The original idea of personalized and tailored drugs, highly specific to a pharmacological target, appeared at the beginning of the XXth century via the “magic bullet” concept (1) brought by Paul Ehrlich (Nobel laureate in 1908). In oncologic field, this postulate aims to optimize the risk-benefit ratio for the patients care. Yet, it was not until the late 1970's that the discovery of hybridoma technology by Georges Kohler and Cesar Milstein (Nobel laureates in 1984) made it possible to apply this theory in clinical practice. The hybridoma technology allowed biochemists and immunologists to produce monoclonal antibodies (mAbs) which specifically recognize antigens on pathologic cells, thereby providing the proof-of-concept for specific immunotherapeutic approaches to treat cancer (2).

With the concomitant progress in nuclear medicine, the radiolabeled monoclonal antibody (mAb) quickly emerged to specifically target antigens of abnormal cells, most often cancerous. The discovery of overexpressed antigens in many hematologic and solid tumors paved the way for the use of gamma emitters such as iodine-131 (¹³¹I) radiolabeled mAbs for tumor detection in nuclear medicine practice (3, 4). In parallel, iodine-131 being also a beta minus emitter, the use of ¹³¹I-mAbs has quickly found its place in therapeutic approaches, mainly in hematologic neoplasms with real efficacy in B-cell lymphoma (5). For almost three decades, mAbs have been labeled with gamma emitters (such ¹³¹I or ¹¹¹In) in planar or Single Photon Emission Computed Tomography (SPECT) imaging procedures. Despite the reliable and confident clinical information, this modality suffers from several drawbacks including poor sensitivity, poor spatial resolution and complex scatter correction due to the collimator.

While Positron Emission Tomography (PET) was revolutionizing diagnostic applications of nuclear medicine with small organic molecules such as fluorodeoxyglucose (FDG) radiolabeled with fluorine-18 (¹⁸F), PET approach seemed unusable with long pharmacokinetic half-life mAbs. The availability in the early 2000s (6–8) of longer half-life isotopes such as iodine-124, copper-64 or zirconium-89 allowed the fruitful association between the high specificity of mAbs with the high resolution of PET leading to the Immuno-PET approach. Comparatively to classical SPECT modality, using of immuno-PET imaging combined several technical advantages such as precise scatter correction, exact correction of attenuation, accurate quantitative information, improved spatial resolution with a better delineation of tumors and organs, and, last but not least, higher sensitivity associated with the capacity to perform a true whole-body imaging in a reasonable time. The recent technological evolution in new PET detectors and reconstruction algorithms (9–11) appeared to be also a key factor for the immuno-PET clinical performance in terms of both spatial resolution and signal-to-noise ratio.

Today, the overall performance of the immuno-PET methodology is carried by different biological or imaging parameters such target accessibility or imaging protocol (time between injection and image acquisition, number of acquisitions). The aim of this paper review is more focused on the pharmacological choices to find the best matching between the radionuclide parameters (physical, chemical, logistic) and the mAb used (specificity, affinity, dose, mAb derivative, pharmacokinetic…). Immuno-PET is a fast-growing approach (12) in many cancer pathologies (13). Currently, from an imaging point of view, the management of patients suffering from a tumor pathology is mainly driven on an initial diagnosis made by conventional techniques combining CT/MRI and ¹⁸F-FDG PET. In this care management, immuno-PET approach finds a particularly effective place to provide some specific information on the tumor phenotype, intra- and inter-tumor heterogeneity and as prognostic/predictive indicator for targeted therapies. It should be noted that in the specific case, where the immuno-PET can be used to select only patients in whom the treatment is likely to provide a benefit, the immuno-PET falls into the scope of companion diagnostic. Some preclinical proof-of-concept will be discussed in mAb derivatives section and a clinical part will illustrate, through some typical examples of results obtained with compounds (mAb and immunological derivatives) allowing antigen targeting in solid tumors, hematological and tumor microenvironment.

One of the immuno-PET key success is based on an appropriate matching between the biological half-life of the immunoprotein and the physical half-life of the isotope to achieve the best contrast in pathologic tissues (14–16). Despite the great variety of radionuclides which are positron emitters, only few of them could be used in immuno-PET. For nuclear medicine applications, the choice of the radionuclide is basically based on three main criteria: physical characteristics, chemical characteristics, and production/logistic. The list of isotopes used (or considered) is summarized in Table 1. Indeed, intact mAbs have a pharmacokinetic half-life of several days with a long circulation time and required a long radionuclide half-life to increase the tumor-to-background ratio such as iodine-124 or zirconium-89. Copper-64 with its intermediate half-life can be used for labeling a large size range of molecules like native mAbs or their smaller derivatives [F(ab')₂, F(ab), minibody, ScFv, nanobody or affibody structures]. Fluorine-18 and gallium-68 with their shorter half-life may be used to label small molecular weight molecules such as mAb derivates or small synthetic molecules for pretargeting approach (15).

Several additional considerations must also be taken to rationalize the choice of a positron emitter. Positron energy range is a key factor for intrinsic resolution of PET modality because it has a direct and significant impact on the positron travel distance before annihilation. As a consequence, a high-energy positron will result in an intrinsic resolution loss. In addition, existence of concomitant γ and/or β⁻ emissions will have a major impact on the staff and/or patient radiation dose. It should be noted, to our knowledge, they are currently no allowed maximal recommended activity limitation per isotopes. Dosimetric studies are on progress and first results obtained with trastuzumab showed an effective dose equivalent to the whole body of 45 and 10.8 mSv, respectively for ⁸⁹Zr and ⁶⁴Cu (17, 18). Of course, these preliminary results are only indicative because, the dosimetric data are also under the dependence of the vector pharmacokinetic profile, As consequences, the dosimetry for smaller mAb derivatives like nanobodies is lower (shorter in vivo stay, due to a fast renal clearance). For example, the dosimetry induced by 185 MBq ⁶⁸Ga-2Rs15d nanobody is close to the dosimetry due to the classical 370 MBq ¹⁸F-FDG injection with, respectively, 4 and 7 mSv (19).

The chemical nature of the isotope presents a direct impact on the immunoprotein radiolabeling (20). Classically in radiopharmaceutical practices, radionuclides may belong to the radiohalogen or radiometal family, Radiohalogens like iodine-131 could be directly radiolabeled to the amino-acid chain on an aromatic residue (such histidine or more usually on a tyrosine) after gentle oxidation of iodide (21). This effective method presents some limits (mAb sensitive to oxidizing environment, low stability of radioactive mAb,…) and indirect radiolabeling could be envisaged. The indirect radiohalogen labeling is based on the use of prosthetic group intermediate (5, 22–25) like Bolton-Hunter reagent, organostanyl compound or iodonium salts (26–29).

Radiometals form very stable coordination complexes with a great variety of ligands, including linear diethylenetriaminepentaacetic acid (DTPA) derivatives, macrocyclic [1,4,7,10-tetraazacyclododecane-1,4,7,10-tetraacetic acid (DOTA) polyaminocarboxylic or desferrioxamine (DFO) derivatives] (30–32). These ligands are transformed in bifunctional chelating agent (BCA) capable of reacting with proteins to form a stable covalent bond with lysine residue, cysteine residue or synthetic bioorthogonal [click-chemistry approach (33)].

Finally, the production way of the isotope has consequences on the cost and in the availability of the radionuclide. The great majority of positron emitters are produced by cyclotron accelerator with a relatively high usual production cost. The half-life of the radionuclide presents a direct impact on the logistic because of decay during the transport. To circumvent this constraint a cyclotron network was built close to nuclear medicine departments for the short half-life isotopes (fluor-18; scandium-44; copper-64). These particular supply chain leads to a particular production schedule without daily availability of all isotopes. To solve this problem, some isotopes like Gallium-68 can be put in generator form (in this case, industrial supplier used germanium-68 with a longer half-life to produce in nuclear medicine department the desired gallium-68 by filiation).

The discovery of hybridoma technology by Köhler et al. (34) originally allowed the production of antigen-directed specific mAbs. This discovery paved the way for immuno-PET development but has appeared to be limited by the immunogenicity due to the murine origin of the amino-acid sequence (35). To circumvent this first hurdle, chimeric then humanized mAbs were designed by molecular biology and enabled less immunogenic approach with a better tolerance profile for the patients (36). The murine, chimeric and humanized mAbs are entire (intact) immunoglobulin G (IgG) with a long pharmacokinetic half-life and high weight (150 kDa) which reduces their capacity to diffuse inside the tumor mass.

To adapt these properties (pharmacokinetic and distribution), biochemists and immunochemists have developed some immunoconjugate derivatives exhibit reduced half-lives and improved tumoral penetration. The overall characteristics of these protein structures are summarized on Table 2.

The first improvement was the use of mAbs fragments like F(ab')₂ and F(ab). These smaller protein fragments consist in the reduction/digestion of initial mAbs by enzymes (pepsin and papain). The F(ab')₂ and F(ab) fragments are smaller (110 and 50 kDa, respectively) than the native mAbs but preserve the affinity and the specificity for the tumor antigen. F(ab')₂ fragments of mAbs were radiolabeled with ⁶⁴Cu in preclinical studies in breast cancer (37) or in lymphoma (38) animal models. Results showed a rapid and sustained uptake in tumor with promising and encouraging perspectives for the clinical evaluation of malignant pathologies while delivering a lower total body radiation dose compared to the entire mAbs. F(ab) fragments radiolabeled with ⁶⁸Ga (39) or ⁶⁴Cu (40) showed similar results with a shorter pharmacological half-life. In terms of biodistribution, F(ab) fragments present the characteristic of having a mass weight lower than the renal size cut-off (30–50 kDa) and consequently of being eliminated by the urinary tract (41). As consequence, F(ab) fragments cause high and persistent localization of the radioactivity in the kidney which could be reduced by a metabolizable linker between F(ab) fragment and isotope part of the construct (42).

Building on these successes, immunochemical engineers have turned to small synthetic proteins which contains only the antigen recognition domain of the initial mAb such minibodies, fusion protein-like single chain variable fragments (ScFv) or nanobodies. Minibodies consist of constant heavy chain (CH) and variable light (VL) and variable heavy (VH) domains, ScFv constructs are made of solely VL/VH domains and nanobodies has only the VH domains. ¹²⁴I and ⁸⁹Zr radiolabeled minibodies validated the feasibility proof-of-concept of immuno-PET in prostate cancer animal model (43). In breast cancer animal model, ScFv fragments radiolabeled with ⁶⁸Ga (44) showed in vivo promising results. On their sides, nanobodies are subjects of intense preclinical research (45–47). These engineered antibody-based proteins demonstrate their capacity to be used in nuclear medicine field with β⁺ emitters for immuno-PET applications (48, 49). To date, the proof-of-concept was validated at a preclinical level and exploratory work in human is started (50).

More recently, synthetic proteins (three-helix scaffold) generated from a phage display library were produced with a high affinity against an antigenic structure. These proteins, named affibodies, have been prepared against different cancer targets for PET applications and were studied in vitro with promising results (51–53).

To optimize the tumor-to-background ratio, recent advances of nuclear medicine demonstrate the great interest of pretargeting approach (54). This technic involves a preliminary injection of a non-radioactive bispecific antibody followed by an injection of a radiolabeled bivalent hapten (peptide or small organic molecule). This system allows to bypass the slow clearance of bispecific mAbs. In this particular case, a bispecific mAb recognizes the tumor antigen with a high specificity by one of its two recognition domains. The second recognition domain is used to bind the radioactive hapten which is injected after the bispecific mAb within an optimal delay. The unbound hapten is rapidly cleared from the blood circulation via the renal system and allows, de facto, an increase of the immuno-PET image contrast. The hapten approach is not the only one feasible for pretargeting and bio-orthogonal reactions based on click-chemistry has recently been considered with successful results in preclinical studies (55, 56). The mAb (entire IgG or derivative) is in vitro prefunctionalized with a chemical clickable function. The intra-venous injection of this compound is followed by the injection of a radioactive compound with clickable complementary function and a specific covalent bond between these two components is in vivo formed. The relatively short pharmacokinetic half-life of the radioactive clickable moiety allows the use of shorter half-life isotopes and a renewed interest for fluorine-18 in immunoPET applications (57).

Today, the immuno-PET field is rapidly progressing and allows to provide essential information for the care management of each patient. To meet the need of personalized medicine era where each patient is unique, immuno-PET provides repeatable, non-invasive whole-body information of biomarkers mapping. Many preclinical developments and clinical proof-of-concepts have been done or are ongoing. The recent clinical developments of immuno-PET have confirmed the potential of mAbs (and their derivatives) as companion diagnostic to determine, whether or not, a patient will respond to a targeted therapy. Nevertheless, the majority of proof-of-concept clinical studies have been performed on small patient samples and require larger studies to confirm their potential as predictive imaging biomarkers. Currently, the main limitation for these clinical studies with great patient numbers are often limited by the costs and the availability limitation of isotopes.

The original phenotypic information provided by immuno-PET currently allows in vivo access of several pieces of information on the tumor aggressivity associated to a patient outcome prognostic. The power of this molecular imaging modality seems to be able to show very informative data on the intra-tumor and intra-patient variability on molecular biomarkers expression and consecutively in phenotypic heterogeneity of the entire disease burden. Moreover, in comparison with “gold-standard” biopsies, immuno-PET allows to acquire in-vivo knowledge of the intra-tumoral penetration/biodistribution of the mAb without inherent-risk of tumor seeding during needle biopsy (105).

Beyond the image information, immuno-PET is attractive to study the in-vivo behavior of antibody-based therapies and for understanding their therapeutic efficacities. Within the scope of theranostic approach, the use of β⁺/β⁻ pair (respectively for diagnostic then therapy) radiolabeling the same mAb is very promising because the same distribution/pharmacokinetic is expected. Immuno-PET allows in theranostic the determination of the patient dosimetry to optimize the cumulated activity and to predict the therapy response by a quantitative measurement of the tumor-antibody uptake. Molecular imaging by immuno-PET with mAbs or their derivatives will play a pivotal role in the close oncologic future to tailor a customized therapy for each patient.

CB, CR, FK-B, and CB-M are nuclear medicine physicians. They have worked for many years on the development of immuno-PET approach in oncology. AL is an endocrinologist with an expertise in immuno-PET approach for neuroendocrine tumors. MC is pharmacist, head of the nuclear oncology research team, with a broad expertise in immunology. MB is radiopharmacist with a long term wide experience of radiopharmaceutical development, mainly for immuno-PET applications. All authors contributed to the article and approved the submitted version.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.