PET Imaging of Receptor Tyrosine Kinases in Cancer

Due to their complex heterogeneity and tendency to spread throughout the body, treating cancers is very challenging. Over the past quarter century, progress in cancer biology has led to the discovery and identification of receptor tyrosine kinases (RTK), which control many fundamental cell behaviors and drive tumor initiation, maintenance, and progression (1). These discoveries have fueled the development of effective targeted therapeutics such as monoclonal antibodies (mAb) and tyrosine kinase inhibitors (TKI), shifting cancer patients' care from traditional empirical treatments to an era of personalized treatment.

Human RTKs contain 20 subfamilies and only approximately half of the RTK families are well understood (2). RTKs are anchored in the cytoplasmic membrane, and gain-of-function mutations and/or overexpression of these RTKs are closely related to growth and proliferation in malignant tissues (2). Of them, vascular endothelial growth factor A and VEGFRs have implicated roles in tumor angiogenesis, and visualization of VEGFR expression in vivo with a radiopharmaceutical may be a viable clinical option for imaging angiogenesis (3). The epidermal growth factor receptor (EGFR) belongs to another family of RTKs and includes three other members (erbB2/HER-2, erbB3/HER-3, and erbB4/HER-4). We previously showed that HER-kinase–targeted imaging agents would enable maximum benefit (i.e., patient stratification, therapeutic response monitoring, and new drug development) in cancer patient management (4). c-Met, another RTK for hepatocyte growth factor (HGF), has been found overexpressed or aberrantly activated in a variety of cancers (5).

Although mAbs targeting the above-mentioned RTKs have become a standard of care for patients with certain mutation-positive tumors, the efficacy of the currently approved mAbs as single agents is very limited; therefore, synergistic regimens containing conventional chemotherapeutic agents and the mAbs are applied (6). In addition, drug resistance almost invariably occurs in cancer patients treated with these therapeutic antibodies. Traditionally, biopsy and immunohistochemistry are performed to determine the RTK status of cancer tissues and to guide subsequent treatment. However, spatial expression levels of RTKs can vary over time and among lesions (1), indicating the urgent need for novel noninvasive approaches to visualize RTKs' plasticity throughout the whole body. Moreover, noninvasive methods that can select patients who will potentially benefit from combinational therapy and predict treatment response will greatly optimize management strategies for cancer patients.

In recent years, molecular imaging has rapidly developed and has been widely used both in clinical setting and in preclinical arenas (7–9). Single-photon emission computed tomography (SPECT) and positron emission tomography (PET) are radionuclide molecular imaging techniques that enable noninvasive evaluation of biochemical changes and expression of molecular targets within living subjects. Although a SPECT probe, ¹¹¹In-labeled trastuzumab, has been used in clinical trials to detect HER2-positive metastatic breast cancer (10), the sensitivity of SPECT is several orders of magnitude lower than that of PET (11). In comparison, PET imaging is of high sensitivity and can be used to investigate physiological and molecular mechanisms of human diseases (11). PET imaging of RTKs with radiolabeled mAbs, denoted as immunoPET, may provide a noninvasive method for assessing the dynamics of RTKs, selecting patients for personalized treatment, predicting response to RTK inhibition therapy, and facilitating drug development (7).

Most PET imaging clinical trials have been focused on using radiolabeled FDA-approved antibodies (12), such as trastuzumab for breast cancer and esophagogastric adenocarcinoma (EGA; refs. 13–17), cetuximab for colorectal and lung cancers (18–20), and bevacizumab for several indications (21–26). PET imaging using the positron emitters ⁶⁴Cu (t_1/2, 12.7 hours) and ⁸⁹Zr (t_1/2, 78.4 hours) for antibody labeling has been extensively studied over the last decade (8, 27). ⁸⁹Zr has a longer half-life, which generally matches the serum half-life of most mAbs and therefore is suitable for antibody imaging (28, 29). A straightforward strategy for generating RTK-specific PET radiotracers has mostly, but not entirely, involved the radiolabeling of therapeutic mAbs. Antibody fragments, nanobodies, and smaller molecule inhibitors have also been investigated for in vivo visualization of RTKs. With shorter biological half-lives, these probes are well suited for fast imaging protocols when labeled with rapidly decaying radioisotopes such as ⁶⁸Ga (t_1/2, 68 minutes) and ¹⁸F (t_1/2, 110 minutes). Our review focuses mainly on the current research in the field of antibody-based immunoPET probes and highlights potential clinical implications of immunoPET in visualizing RTKs and in guiding clinical decisions (Fig. 1).

Dysregulation of HGF/c-Met signaling is implicated in a number of malignancies (30), and synergistic effects of EGF and HGF were noted in non–small cell lung cancer (NSCLC; ref. 31). Moreover, c-Met amplification can drive HER3-dependent PI3K signaling, thereby mediating NSCLC resistance to EGFR inhibitors (32, 33). Currently, c-Met targeted therapy involves small-molecule inhibitors (crizotinib, capmatinib, etc.; refs. 31, 34), and antibodies (emibetuzumab, ficlatuzumab, rilotuzumab, onartuzumab, etc.; refs. 35, 36). Phase II trials in patients with advanced NSCLC demonstrated that the addition of onartuzumab to the EGFR inhibitor erlotinib resulted in an improvement in progression-free and overall survival for c-Met–positive patients (37); however, a phase III study failed to observe the additive therapeutic effect of onartuzumab to erlotinib (38). Despite these disappointing results, suppressing the HGF/c-Met pathway may still have an antitumor effect in other primary tumors and/or overcome the resistance of molecularly targeted therapies (30, 39). Noninvasive PET visualization of c-Met dynamics could therefore potentially facilitate patient stratification and guide c-Met–directed therapies.

Initial attempts to develop nuclear medicine imaging probes for detecting c-Met expression used murine mAbs (40, 41). Based on these studies, taking the step to evaluate immunoPET probes using humanized or fully human anti-c-Met mAbs is relatively straightforward. Using the humanized therapeutic mAb onartuzumab, which inhibits HGF binding and therefore the HGF/MET signaling pathway (35), Jagoda and colleagues initially synthesized ⁸⁹Zr-DFO-onartuzumab and ⁷⁶Br-onartuzumab and found that the former probe exhibited higher uptake and tumor-to-muscle ratios in MKN-45 gastric carcinoma models (42). Li and colleagues developed c-Met–targeting bivalent cys-diabodies and radiolabeled one of the candidates with ⁸⁹Zr. The authors found that uptake of ⁸⁹Zr-labeled H2 cys-diabody was higher in the gefitinib-resistant NSCLC model than in the low c-MET–expressing NSCLC model, indicating that immunoPET could be used to assess c-MET expression levels for both therapeutic and diagnostic purposes (43). Pool and colleagues then reported that the readily translatable probe ⁸⁹Zr-Df-onartuzumab could effectively discriminate erlotinib-induced c-Met upregulation in NSCLC models (44). This study herein highlighted the potential value of ⁸⁹Zr-Df-onartuzumab PET in assessing c-Met upregulation-mediated erlotinib resistance in patients with NSCLC. More importantly, c-Met–directed treatment using NVP-AUY-922, a heat shock protein 90 (HSP90) inhibitor (45), could reduce ⁸⁹Zr-Df-onartuzumab uptake in HCC827 models (Fig. 2A). Rilotumumab (AMG102) is an HGF-binding antibody and can bind to and neutralize HGF, thus preventing its binding to c-Met. However, a phase III trial failed to observe the therapeutic effects of rilotumumab in patients with c-Met–positive gastric or gastroesophageal adenocarcinoma (46). Previously, there were no tools to noninvasively determine the levels of HGF present in the local tumor microenvironment. To this end, ⁸⁹Zr-DFO-AMG102 was developed by Price and colleagues, and this probe could selectively accumulate in tumors with high levels of HGF protein (Fig. 2B and C; ref. 47).

Our team produced a recombinant human HGF and labeled the agent with ⁶⁴Cu. PET imaging revealed specific and prominent uptake of the tracer in c-Met–positive U87MG tumors but significantly lower uptake in c-Met–negative MDA-MB-231 tumors (48). One concern for these tracers based on the HGF ligand is their potential to stimulate tumor growth by activating c-Met (48). Burggraaf and colleagues initially developed a fluorescently labeled peptide (GE-137) for optical imaging of c-Met, and this probe showed high affinity for human c-Met in 15 subjects with high risk of colorectal cancer (49). In addition to optical imaging in assessing c-Met (49, 50), ¹⁸F-AH113804 is a peptide-based c-Met–specific PET imaging probe that has been used to visualize locoregional recurrence of breast cancer in a clinical trial (51).

These results suggest that HGF/c-Met–specific PET imaging may correctly identify patients most likely to benefit from c-Met–targeted therapies or from EGFR inhibition therapy, as it has been reported that c-Met is implicated in acquired resistance to EGFR inhibitors in NSCLC (31, 33).

The VEGF/VEGFR signaling pathway plays an important role in the regulation of angiogenesis. VEGFs bind to three associated transmembrane RTKs known as VEGFR-1, VEGFR-2, and VEGFR-3. Among them, VEGFR-2 is a key receptor involved in the development of blood vasculature and is an attractive target for antiangiogenic tumor therapy (52). Approved antiangiogenic drugs such as bevacizumab, lenvatinib, sorafenib, sunitinib, and pazopanib target this pathway and are associated with modest survival advantages in certain kinds of cancers (53). Bevacizumab and ramucirumab are IgG1 antibodies directed against vascular endothelial growth factor A (VEGF-A) and VEGFR2, respectively. To date, clinical PET studies using ⁸⁹Zr-Df-bevacizumab were performed in patients with breast cancer, neuroendocrine tumors, renal cell carcinoma (RCC), NSCLC, and glioma (21, 22, 26, 54, 55).

Based on fundamental preclinical studies (3, 9, 56, 57), in a clinical feasibility study, Gaykema and colleagues demonstrated that ⁸⁹Zr-Df-bevacizumab could be used to detect primary breast cancer (22). Another pilot clinical study in patients with metastatic RCC demonstrated that ⁸⁹Zr-Df-bevacizumab PET visualized renal tumor lesions with striking heterogeneity between lesions, therefore reflecting differences in vascular characteristics (24). Everolimus, an inhibitor of mammalian target of rapamycin (mTOR), also inhibits VEGF-A expression, but there are not reliable biomarkers to predict efficacy of everolimus in patients with metastatic RCC at the moment. van Es and colleagues investigated the value of ⁸⁹Zr-Df-bevacizumab as a tool to identify everolimus efficacy in 13 patients with metastatic RCC, and reported that everolimus decreased ⁸⁹Zr-Df-bevacizumab tumor uptake in 10 patients who received continuous everolimus treatment and achieved stable disease at 3 months (Fig. 2D; ref. 26). Jansen and colleagues reported that ⁸⁹Zr-Df-bevacizumab had poor uptake in mice bearing diffuse intrinsic pontine glioma (DIPG; ref. 23), implying that treatment with bevacizumab in DIPG patients is justified only after ⁸⁹Zr-Df-bevacizumab immunoPET demonstrates positive VEGF expression in tumor tissue. In spite of these disappointing preclinical data, the same team further reported that, in children with DIPG, five of seven primary tumors showed focal ⁸⁹Zr-Df-bevacizumab uptake while no significant uptake was seen in the healthy brain (58). A further study by Veldhuijzen van Zanten and colleagues reported that ⁸⁹Zr-Df-bevacizumab uptake was correlated with microvascular proliferation in a patient with DIPG and highlighted that ⁸⁹Zr-Df-bevacizumab PET could be used to delineate intralesional heterogeneity (Fig. 2E; ref. 59). These results indicate that ⁸⁹Zr-Df-bevacizumab immunoPET is feasible in children with DIPG and may help select patients with the greatest chance of benefiting from bevacizumab treatment. Notably, ⁸⁹Zr-Df-bevacizumab could also offer a tool to refine clinical management of patients with von Hippel–Lindau disease (25), and neuroendocrine tumors (55). From a clinical perspective, a patient's baseline plasma VEGF-A level is an independent prognostic factor for certain cancer types (60), and resistance to antiangiogenic therapies ultimately leads to patients' relapse and poor survival (53, 61). Therefore, it would be significant to develop personalized therapeutic strategies through immunoPET imaging of angiogenesis biomarkers that might help select proper patients and reflect the response of tumor cells to antiangiogenic drugs.

Ranibizumab is a mAb Fab derivative of bevacizumab and has a higher affinity for all soluble and matrix-bound human VEGF-A isoforms than bevacizumab (62). Nagengast and colleagues initially created ⁸⁹Zr-Df-ranibizumab and reported that VEGF PET imaging using ⁸⁹Zr-Df-ranibizumab allowed serial analysis of angiogenic changes in ovarian tumor models following treatment with the kinase inhibitor sunitinib (63). In addition, several studies have been performed to image tumor vasculature by visualizing VEGFR-2. Meyer and colleagues engineered and radiolabeled single-chain VEGF and further validated the feasibility of imaging VEGFR-1 and VEGFR-2 in an orthotopic murine tumor model (64). We synthesized and characterized a ramucirumab-based PET imaging agent, ⁶⁴Cu-NOTA-RamAb, for mapping VEGFR-2 expression in vivo (65), and found that this technology could visualize VEGFR-2 in nude mice bearing HCC4006 and A549 NSCLC tumor models. Based on our results, a study from Eric and colleagues further suggested that a three-time-point method could be used to assess the in vivo performance of ⁶⁴Cu-NOTA-RamAb (66).

The human epidermal growth factor receptor 2 (HER2) transmembrane oncoprotein is overexpressed in many human tumors, and several HER2-targeting agents have entered clinical practice (4). However, despite this success, responses to these antibodies and small molecules have been hampered by resistance or poor responses. In this setting, molecular PET imaging techniques can help select proper patients for subsequent therapy and elucidate the underlying resistance mechanisms (4, 67–69).

Trastuzumab was the first clinically approved anti-HER2 antibody for breast cancer patients. In preclinical studies, trastuzumab has been used as a targeting agent to map HER2 expression and distribution either by PET imaging or by SPECT imaging (4, 13, 70, 71). In addition, ⁸⁹Zr-Df-trastuzumab PET, rather than ¹⁸F-FDG PET, successfully assessed the pharmacodynamic effects of afatinib (an EGFR-HER2 dual inhibitor) in HER2-positive gastric cancer models (Fig. 3A and B; ref. 72). In clinical practice, HER2 status has been determined by immunohistochemistry, and fluorescence in situ hybridization has been validated to predict the efficacy of the HER2-targeting antibody–drug conjugate trastuzumab emtansine. Because SPECT imaging using ¹¹¹In-trastuzumab discovered new HER2-positive lesions in 13 of 15 patients with breast cancer (10), Dijkers and colleagues developed ⁸⁹Zr-Df-trastuzumab (13), and reported that PET imaging using this probe in breast cancer patients was able to detect both previously known metastatic lesions and lesions which had been previously undetected (73). Recently, ⁸⁹Zr-Df-trastuzumab PET has been investigated as a useful tool to predict response in breast cancer patients treated with trastuzumab emtansine, and to improve the understanding of tumor heterogeneity in breast cancers (14). Indeed, the heterogeneity within and across tumor lesions in a single patient or among different patients is increasingly being recognized, with significant therapeutic implications (74). In support of this concept, Ulaner and colleagues demonstrated that ⁸⁹Zr-DFO-trastuzumab PET/CT detected unsuspected HER2-positive metastases in patients with HER2-negative primary breast cancer (15, 75). It is quite difficult for conventional biopsy techniques, which generally obtain samples from a limited number of lesions, to identify these unique cohorts. This exciting finding may facilitate better management of HER2-negative breast cancer patients, as about 10% to 15% of patients with HER2-negative primary breast cancer may still benefit from HER2-targeted therapy (76). Furthermore, early changes in ⁸⁹Zr-Df-trastuzumab uptake in breast cancer metastases following treatment with heat shock protein 90 inhibitor NVP-AUY922 correlated with changes of lesion size measured on CT images (77). In addition to ⁸⁹Zr-labeled trastuzumab, ⁶⁴Cu-DOTA-trastuzumab is another alternative for optimizing trastuzumab treatment (16).

HER2 is overexpressed in EGA and trastuzumab has been approved by the FDA to treat EGA patients with positive HER2 expression (78). Considering the fact that only a subset of patients with HER2-positive EGA respond to trastuzumab (79), O'Donoghue and colleagues studied the value of ⁸⁹Zr-DFO-trastuzumab in assessing HER2 status in primary and metastatic EGA (17), and reported that this imaging method detected local and metastatic lesions in 80% of the imaged patients. This study indicated that ⁸⁹Zr-DFO-trastuzumab PET has superior advantages over single-site biopsy, because this noninvasive imaging technology can assess variation in levels of HER2 and target engagement in both the primary and metastatic tumor lesions simultaneously. Notably, bifunctional chelators have certain impact on the in vitro stability and in vivo performance of ⁸⁹Zr-labeled trastuzumab and antibody–drug conjugates (80).

Pertuzumab is another HER2-targeting mAb approved by the FDA after a survival benefit was achieved for patients with HER2-positive metastatic breast cancer. Importantly, pertuzumab binds to a HER2 binding site distinct from that of trastuzumab, augmenting the binding and treatment efficacy of each other (81). Thus, it is of great importance to noninvasively detect the biodistribution of pertuzumab-specific binding sites when synergistic regimens containing trastuzumab and pertuzumab are given for breast cancer patients. In this setting, Marquez and colleagues conducted an in vivo study in breast cancer models using ⁸⁹Zr-DFO-pertuzumab and observed significant tracer uptake in MDA-MB-231 xenografts (82). More importantly, pretargeting strategies using unlabeled trastuzumab could further enhance ⁸⁹Zr-DFO-pertuzumab accumulation in tumors. Combination treatments with trastuzumab and pertuzumab have also shown enhanced antitumor effects in xenograft models of human ovarian cancer (83), and the addition of pertuzumab to chemotherapy in patients with platinum-resistant ovarian carcinoma demonstrated favorable trends in progression-free survival (84). Based these preclinical and clinical data, our group synthesized ⁶⁴Cu-NOTA-pertuzumab and broadened the application of the probe to detect both subcutaneous and orthotopic ovarian cancer models (85). These solid data demonstrated that ⁶⁴Cu-NOTA-pertuzumab/⁸⁹Zr-DFO-pertuzumab are effective tools for imaging HER2 expression. More recently, a first-in-human study in patients with HER2-positive breast cancer showed that ⁸⁹Zr-DFO-pertuzumab PET/CT was safe and demonstrated optimal imaging 5 to 8 days after administration of the tracer. Additionally, ⁸⁹Zr-DFO-pertuzumab PET/CT was able to image multiple sites of HER2-positive malignancy including HER2-positive brain metastases (86). Further clinical studies are still needed to validate the efficacies of these two probes, especially in improving patient stratification.

Compared with intact antibodies, antibody fragments and single-domain antibodies (sdAb) have superior imaging characteristics, such as rapid clearance, high target-to-background ratios, reduced radiation dose, and engineered sites for site-specific conjugation (87, 88). However, compared with clinically available antibodies, the immune responses, safety profiles, and efficacy of these novel sdAbs in humans are largely unknown. Positron emitter–labeled nanobodies as probes for evaluating HER2 status have been reported (89–92). One strength of such probes, for example, [¹⁸F]RL-I-5F7 and [¹⁸F]RL-I-2Rs15d (89, 90), is their ability to rapidly cross the intact blood–brain barrier and detect brain metastases from breast cancers 1 hour after injection of the tracers. In comparison, the best time points for detecting brain metastases using either ⁸⁹Zr-Df-trastuzumab or ⁸⁹Zr-DFO-pertuzumab were 4 to 5 days for the former tracer and 5 to 8 days for the latter following tracer administration (73, 86). Another strength is the higher tumor-to-blood and tumor-to-muscle ratios which will result in higher contrast PET imaging (89, 93). A phase I study from Keyaerts and colleagues showed PET imaging using a ⁶⁸Ga-HER2-nanobody is a safe procedure and tracer accumulation in HER2-positive breast carcinoma metastases is high, which warrants further assessment in a phase II trial (94).

EGFR and its relatives (i.e., ErbB2/HER2, ErbB3/HER3, and ErbB4/HER4) are well-known oncogenic drivers in several types of cancers such as lung cancer, breast cancer, and glioblastoma. EGFR was among the first RTKs for which the ligand binding mechanism was studied (95). Inhibitors for this specific RTK, including antibody therapeutics (e.g., cetuximab and panitumumab) and small-molecule TKIs (e.g., erlotinib, gefitinib, lapatinib, and osimertinib), have been the most successful examples of molecularly targeted therapies in cancers (95, 96).

Cetuximab is a chimeric EGFR-specific IgG1 mAb and functions by preventing ligand activation and receptor dimerization (97). It was shown in a phase I first-in-human study that ⁸⁹Zr-Df-cetuximab was safe and well tolerated (20). PET imaging using ⁸⁹Zr-labeled cetuximab may also select NSCLC patients, head and neck cancer patients, and colorectal cancer patients who will potentially benefit from cetuximab treatment (18, 19, 98). In recent years, theranostic approaches using mAbs have represented a rapidly expanding component of cancer treatments (12). Song and colleagues prepared ⁶⁴Cu/¹⁷⁷Lu-labeled versions of cetuximab and assessed the theranostic efficacy in an esophageal squamous cell carcinoma model (99). The authors elaborately found that ⁶⁴Cu-PCTA-cetuximab immunoPET evaluated EGFR expression levels in tumors, and ¹⁷⁷Lu-cetuximab radioimmunotherapy effectively inhibited tumor growth much more thoroughly than that of cetuximab alone (Fig. 3C and D). Thus, the pair of ⁶⁴Cu/¹⁷⁷Lu-cetuximab may provide a tailored treatment strategy for EGFR-positive patients by immunoPET imaging of EGFR expression and by selective therapeutic radiation delivery.

Panitumumab is a fully humanized mAb that binds to EGFR and is FDA approved for use in receptor-expressing colorectal cancers without KRAS mutations (100). Wei and colleagues synthesized ⁸⁹Zr-DFO-panitumumab and this probe accumulated in various subcutaneous athymic nude female xenograft models (101). Lindenberg and colleagues then calculated the maximum dosing for effective imaging with ⁸⁹Zr-DFO-panitumumab in 3 patients, and results from this study indicated that injection of approximately 1 mCi (37 MBq) of ⁸⁹Zr-DFO-panitumumab intravenously was safe and detected lung metastases from colorectal cancer, but imaging of the included 3 patients failed to demonstrate significant radiotracer accumulation in tumors for all time points and failed to detect tumors in the liver due to physiologic hepatic excretion of the probe (102). Future studies may add unlabeled panitumumab to optimize the radiotracer's uptake in tumors (103). More recently, imgatuzumab (GA201), a novel humanized anti-EGFR IgG1 isotype mAb, has been found to efficiently inhibit EGFR pathway activation in head and neck cancer patients and may serve as a potential probe for visualizing EGFR-expressing tumors after radiolabeling with isotopes (104, 105).

Repebody, a novel nonantibody protein scaffold (106), has also been engineered to monitor EGFR expression (107, 108). One advantage of repebody-based PET imaging probes is their small molecular weight (∼30 kDa), which may enable early visualization of various EGFR-expressing solid tumors as early as 1 hour after injection of the probes. Furthermore, due to superior circulation clearance and tissue penetration, those probes could be labeled with positron emitters with shorter half-lives, such as ⁶⁴Cu (t_1/2, 12.7 hours) and ¹⁸F (t_1/2, 110 minutes; ref. 109).

As we previously reviewed (110), PET imaging using bispecific peptide- and antibody-based heterodimers may demonstrate higher targeting efficacy and specificity than their monospecific peers. We generated a bispecific immunoconjugate (denoted as Bs-F(ab)₂) by linking a cetuximab Fab and a TRC105 Fab (targeting CD105), and confirmed that ⁶⁴Cu-NOTA-Bs-F(ab)₂ had a synergistic improvement on both affinity and specificity in visualizing small U87MG tumor nodules (<5 mm in diameter; ref. 111). Similar results were reported by Kwon and colleagues who developed an immunoconjugate targeting HER2 and EGFR and verified its diagnostic efficacy in breast cancer models (112), and by Mahmood and colleagues who prepared PET probes specific to EGFR and HER3 and monitored resistance to PI3K and protein kinase B (PKB, also known as Akt) inhibitors using breast cancer models as well (66).

Unlike other members of the family, HER3 has relatively weak kinase activity but still can form highly activated heterodimers with EGFR and HER2. Activating mutations in HER3 have been identified in multiple cancer types and therefore HER3 is now being examined as a direct therapeutic target (113). Preclinical and clinical PET imaging has also been developed using anti-HER3 antibodies, such as lumretuzumab (114, 115), patritumab (116), and mAb3481 (117), or using peptides (118), and affibodies (119–121). In a pilot clinical trial that included 11 participants, Lockhart and colleagues reported that although administration of ⁶⁴Cu-DOTA-patritumab is safe, the diagnostic efficacy was limited in solid tumors (116). Comparatively, ⁸⁹Zr-Df-lumretuzumab PET could visualize intra- and inter patient tumor heterogeneity and target accessibility, and the most optimal PET conditions were found to be 4 and 7 days after administration of ⁸⁹Zr-Df-lumretuzumab with 100 mg cold antibody (Fig. 3E; ref. 115). In the absence of side effects of lumretuzumab in treating HER3-expressing solid tumors (122), ⁸⁹Zr-Df-lumretuzumab may have potential value in guiding lumretuzumab treatment of patients with HER3-positive solid tumors. More recently, Warnders and colleagues constructed a HER3 signaling–specific probe, ⁸⁹Zr-MSB0010853, and reported that ⁸⁹Zr-MSB0010853 PET could detect HER3-overexpressing H441 NSCLC cancer models in vivo (123). Although development of therapeutic HER3 antibodies is still in an early phase (122, 124), these preliminary results shed light on the feasibility of noninvasive HER3-specific PET imaging for mapping HER3 expression and distribution of HER3 antibodies.

Apart from the above RTKs, there are 17 more RTK subfamilies and nearly half of these RTK families are poorly understood (2). One example is the platelet-derived growth factor receptor-alpha (PDGFR-α), which has been reported to be involved in tumor angiogenesis and maintenance of several cancer types including thyroid cancer (125–127). Importantly, PDGFR-α is associated with lymph node metastases in papillary thyroid carcinoma (PTC; ref. 127). Michael and colleagues described a PDGFRα-specific immunoPET probe—⁶⁴Cu-NOTA-D13C6—for PTC and demonstrated that PET imaging using this agent could delineate PDGFRα expression in PTC models in vivo and could therefore potentially serve as a novel and promising radiotracer for imaging PDGFRα-positive metastasis from PTC (128). Another example is the insulin-like growth factor-1 receptor (IGF-1R), which has a well-established role in malignant tumors and has become a promising target for cancer therapy and imaging (129, 130). As R1507 is a mAb directed against IGF-1R, Heskamp and colleagues developed ¹¹¹In-R1507 and ⁸⁹Zr-Df-R1507 and reported that both of the tracers could be applied to noninvasively determine IGF-1R expression in vivo in breast cancer xenografts (131). Our group also developed an IGF-1R–specific PET probe (denoted as ⁸⁹Zr-DFO-1A2G11) and demonstrated highly specific uptake of the tracer in IGF-1R–positive pancreatic tumor models (132), implying the potential value of this probe in identifying patients that may benefit from anti–IGF-1R therapy.

In addition, RTKs are important drug targets, and several small-molecule inhibitors have been developed and approved by the FDA in addition to mAbs. Actually, development of radiolabeled TKIs and their analogues is under preclinical and clinical translational research (133), and various radiolabeled TKIs have been developed to image tumors in both preclinical and clinical studies for several RTKs, such as EGFR (134, 135), HER2 (136), VEGFR (137), tropomyosin receptor kinase (Trk; ref. 138), stem cell growth factor receptor (SCFR; ref. 139), and IGF-1R (129, 140, 141).

Currently, the role of spatial deregulation of RTKs in tumorigenesis may be vastly underappreciated, and the development of high-resolution immunoPET techniques for detecting RTK activity in normal and tumor tissues will be essential for studying the role of spatial RTK patterning in the heterogeneous tumor environment. In this state-of-the-art review, we presented typical examples of deregulated RTKs in cancers and summarized strategies to prepare noninvasive PET imaging probes specific for these targets. The current review demonstrates that noninvasive immunoPET could accurately delineate RTK status within a tumor or across multiple tumors within a patient. Visualization of RTK phenotype using immunoPET could facilitate selection of patients for targeted therapies and response assessment/prediction. In addition, a greater understanding of the spatial expression of RTKs with the help of immunoPET could affect ongoing efforts to develop therapeutic drugs targeting RTKs.

In 2017, the development of antibody therapeutics proceeded at a fast pace, and this is expected to continue in the coming years. Several RTK-targeting therapeutic antibodies [such as margetuximab, (vic-)trastuzumab duocarmazine, and DS-8201 for HER2, and depatuxizumab mafodotin for EGFR] are undergoing evaluation in late-stage clinical studies of patients with cancer (142). For mAb-based immunoPET probes, ⁸⁹Zr is utilized in clinical trials much more extensively than any other radiometals. DFO and DFO-based bifunctional chelators are the most commonly used chelators for ⁸⁹Zr-radiolabeling (143), and there are two simple protocols that can be followed to do ⁸⁹Zr radiolabeling (144, 145). However, DFO is not an optimal chelator for ⁸⁹Zr coordination because of ⁸⁹Zr transchelation, which will cause high bone uptake of ⁸⁹Zr. This was exemplified by a recent study by Vugts and colleagues, in which the authors reported that, when compared with DFO, the bifunctional isothiocyanate variant of desferrioxamine (denoted as octadentate DFO*) was advantageous for ⁸⁹Zr labeling because ⁸⁹Zr-DFO*-trastuzumab had enhanced stability in in vitro studies and superior performance in in vivo studies over ⁸⁹Zr-DFO-trastuzumab (80). Therefore, there is still room for future studies to optimize radiolabeling strategies when developing ⁸⁹Zr-immunoconjugates (143). With the advent and maturation of click chemistry (146), future studies may further harness this powerful method to synthesize and assess antibody-based PET imaging probes with improved in vivo performance. In addition to singly targeted imaging probes, antibody-based heterodimers or dual-targeting probes may have higher targeting efficacy and superior specificity than their corresponding monospecific peers (110, 147).

Generally, adequate tissue penetration ability, which is inversely proportional to the size of an imaging probe, and high signal-to-noise ratio, which is closely related to the circulation of a probe, are two important properties of an immunoPET probe. sdAbs, also known as nanobodies or heavy chain–only antibodies bearing a variable region (VHH), hit the sweet spot and have emerged as substitutes for their full-size counterparts in diagnostic or therapeutic applications (87). Enzymatic methods such as sortase have been used to append metal chelators to enable labeling sdAbs with ⁶⁴Cu or ⁸⁹Zr (148, 149). Because a first-in-human trial demonstrated the success of ⁶⁸Ga-HER2-nanobody in a clinical setting (94), and it is easier for immunoPET probes derived from sdAbs to traverse the blood–brain barrier (90), future studies may further design noninvasive means to detect RTKs through the use of sdAbs (150).

In the development of antibody-based diagnostic, therapeutic, or theranostic probes, we should pay attention to the fact that the immunodeficient strains of animals we use in preclinical studies may impact the in vivo fate and performance of the investigated immunoPET probes. For example, a recent study reported that immunoPET radiotracers had inefficient tumor targeting and high off-target binding to the spleen in highly immunodeficient mouse models (151). This was consistent with a previous study which reported that antibody–drug conjugates had limited antitumor activity in NSG mice (152). In this setting, pretargeting strategies may also be harnessed for optimizing both the imaging quality and therapy effects in the coming future (151, 153, 154).

Besides assessing RTKs and selecting proper patients for subsequent molecularly targeted therapy using either small molecules or antibodies, noninvasive imaging of RTKs could facilitate image-guided radionuclide therapy. Actually, therapy of B-cell lymphomas with radiolabeled mAbs has produced impressive clinical results because two radiolabeled anti-CD20 antibodies, ¹³¹I-tositumomab and ⁹⁰Y-ibritumomab tiuxetan, were approved by FDA more than a decade ago (155, 156). RTKs are ideal candidates for investigating and performing radioimmunotherapy (157, 158). Therefore, radiolabeled antibodies or antibody fragments targeting markers included in the current review and other potential neoantigens may provide another therapeutic option for patients with cancer in the era of molecularly targeted therapy and precision medicine.

We firmly believe that immunoPET will allow for better cancer patient management in the clinical setting with the fast development of this field. Considering most data with immunoPET probes have been generated in preclinical stages or in small groups of patients, further studies are still needed to push clinical translation of some of these promising probes and to confirm the value of clinically reported probes.

No potential conflicts of interest were disclosed.

This work was partially sponsored by the Ph.D. Innovation Fund of Shanghai Jiao Tong University School of Medicine (No. BXJ201736) and the China Scholarship Council (No. 201706230067) to W. Wei, the Shanghai Key Discipline of Medical Imaging (No. 2017ZZ02005) to Q.Y. Luo, the American Cancer Society (125246-RSG-13-099-01-CCE), and the NIH (P30CA014520) to W. Cai. E.B. Ehlerding was partially sponsored by the NIH (T32CA009206 and T32GM008505).