an antibody for radioimmunotherapy (RIT), or incorporation of 90 Y into a resin or glass microsphere for radioembolization (RE) therapy. The high-energy − particle emission produces a continuous spectrum bremsstrahlung radiation which can then be imaged using conventional nuclear medicine imaging systems such as planar gamma cameras, SPECT, and SPECT/CT (lower left).
Although the vast majority of 90 Y decays are − emitting, 32 per million 90 Y decays result in internal pair production that can be readily imaged using conventional PET/CT and PET/MRI systems (lower right). The high-energy − particle emission also produces continuous spectrum light photons or Cerenkov luminescence which can then be imaged using existing bioluminescence imaging systems (upper right). These 3 noninvasive imaging approaches allow for simultaneous diagnostic assessment/localization of the therapeutic 90 Y radioactivity.
increasing source distances to the camera), type of collimation employed (i.e., low, medium, or high-energy collimators), and image processing. In particular, attenuation coefficients may not be constant for the range of photon energies acquired by the gamma camera. Likewise, lower energy bremsstrahlung photons are more likely to scatter than highenergy photons. On the other hand, higher energy photons are more likely to penetrate collimator septae and detector crystals which degrade image quality and limit quantification [6,[33][34][35][36][37]. No standardized imaging protocol was used for these early 90 Y bremsstrahlung imaging studies. Subsequent efforts to optimize planar 90 Y bremsstrahlung imaging have used Monte Carlo simulation modeling [35] and these efforts support the use of medium or high-energy parallelhole collimators and energy windows ranging from 50 to 200 keV. Quantification of 90 Y bremsstrahlung radioactivity is likewise challenging but advances in both qualitative 90 Y bremsstrahlung imaging and quantitative 90 Y bremsstrahlung imaging have been described using optimized It should be noted though that planar quantification is a two-dimensional (2D) assessment of 90 Y radioactivity with limited potential for distinguishing overlapping sources of 90 Y radioactivity [38]. Compared to planar imaging, the application of single photon emission computed tomography (SPECT) to 90 Y bremsstrahlung imaging allows for improved three-dimensional (3D) visualization and anatomic discrimination of discrete adjacent foci of 90 Y radioactivity as well as improving the potential for quantification [6]. The use of medium-and high-energy parallel-hole collimation is again supported to optimize camera sensitivity for 90 Y bremsstrahlung photons but, like planar imaging, SPECT cannot distinguish between primary and scattered bremsstrahlung photons and this limits quantitation [2,45]. The fusion of 90 Y bremsstrahlung SPECT with X-ray computed tomography (CT) allows for attenuation correction and 3D anatomical localization of SPECT findings (i.e., SPECT/CT) [38]. This represents another distinct advantage over bremsstrahlung 2D planar and 3D SPECT only imaging [46].
Given that SPECT imaging requires much more time than planar imaging approaches, planar 90 Y bremsstrahlung imaging can be more readily adopted for whole-body assessment of 90 Y distribution [38]. On the other hand, bremsstrahlung SPECT imaging may allow for improved quantification when compared with planar approaches and better 3D dose assessment of localized 90 Y radioactivity [36]. Recently, bremsstrahlung SPECT/CT imaging has been the imaging modality of choice for qualitative post-90 Y RE assessment of liver radioactivity but image quality is still less than ideal [14,65].

Internal Pair Production
Although the vast majority of 90 Y decays result in therapeutic − particle emission, 32 per million decays result in internal pair production that produces annihilation radiation that can be also imaged in vitro using positron emission tomography (PET) imaging systems [66][67][68]. While this rate of internal pair production is very small, there is a detectable peak of 511 keV photons which exceeds the continuous spectrum of bremsstrahlung photons and these 511 keV photons can be detected and imaged using conventional PET imaging [66]. PET detection of 90 Y internal pair production represents a promising approach for even more accurate 90 Y quantification in vitro and in vivo by minimizing the previously noted challenges associated with 90 Y bremsstrahlung imaging [67]. These observations led to the first clinical case report, in 2009, of PET/CT imaging of 90 Y radioactivity following 90 Y-labeled resin microsphere RE for colorectal liver metastases, which demonstrated the feasibility of imaging 90 Y in vivo using an existing conventional PET/CT system [50]. The detected intrahepatic 90 Y radioactivity correlated well with the targeted intrahepatic lesion. Likewise, quantitative assessments of 90 Y radioactivity in phantoms could also be performed with further improvement in quantitative accuracy using Time-of-Flight (ToF) PET reconstruction [44,69,70]. ToF PET imaging demonstrates some advantages in 90 Y radioactivity assessment when compared with non-ToF PET imaging systems [71] and 90 Y bremsstrahlung SPECT/CT imaging [40,51]. Subsequently several other clinical studies have described posttherapy 90 Y internal pair production PET imaging for patients following direct injection of 90 Y [47], intravenous administration of 90 Y-labeled RIT [54] and PRRT [47], and intra-arterial administration of 90 Y-labeled microspheres (resin [14, 20, 28, 44, 51-55, 58, 70, 72-77], glass [54,61,62,78], or not specified [64,79]).
Image quality for 90 Y internal pair production is limited by its very small branching fraction (i.e., 32 per million decays) and therefore necessitates longer acquisition times than traditional positron-emitting radioisotopes (e.g., Fluorine-18 ( 18 F) which has a branching fraction of 967 per 1000 decays). It was also noted that measureable background radioactivity was dependent upon the PET imaging system used. The presence of a small fraction of radioactive Lutetium-176 ( 176 Lu) within the detection crystals (i.e., lutetium yttrium orthosilicate or LYSO or lutetium oxyorthosilicate or LSO) of PET imaging systems contributes to this measureable background radioactivity [69,78]. This requires that 176 Lu background radioactivity be corrected for in order to obtain any accurate 90 Y radioactivity assessment using these PET systems [78]. The 176 Lu background radioactivity is not present on PET imaging systems which utilize bismuth germinate (BGO) detector crystals [66] and the BGO PET can provide 90 Y radioactivity quantification [80]. It has been reported that BGO PET systems may be less accurate for 90 Y radioactivity quantification when compared with LYSO-dependent PET systems due to the slower response rate and poorer contrast performance of BGO PET systems [71]. There are no reported clinical instances of PET detector saturation from 90 Y bremsstrahlung radiation.
Despite the low branching fraction for 90 Y and background radioactivity of some PET imaging systems, PET/CT imaging demonstrates better spatial resolution and image contrast than bremsstrahlung imaging (planar, SPECT, and SPECT/CT) [28,44,51] and clinically demonstrates improved detection of nontarget 90 Y radioactivity compared with even bremsstrahlung SPECT/CT [14]. Although 90 Y internal pair production imaging has been studied in vitro and in vivo using a variety of different PET imaging systems, different acquisition times, and different reconstruction algorithms, no standardized or consensus imaging protocol has been described for 90 Y PET/CT imaging studies to date. Table 2 details some of the acquisition and image reconstruction parameters used for clinical 90 Y internal pair production PET imaging studies. In 2013, Kao et al. [14] described a diagnostic reporting approach for 90 Y PET/CT imaging following RE therapy in order to (1) confirm successful deposition of the 90 Y microspheres within the target lesion(s) and (2) detect any nontarget 90 Y radioactivity. In this study, 90 Y PET/CT imaging was consistently superior to 90 Y bremsstrahlung SPECT/CT imaging in the qualitative assessment of post-RE patients, especially in the detection of nontarget 90 Y radioactivity [58].

Cerenkov Luminescence
Another innovative approach for imaging of 90 Y is real time detection of Cerenkov radiation (CR), that is, ultraviolet and visible light emitted in the presence of high-energy − particle and positron-emitting radionuclides [81][82][83]. CR is produced when electrons or positrons travel faster than the speed of light through an aqueous medium (i.e., cells, tissues, and organs). As these high-energy charged particles travel through water, they disrupt the local electromagnetic field in the water. Electrons in the atoms of the water molecules will be displaced, and the atoms become polarized by the passing electromagnetic field of the − particle or positron. Visible and ultraviolet light photons are emitted as the displaced electrons in the water molecules restore themselves to equilibrium and these light photons can be detected with existing high-sensitivity bioluminescence imaging systems. This optical imaging of CR has been designated as Cerenkov luminescence imaging (CLI) [84]. Detectable CLI signals have been described in vitro for a number of positronemitting radioisotopes (e.g., 18 F, Gallium-68, or 68 Ga) and − particle emitting radioisotope (e.g., 90 Y and 131 I) [85][86][87]. To date, 90 Y is the most efficient medical radioisotope for Cerenkov luminescence production [85]. In preclinical studies, in vivo CLI has been performed in mouse models following intravenous administration of 90 Y salt solution [85] and 90 Y-labeled peptide [85,88]. This novel optical imaging approach for noninvasively detecting 90 Y radioactivity in vitro and in vivo presents many exciting opportunities. High spatial resolution images of 90 Y radioactivity using CLI can be obtained within seconds as opposed to several minutes with conventional planar, SPECT, and PET imaging systems. CLI systems also allow for imaging multiple animals simultaneously as opposed to individually using micro-SPECT/PET imaging systems. These CLI systems are also much less expensive when compared with conventional-or micro-SPECT/PET imaging systems. This CLI approach for the preclinical development of targeted 90 Y theranostics (e.g., nanoparticles, microspheres, colloids, peptides, and antibodies) will be tremendously enabled for researchers and clinicians. Clinical proof-of-concept (i.e., human Cerenkography) has recently been described for radiotherapy using 131 I [89]. To date, no clinical applications for 90 Y Cerenkography have been described in the literature.

Challenges and Future Directions for 90 Y Imaging
One current challenge for 90 Y imaging is the lack of consensus guidelines for the technical acquisition, imaging reconstruction, and qualitative/quantitative interpretation of planar, SPECT, and PET imaging by the nuclear medicine community (e.g., Society of Nuclear Medicine and Molecular Imaging (SNMMI) and European Association of Nuclear Medicine (EANM)). An initial consensus guideline would establish the basis for future imaging studies to design, develop, and optimize 90 Y imaging approaches and reporting. Likewise, a consensus guideline would describe relevant imaging signs following 90 Y radiotherapy for imagers [63]. Another closely related challenge is that the vast majority of nuclear medicine imaging systems in place around the world are not currently designed or specifically optimized for 90 Y imaging applications. While some manufacturers have provided assistance and expertise to adapt existing imaging systems for 90 Y imaging [46], most imaging centers may have to internally customize imaging protocols with little guidance or validation. It is critical that professional organizations, nuclear medicine physicians, and researchers continue to interface and actively engage the imaging system manufacturers to develop and optimize specific protocols for more consistent and comparable 90 Y image acquisition, image reconstruction, and, ideally, quantification. In addition, new technical advances incorporated into the state-of-the-art PET/CT imaging systems like digital PET/CT and continuous bed motion PET acquisition will need to be methodically assessed for advantages and limitations. Although a single case report on respiratory-gated PET/CT imaging for 90 Y RE has been described [79], the advantages and limitations of respiratory-gated 90 Y PET imaging will also need to be addressed. Recently, the trend in 90 Y imaging has largely focused on 3D modalities like SPECT/CT and PET/CT (Figure 2). The majority of the literature relates to 90 Y radioactivity imaging for post-RE assessment of 90 Y-labeled resin microspheres using bremsstrahlung SPECT/CT and, more recently, internal pair production PET/CT. There are fewer reports related to the post-RE assessment of 90 Y-labeled glass microspheres and even less related to 90 Y imaging assessment of direct injection radiotherapies, RIT and PRRT. For the near future, 90 Y internal pair production PET/CT will likely be compared with 90 Y bremsstrahlung SPECT/CT imaging (i.e., a reference imaging standard). Although PET/CT imaging systems are more readily accessible today, 90 Y PET imaging may be more challenging to incorporate into routine clinical workflows due to the low branching fraction and corresponding low count rates for 90 Y (i.e., it requires longer acquisition times per bed position than more traditional BioMed Research International 7 18 F-fluorodeoxyglucose PET/CT imaging studies) [61]. There is a single case report for 90 Y imaging with PET integrated with magnetic resonance imaging (MRI) [62]. Given that even fewer PET/MRI imaging systems are available than PET/CTs, it will be important that future studies address the advantages and limitations of PET/MRI imaging over PET/CT.
Review of current literature suggests that 90 Y bremsstrahlung SPECT/CT imaging will continue in the future as (1) a reference standard for comparing different 90 Y imaging modalities and (2) a more widely accessible imaging modality for qualitative assessment of 90 Y radioactivity. As such, continued technical and methodological advances will likely improve SPECT/CT image quality, consistency, and quantification. Although 90 Y bremsstrahlung imaging is better with SPECT/CT than planar imaging, planar imaging approaches may represent a more accessible and less expensive qualitative imaging modality capable of performing faster whole-body assessment of 90 Y radioactivity than existing SPECT/CT technology. If any gross irregularity is detected with qualitative planar imaging, the patient could be referred for SPECT/CT or PET/CT assessment. The ever-present limitation of 2D planar bremsstrahlung imaging of 90 Y radioactivity is the inability to resolve adjacent foci of 90 Y radioactivity in target and nontarget tissues. In terms of patient safety and quality control/assurance during 90 Y radiotherapy administration (e.g., direct cavity injection, intravenous and intra-arterial), planar bremsstrahlung imaging may play an important role in the future to document successful administration, confirm systemic circulation for nonembolic agents, and exclude any focal soft tissue extravasation or nontarget 90 Y radioactivity. To this end, it has been recently proposed to optimize conventional Anger camera technology for interventional 90 Y bremsstrahlung imaging applications [90].
Another exciting potential imaging modality for 90 Y assessment is CLI. This technology may help to facilitate rapid and more cost-effective preclinical development of a wide array of targeted 90 Y-labeled theranostic agents. One challenge for clinical implementation for CLI is the current requirement for no ambient light within the field of view of the CLI system (i.e., the sample, specimen, or subject must be imaged in total darkness). Ambient light can saturate the highly sensitive CLI imaging system and obscure the true Cerenkov luminescence emissions. Despite this limitation and challenge, human Cerenkography following 131 I radiotherapy has already been described [89]. Future studies will also determine the feasibility and practicality of incorporating this optical imaging technology into qualitative clinical assessment of radiotherapy administration (i.e., during and after direct injection into a body cavity or space, intravenous or intra-arterial administration) as well as in vivo/ex vivo assessment of posttherapy 90 Y-labeled target or nontarget lesions using CLI-capable endoscopes and specimen analyzers.
An international collaborative project (metrology for molecular radiotherapy or MetroMRT) has been initiated to address the known variability in absorbed dose for patients following radiotherapy, including 90 Y [91]. Recently, an approach for developing a primary standard for 90 Ylabeled resin microspheres was described [92]. This approach involves the complete dissolution of the 90 Y-labeled resin microspheres within the source vial in order to obtain a more homogeneous 90 Y activity distribution followed by primary measurement of the triple to double coincidence ratio (TDCR) of the sample using both Cerenkov and liquid scintillation detection techniques. The goals for the MetroMRT project as well as other future collaborations will be to develop and validate new approaches for accurately calibrating, assessing, quantifying, and verifying patient dosimetry related to targeted molecular radiotherapy. Such approaches that are ultimately traceable to a primary standard will enable more accurate individual patient dosimetry.
Recognizing and addressing the challenges for multimodality 90 Y imaging will impact future prospective clinical trials which investigate the efficacy and safety of new 90 Y theranostics. The long-term value for improved qualitative and quantitative 90 Y imaging will be in confirming targeted delivery of the theranostic agent, evaluating nontarget radioactivity, estimating the absorbed dose to the target lesion(s) and nontarget tissue(s), evaluating and predicting treatment response, assessing the predictive power of existing non-90 Y surrogate imaging agents, and promoting personalized medicine. 90 Y is a theranostic agent which has been used clinically for direct radiation therapy, RIT, PRRT, and RE but it has been and remains a challenging radiotracer in terms of conventional nuclear medicine imaging approaches. The utilization of 90 Y targeted radiotherapies is anticipated to increase. There is continued interest in developing and validating noninvasive imaging strategies to assess both targeted 90 Y radioactivity and nontargeted 90 Y radioactivity that are readily accessible, easy to implement, easy to interpret, and reported in a concise and consistent manner. In general, the 90 Y imaging approaches discussed in this review are compatible with a theranostic paradigm [93]. Intraprocedural and postprocedural imaging can assess the adequacy of targeted 90 Y delivery and provide absorbed dose estimates for the target(s) and nontarget tissues. These novel imaging approaches have the potential to further improve the efficacy of targeted 90 Y radiotherapies, provide objective treatment monitoring and assessment, and ensure patient safety. Further innovations in qualitative and quantitative nuclear medicine imaging of 90 Y radioactivity will continue to impact posttherapy patient management in this era of personalized medicine. The potential for optical imaging of 90 Y radioactivity in vitro and in vivo (and potentially ex vivo) using Cerenkov luminescence may promote more timely and cost-effective preclinical development of targeted theranostics. Clinical and interventional applications for 90 Y CLI are also likely to evolve.