Approaches to Imaging Immune Activation Using PET

Assessment of Treatment Response

¹⁸F-FDG PET/CT is commonly used for staging and assessment of treatment response in several malignancies (9). Response assessment criteria such as the Lugano criteria for lymphomas and adaptations of PERCIST for solid tumors are used for interpretation in the setting of conventional therapies such as cytotoxic chemotherapies and radiation therapy (15–19). However, there are challenges to assessing response to immunotherapies as they may be associated with certain unique response patterns due to the underlying mechanisms of action of the therapeutic agent (4). Other than the standard response patterns seen with conventional therapies, immunotherapies have additional imaging patterns such as pseudoprogression and hyperprogression.

Pseudoprogression refers to the apparent impression of progressive disease by increasing size or metabolic activity of preexisting lesions or the appearance of new lesions with otherwise clinical improvement. However, when these patients are followed up without a change in therapy, subsequent imaging, typically obtained 6–8 wk after therapy, shows overall regression of the disease (Fig. 2). This pattern is relatively more common with cytotoxic T-lymphocyte–associated antigen 4 inhibitors, such as ipilimumab, compared with the anti–programmed death protein 1 (PD-1) and its ligand, programmed death ligand 1 (PD-L1) inhibitors, and is seen more often with combination immunotherapies than with single-agent immunotherapy (20). The underlying mechanism of this imaging pattern is postulated to reflect the increased infiltration and activation of immune cells in the TME, promoting cytokine release and local edema, which manifests as an increase in the size of the lesions on imaging. Further, the increased and activated immune-cell population leads to an overall increase in the glucose metabolism in the TME, manifesting as increasing SUVs on an ¹⁸F-FDG PET/CT image. Even previously undetectable sites of disease may now be apparent because of this increased immune-cell infiltration and activation (9,21). Over time, these activated immune cells exert their cytocidal effects, leading to an overall reduction in the size and metabolic activity of the tumors, which can then be captured on a follow-up ¹⁸F-FDG PET/CT. It is therefore important to not directly attribute such imaging appearances to progressive disease and remain cognizant of pseudoprogression, which should be further confirmed or excluded on a follow-up ¹⁸F-FDG PET/CT scan. In contrast, the prevalence of pseudoprogression is typically around 10%, which implies that most progressive disease patterns seen on imaging still represent a true progressive disease, especially with clinical decline (2,22–24).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 2.

63-y-old woman with urethral melanoma being treated with pembrolizumab developed metastatic right pulmonary nodule (black arrow), and her treatment was changed to combination ipilimumab (Ipi) and nivolumab (Nivo). Her follow-up ¹⁸F-FDG PET/CT at 2.5 mo showed new lymphadenopathy involving bilateral axillary, retroperitoneal, iliac and inguinal stations (red arrows), new ¹⁸F-FDG uptake in endometrium (blue arrow) in addition to enlarging pulmonary nodule (black arrow). There was also increased ¹⁸F-FDG uptake in spleen and bilateral renal cortices (brown arrows). Taken together, these findings were suggestive of immunoreactive process contributing to pseudoprogressive pattern, with immunotherapy-related nephritis and likely endometrial inflammation. Patient continued treatment, and pseudoprogression was confirmed on follow-up ¹⁸F-FDG PET/CT obtained 6 wk later and ongoing complete response 18 mo later.

Hyperprogression represents a markedly rapid and fulminant disease progression with clinical worsening in patients being treated with immune-targeted therapies (Fig. 3) (25). Although there are varying definitions of hyperprogression, the common elements include the rate and degree of tumor growth. The pathogenesis of hyperprogression is incompletely understood, but proposed mechanisms include the modulation of TME with activation of the regulatory T-cells and macrophage subtypes to promote a locally immunosuppressed environment (26,27). Identification of hyperprogression is important, as these patients need to be switched to alternative treatments with cessation of immunotherapy. Hyperprogression has also been associated with poor overall survival, worse than that seen with conventional progressive disease (28). The incidence of hyperprogression has varied across studies, mainly due to nonuniform definitions, and a metaanalysis reported an incidence ranging from 5.9% to 43.1% across 24 studies with 3,109 patients (25).

Despite these atypical response patterns with immunotherapy, response assessment using ¹⁸F-FDG PET/CT continues to be a robust predictor of patient outcomes, and multiple response criteria have been proposed to tackle these challenges (Table 1) (6,29). PET-derived quantitative parameters such as metabolic tumor volume, total lesion glycolysis, and SUV ratios, measured at baseline and follow-up, have been associated with overall survival of patients with metastatic melanoma receiving immunotherapy (6,29). ¹⁸F-FDG PET/CT has also been applied for response assessment with immunotherapy in a neoadjuvant setting, and although the literature is currently immature, the results seem promising (30,31). The use of ¹⁸F-FDG PET/CT at baseline, during therapy (interim assessment), and at the end of therapy is also recommended by the joint European Association of Nuclear Medicine/Society of Nuclear Medicine and Molecular Imaging/Australian and New Zealand Society of Nuclear Medicine procedure standards (32–34). These guidelines also recommend extraction of volumetric parameters as previously described, which can aid in the assessment of response and have important prognostic value.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 3.

71-y-old woman with previously excised left lower extremity melanoma developed hepatic metastases (red arrow), left inguinal lymph nodes (black arrow), and subcutaneous deposits. Around 3 wk later, with 1 cycle of relatlimab and nivolumab therapy completed, patient worsened clinically, with abdominal pain and distention, increased fatigue, and dyspnea. Follow-up ¹⁸F-FDG PET/CT showed marked increase in hepatic disease burden (red bracket). Hyperprogression was suggested, and immunotherapy was discontinued. Patient declined additional therapy and unfortunately died approximately 10 mo later. During 4-wk period between 2 ¹⁸F-FDG PET/CT studies, her hepatic metabolic tumor volume had 5-fold increase (baseline, 466 cm³; follow-up, 2,703 cm³). Note also reduced FDG uptake in brain (brown arrow), likely secondary to tumor sink-effect.

irAEs

Immune-targeting therapies are associated with certain adverse effects, termed irAEs, several of which produce characteristic imaging patterns on ¹⁸F-FDG PET/CT (Table 2) (35–37). Recognition of these irAEs is useful as they may impact management, including additional treatment with corticosteroids, and temporary or even permanent cessation of immunotherapy based on the severity of the effect. Occasionally, irAEs may manifest at a subclinical stage on ¹⁸F-FDG PET/CT, preceding any clinical signs or symptoms. Thus, familiarity with the imaging manifestations is important, so that these findings can be communicated promptly and appropriate management instituted to potentially reduce the severity or prevent the occurrence of the clinically manifest event (Figs. 2 and 4).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 4.

Key manifestations of irAEs on ¹⁸F-FDG PET/CT. (A) Immune-related thyroiditis in patient with metastatic melanoma treated with immune checkpoint inhibitors (ICIs) presenting as diffuse increased ¹⁸F-FDG uptake in thyroid with resolution 3 mo later. (B) Immune-related sarcoidlike reaction in patient with metastatic melanoma treated with ICIs, presenting with new and worsening ¹⁸F-FDG–avid symmetric mediastinal and bilateral hilar lymphadenopathy with resolution of known disease sites, and spontaneous resolution on follow-up ¹⁸F-FDG PET/CT 5 mo later. (C) Immune-related colitis in patient with metastatic melanoma treated with ICIs presenting with new diffuse increased ¹⁸F-FDG uptake in colon and multiple episodes of diarrhea, with spontaneous resolution 2 mo later. (D) Immune-related hypophysitis in patient with metastatic melanoma presenting as diffuse increased ¹⁸F-FDG uptake in the sella, correlating with new enlargement of pituitary on MRI and resolution after treatment with prednisone. (E) Immune-related pneumonitis in patient with metastatic melanoma presenting as diffuse increased ¹⁸F-FDG uptake throughout both lungs, necessitating cessation of ICI and initiation of steroids, which led to resolution 5 mo later. (F) Immute-related cholangitis presenting as new, increased ¹⁸F-FDG uptake along bile ducts that appear dilated and thickened on MRI in patient treated with ICIs. (G) Immune-related arthritis: 2 examples of hip and knee arthritis, with new diffuse increased ¹⁸F-FDG uptake at hip (left panel) and knee (right panel) joints after initiation of ICIs and correlating with new and increasing pain at these sites. Ipi = ipilimumab; Nivo = nivolumab.

Immune Cells

T-lymphocytes, a central component of the adaptive immune response, play a crucial role in recognizing and attacking pathogens, tumors, and other abnormal cells and are a key effector in the cellular antitumor response. Given their significance, developing specific radiopharmaceuticals to track their activation, migration, and proliferation noninvasively can provide valuable insights for predicting response to therapy and longitudinal tracking of spatial changes in the T-cell population (46).

The approaches to image T-cells include radiolabeled antibodies, antibody fragments, and other small molecules designed to bind specifically to surface markers or subcellular components of T-cells (47,48). As a T-cell surface marker, CD3 is a global imaging marker for the total T-cell population. Thus, noninvasive anti-CD3–targeted imaging can partly segregate the T-cells enriched (irrespective of the type) from T-cell–poor TMEs. CD3-targeting PET imaging using ⁸⁹Zr-DFO-CD3 has been used in models of colon cancer to predict response to anticytotoxic T-lymphocyte–associated antigen 4 immunotherapy noninvasively (49). A limitation of this approach is that pan-T-cell markers such as CD3 do not provide adequate information about the cellular subtypes and their functional status within the TME. For example, regulatory T-cells are involved in immune-modulatory pathways that promote cytotoxic T-cell exhaustion, promoting a protumorigenic environment. This information cannot be reliably captured with imaging of surface markers such as CD3, in the absence of more specific markers of the T-cell subtype (50). Antibody or antibody fragment–based targeting of the CD8 receptor (e.g., with ⁸⁹Zr-Df-IAB22M2C) can help in the visualization of CD8+ T-cell infiltration in the TME, thus informing on the downstream response to immunotherapy (51). The main drawbacks of this imaging technique include the ubiquitous presence CD8+ T-cells in lymphoid tissues and the inability to differentiate activated from exhausted T-cells.

Immune Cell Function

As T-cells are distributed throughout the body, imaging techniques have been developed to capture intracellular signaling molecules that indicate T-cell activation. Granzyme B, a serine protease released by activated CD8+ T-cells and natural killer cells, is a potent driver of apoptosis, and its imaging has been evaluated in preclinical and clinical models (48,52). Granzyme B PET imaging allows for the early detection of T-cell activation and has shown potential in predicting antitumor responses to immune checkpoint inhibitor therapies (53). Preliminary clinical data show that ⁶⁸Ga-grazytracer PET/CT (targeting Granzyme B) shortly after initiating immune checkpoint inhibitor therapy can predict response to treatment. Patients with a positive ⁶⁸Ga-grazytracer PET/CT showed a more favorable response to therapy compared with those with a negative scan (53). In addition, a preclinical study demonstrated that Granzyme B–targeted PET imaging could noninvasively detect irAEs, potentially improving the diagnosis and management of these toxicities (54). Another study using ⁶⁸Ga-NOTA-GZP PET/CT for imaging of Granzyme B in mice models of inflammatory bowel disease showed that the tracer activity was high in animals with active inflammation versus those without and declined over time in responders after administration of tumor necrosis α inhibitor (55). Another approach to imaging activated T-cells is using 2′-deoxy-2′-[¹⁸F]fluoro-9-β-d-arabinofuranosyl guanine (¹⁸F-AraG), a radioactive analog of arabinosyl guanine, which is a substrate for deoxynucleotide kinases and thus accumulates preferentially in activated T-cells (56). The nucleoside salvage pathways, mediated by deoxycytidine and deoxyguanosine kinases, are crucial for replenishing the nucleotides in activated and proliferating T-cells, and thus, AraG is preferentially taken up by these cells (57). ¹⁸F-AraG has been shown to delineate activated CD8+ T-cells in animal tumor models after initiating immune-priming chemotherapies (57). Clinical trials are under way to further study the changes in the biodistribution of ¹⁸F-AraG before and after initiation of treatment in patients with various malignancies with the broad goal of predicting treatment outcomes (NCT04678440, NCT03142204).

Interferon-γ (IFN-γ) and IL-2 are cytokines released by several T-cell subtypes. IL-2, a cytokine central to T-cell proliferation and activation, has been a target of interest for imaging because it reflects the activity of activated T-cells (58). The binding of IL-2 to its receptor leads to the activation and differentiation of T-cells; thus, imaging this axis is of clinical relevance (58). Radiolabeled IL-2, such as ¹⁸F-FB-IL2, has been investigated for tracking T-cell activation in vivo, especially in cancer immunotherapy (59). Although imaging was safe and feasible, serial ¹⁸F-FB-IL2 PET/CT in a small set of patients with metastatic melanoma did not predict response to immune checkpoint inhibitor therapy (59). Similarly, IFN-γ, a key cytokine secreted by activated T-cells (type 1 T-helper cells and cytotoxic T-cells) and natural killer cells, plays a crucial role in antitumor immunity and host defense against infections (60). ⁸⁹Zr-anti-IFN-γ PET in mice models has been shown to characterize the status of T-cells in the TME, with increased tracer accumulation correlating with a strong antitumor immune response and, conversely, a low tracer activity correlating with a dysfunctional and exhausted T-cell population with increased expression of PD-1 on the tumor cells (50).

PET imaging of cancer-associated fibroblasts using radiolabeled fibroblast activation protein inhibitors has shown clinical promise in various oncologic and nononcologic inflammatory diseases (61). Similarly, PET imaging of tumor-associated macrophages by targeting CD206, a mannose receptor expressed on the immunosuppressive M2-like macrophages, can predict response to immunotherapy, with a high PET signal indicating a poor outcome (62).

Immune Checkpoints as Targets

Immune checkpoint proteins, such as PD-1, PD-L1, and cytotoxic T-lymphocyte–associated antigen 4, are intricately involved in maintaining an immunosuppressive, protumorigenic TME. Therapeutic inhibition of these proteins using immune checkpoint inhibitors has led to significant improvements in clinical outcomes of patients with several different malignancies, and thus, noninvasive imaging of these targets has been of clinical interest (5,7). PD-1 is expressed most prominently on the surface of activated T-cells and to a lesser extent in other immune cells, such as natural killer cells, macrophages, and dendritic cells. PD-L1 is also expressed on the surface of immune cells but is also expressed by tumor cells as a mechanism of evading immune surveillance. In vivo characterization of the PD-1/PD-L1 axis can help assess the heterogeneity of expression of these proteins in the TME, both spatially and temporally (using serial imaging obtained longitudinally), information that is vital for therapeutic efficacy and cannot be captured reliably using immunohistochemistry. The PD-1/PD-L1 axis can be visualized noninvasively using radiolabeled antibodies or small molecules (63,64). Preclinical and preliminary clinical data suggest that whole-body quantification of PD-1 and PD-L1 is feasible using an adnectin-based radiopharmaceutical (¹⁸F-BMS-986192) and a radiolabeled antibody (⁸⁹Zr-nivolumab), respectively (63). In vivo assessment of PD-L1 status in patients with breast, bladder, or lung malignancies using ⁸⁹Zr-Atezolizumab showed an overall heterogeneity in tracer activity across the tumor types as well as within the same lesions in several patients (65). This study also showed a positive correlation between favorable treatment response to atezolizumab therapy and the increasing intensity of tracer uptake across the lesions, although in a relatively small group of patients (n = 22). A challenge with PD-1/PD-L1 imaging–based prediction of response to immune checkpoint inhibitor therapy is the expression of these proteins on other immune cells, including the immunosuppressive regulatory T-cells, which limits the selective characterization of effector T-cells (5). Moreover, as with the assessment of T-cell markers, PD-L1 expression is seen in many nontumor tissues such as the spleen, making its use in patient selection for PD-1/PD-L1 targeting therapeutic agents difficult.

Reporter Gene and Chimeric Antigen Receptor (CAR) T-Cell Imaging

A reporter gene is an artificially introduced sequence of nucleic acids used to detect or “report” the functionality or expression of the gene of interest. This information is valuable in elucidating the mechanisms, tracking the temporal changes, and predicting the success of gene therapies (66–69). These reporter genes may be selected on the basis of their ability to express a target that can be visualized on PET imaging using a specific radiopharmaceutical (67). Ideally, this target should not be expressed in the native host or expressed only at limited, well-defined sites so that its detection by imaging of the gene of interest is unequivocal and unambiguous. These reporter systems should also be nontoxic and without any significant downstream physiologic activity other than expression of the detectable target (68).

Traditional imaging reporter gene systems express targets belonging to 1 of these 3 categories: enzymes, receptors, or transporters (68). The most widely used enzyme-expressing reporter gene is the Herpes simplex virus 1–thymidine kinase system. This reporter gene expresses the enzyme thymidine kinase, which can be imaged using different PET-based probes, including 9-(4-[¹⁸F]fluoro-3-hydroxymethylbutyl)guanine (70). Reporter genes expressing receptors, such as somatostatin receptor or prostate-specific membrane antigen, can be imaged using established PET radiopharmaceuticals such as ⁶⁸Ga/⁶⁴Cu-DOTATATE and ⁶⁸Ga-PSMA-11 or ¹⁸F-DCFPyL, respectively (71,72). A reporter gene expressing a transporter includes the well-characterized sodium iodide symporter, which can be imaged with the PET probe, ¹²⁴I-sodium iodide (73).

CAR T-cells are modified T-cells with a chimeric extracellular antibody construct specific to a tumor antigen that, on binding to the target, initiates T-cell activation. Despite the success of CAR T-cell therapy in various hematologic malignancies, relapses are frequent and partly related to a lack of CAR T-cell persistence or immune-cell exhaustion (74). Thus, noninvasive in vivo imaging that can localize and monitor CAR T-cell activity longitudinally is valuable for predicting treatment response. CAR T-cells can be tracked using the previously described reporter genes by adding a reporter gene construct with the chimeric T-cell receptor (75,76). A recent study evaluated the feasibility of antigen-based PET imaging of the CAR T-cells using the ectodomain of CD19, a common B-cell marker and target for several CAR T-cells (74). The purported benefits of antigen-based imaging, compared with reporter gene imaging, are the low immunogenic potential of the technique, no requirement of alterations in the CAR T-cell model, and the feasibility of real-time imaging at multiple time points. Although still early in development, these imaging technologies are promising candidates for future clinical translation.