Redefining PET Imaging Through Nuclear Properties, Production Technologies and Scalability of Diagnostic Radionuclides

Abstract

This review provides a critical and forward-looking analysis of established PET positron-emitting radionuclides—¹¹C (carbon-11),¹³N(nitrogen-13), ¹⁵O(oxygen-15), ¹⁸F(fluorine-18), ⁶⁸Ga (gallium-68),⁸²Rb(rubidium-82)—alongside some less widely adopted positron emitters—⁴⁴Sc (scandium-44), ⁶⁴Cu (copper-64), ⁸⁶Y (yttrium-86), ⁸⁹Zr (zirconium-89), ¹²⁴I(iodine-124)—examining the scientific, technological and operational factors influencing their clinical translation and applicability. Particular emphasis is placed on the role of nuclear properties as a key factor in radionuclide selection and development. For each radionuclide, the relevant aspects, including nuclear decay characteristics, production routes and logistical modalities, are discussed in terms of their impact on PET diagnostic performance and sustainability. The review summarizes recent technological advances designed to mitigate supply chain limitations that affect established positron emitters and discusses critical challenges related to other promising PET radionuclides, such as production scalability and dosimetric implications. Finally, ongoing developments in hybrid imaging platforms and multiparametric PET systems are briefly addressed, illustrating how these innovations are redefining diagnostic accuracy and accelerating the evolution of PET toward increasingly personalized clinical strategies.

1. Introduction

Molecular imaging plays a central role in modern diagnostic medicine, enabling the non-invasive visualization and quantitative assessment of biological processes in vivo [1]. Among current imaging modalities, positron emission tomography (PET) has emerged as a pivotal technique, offering unparalleled sensitivity and the ability to provide quantitative functional information [2,3]. PET imaging relies on the administration of radiotracers, i.e., biologically active molecules labeled with positron-emitting radionuclides. These compounds are designed to participate in specific biochemical processes, enabling the reconstruction of three-dimensional images that characterize the spatial and temporal distribution of the tracer and depict the associated molecular activity.

Over the past decades, PET has experienced continuous growth, driven by advances in detector technology, image reconstruction algorithms and radiopharmaceutical development [4]. By enabling the accurate in vivo evaluation of metabolic activity, receptor expression and molecular pathways, PET has become an essential tool in both medical practice and biomedical research, facilitating early disease detection, characterization and therapy monitoring across a wide range of clinical applications [5].

In addition to its high sensitivity, PET offers the key advantage of enabling quantitative measurement of tracer uptake, most commonly expressed as standardized uptake values (SUVs), even if other quantitative metrics are also widely used depending on the application, such as total volume of distribution (VT) and binding potential (BP), particularly in brain PET studies. Accordingly, in recent years, diagnostic PET imaging has expanded its clinical impact well beyond traditional oncological applications [6] to include cardiology, neurology, inflammation and infection imaging, as well as emerging fields such as immuno-imaging and cell tracking [7].

The use of PET techniques in combination with more conventional imaging modalities, such as Computed Tomography (CT) [8] and Magnetic Resonance Imaging (MRI) [9], offers the potential to further improve lesion localization and diagnostic accuracy. Through the quantitative integration of complementary molecular, morphological and functional data within a single acquisition, multiparametric imaging may enable detailed insights into tumor pathophysiology, receptor expression, metabolism and microenvironmental features of diseased tissues [10]. It also would allow the assessment of pharmacodynamic drug responses and precise characterization of spatial and temporal tumor variability, helping to identify potential therapeutic targets. However, these capabilities remain only partially exploited in current clinical practice, where image evaluation still relies on a limited set of parameters. At the same time, the extensive experience gained from PET/CT clinical studies is supporting the development of total-body PET/CT systems, which will enable more comprehensive and dynamic imaging across the entire patient.

Recently, positronium imaging has been proposed as a novel PET-based approach based on the properties of positronium atoms formed in biological tissues. These properties, sensitive to the size of intramolecular voids and local oxygen concentration, provide complementary information beyond conventional anatomical, morphological and metabolic imaging. In particular, mean ortho-positronium lifetime imaging has attracted increasing interest, with the first ex vivo and in vivo human studies demonstrated using dedicated systems such as the Jagiellonian Positron Emission Tomograph [11]. Such technology enables the simultaneous detection of annihilation photons and prompt γ-rays from β⁺/γ emitters for the reconstruction of positronium formation and decay dynamics.

Beyond multiparametric approaches, PET is increasingly undergoing a shift to a more quantitative approach through radiomics, which represents a natural evolution of molecular imaging toward data-driven analyses [12]. Radiomics is based on the high-throughput extraction of a large number of images that quantitatively capture tumor heterogeneity, texture, shape and intensity patterns, transforming medical images into high-dimensional datasets suitable for advanced computational analysis. This concept reflects the shift in molecular imaging toward integrative and computational methodologies, where images are combined with clinical and genomic data to obtain a comprehensive disease characterization [13]. By converting images into high-dimensional datasets, PET radiomics represents a key contributor to the clinical translation of precision medicine, although its implementation still faces challenges mainly related to validation and standardization across the many different imaging platforms [14].

Taken together, the expanding methodological approaches and the increasing technological diversifications provide important insights into the strengths and limitations of radionuclides for specific diagnostic applications, emphasizing the growing dependence of imaging performance on radiotracer properties. This has renewed interest in both established and emerging PET radioisotopes and underscores the need to reassess radionuclide selection.

In fact, although PET imaging is often discussed in terms of radiopharmaceutical design and biological targeting, the diagnostic performance of this modality is largely determined by the choice of radioisotope. The nuclear properties of the employed radioisotope remain a determining factor directly influencing image quality, patient dosimetry and compatibility with targeting vectors used for radiolabeling in both traditional and next-generation clinical applications. Therefore, assessing candidate radioisotopes for PET involves not only their performance in preclinical and clinical studies, but also a comprehensive analysis of their nuclear characteristics, production feasibility and radiochemical behavior.

Beyond the well-established radionuclides currently employed in routine clinical practice, a broad spectrum of alternative radionuclides has been explored in recent years. Such initiatives are motivated by constraints in current supply chains, the need for nuclides better suited to novel biological targets and the expanding interest in personalized and theranostic strategies [15]. Concurrently, improvements in accelerator technology, reactor-based production methods and target design are facilitating the investigation of non-conventional radionuclides, whose nuclear characteristics can provide unique benefits for specific imaging applications or combined diagnostic–therapeutic use.

This paper reviews the PET radioisotopes proposed and investigated for diagnostic applications, with a specific emphasis on the underlying nuclear properties and their impact on production routes and imaging performance. In contrast to the existing literature, this review highlights the often underexplored role of nuclear characteristics in guiding radionuclide selection and development. The aim is to provide a structural framework supporting the rational design of diagnostic radionuclides for current and future PET imaging applications.

The preparation and radiolabeling of radiopharmaceuticals, including the design and synthesis of targeting vectors, are deliberately not illustrated here, as they involve complex multidisciplinary aspects spanning radiochemistry, pharmaceutical formulation, bioconjugation chemistry and quality control, which are beyond the scope of the present review. For comprehensive discussions of radiopharmaceutical preparation and vector design, readers are therefore referred to specialized textbooks and reviews [16,17].

2. Fundamentals of Pet Imaging

From a physics perspective, the selection of PET radionuclides is governed by key nuclear properties directly relevant for medical imaging, namely the decay mode, the half-life and the energies of the emitted particles and photons [18].

The decay mode describes the specific nuclear transformation by which an unstable nucleus releases energy. For PET radionuclides, this corresponds to β⁺-decay (positron emission), which enables the production of annihilation photons used for the image. In some radionuclides, β⁺-decay may compete with electron capture (EC), which represents an alternative decay pathway.

The physical half-life (t_1/2) is defined as the time required for half of the radioactive nuclei in a given sample to decay. This parameter determines, for each PET radionuclide, the temporal window available for radiopharmaceutical synthesis, biological uptake and image acquisition.

The energies of the emitted β⁺-particles, together with those of any associated γ-photon emission, if present, directly affect tissue penetration, spatial resolution, detection response and patient safety. Therefore, such energies play a critical role in imaging performance and dosimetric considerations. It should be noted that β⁺-emissions are not monoenergetic, but show a continuous kinetic energy distribution extending from zero to a defined maximum (E_max), with E_ave representing the average energy of the emission.

The PET imaging technique exploits the properties of positrons emitted by radiotracers to probe specific metabolic and biochemical pathways [19]. Following radioactive β⁺-decay, the emitted positron travels a short path in tissue, typically on the order of a few millimeters, before annihilating with an electron and generating a pair of 511 keV γ-photons emitted at approximately 180°. Coincident detection of the two annihilation photons by a ring of scintillation detectors surrounding the patient allows localization of the annihilation event along the detector line of response.

Modern PET systems employ arrays of pixelated high-density scintillation crystals, such as LSO (Lutetium oxyorthosilicate) or LYSO (Lutetium–Yttrium oxyorthosilicate), coupled to photodetectors to efficiently convert γ-photons into electrical signals [20]. The high stopping power, fast decay and high light yield of such scintillators enable efficient detection and good timing resolution. On the other hand, optical readout performed using either conventional vacuum photomultiplier tubes (PMTs) or solid-state silicon photomultipliers (SiPMs), the latter particularly suitable for time-of-flight PET, ensures high timing performance and compact design. Recent advances in multiphoton PET scanners have further enhanced these capabilities by enabling the coincident registration of the two 511 keV annihilation photons and of any associated prompt γ-radiation [21]. This emerging technology further facilitates multi-tracer imaging approaches and supports the development of novel techniques such as positronium imaging [11].

The positron energy is an important parameter because it determines the range of the particle and its linear energy transfer (LET) in tissue. Higher positron energies correspond to longer ranges and lower LET, while lower energies lead to more localized energy deposition with higher LET. For PET imaging, β⁺-emitting radionuclides with appropriate half-lives and low energies are preferred to optimize spatial resolution and quantitative accuracy, while minimizing unnecessary dose. Additionally, radionuclides without concomitant prompt γ-ray emission are preferred to reduce background and to avoid additional radiation dose to the patient.

However, the non-monochromatic nature of β⁺ emission, combined with its low LET in tissue (typically <1 keV/μm), intrinsically limits the spatial resolution of PET. During their propagation, positrons deposit energy along their tracks in tissue, a process which can be described in terms of LET. The continuous positron energy spectrum results in a distribution of positron path lengths before annihilation, while the low LET leads to a relatively diffuse energy deposition, affecting detector response and dosimetric modeling.

These factors collectively may degrade image quality and quantitative accuracy, making sophisticated image reconstruction techniques necessary. PET images are therefore reconstructed from a large number of collected coincidence events using analytical or iterative tomographic algorithms. To enhance spatial localization and signal-to-noise ratio, modern PET systems often integrate coincidence data with time-of-flight (TOF) information for attenuation and scatter corrections [22].

3. Production Facilities for Positron-Emitting Radionuclides

Although PET technology originated in the late 1970s, it became a widely feasible procedure only in the 1990s, owing to technological advances and the commercialization of nuclear facilities suitable for the production of the radionuclides essential for PET imaging [23].

Positron-emitting radionuclides for PET imaging can be obtained by using several types of nuclear infrastructures, each characterized by specific capabilities and limitations in the production of different radionuclides [24]. Overall, the choice of production infrastructure depends on the radionuclide of interest, the required molar activity, the desired production scale and the intended clinical or research application [25].

Historically, many radionuclides for nuclear medicine, including several for PET applications, have been produced in research reactors by means of neutron capture or (n,p) reactions in appropriate target materials. Reactor production generally results in lower radionuclidic purity compared to other routes, such as the cyclotron one, due to the presence of co-produced radionuclides in the neutron-irradiated target. Nuclear reactors are presently employed for radionuclides that are difficult to generate efficiently via charged-particle reactions, or when large-scale batch production is required. Examples include the relatively long-lived radionuclides ⁶⁴Cu and ¹²⁴I [26]. Indeed, the high-yield, low-cost and easy target-preparation reactor-based method is not suitable for very short-lived radionuclides, due to significant decay losses occurring during transport from centralized production facilities to hospitals and imaging centers. Moreover, the current supply of PET nuclides still largely relies on research reactors built in the mid-20th century and most of these facilities are expected to cease operation before 2030.

Altogether, these considerations highlight the limitations associated with reactors in the supply of positron emitters and motivate the growing shift toward cyclotron-based production methods [27].

Cyclotrons play a central role in the production of PET radionuclides, enabling the routine production of fundamental positron emitters such as ¹¹C, ¹³N, ¹⁵O, ¹⁸F and ⁶⁸Ga [28].

A cyclotron is a circular particle accelerator that uses a static magnetic field and a radiofrequency oscillating electric field to accelerate charged particles, typically protons (p) and occasionally deuterons (d), to energies in the 10–30 MeV range, thereby inducing nuclear reactions in solid, liquid or gaseous targets. Cyclotrons provide high specific activity radionuclides and no-carrier-added products, offering at the same time broad flexibility in target material selection. In order to further increase production yields and specific activity, isotopically enriched target materials are commonly used.

Modern cyclotrons operating in the 16–18 MeV proton energy range are sufficiently compact to be installed in hospitals or regional production centers, allowing on-site generation of clinically required tracers. This configuration makes cyclotron-based production particularly well-suited for the synthesis of short-lived radiopharmaceuticals in both hospital-based and commercial facilities. According to IAEA [29], there were over 1300 cyclotron facilities operating worldwide in 2023. In Europe alone, approximately 348 cyclotrons are operational, many of which are optimized for routine medical isotope production [30]. The global number of cyclotrons continues to grow, particularly in developing regions [31], driven primarily by the increasing deployment of compact medical cyclotrons that facilitate on-site PET radionuclide production [32].

PET radionuclides can also be produced using linear accelerators (Linacs), devices that accelerate charged particles to high energies by means of oscillating electric fields. In PET technology, Linacs are primarily employed to generate high-purity, short-lived positron-emitting radionuclides through the irradiation of suitable target materials by precisely controlled proton beams. This approach is particularly advantageous for the production of very short-lived isotopes, such as ¹⁵O and ¹³N, as well as of radionuclides that require particle energies higher than those provided by conventional cyclotrons [33]. While Linac-based production is less common in routine clinical PET practice, it plays a significant role in research settings and in large-scale radionuclide generation.

In addition to direct charged-particle reactions, Linacs can also induce photonuclear reactions. These are nuclear processes triggered by bremsstrahlung γ-rays that are generated when high-energy electron beams impinge on high-Z converter targets. The photonuclear approach offers an alternative route, expanding the range of producible radionuclides and enabling access to radionuclides that may be difficult to obtain through conventional cyclotron-based methods [34]. It offers also other potential advantages, including flexibility in target materials and the utilization of existing electron accelerator infrastructures. However, photonuclear reaction cross-sections are generally lower than those of charged-particle-induced reactions, leading to reduced production yields and low specific activities. The achievement of clinically relevant activities requires high electron beam power, which increases energy consumption and overall infrastructure demands. Consequently, at present, photonuclear production is considered a complementary pathway rather than a competitive alternative to conventional cyclotron-based PET radionuclide production [35].

Recently, innovative methods for the production of medical radionuclides have been proposed based on high-intensity, high-repetition-rate laser systems. Batani et al. [36] demonstrated that laser-driven processes can generate p- and α-particle beams capable of inducing nuclear reactions in suitable targets, leading to the production of important PET nuclides such as ¹¹C, ¹⁸F and ⁴⁴Sc. In their initial experiments, yields on the order of 10⁶ atoms of ¹¹C and 10⁴ atoms of ⁴⁴Sc per laser shot were reported [36].

The laser-based approach offers the potential for much more compact and modular systems compared to conventional cyclotrons and reactors, which could facilitate the localized generation of short-lived PET radionuclides directly in medical centers. However, current yields are far below those achievable with cyclotrons or nuclear reactors, whereas technical challenges related to target engineering, beam flux optimization and radioisotope purity remain significant.

In summary, among infrastructure-based production techniques, cyclotrons currently dominate clinical PET radionuclide production due to their flexibility, reliability and high specific activity, whereas nuclear reactors and linear accelerators complement this capability for specialized isotopes or research-scale production [25]. Laser-driven systems could represent a highly innovative approach with the potential to decentralize radionuclide generation, but they are still at an early research stage.

An infrastructure-independent, flexible solution for the on-demand continuous production of short-lived isotopes in decentralized settings is provided by radionuclide generators. First proposed by G.I. Gleason in 1960 [37], these systems rely on a long-lived parent radionuclide immobilized on a solid matrix. The parent nuclide is typically produced at centralized infrastructures, incorporated into generators and then shipped to end PET users [38]. Through radioactive decay, the parent continuously generates a daughter radionuclide, which can be selectively eluted when needed.

The radionuclide generators, especially advantageous for short-lived radionuclides, enable repeated elutions over time, ensuring a reliable on-demand supply of the specific positron-emitting nuclide. This configuration offers portability, predictable activity and ease of use in clinical settings, making it particularly valuable for hospitals or centers lacking direct access to production facilities. Modern generator systems are integrated with automated infusion and quality control modules, which guarantee radiochemical purity, sterility and accurate dosing, enabling rapid and direct delivery to the patient.

An example of radionuclide generators widely used for PET is the ⁶⁸Ge/⁶⁸Ga system, which enables a routine availability of gallium-68 without the need for an on-site cyclotron. Other generator systems used for medical applications include ⁸²Sr/⁸²Rb, ⁹⁹Mo/^99mTc, ⁹⁰Sr/⁹⁰Y, ¹⁸⁸W/¹⁸⁸Re and the newly proposed ⁴⁴Ti/⁴⁴Sc system, which appears particularly promising for hybrid PET/CT imaging [38].

4. Radionuclides for Pet Imaging

PET imaging relies on several positron-emitting radionuclides selected for their decay characteristics, compatibility with the chemical and physical properties of targeting pharmaceutical vectors and suitability for specific diagnostic applications [39].

The radionuclides most commonly used for PET include ¹⁸F, ¹¹C, ¹³N and ¹⁵O, with ⁶⁸Ga and ⁸²Rb also well-established in specific clinical settings. Some other, less widely adopted positron emitters, such as ⁴⁴Sc, ⁶⁴Cu, ⁸⁶Y, ⁸⁹Zr and ¹²⁴I, are gaining increasing attention owing to their favorable physical characteristics and potential for advanced specialized or niche diagnostics.

The following sections provide an overview of the nuclear decay data [40] of the PET radionuclides listed above, presented in order of increasing atomic number (Z), together with their production routes and clinical applications.

4.1. ¹¹C (Carbon-11)

¹¹C is a radioactive isotope of carbon with a half-life of 20.34 min. It decays by β⁺-emission (E_max = 960.4 keV; E_ave = 385.7 keV) to stable ¹¹B. Because carbon is a fundamental constituent of organic molecules, ¹¹C can be incorporated into biological compounds without significantly altering their chemical or biological behavior. This makes ¹¹C greatly valuable for tracing in vivo metabolic pathways, enzyme activity, neurotransmitter systems and receptor expression levels. ¹¹C-labeled tracers enable highly selective and quantitative PET imaging in neurology, oncology, and cardiology. Common examples of radiolabeled ¹¹C-based PET tracer include:

-

[¹¹C]C₁₉H₂₆ClN₃O₂ ([¹¹C]raclopride), a dopamine D₂/D₃ receptor antagonist, largely used for PET imaging of striatal dopamine receptors. It acts as a reversible competitive antagonist, enabling non-invasive quantification of receptor availability, occupancy and dopamine release dynamics in vivo. In addition, the imaging of dopamine receptors by this ¹¹C-based PET radiotracer plays a crucial role in biological psychiatry, providing insights into the pathophysiology of neuropsychiatric disorders such as schizophrenia, including mechanisms related to receptor internalization and the dysregulation of dopaminergic signaling [41];

-

[¹¹C]PiB(Pittsburgh Compound-B), a PET tracer for imaging amyloid-β (Aβ) plaques in Alzheimer’s disease [42]. This thioflavin-T derivative binds with high affinity to fibrillar Aβ aggregates, enabling both qualitative visualization and quantitative assessment of amyloid burden in vivo, and facilitating differential diagnosis of dementia [43];

-

[¹¹C]C₂H₃O₂⁻ ([¹¹C]acetate), used for PET imaging of myocardial metabolism and selected malignancies [44]. It probes oxidative and lipid metabolism in cardiac and oncological applications, with uptake linked to tumor metabolic phenotypes and hypoxia-related molecular signatures [45].

¹¹C is produced in medical cyclotron through nuclear reactions on high-pressure nitrogen gas targets, most commonly via the ¹⁴N(p,α)¹¹C reaction in which high-energy protons interact with N₂ to generate ¹¹C nuclei. Trace amounts of oxygen or hydrogen influence the chemical form of the produced carbon nuclide, yielding [¹¹C]CO₂ or [¹¹C]CH₄, respectively [46]. These primary products serve as precursors for subsequent radiochemical synthesis of a wide variety of ¹¹C-labeled PET radiotracers. This production route is favored because of its high cross-section at clinically accessible proton energies, efficient yields and the relative ease of the target system.

The short half-life of ¹¹C, while advantageous for minimizing patient radiation dose, imposes stringent requirements on rapid isotope production, radiochemical synthesis, quality control and clinical administration. As a result, the use of ¹¹C-labeled radiotracers is confined to research and clinical settings equipped with cyclotrons and necessitates a strong coordination between production, radiochemistry and PET imaging.

4.2. ¹³N (Nitrogen-13)

The positron-emitting ¹³N (E_max = 1197.9 keV; E_ave= 492.2 keV) is a radionuclide with a short physical half-life (t_½ = 9.97 min), therefore suitable for applications which require rapid tracer kinetics and repeated imaging protocols. ¹³N is used in PET in different chemical forms, including [¹³N]NH₃ ([¹³N]-ammonia) and [¹³N]N₂ (¹³N-nitrogen gas). Among these, [¹³N]ammonia is the most widely used clinically, particularly for myocardial perfusion PET imaging, owing to its favorable biokinetic properties and efficient intracellular trapping after metabolic conversion. These characteristics allow high-quality perfusion images and enable quantitative assessment of both myocardial blood flow and flow reserve, providing high diagnostic accuracy for the evaluation of coronary artery disease (CAD) and microvascular dysfunction [47]. It has been demonstrated that [¹³N]ammonia PET/CT myocardial perfusion imaging detects obstructive CAD with high accuracy, demonstrating and specificity of approximately 90% when compared with coronary angiography as the reference standard [48]. In addition to perfusion assessment, [¹³N]ammonia allows high-resolution, quantitative PET measurement of myocardial flow reserve, which improves the detection of significant CAD, even in the presence of normal myocardial perfusion imaging [49]. Moreover, this radiopharmaceutical can also provide deeper insights into myocardial function and microvascular abnormalities [50].

The ¹³N is typically produced in a cyclotron by high-energy (10–20 MeV) proton irradiation of ¹⁶O-enriched water targets via the nuclear reaction ¹⁶O(p,α)¹³N. The produced ¹³N is subsequently processed to form ¹³N-labeled ammonia or other radiotracers suitable for clinical PET imaging.

However, the physical half-life of approximately 10 min requires on-site cyclotron production and rapid radiochemical synthesis, which has historically limited its use to centers equipped with dedicated infrastructure. Recent advances in compact cyclotron technology and fully automated production systems are improving the feasibility of bedside [¹³N]ammonia production, expanding clinical access to this modality. These developments enable high-quality myocardial perfusion imaging with excellent tracer kinetics and the possibility of dynamic quantitative PET studies. Clinical implementation of compact superconducting cyclotron systems has demonstrated the practical feasibility of this approach [51].

4.3. ¹⁵O (Oxygen-15)

¹⁵O is a radionuclide characterized by a very short physical half-life (t_½ =122.27 s), which decays to stable ¹⁵N, emitting β⁺-particles with E_max = 1732 keV and E_ave= 735.3 keV. These relatively high energies result in a longer positron range in tissue (~1.5 mm mean, up to 4–5 mm) compared with ¹⁸F (~0.6 mm), which slightly limits spatial resolution but enables excellent quantitative functional imaging [52]. ¹⁵O is commonly used as a simple physiological tracer that mimics natural substrates, making it particularly suitable for quantitative PET studies. An example is [¹⁵O]H₂O ([¹⁵O]-water), a freely distributed and metabolically inert radiopharmaceutical considered the gold standard radiotracer for quantitative PET imaging of perfusion in heart, brain, and other organs [53,54].

Other major ¹⁵O-labeled radiopharmaceuticals include:

-

[¹⁵O]O₂ ([¹⁵O]oxygen), used to assess oxygen metabolism and the cerebral metabolic rate of oxygen

-

[¹⁵O]CO ([¹⁵O]carbon monoxide, used for blood volume measurements

-

[¹⁵O]CO₂ ([¹⁵O]carbon dioxide), used for ventilation and perfusion studies.

Due to their rapid kinetics, versatile chemistry and very low metabolic trapping, the ¹⁵O-labeled tracers are highly suitable and widely used for dynamic PET imaging and absolute quantification, especially in cardiology and neuroscience [55,56]. They provide a unique tool for cerebral PET imaging, enabling quantitative assessment of brain oxygen utilization through precise measurements of the cerebral metabolic rate of oxygen and oxygen extraction fraction [52]. This approach is particularly valuable for investigating neurovascular disorders and, when integrated with MRI, offers unique insights into stroke and other cerebral pathologies affecting cerebral metabolism [57].

With regard to the ¹⁵O production, it is most efficiently generated in a medical cyclotron via the ¹⁴N(d,n)¹⁵O nuclear reaction using deuteron irradiation of nitrogen gas targets [58]. This reaction has a relatively low threshold energy of approximately 2.4 MeV, and its cross-section increases rapidly with deuteron energy, reaching a maximum in the range of 5–8 MeV, which is well within the typical energy range of clinical cyclotrons [59].

The alternative ¹⁵N(p,n)¹⁵O reaction requires isotopically enriched ¹⁵N targets, because the natural abundance of ¹⁵N is only ~0.37% and the use of natural nitrogen (¹⁴N) leads to substantially lower yields at comparable projectile energies. The need for highly enriched gas targets and gas-recovery systems significantly increases operational complexity and cost, making this route less attractive for routine clinical production [60]. The ¹⁴N(d,n)¹⁵O pathway is therefore the preferred method for routine ¹⁵O production, as it allows the use of natural nitrogen gas targets and provides high yields with moderate beam energies. Nevertheless, deuteron beams are not available on all medical cyclotrons, and proton-induced production via the ¹⁵N(p,n)¹⁵O reaction remains a viable alternative in facilities lacking deuteron beams [31].

A proposed alternative to the conventional cyclotron-based production route is the ¹⁶O(γ,n)¹⁵O reaction driven by an electron linear accelerator (e-Linac). This photonuclear approach can use simple oxygen or water targets and may offer a cost-effective and flexible means of producing ¹⁵O for specialized PET applications. However, the broad bremsstrahlung photon spectrum generated by e-Linacs, together with the relatively low photonuclear cross-section, requires increasing both beam energy and irradiation time to maximize ¹⁵O production while minimizing unwanted contaminants [61].

It should be noted that, due to the extremely short half-life of ¹⁵O, all production routes require rapid and fully automated radiochemical processing to generate the ¹⁵O-labeled pharmaceuticals. In this context, recent methodological advances have introduced effective strategies to minimize the time between their production and administration. For example, Corsaut et al. [62] demonstrated the direct delivery of [¹⁵O]O₂, produced via the ^14N(d,n)¹⁵O reaction in a medical cyclotron, to a hybrid PET/MR scanner with quality control following cGMP standards. Similarly, Iguchi et al. [63] proposed automated radiosynthesis systems enabling sequential production of multiple ¹⁵O-labeled gases, including [¹⁵O]O₂, [¹⁵O]CO and [¹⁵O]CO₂, at time intervals compatible with clinical applications.

Together, these developments demonstrate that automation and direct delivery systems can substantially streamline ¹⁵O workflows, improving the feasibility of dynamic and quantitative PET imaging in clinical settings.

4.4. ¹⁸F (Fluorine-18)

¹⁸F is the most widely used β⁺-emitter for PET imaging, as it best corresponds to the decay characteristics and chemical properties necessary to achieve optimized imaging conditions. ¹⁸F undergoes 100% β⁺ decay and is characterized by a relatively low positron emission energy (E_max = 633.5 keV, E_ave = 249.8 keV). As a result, positron annihilation takes place in close proximity to the emission site, leading to improved PET image resolution. The half-life of 109.77 min allows for straightforward preparation and administration of radiopharmaceuticals without causing excessive radiation exposure to the patient.

Moreover, the unique chemistry of fluorine allows both electrophilic and nucleophilic fluorination strategies, enabling the inclusion of ¹⁸F into a variety of molecular frameworks and driving the development of a wide range of highly specific ¹⁸F-labeled radiopharmaceuticals [64,65].

Among the most widely used ¹⁸F-labeled tracers are 2-[¹⁸F]fluoro-2-deoxy-D-glucose ([¹⁸F]FDG) for glucose metabolism imaging [66], [¹⁸F]NaF for bone imaging [67], 3′-deoxy-3′-[¹⁸F]fluorothymidine ([¹⁸F]FLT) as a marker of cellular proliferation [68] and 3,4-dihydroxy-6-[¹⁸F]fluoro-L-phenylalanine ([¹⁸F]DOPA) for neurotransmitter and neuroendocrine tumor imaging [69].

In non-small cell lung cancer, [¹⁸F]FDG-based radiomics improves the detection of metastases, enables discrimination among histopathological subtypes, characterizes driver mutations and allows assessment of treatment response [70]. Beyond oncology, [¹⁸F]FDG is also employed in advanced PET procedures to investigate cerebral energy metabolism and regional glucose utilization in brain tissue. In this context, dynamic [¹⁸F]FDG PET enables the detection of metabolic alterations associated with neurological disorders [71].

In recent years, novel ¹⁸F-labeled peptides and targeted tracers, such as [¹⁸F]SiTATE for somatostatin receptor imaging, have further broadened the clinical applications of PET imaging [72,73].

The significant expansion of PET imaging beyond its traditional applications is clearly illustrated by the use of [¹⁸F]FDG, which remains the cornerstone radiotracer in clinical oncology and metabolic imaging, also in the cardiovascular field. As reported in [74], the development of novel ¹⁸F-labeled tracers targeting myocardial sympathetic nerve density has demonstrated the feasibility of quantitatively assessing cardiac autonomic function and improving cardiovascular risk stratification. Alongside these applications, within a multimodality imaging framework, [¹⁸F]FDG PET/CT represents a valuable tool for detecting extracardiac infectious foci and for evaluating infections involving cardiac implantable electronic devices and ventricular assist devices [75].

Based on this progress, the recent first regulatory approval of C₁₈H₂₂Cl [¹⁸F]N₂O₃ (Flurpiridaz F-18, commercial name FLYRCADO™) represents a major milestone in PET myocardial perfusion imaging, allowing for the identification of coronary artery disease under both resting and stress conditions [76]. Furthermore,¹⁸F-labeled compounds have found applications in neurodegenerative disease research, particularly for the targeting of amyloid-β plaques. Such investigations include both in vivo PET imaging and ex vivo studies on postmortem brain tissues using autoradiography techniques. Alongside a number of well-established ¹⁸F-based amyloid PET tracers, such as [¹⁸F]florbetapir, [¹⁸F]flutemetamol and [¹⁸F]florbetaben, the PET tracer [¹⁸F]Florapronol, also known as [¹⁸F]FC119S, has also been investigated [77]. Within the field of brain pathologies, recent preclinical studies have further shown that ¹⁸F-labeled tryptophan-derived radiotracers exhibit favorable radiochemical properties and promising imaging performance [9].

The widespread use of ¹⁸F-labeled compounds in medical imaging relies fundamentally on the efficient and reliable production of the radionuclide itself. ¹⁸F can be generated either in form of fluoride ions (F⁻) or of molecular fluorine (F₂) through different nuclear reaction pathways. Among these, the most commonly employed method to prepare ¹⁸F in its nucleophilic anionic form involves proton bombardment of ¹⁸O-enriched water targets through the ¹⁸O(p,n)¹⁸F reaction, which proceeds with a relatively high cross-section and has become the standard approach in clinical PET radiochemistry.

The achievable ¹⁸F activity strongly depends on the cyclotron parameters and operational conditions. At ~18–20 MeV proton energy, hospital-based PET cyclotrons routinely produce ¹⁸F in the tens of GBq range per batch, with radiochemical yields in the 30–70% range. Depending on the efficiency of the ¹⁸F labeling chemistry, this production method can provide very high molar activity, with yields exceeding 70%. For instance, as reported in [78], a [¹⁸F]FDG activity of approximately 900 mCi at the end of synthesis (EOS), corresponding to a 72% non-decay-corrected yield, was achieved over 235 independent production runs.

In contrast to decentralized, on-site batch production, Silicon Fluoride Acceptor (SiFA) chemistry enables the centralized manufacturing of peptide-based ¹⁸F radiotracers through fully automated syntheses of clinically validated compounds such as [¹⁸F]SiTATE and [¹⁸F]Ga-rhPSMA-7/-7.3. The SiFA labeling approach, based on isotopic exchange at the Si–F bond, allows rapid radiolabeling under mild conditions and without the need for complex synthetic procedures or time-consuming purification steps [79]. The SiFA processes, that combine operational simplicity with high radiochemical yields and reproducibility, can yield large batches of ⁸F radiotracers with activities between 5 and 14 GBq/µA, making them suitable for distribution to multiple clinical centers. Compared to conventional ¹⁸F-labeling strategies, SiFA chemistry reduces synthesis time and improves production efficiency, thereby supporting the scalability of routine clinical applications and facilitating broader access to PET radiopharmaceuticals [79].

An alternative production route for ¹⁸F involves deuteron irradiation of naturally abundant (90.5%) ²⁰Ne targets. In this case, ¹⁸F is produced through the ²⁰Ne(d,α)¹⁸F reaction. However, measured cross-sections are approximately six times lower than those reported for the standard ¹⁸O(p,n)¹⁸F reaction, which reaches peak values of several hundred millibarns [80]. Consequently, the achievable ¹⁸F yields are substantially reduced, generally limiting production to a few tens of MBq/µA·h. Moreover, the generated [¹⁸F]fluorine is typically characterized by low molar activity, since carrier fluorine must be added to recover the radioactivity from the cyclotron target material. In addition, in radiolabeling applications, only up to approximately half of the available radioactivity can be effectively incorporated in a mono-radiofluorination step, while the remaining fraction is not utilized. These factors significantly constrain the scalability and routine clinical applicability of the ²⁰Ne(d,α)¹⁸F route compared with proton-based production methods [81].

A newly emerging approach for ¹⁸F production exploits photonuclear reactions induced by high-energy bremsstrahlung γ-rays. This route, built upon a nuclear process first explored several decades ago using the Frascati Electron Synchrotron [82], is now implemented using modern high-power Linacs. In this updated approach, ¹⁸F nuclei are generated through the ¹⁹F(γ,n)¹⁸F reaction, triggered by high-energy γ-rays impinging on CaF₂ nanoparticles, offering a complementary pathway to cyclotron-based production methods [83]. Recent experimental studies have demonstrated that, by optimizing the particle size under right irradiation conditions (10 kW accelerator at ~35 MeV electron energy), after 4 h of irradiation, it is possible to achieve ¹⁸F activities on the order of 1 Ci [83]. Although the overall yields and energy requirements remain less favorable than proton-induced reactions in terms of molar activity and scalable output, the photonuclear route may provide a complementary production option, especially where high-power electron accelerators are available.

4.5. ⁴⁴Sc (Scandium-44)

⁴⁴Sc decays predominantly via β⁺ emission (~94%) with half-life of 3.97 h and E_max = 1473 keV. It also emits 1157 keV γ-rays (99%), a feature that opens the possibility of triple-coincidence PET imaging (β⁺ annihilation photons plus prompt γ-rays) [84]. This multiphoton approach, which may also find application also in the positronium imaging technique, can improve image contrast, reduce background events and enable more accurate activity quantification and dosimetry. Moreover, the relatively long half-life of ⁴⁴Sc enables delayed imaging over extended periods and facilitates centralized production and distribution, while the moderate positron energy contributes to high spatial resolution of the images. These characteristics make ⁴⁴Sc particularly well-suited for molecular imaging in oncology, especially for radiolabeled peptides and antibodies targeting tumor-associated receptors, and for imaging biomolecules with slow pharmacokinetics, such as monoclonal antibodies and larger peptides [85].

Most ⁴⁴Sc radiopharmaceuticals are based on DOTA- and NODAGA-chelated vectors, including somatostatin analogs (e.g., [⁴⁴Sc]DOTATATE, [⁴⁴Sc]DOTATOC) for neuroendocrine tumors and PSMA ligands (e.g., [⁴⁴Sc]PSMA-617, [⁴⁴Sc]-PSMA-I&T) for prostate cancer. ⁴⁴Sc-DOTA complexes have demonstrated excellent in vitro and in vivo stability, maintaining intact radiolabeled peptides such as DOTATOC over multiple half-lives and resisting transchelation, confirming the robustness of DOTA chelation under physiological conditions [86].

An additional advantage of ⁴⁴Sc lies in its role in theranostics when paired with the β⁻-emitter ⁴⁷Sc. The ⁴⁴Sc/⁴⁷Sc pair provides a fully matched, clinically relevant diagnostic–therapeutic platform [87].

Currently, ⁴⁴Sc is produced via the ⁴⁴Ca(p,n)⁴⁴Sc nuclear reaction in clinical cyclotrons, using enriched calcium targets. This method yields high radionuclidic purity and is compatible with existing medical cyclotrons. Less common production routes include proton or deuteron irradiation of other calcium isotopes.

Another viable approach would be the ⁴⁴Ti/⁴⁴Sc generator system, where long-lived ⁴⁴Ti (half-life 59.1 y) serves as a parent nuclide [88]. While generators offer on-site availability without a cyclotron, the production of ⁴⁴Ti requires high-energy proton irradiation and complex radiochemical separation procedures. Furthermore, the long half-life of ⁴⁴Ti poses challenges in waste management and regulatory handling, which currently limits the widespread implementation of ⁴⁴Ti/⁴⁴Sc generators [89]. Radiochemical separation and labeling protocols have been established for both cyclotron-produced and generator-derived ⁴⁴Sc [90].

Overall, even if ⁴⁴Sc is not yet widely established in routine practice, it represents a promising PET radionuclide due to its favorable half-life, convenient decay properties and compatibility with generator systems, making it a versatile candidate for preclinical and clinical imaging applications.

4.6. ⁶⁴Cu (Copper-64)

This versatile copper radioisotope is gaining increasing importance in nuclear medicine and personalized theranostics. The imaging and therapeutic potential of ⁶⁴Cu (t_1/2 = 12.7 h) arises from its complex decay scheme, which includes 17.5% β⁺-emission (E_max = 652.6 keV, E_ave = 278.0 keV), 37.5% β⁻ emission (E_max = 579.6 keV) and 44% electron capture (EC). The 44% EC pathway produces characteristic low intensity 1345.8 keV photons and is moreover accompanied by the emission of high-LET Auger electrons [91]. ⁶⁴Cu’s decay properties make it suitable for delayed imaging and longitudinal studies, particularly with slowly circulating biomolecules.

For PET imaging, ⁶⁴Cu can be conjugated to a wide range of biomolecules, including peptides and antibodies. Key applications include cancer imaging, where ⁶⁴Cu-labeled agents target tumor-specific receptors, such as somatostatin receptors ([⁶⁴Cu]DOTA-TATE), HER2 ([⁶⁴Cu]trastuzumab) and integrins ([⁶⁴Cu]RGD peptides). In cardiology, [⁶⁴Cu]PTSM is used for myocardial perfusion and hypoxia imaging. Additionally, [⁶⁴Cu]ATSM and [⁶⁴Cu]GTSM compounds can be employed to investigate brain function, including studies of cerebral perfusion, neuroinflammation, and oxidative stress [92].

⁶⁴Cu can be produced via cyclotron or reactor methods [93]. Cyclotron production, that requires the use of expensive ⁶⁴Ni-enriched targets and of high-intensity proton beams to trigger the ⁶⁴Ni(p,n)⁶⁴Cu reaction, provides high specific activity suitable for radiopharmaceutical synthesis.

In addition to the cyclotron route, clinical-grade ⁶⁴Cu suitable for PET imaging can be produced in nuclear reactors through neutron capture on ⁶³Cu, or neutron-induced ⁶⁴Zn(n,p)⁶⁴Cu reactions [94]. However, the ⁶³Cu(n,γ)⁶⁴Cu route yields very low specific activity due to the large non-radioactive Cu matrix, while the ⁶⁴Zn(n,p)⁶⁴Cu reaction has a low spectrum-averaged cross-section and significant co-production of other radionuclides unless highly enriched targets are used. Concerning the ⁶⁵Cu(n,2n)⁶⁴Cu reaction induced by a high-brilliance 14 MeV neutron field [95], the activation product ⁶⁵Ni was found in the irradiated sample with an activity only about two orders of magnitude lower than that of ⁶⁴Cu. As a result, reactor-based production methods are less commonly employed for clinical PET applications.

After nuclear production, ⁶⁴Cu is separated from the target material by ion-exchange chromatography or solvent extraction and subsequently complexed with chelators such as DOTA, NOTA, or TETA to form stable radiopharmaceuticals for in vivo imaging. These processes require specialized chemical handling to minimize radiation exposure, making ⁶⁴Cu an expensive and relatively scarce radionuclide for routine use [93].

4.7. ⁶⁸Ga (Gallium-68)

Due to the high positron yield (≈89 %) and the favorable decay properties (t_1/2 = 67.71 min, E_max = 1899.0 keV, E_ave = 829.5 keV), ⁶⁸Ga has emerged as an important positron emitter for PET imaging. The relatively short mean range of the emitted β⁺-particles in soft tissue (~1mm) contributes to good image quality while delivering acceptable radiation dose to patients [96]. The more commonly used ⁶⁸Ga PET tracers include [⁶⁸Ga]PSMA, [⁶⁸Ga]Ga-DOTA-TOC and [⁶⁸Ga]DOTATATE.

[⁶⁸Ga]PSMA PET is a highly sensitive and specific imaging modality for prostate cancer, enabling accurate detection, disease staging and follow-up evaluation even at low (<1 ng/mL) PSA levels [97]. By visualizing PSMA expression in vivo, it helps in selecting patients suitable for PSMA-targeted therapies [98]. The combination of [⁶⁸Ga]PSMA PET with MRI further enhances anatomical localization and biopsy accuracy [99].

One of the most prominent clinical applications of ⁶⁸Ga PET is the imaging of somatostatin receptor–positive gastroenteropancreatic neuroendocrine tumors. This is achieved using radiopharmaceuticals that combine the ⁶⁸Ga radionuclide with somatostatin analogs, such as DOTA-TOC. Notably, [⁶⁸Ga]DOTA-TOC was the first ⁶⁸Ga-labeled PET radiopharmaceutical to receive clinical approval by the U.S. FDA in 2019 [100]. Furthermore, recent studies have demonstrated the clinical impact of [⁶⁸Ga]DOTATATE PET/CT, confirming its significant influence on therapeutic strategies for patients with gastroenteropancreatic neuroendocrine tumors [101]. Together, these ⁶⁸Ga-labeled tracers have established PET as a highly sensitive and specific tool for both the detection and therapeutic planning of somatostatin receptor–positive neuroendocrine tumors.

A common route for the production of ⁶⁸Ga involves proton irradiation of enriched ⁶⁸Zn targets via the ⁶⁸Zn(p,n)⁶⁸Ga reaction. Nevertheless, ⁶⁸Ga is more frequently obtained on-site through a ⁶⁸Ge/⁶⁸Ga radionuclide generator system [102]. The artificially produced parent radionuclide ⁶⁸Ge (t_1/2 = 271 d) is itself produced in high-energy accelerators, mainly via proton-induced spallation reactions on gallium targets (e.g., ⁶⁹Ga(p,2n)⁶⁸Ge). By relying on a generator system, ⁶⁸Ga-labeled tracers can be obtained locally without cyclotron infrastructure, enhancing their accessibility in clinical practice.

Both theoretical and experimental studies have evidenced that ⁶⁸Ga can also be obtained via the photonuclear production mechanisms on stable gallium isotopes through the ⁶⁹Ga(γ,n)⁶⁸Ga channel [103]. However, this alternative nuclear route is not being widely adopted for routine ⁶⁸Ga production due to low yields and practical constraints, and is mainly used to obtain valuable nuclear data rather than as a production method.

In conclusion, the short-lived, generator-produced ⁶⁸Ga is a PET radionuclide that offers versatility in labeling peptides and small molecules, and complements traditional tracers like ¹⁸F.

4.8. ⁸²Rb (Rubidium-82)

⁸²Rb is a radionuclide characterized by an extremely short physical half-life (t_1/2 = 1.26 min), which decays to stable ⁸²Kr by emitting β⁺-particles of E_max = 3350 keV and E_ave = 1481 keV. Its high positron energy results in a relatively long positron range in tissue, while the very short half-life enables rapid dynamic imaging and repeated measurements with minimal radiation dose to the patient. In clinical practice, ⁸²Rb is used almost exclusively as [⁸²Rb]RbCl, which behaves as a potassium analog and is actively transported into myocardial cells by means of Na⁺/K⁺-ATPase pumps. This physiological uptake mechanism renders ⁸²Rb a highly effective tracer for myocardial perfusion imaging, enabling dynamic assessment of myocardial blood flow and myocardial flow reserve in patients with CAD. Moreover, the rapid kinetics of ⁸²Rb allows accurate quantification of myocardial blood flow. Although some exploratory studies have investigated ⁸²Rb uptake in tumors and brain perfusion, its clinical use remains primarily cardiac.

Unlike most PET radionuclides, ⁸²Rb is not produced directly in a cyclotron but is obtained from an ⁸²Sr/⁸²Rb generator system, which enables on-demand production of ⁸²Rb at the imaging site. The ⁸²Sr/⁸²Rb generator exploits the β⁺-decay of ⁸²Sr (t_½ = 25.4 d) to produce ⁸²Rb. An advantage is that, owing to its very short half-life, ⁸²Rb can be eluted repeatedly every few minutes with saline solution, as [⁸²Rb]RbCl. The parent radionuclide ⁸²Sr is produced through the irradiation of rubidium targets via reactions, such as ⁸⁵Rb(p, 4n)⁸²Sr and ⁸⁷Rb(p,6n)⁸²Sr, which require high proton energies (typically 60–100 MeV) and high-current beams available only at large research or industrial cyclotron facilities. After irradiation, ⁸²Sr is chemically separated and loaded onto a generator column [104].

Whereas ⁸²Rb PET offers significant advantages for myocardial perfusion imaging, including no need for an on-site cyclotron, its clinical use is limited by several factors. These include reduced spatial resolution due to the high positron energy, the high cost of ⁸²Sr production and generator systems, and tracer chemistry restricted to simple ionic forms.

4.9. ⁸⁶Y (Yttrium-86)

The physical half-life of 14.74 h makes ⁸⁶Y a positron-emitting radionuclide suitable for PET imaging of compounds with intermediate/slow pharmacokinetics. However, in addition to the 33% β⁺ emission (E_max = 1481 keV, E_ave = 650 keV), ⁸⁶Y also undergoes EC, accompanied by the emission of multiple prompt high-energy γ-rays (443 keV, 628 keV, 1027 keV and 1854 keV), which can affect image quality and require appropriate correction strategies.

⁸⁶Y is mainly employed as a PET imaging surrogate for ⁹⁰Y, a clinically established β⁻-emitting therapeutic radionuclide. Owing to their identical chemical behavior,⁸⁶Y-labeled compounds provide a reliable model for ⁹⁰Y-labeled therapeutic vectors. This enables quantitative, non-invasive evaluation of biodistribution, pharmacokinetics and patient-specific dosimetry prior to ⁹⁰Y-based therapy. Such a theranostic pair is particularly relevant for peptide receptor radionuclide therapy (PRRT), radioimmunotherapy and other targeted radiotherapies, where ⁸⁶Y-based imaging allows optimization of ⁹⁰Y treatment planning [105]. To be used for PET diagnostics, ⁸⁶Y is incorporated into a wide range of radiopharmaceuticals, including [⁸⁶Y]DOTA-conjugated peptides, [⁸⁶Y]labeled monoclonal antibodies and [⁸⁶Y]DOTA-based small molecules and nanoparticles.

⁸⁶Y is commonly produced in cyclotrons via proton irradiation of strontium targets highly enriched in ⁸⁶Sr, through the ⁸⁶Sr(p,n)⁸⁶Y nuclear reaction [106]. After irradiation and chemical separation from the target material using selective resins, the purified ⁸⁶Y is formulated and radiolabeled with DOTA-based ligands under controlled pH and temperature conditions to obtain stable radiopharmaceuticals.

However, the overall production process is costly, due to the requirement for the enriched ⁸⁶Sr targets, and also technically complex. In fact, stable radiolabeling necessitates the use of strong macrocyclic chelators, such as DOTA and its derivatives, to ensure high in vivo stability of the trivalent yttrium complexes [107].

4.10. ⁸⁹Zr (Zirconium-89)

⁸⁹Zr is a long-lived radionuclide (t_1/2 = 78.36 h) which decays to the stable ⁸⁹Y predominantly via EC (77.2%), with a minor contribution from β⁺ emission (E_max = 902 keV, E_ave = 396 keV). The decay scheme is characterized by the emission of 909 keV γ-ray (94%), which must be considered for dosimetry and radiation protection. ⁸⁹Zr is used mainly in immunoPET imaging, particularly for labeling monoclonal antibodies (mAbs) and other slowly accumulating biomolecules, enabling non-invasive assessment of target expression, pharmacokinetics and dosimetry in oncology and theranostics [108].

Most ⁸⁹Zr-based tracers are chelated with desferrioxamine (DFO) derivatives (e.g., [⁸⁹Zr]DFO-trastuzumab) [109,110]. In addition to being conjugated to peptides, other ⁸⁹Zr-labeled entities, such as nanoparticles and cells, have also been investigated in preclinical and clinical PET studies [111].

⁸⁹Zr is produced in a cyclotron via the ⁸⁹Y(p,n)⁸⁹Zr reaction using natural yttrium targets, and subsequently is formulated as ⁸⁹Zr-oxalate for biomolecule radiolabeling, predominantly with DFO-based chelators [112].

For the production of ⁸⁹Zr, researchers have also explored some photonuclear reactions induced by bremsstrahlung γ-rays [113]. Among the photonuclear processes investigated, the ⁹²Mo(γ,p2n)⁸⁹Nb → ⁸⁹Zr pathway is considered the most promising for the production of ⁸⁹Zr. This synthesis route begins with the emission of one proton and two neutrons from a ⁹²Mo target nucleus, producing the intermediate radionuclide ⁸⁹Nb, which decays almost exclusively via EC (t_1/2 = 2.03 h) to form the positron-emitting ⁸⁹Zr. The two-step process provides high ⁸⁹Zr radionuclidic purity and allows the use of natural molybdenum targets, making it highly favorable for medical applications [114].

As a potential alternative route for ⁸⁹Zr production, the ^natSr(α,xn)⁸⁹Zr reaction has been proposed. However, this approach requires α-beams and results in the co-production of significant quantities of radionuclidic impurities, such as ⁸⁸Zr and ⁸⁶Zr [115].

Compared to other long-lived PET isotopes, ⁸⁹Zr matches the biological half-life of antibodies, enabling imaging even several days post-injection and making it uniquely suited for immunoPET and theranostic applications. Moreover, ⁸⁹Zr production is relatively straightforward and cost-effective. However, its use presents several drawbacks, including high γ-ray emission and a relatively long physical half-life, both of which increase patient radiation dose. In addition, a major critical issue is the in vivo low stability of the chelate, since released ⁸⁹Zr tends to accumulate in bones.

4.11. ¹²⁴I (Iodine-124)

Owing to its long half-life (t_1/2 = 4.18 d) and decay mode, predominantly EC (78%) with a significant contribution (23%) from β⁺ emission (E_max = 2137 keV, E_ave = 820 keV), along with associated γ-ray emissions at 603 keV (63%), 723 keV (10%) and 1691 keV (11%), ¹²⁴I is well-suited for PET imaging of biological processes with slow biokinetics. The extended half-life enables delayed imaging and precise dosimetry, making ¹²⁴I particularly valuable for thyroid cancer imaging and immunoPET applications. Moreover, it is suitable for radiopharmaceuticals requiring time-consuming syntheses and labeling protocols [116].

¹²⁴I can be incorporated into monoclonal antibodies, peptides, and small molecules, enabling diagnosis and therapy planning. Commonly employed ¹²⁴I-based tracers include ¹²⁴I-labeled monoclonal antibodies for immunoPET (e.g., [¹²⁴I]cG250 and other antibody-based probes), ¹²⁴I-labeled peptides for receptor-targeted imaging, and ¹²⁴I-labeled thyroxine analogs for thyroid imaging and dosimetry studies prior to ¹³¹I therapy.

¹²⁴I is typically obtained in a cyclotron using expensive highly enriched ¹²⁴Te targets via the ¹²⁴Te(p,n)¹²⁴I reaction. After irradiation, ¹²⁴I is chemically separated from the tellurium target and formulated as Na [¹²⁴I] for use in radiopharmaceutical synthesis. This cyclotron-based production offers the advantage of high specific activity. However, the production yield is rather low and radionuclidic purity can be adversely affected by the presence of isotopic carriers in the target material and by competing nuclear reactions, which limit the overall production efficiency [117].

An alternative cyclotron-based production route that has been proposed is the ¹²⁴Te(d,2n)¹²⁴I reaction, which provides a relatively higher production yield compared with the ¹²⁴Te(p,n)¹²⁴I reaction. However, this approach requires the use of deuteron beams, which are not always readily available, and generates other iodine isotopes, notably ¹²⁵I, at levels up to ~1.7% [118].

Due to these constraints, ¹²⁴I, like other longer-lived positron-emitting radionuclides, is often produced in nuclear reactors via the neutron capture reaction ¹²³Te(n,γ)¹²⁴Te, followed by β⁻ decay of ¹²⁴Te to ¹²⁴I. This indirect production route offers relatively high yields and simplifies target processing compared to direct iodine irradiation [25].

Although its production is relatively expensive and necessitates stringent radiation shielding due to the emission of high-energy γ-photons, ¹²⁴I remains a highly valuable PET radionuclide for long-term imaging, kinetic studies and theranostic applications, including immunoPET and the development of precision medicine.

Table 1. Half-life, daughter nuclide, percentage of β⁺ decay, maximum and average β⁺ energies, energy and intensity of the main photons emitted following EC decay, commonly used production routes. Decay data taken from Ref. [40].

Radio Nuclide	T1/2	Daughter Nuclide	Iβ+ (%)	Eβ+ (keV) (Emax; Eave)	Eγ(keV) (%)	Production Routes
11C	20.34 min	11B	99.8	max: 960.4 ave: 385.7	-	14N(p,α)11C
13N	9.97 min	13C	99.8	max: 1197.9 ave: 492.2	-	16O(p,α)13N
15O	122.7 s	15N	99.9	max: 1732.0 ave: 735.3	-	14N(d,n)15O 15N(p,n)15O 16O(γ,n)15O
18F	109.73 min	18O	100	max: 633.5 ave: 249.8	-	18O(p,n)18F 20Ne(d,α)18F 19F(γ,n)18F
44Sc	3.97 h	44Ca	94	max: 1473.7 ave: 630.1	1157 (~100)	44Ca(p,n)44Sc 44Ti/44Sc generator
64Cu	12.7 h	64Ni	17.5	max: 652.6 ave: 278	1345 (~0.5)	64Ni(p,n)64Cu 63Cu(n,γ)64Cu 64Zn(n,p)64Cu 65Cu(n,2n)64Cu
68Ga	67.71 h	68Zn	89	max: 1899.0 ave: 829.5	1077 (3.2)	68Ge/68Ga generator 68Zn(p,n)68Ga 69Ga(γ,n)68Ga
82Rb	1.26 min	82Kr	95.4	max: 3350 ave: 1481	-	82Sr/82Rb generator
86Y	14.74 h	86Sr	33	max: 1481 ave: 650	443 (17) 628 (32) 1027 (82) 1854 (16)	86Sr(p,n)86Y
89Zr	78.36 h	89Y	23	max: 902 ave: 396	909 (94)	89Y(p,n)89Zr 92Mo(γ,p2n)89Nb 89Nb → 89Zr
124I	4.77 d	124Te	22.7	max: 2137 ave: 829	603 (94), 723 (10) 1691 (11)	124Te(p,n)124I 124Te(d,2n)124I 123Te(n,γ)124T 124Te → 124I

below summarizes the main characteristics of nuclear decays and the primary production routes of the radioisotopes discussed in this review. The Table reports, for each radionuclide, the half-life, daughter nuclide, percentage of β⁺ decay, maximum and average β⁺ energies, and, when applicable, the energy and intensity of the main photons emitted following EC decay. Table 1 also indicates the most commonly used production routes.

In concluding this section, it should be emphasized that, beyond the most commonly used PET radionuclides, several other positron emitters, such as potassium-38, copper-62, manganese-52m, iron-52, selenium-73, bromine-75, bromine-76, rubidium-81 and rubidium-82m, are currently being investigated. However, their broader application remains limited at present due to factors such as production complexity, short or non-optimal half-lives or challenging radiochemistry, and they have therefore not yet achieved widespread clinical adoption [119].

5. Final Considerations

In the last two few decades PET has evolved into a well-established and indispensable imaging modality for the diagnosis and management of oncological, neurological and cardiological disorders. Its integration into routine clinical practice reflects not only its high diagnostic accuracy but also its unique ability to deliver quantitative, functional and molecular information beyond the capabilities of traditional imaging techniques. The remarkable progress in radiopharmacy, driven by the development of novel pharmaceuticals able to target specific molecular pathways, has further expanded the clinical impact of PET. By enabling the non-invasive visualization, characterization and quantification of biological processes at the molecular and cellular levels, PET has rapidly become the method of choice for disease detection, therapy planning and treatment monitoring across a wide spectrum of pathologies.

Nevertheless, despite these advances, PET remains an expensive technology, substantially owing to the technical challenges and operational constraints in radionuclide production. Although significant technological improvements have been achieved in accelerator design, targetry and radiochemical processing, the production of positron emitter radionuclides remains a major infrastructural and economic bottleneck which limits the large-scale implementation of PET technology.

As highlighted throughout this review, key parameters such as half-life, decay characteristics and production cross-sections of positron emitter nuclides critically determine the complexity, scalability and long-term sustainability of their manufacturing processes. Consequently, the selection of a PET radionuclide for clinical or research applications depends not only on biological targeting and on the chemistry involved in separation, chelation, and radiotracer labeling, but also on multiple physical and logistical factors, from the nuclear properties of the nuclides to practical clinical considerations.

The first parameter to be considered when selecting a positron emitter is its half-life, which determines the suitability of the nuclide for specific imaging protocols and clinical use. Short-lived radionuclides, such as ¹¹C, ¹³N, ¹⁵O, and ¹⁸F, are ideal for rapid studies due to their fast decay and minimal radiation burden, making them well-suited for high-throughput clinical and research setting.

One of the main historical challenges associated with short-lived radionuclides is the complexity of tracer preparation and the need for prompt administration. The development of automated radiosynthesis platforms capable of sequential production of labeled compounds at intervals suitable for clinical practice has greatly improved operational efficiency. Automated synthesis modules, operating in strict compliance with Good Manufacturing Practice, ensure high radiochemical purity and process reproducibility within the time windows imposed by radioactive decay. By minimizing the delay between production and patient injection, these systems effectively mitigate constraints imposed by short half-lives. Overall, automation and direct delivery systems have substantially simplified the workflow of short-lived PET radionuclides, facilitating the use of such dynamic and quantitative imaging techniques.

Unlike short-lived PET radionuclides, the longer-lived ones are less constrained by production and timing issues, making them easier to handle. Longer-lived PET radionuclides, including ⁶⁴Cu, ⁸⁹Zr and ¹²⁴I, provide advantages for labeling long-circulating biological vectors such as antibodies and peptides, allowing imaging over time scales of several days.

While the clinical utility of well-established radionuclides such as ¹⁸F, ¹¹C, ¹³N, ¹⁵O and ⁶⁸Ga is widely recognized, a growing number of emerging PET radionuclides, including ⁴⁴Sc, ⁶⁴Cu, ⁸²Rb, ⁸⁶Y, ⁸⁹Zr and ¹²⁴I, are attracting increasing attention within the medical research community, owing to their potential to provide new opportunities for advanced and more versatile diagnostic applications.

However, the use of these emerging radionuclides involves a series of nuclide-specific challenges.

Cyclotron production of ⁴⁴Sc and ⁶⁴Cu requires enriched calcium and nickel targets, respectively, with ⁶⁴Cu supply further limited by a complex radiochemistry and the need for strict radiation safety protocols. ⁸⁶Y emits multiple high-energy γ-rays that, due to cascade coincidences and increased scatter, complicate image acquisition and analysis as well as dosimetry.

⁸⁹Zr, beyond the technical limitation represented by the emission of γ-rays, requires careful chelation chemistry to avoid in vivo release and bone accumulation. ¹²⁴I has a long half-life and emits multiple high-energy γ-rays which increase patient radiation dose and pose challenges for accurate quantitative imaging and precise dosimetry.

Conversely, the extremely short-lived ⁸²Rb requires automated radiosynthesis platforms and dedicated administration systems, which are not always available in all clinical settings. In practice, ⁸²Rb is typically obtained on-site by elution from an ⁸²Sr/⁸²Rb generator. This example highlights the broader relevance of radionuclide generators, which deserve particular attention given the current and anticipated reduction in production facilities associated with the planned decommissioning of several nuclear infrastructures. In this context, generator-based strategies may represent a reliable and practical source of PET isotopes independent of large-scale installations.

Nevertheless, large-scale infrastructures such as nuclear reactors, cyclotrons and linear accelerators remain essential for the production of most PET isotopes.

Nuclear reactors enable the production of selected longer-lived isotopes in quantities sufficient for multiple patients, although their access remains limited and transport to medical facilities can be logistically challenging. Cyclotrons and Linacs provide high-purity radionuclides, effectively supporting clinical demand. However, they require substantial capital investment and highly specialized personnel for operation and maintenance.

Despite these constraints, both centralized large-scale production and generator-based approaches contribute significantly to the supply chain of PET radioisotopes. However, the current production capacity remains insufficient to guarantee the widespread and consistent availability of positron-emitting isotopes for clinical PET imaging.

At the same time, a notable advantage has emerged with developments in compact cyclotron systems, which are increasingly deployed directly in clinical environments. These systems enable the on-site production of short-lived isotopes, reducing dependence on complex distribution networks and minimizing decay losses during transport. Overall, compact cyclotrons represent a significant advancement toward more decentralized and flexible production strategies for PET radionuclides.

Beyond advances in compact cyclotron systems and automated synthesis modules, which help overcome challenges associated with short half-lives and enable rapid radiotracer delivery, the evolution of PET is also linked to the recent implementation of hybrid imaging platforms such as PET/CT and PET/MRI. These multimodal systems are expanding the role of PET in advanced functional imaging and translational research.

Future developments are expected to increasingly evolve toward integrated multiparametric PET-based systems, further supported by the growing adoption of Artificial Intelligence to guide acquisition protocols and improve the interpretation of complex imaging data. In this evolving scenario, PET is expected to play a key role also in radiomics, by enabling the translation of complex imaging data into high-dimensional quantitative descriptors that can be integrated with clinical and genomic information for more comprehensive disease characterization and predictive modeling.

As highlighted throughout this review, enhancing accessibility, reducing costs and expanding the clinical impact of PET imaging require sustained innovation in nuclear technology together with the selection and production of radionuclides with suitable decay characteristics, streamlined radiochemistry and optimized logistical frameworks.

Alongside these advances, continued progress in the development of novel radiopharmaceuticals, coupled with developments in automation and hybrid imaging technologies, is significantly enhancing the efficiency, reproducibility, and clinical impact of PET, paving the way for increasingly sophisticated and personalized molecular imaging paradigms.

Taken together, the technological and pharmaceutical developments place PET as a cornerstone in the evolving landscape of precision medicine, with the potential to reshape diagnostic pathways, enable earlier diagnosis and govern targeted therapeutic interventions.