Immune PET Imaging

Abstract

FDG PET/CT is sensitive to the metabolic, immune-related, and structural changes that can occur in tumors in the setting of cancer immunotherapy. However, unique mechanisms of immune checkpoint inhibitors (ICIs) can occasionally make response evaluation challenging, as tumors and inflammatory changes are both FDG avid. We discuss these response patterns and additional sequelae of ICI immunotherapy such as immune-related adverse events. We also review new immune-specific PET imaging probes that are either at the preclinical stage or in early clinical trials, which may help guide the clinical management of cancer patients treated with immunotherapy, and will likely have applications outside of oncology for other diseases in which the immune system plays a role.

INTRODUCTION

To promote their own proliferation and survival, cancer cells are known to escape immune surveillance and suppress the immune response.^1,2 Although the processes by which tumors escape immune surveillance are not completely understood, some of these mechanisms have been elucidated.³ For example, tumors can express T cell suppressor proteins, either constitutively or in response to the initial immune response in the tumor microenvironment.⁴ As a result, drugs that target suppressors of cytotoxic T cells have been seen as attractive tools in immunotherapy.

Exemplary immunotherapeutic agents that have demonstrated survival benefit include immune checkpoint inhibitors such as ipilimumab, pembrolizumab, and nivolumab, which are usually administered every 2–3 weeks.^5–7 Ipilimumab inhibits cytotoxic T lymphocyte-associated protein 4 (CTLA-4) by preventing CTLA-4 binding to the B7 ligand on antigen presenting cells (APCs) or even tumors.^8–11 Nivolumab and pembrolizumab inhibit the membrane protein programmed cell death-1 receptor (PD-1) on cytotoxic T cells, preventing PD-1 binding to programmed cell death ligand 1 or 2 (PD-L1/PD-L2) (Figure 1).^12,13 Similarly, PD-L1 inhibitors such as atezolizumab, avelumab, and durvalumab prevent PD-1 binding on cytotoxic T cells. In the absence of therapy, signaling through CTLA-4 or the PD-1 axis leads to suppression of cytotoxic T cell function and persistence of tumors. Treatment with checkpoint inhibitors blocks these inhibitory signals, and leads to activation of the immune response with T cell activation and expansion.

Figure 1.

Illustration highlighting the interaction between CTLA-4 (on a T cell) and B7 (on an antigen presenting cell or tumor), and interaction between PD-1 (on a T cell) and PD-L1 (on an antigen presenting cell or tumor). Inhibitors of CTLA-4 (for example, ipilimumab) block the interaction between CTLA-4 and B7. Nivolumab and pembrolizumab are examples of immunotherapy agents that block PD-1, preventing the interaction between PD-1 and PD-L1. These inhibitors enhance anti-tumor activity through the aforementioned blockades.

APC = antigen presenting cell, TCR = T cell receptor, MHC = major histocompatibility complex, CD28 interacts with B7 to generate a co-stimulatory signal to T-cells. Yellow “oval” insert between MHC and TCR indicates processed peptide presented by MHC to TCR/T cell.

Immunotherapies have been used to re-engage and augment the immune response against a variety of malignancies such as melanoma, non-small cell lung cancer, renal cell cancer, urothelial cancer, head and neck squamous cell cancer, Merkel cell carcinoma, and Hodgkin lymphoma, and the role of immunotherapy in the treatment of cancer continues to expand.^{5–7,14–19} Although immunotherapy is generally associated with more frequent durable responses compared to chemotherapy or targeted therapy, more than 50% of patients do not respond.^20–22 Given non-redundancy in immune checkpoint pathways, combination immunotherapy has been utilized to increase efficacy, although combination therapies are also associated with greater toxicity.^7,23–27

Imaging is routinely used for diagnosis, staging, treatment planning and response assessment in oncology. Ultrasound, computed tomography (CT), and magnetic resonance imaging (MRI) can identify potential tumor lesions and assess changes in the size and density of lesions after treatment. In addition, functional imaging using positron emission tomography (PET) with 2-deoxy-2[¹⁸F]fluoro-D-glucose (FDG) is routinely used due to its high sensitivity for detecting malignancy and characterizing tumor metabolism. In this article, we discuss the role of FDG PET/CT in the assessment of treatment response after cancer immunotherapy, and identify a few approaches that utilize FDG PET/CT to evaluate the immune response. We also review new immune-specific PET imaging probes that are just beginning to be explored in early phase clinical trials.

ATYPICAL RESPONSE PATTERNS IN IMMUNOTHERAPY

The novel mechanism of action of immunotherapies, with immune and T cell activation, has the potential to lead to unusual patterns of response, such as pseudoprogression or hyperprogression, which are discussed below. However, it is important to note that these atypical responses are quite rare, and the vast majority of patients treated with current immunotherapy regimens have typical response patterns. In addition, it is important to be aware of potential immune-related adverse events (irAE), which can result in misleading findings on imaging.

Pseudoprogression

For some patients on immunotherapy, tumors can transiently increase in size, or new lesions may be seen.²⁸ If follow-up evaluation shows resolution of the new lesions, and decreasing size or resolution of the lesions that had previously grown, this is termed pseudoprogression (Figure 2), and can be early or delayed.²⁹ This phenomenon likely occurs as a result of tumor infiltration by immune cells, which has been confirmed by biopsy in a few cases.^30–32 Many of these transiently increased/new lesions will also be avid on FDG PET/CT,³³ and in some patients, pseudoprogression may be associated with clinical symptoms.³⁴

Figure 2.

Schematic comparing response patterns following immunotherapy. Hyperprogression (red line) indicates a rapid increase in disease burden following immunotherapy, such that disease progresses at a significantly faster rate when compared to the pre-immunotherapy period. In routine progression (yellow line), tumor growth is grossly unchanged or only slightly diminished after initiation of immunotherapy. In pseudoprogression (blue line), tumors initially increase in size, however, subsequent anatomic or metabolic imaging demonstrates a decrease in disease burden. The green curve represents a typical response pattern, with tumor shrinkage following treatment.

Pseudoprogression has mainly been reported in melanoma patients treated with ipilimumab (occurring in up to 15% of cases), and appears much rarer with the use of anti-PD-1/PD-L1 agents.^35–37 Billan and colleagues have compiled frequencies of pseudoprogression in pooled studies and clinical trials where anti-PD-1 axis immunotherapy agents were used to treat different cancers, and frequencies range from 1.3 to 9.3%.²⁸ In the largest analysis to date involving 19 clinical trials and 2400 participants, nivolumab and pembrolizumab were investigated in various advanced solid tumors, and pseudoprogression was observed in 6.3% of patients.³⁸ Thus, morphologic increase in tumor volume or metabolic activity on FDG PET/CT is much more likely to reflect true progressive disease.

Interestingly, pseudoprogression has also be seen with chimeric antigen receptor (CAR) T cell therapy when patients are imaged early; a case report of a patient with relapsed B-cell acute lymphoblastic leukemia noted pseudoprogression of extramedullary disease on MRI at 16 days post CAR T cell treatment, with subsequent response on day 30.³⁹ In contrast, when patients with lymphoma were imaged at 1 month post CAR T cell therapy via FDG PET/CT, no evidence of pseudoprogression was identified even in patients that had cytokine release syndrome, suggesting that pseudoprogression should not be a confounding factor for routine follow-up scans in patients treated with CAR T cell therapy.⁴⁰

Hyperprogression

Hyperprogression of cancer after the initiation of immune checkpoint inhibitors is a recently described response pattern in a subset of patients receiving PD-1/PD-L1 axis inhibitors.⁴¹ Hyperprogression is considered to be a therapy-induced acceleration of tumor growth kinetics (Figure 2), and has been defined as treatment failure of less than 2 months, or a two-fold or greater increase in tumor burden/growth rate during immunotherapy.⁴² However, the existence of hyperprogression continues to be controversial, given that it is difficult to establish if rapid progression is due to the natural history of the disease or an immunotherapy-induced process.^42,43

In several reports, pre-baseline, baseline, and post-treatment scans were utilized, so that the tumor growth rate during immunotherapy could be compared with the growth rate prior to immunotherapy. Champiat and colleagues used this approach to show that 12/131 (9%) patients who received anti-PD-1/PD-L1 immunotherapy could be classified as hyperprogressors.⁴¹ In another study where pre-baseline, baseline and post-therapy scans were available, 6/155 (4%) of patients experienced hyperprogression.⁴⁴ However, additional studies are needed to understand the biology driving hyperprogression, and provide more evidence for this controversial phenomenon.

Immune-Related Adverse Events

Immunotherapeutic agents can cause off-target side effects known as immune-related adverse events (irAEs), which result from inflammation of various organs/organ-systems. irAEs usually occur within 12 weeks of immunotherapy initiation, and commonly occur in the skin and gastrointestinal tract (Figure 3), although the pancreas (Figure 4), thyroid gland (Figure 5), pituitary gland, liver, lung (Figure 6), heart, and joints may also be affected.^45,46 Sarcoid-like reactions can also occur as a manifestation of irAE (Figure 7).⁴⁷ Although incidence rates vary by organ-system, irAEs may occur in over 50% of patients, and they appear to be more common in patients on anti-CTLA-4 monotherapy and combination immunotherapy.^45,46,48 Fatality is rare and ranges from 0.3 to 1.3% in patients treated with PD-1/PD-L1 and CTLA-4 inhibitors, and is more frequently attributable to colitis-related toxicity in patients treated with ipilimumab, and pneumonitis when patients receive anti-PD-1/PD-L1 therapy.⁴⁹

Figure 3.

Colitis in a 53-year-old woman with poorly differentiated adenocarcinoma of the lung treated with durvalumab (anti-PD-L1 therapy). MIP (A), CT (B) and fused FDG PET/CT images (C) acquired 3.5 months after treatment initiation revealed marked FDG uptake in the descending and sigmoid colon (arrowheads) with associated wall thickening and fat stranding consistent with colitis. The patient had bloody diarrhea at the time of the scan; sigmoidoscopy performed a week later showed acute colitis. Symptoms resolved following treatment with prednisone. Follow-up MIP (D), CT (E) and fused FDG PET/CT images (F) obtained 5 months later demonstrated marked improvement.

Figure 4.

Pancreatitis in a 57-year-old woman with metastatic anal cancer on nivolumab. Fused FDG PET/CT (A) and CT (B) images acquired 7.5 months after initiation of treatment demonstrate increased FDG uptake in an edematous pancreas (arrows); the patient had diarrhea at the time of the scan. Nivolumab was held for one cycle and pancreatic enzyme supplementation was started. A follow up contrast-enhanced CT (C) performed 9 months later showed resolution with interval atrophy of the pancreas. Chronic right-sided hydronephrosis is also seen.

Figure 5.

Thyroiditis in a 56-year-old woman with metastatic melanoma treated with ipilimumab / nivolumab. CT (A) and fused FDG PET/CT images (B) acquired 2.5 months after treatment initiation revealed marked FDG uptake in the thyroid gland consistent with thyroiditis; the patient had thyrotoxicosis at the time of the scan, which was followed by persistent hypothyroidism requiring levothyroxine replacement. Follow-up CT (C) and fused FDG PET/CT images (D) obtained 1 year later demonstrated resolution of the abnormal uptake in the thyroid gland.

Figure 6.

Pneumonitis in a 61-year-old man with metastatic squamous cell carcinoma of the left tonsil on pembrolizumab. CT (A) and fused FDG PET/CT images (B) acquired 7.5 months after initiation of treatment demonstrate FDG-avid nodular opacities in the lungs in a peribronchovascular distribution consistent with pneumonitis; at the time of the scan the patient had shortness of breath and cough. Pembrolizumab was held and steroids were initiated, and the patient’s symptoms improved. Follow-up CT (C) and fused FDG PET/CT images (D) acquired 4 months later demonstrate resolution of the pneumonitis.

Figure 7.

Sarcoid-like reaction in a 77-year-old woman with metastatic melanoma treated with pembrolizumab. Oblique MIP (A), CT (B) and fused FDG PET/CT images (C) acquired 13 months after treatment initiation revealed marked FDG uptake in mediastinal and bilateral hilar lymph nodes (arrow) which were new from the prior study and consistent with a sarcoid-like reaction. At the time of the scan there was also a new FDG avid right forearm mass (arrowhead); biopsy of the mass revealed granulomatous inflammation and no tumor. The mass persisted for an additional 12 months (including 5 months of pembrolizumab and 7 months of no therapy) and then spontaneously resolved.

irAEs can manifest on imaging in a range of organs and organ-systems, can precede clinical symptoms, and may even mimic metastatic disease,^45,50–52 therefore it is important that radiologists are aware of this entity so that is included in the differential diagnoses for patients on immunotherapy. On FDG PET/CT, irAEs manifest as increased FDG uptake in the involved organs, and subsequent decreased uptake suggests resolution of acute inflammation.^53–55 irAEs may also predict response to immunotherapy,^56,57 although this may be organ/system dependent.⁴⁶

RESPONSE EVALUATION

Response criteria in solid tumors (RECIST) and other metrics are routinely used to assess response to cancer therapy.^58–61 However, the observation of pseudoprogression in a subgroup of patients treated primarily with ipilimumab motivated the development of new criteria for response assessment in the setting of cancer immunotherapy, in order to distinguish true progressive disease from pseudoprogression. In the majority of these new immune-related response criteria such as irRC, iRECIST and iPERCIST,^33,37,62,63 an increase in size of lesions and/or appearance of new sites of disease on the first follow-up (relative to baseline imaging) reflects unconfirmed progressive disease (UPD). If follow-up anatomic imaging and/or FDG PET/CT after ≥4 weeks demonstrates no improvement or even worsening of disease, patients are classified as confirmed progressive disease (CPD). In the modified Lugano criteria for immunotherapy in lymphoma, biopsy or subsequent imaging can be performed.⁶⁴ In addition, some investigators have combined anatomic and molecular imaging criteria to characterize response,⁶⁵ while others have used thresholds of lesion size and number to determine progressive disease.^66,67 Despite this wide variety of new immune-related response criteria, RECIST remains the primary method of response assessment for most clinical trials, including immunotherapy trials, with immune-related response criteria used for exploratory endpoints.

IMMUNE IMAGING WITH FDG PET/CT

FDG is known to be taken up by activated immune cells. In clinical FDG PET/CT scans this is reflected in inflammatory conditions such as infection, rheumatoid arthritis, and sarcoidosis, which demonstrate elevated FDG uptake.^68–70 Additionally, in vitro studies have demonstrated markedly increased uptake in activated T cells compared to unstimulated T cells.⁷¹ In the routine clinical setting FDG activity in immune cells cannot be discriminated from FDG activity in tumor cells. However, if a baseline FDG PET/CT is compared with an early post-treatment FDG PET/CT over a short interval that minimizes changes in the tumor, any increase in FDG uptake should reflect tumor infiltration by activated immune cells. This metabolic “flare” phenomenon has been demonstrated in a preclinical mouse tumor model and reported in a few clinical cases, and is potentially an earlier and more sensitive measure of response to cancer immunotherapy.^72–74 In fact, a recent clinical trial demonstrated that a metabolic flare could be detected in 2/16 (13%) patients with melanoma on pembrolizumab as early as 6–7 days post therapy, with dramatic increases in tumor maximum standardized uptake value (SUV_MAX) that more than doubled and predicted a complete response to therapy; no tumor flare was seen in nonresponders.⁷⁵ Future studies will need to test this approach in a larger cohort of patients, and explore the optimal posttreatment imaging time.

Other approaches to use FDG PET/CT imaging to predict response to immunotherapy have also been explored. For example, two studies have reported that an increased ratio of mean standardized uptake value (SUV_MEAN) of bone marrow to liver (BLR) on baseline FDG PET/CT has been associated with decreased survival after anti-PD-1 immunotherapy in the setting of metastatic melanoma.^76,77 This bone marrow hypermetabolism in patients with cancer is hypothesized to reflect a systemic inflammatory response, which leads to immunosuppression and is associated with cancer progression. Additional support for this hypothesis is provided by a significant positive correlation between FDG uptake in bone marrow and serum inflammatory markers including the white blood cell count and C-reactive protein.⁷⁶

During immunotherapy, activation of the immune system can cause infiltration of lymphoid organs by immune cells. Sarcoid-like reaction, although considered to be an irAE, has been shown to reflect nodal infiltration by immune cells post-immunotherapy, and such nodal infiltration corresponds to associated FDG-avidity (Figure 7).⁷⁸ In a recent study, all patients with FDG-avid sarcoid-like reactions following immunotherapy demonstrated positive response.⁷⁹ Pseudoprogression also appears to indicate infiltration of tumors by immune cells which are FDG avid.^31,80

FDG PET/CT has also been used to visualize the immune response following vaccination. Increased FDG uptake has been seen in ipsilateral axillary lymph nodes following the influenza vaccine for up to 2–4 weeks, with the highest uptake seen within the first week after the vaccine.^81,82 In one case report, transiently increased FDG activity was also seen in the spleen at 2–3 days post vaccination, which resolved 12 days later.⁸³ A similar pattern of increased FDG uptake in the deltoid muscle and ipsilateral axillary lymph nodes has been seen for up to several weeks following COVID-19 vaccination (Figure 8).⁸⁴ These cases underscore the need for an accurate patient history, to ensure that FDG avid reactive axillary lymph nodes are not mistaken for metastatic disease.

Figure 8.

Imaging findings of COVID-19 vaccination. MIP (A), CT (B), and fused FDG PET/CT images (C) acquired 2 days after COVID-19 vaccination in the right arm revealed increased FDG uptake in the right deltoid muscle (arrowhead) and markedly increased uptake in right axillary lymph nodes (arrows) and a supraclavicular lymph node, which were normal in size. These findings were consistent with reactive changes from COVID-19 vaccination in a 70-year-old woman with a history of treated lung cancer and no evidence of recurrent disease for 4+years.

NEW IMMUNE-SPECIFIC PET IMAGING PROBES

Since FDG accumulates in both tumor cells and activated immune cells, FDG uptake can be nonspecific. In order to overcome this limitation, PET probes with higher specificity for immune-related targets are needed, which can be grouped into two different categories: 1) imaging probes that target general immune-related markers, or 2) probes designed to target markers that are more uniquely expressed in the setting of immune activation.

Given that tumor-infiltrating CD8+ T cells are predictive of response to immunotherapy,⁸⁰ whole-body CD8 PET/CT imaging is of interest, as it has the potential to allow non-invasive assessment of temporal changes in CD8+ T cell concentration in tumors, both before and after immunotherapy. Although the majority of immune-specific probes are in preclinical development, a few are in early phase clinical trials. For example, a ⁸⁹Zr-labeled anti-CD8 minibody (⁸⁹Zr-Df-IAB22M2C) is currently in a phase 2 trial [NCT03802123] as a PET probe for imaging CD8+ T cells in patients with metastatic solid tumors, with the goal of correlating CD8 signal on PET/CT imaging to CD8+ T cell infiltration from biopsy samples, and response to cancer immunotherapy. Results from the phase 1 trial of ⁸⁹Zr-Df-IAB22M2C demonstrated tracer uptake in tumors (Figure 9) and CD8 rich tissues (e.g. spleen, bone marrow, lymph nodes) with maximum uptake at 24–48 hours post injection and low background activity in non-T cell rich tissues (e.g. muscle, heart).⁸⁵ In preclinical models of cancer immunotherapy, CD8- and CD3-specific imaging agents have both demonstrated greater trafficking and/or a more central distribution of tumor infiltrating T cells in responders versus nonresponders, which supports the potential utility of these agents as an early measure of response to immunotherapy.^86–89 In addition, imaging agents that target immune cells have the potential to serve as noninvasive predictive biomarkers by differentiating patients with “hot” versus “cold” tumors, and their likelihood of responding to immunotherapy.⁹⁰ CD8 PET/CT imaging could also be helpful in distinguishing pseudoprogression from treatment failure, and may complement FDG PET/CT as a problem-solving tool when immune-related changes need to be isolated from tumor growth.

Figure 9.

67-year-old man with metastatic hepatocellular carcinoma (HCC) treated with nivolumab. A CD8 PET/CT scan acquired 14 days after starting immunotherapy demonstrated increased tracer activity in the primary tumor (arrow; SUV_MAX = 22.9) on the coronal PET (A) and fused PET/CT images (B), suggestive of tumor infiltration by CD8+ T cells and a productive anti-tumor immune response; physiologic tracer activity is seen in the spleen, liver, bone marrow, and kidneys. Follow-up imaging demonstrated a partial response to therapy, which has lasted 3+ years, with an associated drop in alpha-fetoprotein from 33.2 ng/mL at baseline to 1.4 ng/mL at 3 years.

Other immune-specific imaging agents that are in clinical trials include probes that target PD-L1 (⁸⁹Zr-atezolizumab and ¹⁸F-BMS-986192) and PD-1 (⁸⁹Zr-nivolumab).^91–93 Given that PD-L1 expression levels have been shown to be a positive (albeit imperfect) predictive biomarker for patients undergoing immune checkpoint blockade therapy, these agents have the potential to serve as a noninvasive measure of PD-L1 expression, which would have particular utility in patients with lung cancer where assessment of PD-L1 expression is required prior to first-line treatment with anti-PD-1 therapy.⁹⁴ A wide variety of other imaging probes that target immune-related markers such as CD4, CTLA-4, CD11b, CD47, VLA-4, and CXCR4 (and other chemokine receptors and ligands) are in preclinical development and may also prove to have utility in the setting of cancer immunotherapy in the future.^95,96

Imaging agents that target immune activation are also being developed, which will be helpful in distinguishing activated immune cells present in the tumor microenvironment from quiescent immune cells. These include agents that are specific for key enzymes involved in T lymphocyte and other immune cell activation and proliferation (¹⁸F-FAC, ¹⁸F-CFA, and ¹⁸F-AraG), which are in early phase clinical trials, and have been used preclinically for detecting the location of activated T cells, monitoring graft-versus-host-disease, and evaluating auto-immune disorders.^95,96 However, clinical data on the utility of these imaging agents in the setting of cancer immunotherapy have not yet been published. Other probes that are specific for activated immune cells include agents that target granzyme B (⁶⁸Ga-NOTA-GZP), IL-2 (¹⁸F-FB-IL2), OX40 (⁶⁴Cu-DOTA-AbOX40), and ICOS (⁸⁹Zr-DFO-ICOS mAb).⁹⁶ PET imaging probes that target granzyme B, OX40, and ICOS have all been tested in preclinical models of cancer immunotherapy, and demonstrated increased tumor uptake in responders versus nonresponders, suggesting that they could serve as an early measure of response.^97–99 In addition, granzyme B PET/CT imaging using ⁶⁸Ga-NOTA-hGZP is currently in a phase I clinical trial (NCT04169321).

The aforementioned studies indicate that the immune-specific PET imaging toolbox is likely to expand, and will provide information that supplements FDG PET/CT and anatomic imaging.¹⁰⁰ These new imaging tools have the potential to have a major impact on patient management in the setting of cancer immunotherapy, and will likely have applications outside of oncology for other conditions in which the immune system plays a role, such as autoimmune and inflammatory diseases, transplant rejection, and infection.

Key Points:

• Immunotherapy causes infiltration of tumors by immune cells and in rare cases is associated with unique response patterns such as pseudoprogression.

• FDG PET/CT is frequently used to assess response to immunotherapy, and while it cannot distinguish immune-related activity from tumor growth it has the potential to provide insight into the immune response.

• New probes for PET imaging of the immune system are likely to be helpful in predicting response to cancer immunotherapy and separating immune-related changes from progressive disease.