SPECT/CT: Standing on the Shoulders of Giants, It Is Time to Reach for the Sky!

High-Level Features Coming Together

From approximately 2010 onward, high-end multislice CT variants have been part of all SPECT/CT portfolios. At the same time, iterative reconstruction techniques for CT were gradually introduced, reducing ionizing radiation doses up to 80% without loss of image quality and solving, in part, one of the initial barriers in the implementation of SPECT/CT (18). Around 2013, SPECT/CT as a hybrid modality had come of age, available with a full set of technical capabilities (enabling correction for attenuation, scatter, partial volume, and motion). This brought the potential for robust absolute quantification to clinical practice (Fig. 4), with reported accuracies for ^99mTc imaging to within ±5% of the true radionuclide concentration (19,20). Ironically, this technologic progress was briefly overshadowed by a global ^99mTc shortage that created doubt on the viability of SPECT/CT once more, as further discussed below.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 4.

Summary of approximate time of introduction of milestones in whole-body SPECT/CT technology, showing incorporation of subsequent innovations in CT and SPECT technology. Although some novel CT technologies were quickly incorporated into SPECT/CT systems (e.g., iterative reconstruction), others showed significant delays in their translation into hybrid scanners. For SPECT innovations, it is interesting to note parallel development of dedicated cardiac gantry designs that saw earlier introduction of breakthroughs, such as CZT detectors, compared with whole-body capable devices. DECT = dual-energy CT; IR = iterative reconstruction; MAR = metal artifact reduction.

Precision Medicine as Breakthrough Application

A renewed interest in radionuclide treatments using ⁹⁰Y and ¹⁷⁷Lu in oncology under the paradigm of precision medicine (popularized in nuclear medicine using the term theranostics) strongly relies on molecular tumor characterization and individualized treatment planning. Although ⁹⁰Y has been one of the main radionuclides used for therapy, it has recently been surpassed by ¹⁷⁷Lu in theranostics applications. The separate γ- and β-decay modes of ¹⁷⁷Lu allow for imaging of the activity distribution and calculation of the dose distribution. It has now been recognized that individualized dosimetry is favored over fixed, or patient-weight– or body-surface-area–based, administered activities. Individualized dosimetry requires absolute activity measurements in order to deliver target doses that are as high as safely attainable and is possible only after correcting for degrading physical phenomena such as photon scatter, photon attenuation, partial-volume effect, and detector dead time (21). Increasingly accurate correction techniques and iterative reconstruction methods have improved the capabilities of modern SPECT/CT systems to achieve high-level quantitative images, comparable to PET/CT.

Tools to process an image in activity per unit volume have now become available, with acceptable accuracies in most organs and moderate-sized tumors. The in vivo accuracy of solid-state SPECT/CT quantification of ¹⁷⁷Lu-prostate-specific membrane antigen was recently reported to be within 20%, with challenges remaining in quantifying smaller structures (22). It is expected that quantification and dosimetry applications will drive further applications of SPECT/CT, as seen in literature citations today (Fig. 2A).

Leveraging the Full Diagnostic Power of SPECT/CT

The improved capabilities of the embedded CT systems are benefitting the use of SPECT/CT in, for example, bone disease (Fig. 2A). In addition, the novel multimodal SPECT reconstruction techniques can improve image resolution and have shown benefit in assessing uptake in small osseous structures (23). The availability of metal artifact reduction algorithms to reduce the degradation of CT images caused by photon-deprivation and beam-hardening artifacts has further improved the image quality of SPECT/CT. Another enhancement now available on SPECT/CT is dual-energy CT, which uses virtual monochromatic reconstructions for high-energy photons (140 keV) that are less susceptible to metal artifacts (24). These breakthroughs have improved, among others, the assessment of painful arthroplasties, a diagnostic challenge increasing in prevalence due to changes in demographics. Moreover, SPECT/CT is less susceptible than MRI to image-degrading artifacts in patients after arthroplasty, even with the recent improvements in MRI metal artifact–reducing sequence technology.

The painful total-knee arthroplasty has been a successful model for developing the clinical application of high-end bone SPECT/CT. A consistent and relevant impact on patient management has been demonstrated, with the potential for health-care cost savings, making SPECT/CT part of the routine diagnostic algorithm for patients with pain after primary total-knee arthroplasty according to some leading groups (25,26). Exciting results are emerging in other postoperative skeletal settings, including the spine, the hip, and hand or foot pain as well (27–29). Further progress will require novel radiopharmaceuticals offering improved pharmacokinetics (e.g., next-generation radiobisphosphonates) or visualizing specific causes such as infection (e.g., ^99mTc-UBI-29-4), as further discussed below.

Despite the once gray-sky scenarios predicted by some for SPECT/CT imaging, current trends show a strong growth in the acceptance of state-of-the-art SPECT/CT devices. Recent data from Europe show a 22% increase in the number of installed SPECT/CT scanners in France from 2015 to 2018, with similar data from the United Kingdom and Germany available (30). Although SPECT/CT has come of age and is a reliable hybrid imaging modality ready for prime time, some unique challenges will need to be addressed. In particular, there are challenges regarding the use of SPECT/CT in young patients (e.g., for the assessment of low back pain)—an application that requires even more attention to exposure to ionizing radiation and regional issues with reimbursement pending further evidence of cost effectiveness.

Innovation in Detector and Collimator Design

Over the last few years, digital solid-state SPECT detector technology has become commercially available. Cadmium-zinc-telluride (CZT) detectors have excellent count sensitivity, system resolution, and energy resolution, enabling significant reductions in administered activities or acquisition time, as well as facilitating dynamic SPECT. Their superior energy resolution also facilitates dual-isotope imaging through improved separation of the photon peaks in the detected energy spectrum. The higher stopping power, on the other hand, allows for thinner detectors and a higher imaging resolution, on the order of 2.5 mm or even the submillimeter range for preclinical systems (31). Although these capabilities have been known for many years, issues such as slow and small crystal growth and crystalline defects hampered its cost-effective large-scale production—issues that have been solved only recently (32).

CZT itself is not a prerequisite of higher-resolution systems and has intrinsic detector efficiencies similar to traditional systems. Yet, the smaller footprint and lighter weight of these detectors facilitated the construction of gantry designs optimized for specific anatomic regions, focusing primarily on cardiac imaging. The improved spatial resolution and sensitivity with novel detector arrangements in CZT cardiac cameras (33) were facilitated by the introduction of more sophisticated reconstruction algorithms (e.g., Bayesian reconstruction) that could incorporate prior information about the expected characteristics of the image (34). For example, in one comparison between 2 dedicated cardiac CZT SPECT systems and the conventional dual-head Anger camera, both the central spatial resolution (6.7–8.6 vs. 15.0–15.3 mm) and the count sensitivity on cardiac phantom images (460–850 vs. 130 counts s⁻¹ MBq⁻¹) were superior with the CZT systems (33). At present, this technology is being incorporated into general-purpose dual- and multihead SPECT/CT system designs (Fig. 4) (35,36). Most importantly, reducing image acquisition times and lowering the injected radionuclide activity are now possible for cardiac studies (37) and are being explored for a wider range of nuclear medicine studies.

The availability of compact detectors with higher intrinsic spatial resolution has enabled the implementation of novel collimator designs in the clinic (37,38). This has been made possible by the use of tungsten and 3-dimensional printing techniques, delivering higher stopping power, smaller collimator bores, and reduced septal thickness. These new materials make possible new collimator geometries that, until now, had been beyond reach or impractical to implement on a large scale. Examples include minifying multiple-pinhole collimators (allowing more projections per detector) and (multiple-slit) slit-slat collimators (combining the advantages of pinhole and parallel-hole collimators). Both these approaches may be the basis for developing fully stationary detector systems, which could represent the next revolution in SPECT gantry design. The resulting clinical SPECT images achieve resolutions down to 3 mm, with exciting preliminary clinical data in brain and cardiac imaging (39). However, the technique has not yet been implemented in hybrid SPECT/CT system designs.

The Next Major Leap Forward

A true paradigm shift in SPECT/CT design was unveiled in 2017, when the first 360° ring-shaped gantry was introduced, equipped with 12 CZT-based elongated detectors that can move inward and outward to come as close to the patient as possible. These detectors are equipped with novel tungsten parallel-hole collimators and move in a swiveling motion to provide a unique scanning geometry (35). Although the clinical experience with this new device is still limited and the additional cost associated with this new design remains to be justified, preliminary results in bone imaging show that a whole-body SPECT acquisition of 20 min offers significant improvements in sharpness and contrast and can replace current bone-scanning protocols (40). This advance is also signaling an important and inevitable milestone in nuclear medicine practice: the end of the planar imaging era. Indeed, now is the time to let go of the paradigm of the hybrid γ-camera as the jack of all trades, master of none.

Need for Standardization

The recent technologic progress has created considerable variability among the available SPECT/CT systems, with respect to not only their physical designs but also the algorithms used for image correction and reconstruction. Such discrepancies may result in inconsistent or inaccurate results, limiting reproducibility, as seen early on with PET (41). Only rigorous standardization provided the solution to achieve reproducible, consistent, and accurate quantification. This required a common standard for quality control and quality assurance, as outlined in protocols by North American and European professional societies and groups (42,43).

Multiple collaborative groups have started ambitious projects aimed at standardizing quantitative SPECT/CT and creating protocols to determine the accuracy and uncertainties of specific dosimetry platforms (11,44). However, these have been hampered by a tendency to sacrifice user control of application settings and transparency in closed-source workstation solutions using proprietary algorithms in favor of easy-to-use interfaces. Other issues that complicate standardization are the increasing design differences between systems, the need to define the acceptable minimum level of image quality and, for the purpose of image quantification, the need to validate the use of (often small) phantoms as accurate surrogates of patient geometry.

These observations should prompt a call to action from major professional societies and international bodies such as the Society of Nuclear Medicine and Molecular Imaging, the European Association of Nuclear Medicine, and the International Atomic Energy Agency to launch initiatives supporting the standardization of SPECT/CT quantification. Indeed, standardization and traceability are key requirements to harness the full potential of theranostic applications and provide truly individualized and high-quality patient care.

Strength and Challenges of SPECT Radionuclides

Improvements in SPECT/CT technology alone are not enough, as further adoption of the technique is critically linked to the development of novel SPECT radiopharmaceuticals. These remain highly attractive tools for imaging both in clinical practice and in preclinical research. A multitude of single-photon emitters is available (e.g., ^99mTc, ¹¹¹In, and ¹²³I) with half-lives longer than those of commonly used PET radionuclides, facilitating their distribution to more remote locations. Also, SPECT tracers are relatively inexpensive, often being available at a tenth of the cost of PET agents. Today, ^99mTc is used in approximately 85% of all nuclear medicine diagnostic procedures, accounting for 30 million examinations worldwide every year (45). However, its production depends on the use of highly enriched uranium targets in nuclear research reactors, making the global supply chain complicated and particularly vulnerable to reactor shutdowns. First in 2009, and then in 2012 and 2013, planned and unplanned shutdowns have caused extended global shortages. To safeguard future supply, providers have been encouraged to recover their full costs in order to support the transition to low-enriched-uranium targets and investments in sufficient production and reserve capacity in case of unplanned outages. At the same time, alternative or supplementary production technologies are also being developed, including neutron activation of ⁹⁸Mo targets in research reactors, ⁹⁹Mo production with linear accelerators, and direct production of ^99mTc on cyclotrons (46).

Beyond ^99mTc, SPECT radionuclides with an extended half-life (i.e., ¹¹¹In and ¹²³I) enable the use of lengthier and more complex radiolabeling procedures, as well as longer imaging protocols. A typical example is infection imaging using ¹¹¹In-labeled white blood cells, a modality that benefits from delayed acquisitions to improve specificity (47). A unique strength of SPECT is dual-isotope imaging, which enables simultaneous detection of tracer distribution in space and over time in a single imaging session, thus reducing acquisition times (and therefore patient discomfort), image artifacts, and quantitation errors due to patient motion. This approach is particularly well suited for the study of interconnected functional, metabolic, chemical, and biologic processes. This methodology was recently used to study Takotsubo cardiomyopathy, showing that dual-isotope SPECT with the perfusion tracer ^99mTc-MIBI and the fatty-acid metabolism imaging agent ¹²³I-BMIPP (¹²³I-betamethyl-p-iodophenyl-pentadecanoic acid) combined with cardiac CT could visualize the typical changes in both processes caused by vasospasms in the distal left anterior descending coronary artery (48).

Innovation in SPECT Tracers for Imaging and Treatment

Therapeutic drug development is targeting ever more specific pathways or processes using small-molecule inhibitors or monoclonal antibodies. Ideally, tracer development can mirror these versatile approaches to derive predictive diagnostic radiopharmaceuticals or to exploit the highly specific targeting properties of these compounds to produce radionuclide therapies.

The small-molecule prostate-specific membrane antigen inhibitor radiotracers have become an important class of compounds in the management of prostate cancer. One of the first compounds of this class to be developed was the radioiodinated (¹²³I) tracer MIP-1095, which has successfully been used for targeted tumor therapy (49). Uniquely, MIP-1095 can be used as a theranostic agent for SPECT (¹²³I) and radionuclide therapy (¹³¹I). In addition, the ^99mTc-analog MIP-1404 (also known as ^99mTc-trofolastat), prepared using the technetium tricarbonyl core, has been proposed for SPECT/CT imaging, has shown good tumor targeting ability and favorable biodistribution, and has recently entered a phase 3 study to support possible market introduction (50).

For the development of monoclonal antibody or peptide-based imaging agents, nonmetallic radionuclides such as ¹²³I have frequently been labeled using direct electrophilic aromatic substitution on histidine or tyrosine amino acid residues. Unfortunately, the required labeling conditions are often too harsh for antibodies and may damage their structure and affinity. This problem has recently been overcome by the introduction of reactive prosthetic groups, such as the novel radioiodinated BODIPY dual functional agent, allowing labeling under milder conditions. Proof-of-concept data demonstrated successful SPECT imaging with ¹²³I-trastuzumab of HER2-positive tumors (51).

For metallic SPECT radionuclides such as ^99mTc and ¹¹¹In, bifunctional chelating groups have been the cornerstone for labeling antibodies and peptides. The versatile chemistry of ^99mTc, with its stable and readily accessible oxidation states, is characterized by chemically robust core structures that can be exploited as platforms for radiopharmaceutical design. One such well-known approach uses the ^99mTc(V)-organohydrazino (HYNIC) core to derivatize a wide range of different targeting molecules (52). A promising example is the development of ^99mTc-duramycin as a SPECT tracer for apoptosis, enabling imaging of the induction of cell death early after chemotherapy and radiotherapy. Providing a very early readout of treatment effect, it could potentially be used to tailor cancer treatments more quickly, sparing toxicity and cost in nonresponding patients (53). Another interesting target in oncology and cardiovascular medicine is angiogenesis, with ^99mTc-HYNIC-l(arginine-glycine-aspartate) or ^99mTc-maraciclatide integrin α_vβ₃ imaging demonstrating the ability to visualize tumors and treatment response in cancer models (54). Integrin imaging has also been used to visualize inflamed atherosclerotic plaques, with the potential to noninvasively identify high-risk plaques in patients (55). Finally, expression of the c-Met receptor in non–small cell lung cancer could recently be demonstrated with ^99mTc-HYNIC-c-Met–binding peptide SPECT, providing essential information on a signaling pathway that contributes to cancer progression and may mediate acquired resistance to epidermal growth factor receptor–targeted therapy (56). A multitude of SPECT/CT applications and their potential advantages can be envisioned, targeting a wide range of important predictive tumor patterns aiding treatment selection and available in convenient kit formulations.

A drawback of the HYNIC platform is that its chemistry is often complex, hampering clinical translation because of regulatory issues. A possible solution is the introduction of the organometallic [^99mTc(CO)₃]⁺ core. It is a versatile building block, and the immediate water-soluble precursor [^99mTc(CO)₃(H₂O)₃]⁺ can easily be obtained from homemade or commercially available kits. A major advantage of labeling with the ^99mTc-tricarbonyl core is the very wide variety of ligands that bind efficiently to Tc(I), forming highly robust complexes. Recently, fibroblast-activating protein inhibitors were successfully labeled using this method, enabling SPECT/CT imaging of cancer‐associated fibroblasts, present in many malignancies. Because fibroblast-activating protein is typically not expressed in healthy tissues, it may represent an important target for theranostic applications (57).

Single-domain antibodies have also proven to be particularly suited for ^99mTc labeling via tricarbonyl chemistry, with many exciting possible applications, including the noninvasive quantification of immune checkpoint (programmed death-ligand 1) expression in tumors to guide immunotherapy (58).

Progress is also being made in improving surrogate radionuclides used for pretherapy SPECT/CT imaging and dosimetry studies. Although ¹¹¹In is still frequently used for this purpose, its chemistry is slightly different from that of the radiolanthanides typically coordinated in these ligands (e.g., ¹⁷⁷Lu). This difference may introduce errors in assessing the tissue distribution and could be solved using the radiolanthanide ¹⁵⁵Tb (59). The administration of ¹⁵⁵Tb alongside a therapeutic terbium isotope, such as ¹⁴⁹Tb or ¹⁶¹Tb, would represent the ideal theranostic pair because of the identical chemical properties. Similarly, the cyclotron-produced SPECT radioisotope ²⁰³Pb is attracting attention as the imaging surrogate for the therapeutic ²¹²Pb, as both can be directly imaged using SPECT/CT (60). These new diagnostic–therapeutic pairs may greatly facilitate the calculation of the dosimetry of α- and β-emitter–labeled radiopharmaceuticals. It is clear that recent innovations in radiopharmaceutical and chemistry techniques have created a promising and versatile landscape of powerful labeling methods that show great opportunity in developing the next generation of SPECT/CT radiopharmaceuticals.