Imaging inflammation and its resolution in health and disease: current status, clinical needs, challenges, and opportunities

Abstract

Inflammation is a normal process in our body; acute inflammation acts to suppress infections and support wound healing. Chronic inflammation likely leads to a wide range of diseases, including cancer. Tools to locate and monitor inflammation are critical for developing effective interventions to arrest inflammation and promote its resolution. To identify current clinical needs, challenges, and opportunities in advancing imaging-based evaluations of inflammatory status in patients, the U.S. National Institutes of Health convened a workshop on imaging inflammation and its resolution in health and disease. Clinical speakers described their needs for image-based capabilities that could help determine the extent of inflammatory conditions in patients to guide treatment planning and undertake necessary interventions. The imaging speakers showcased the state-of-the-art in vivo imaging techniques for detecting inflammation in different disease areas. Many imaging capabilities developed for 1 organ or disease can be adapted for other diseases and organs, whereas some have promise for clinical utility within the next 5–10 yr. Several speakers demonstrated that multimodal imaging measurements integrated with serum-based measures could improve in robustness for clinical utility. All speakers agreed that multiple inflammatory measures should be acquired longitudinally to comprehend the dynamics of unresolved inflammation that leads to disease development. They also agreed that the best strategies for accelerating clinical translation of imaging inflammation capabilities are through integration between new imaging techniques and biofluid-based biomarkers of inflammation as well as already established imaging measurements.—Liu, C. H., Abrams, N. D., Carrick, D. M., Chander, P., Dwyer, J., Hamlet, M. R. J., Kindzelski, A. L., PrabhuDas, M., Tsai, S.-Y. A., Vedamony, M. M., Wang, C., Tandon, P. Imaging inflammation and its resolution in health and disease: current status, clinical needs, challenges, and opportunities.

Chronic unresolved inflammation is implicated in cancer development and other diseases. Unlike acute inflammation, chronic inflammation often does not exhibit obvious signs and symptoms and is often neglected until the disease is clinically apparent. Monitoring the inflammatory status of these processes can provide an opportunity for clinical interventions to prevent disease development and progression. Current clinical evaluation for inflammatory status in an individual patient relies on immunologic tests of biologic fluids and pathologic analyses of biopsied tissue samples. Although biofluids can be collected repeatedly from a patient, the assays for inflammatory signatures do not yield information about the origin or location of the inflammation process. Tissue biopsies are invasive procedures and provide information on the tissue only at the time and site of tissue removal. Thanks to recent advances in imaging, there are now techniques that can assess tissue immunologic status in vivo. Many of these imaging approaches can potentially be adapted to become clinical tools for assessing the inflammatory status of an individual patient. The goal of the “Trans–National Institutes of Health (NIH) Workshop on Imaging Inflammation and Its Resolution in Health and Disease” held on June 10–11, 2019 in Rockville, MD, USA was to align research and development efforts by the imaging community toward the application and development of in vivo imaging-based tools and techniques in assessing and monitoring inflammation for clinical decision making. This workshop brought together practicing clinicians, clinical researchers, and members of the imaging community to discuss the challenges, needs, and opportunities in advancing imaging evaluation of inflammation for early intervention, treatment planning, and improved patient outcomes.

THE ENDOGENOUS MEDIATORS OF INFLAMMATION AND ITS RESOLUTION

Dr. Charles Serhan (Brigham and Women’s Hospital, Boston, MA, USA) introduced the concept of uncontrolled inflammation as the integral component of many diseases and that the anti-inflammatory process does not equate to the proresolution process. Anti-inflammatory agents, such as nonsteroidal anti-inflammatory drugs and cyclooxygenase-2 inhibitors, are resolution-toxic, as they lower the amplitude of inflammation but prolong the time needed for resolution. Acute inflammation occurring in response to infection and injury is often a protective mechanism, with resolved inflammation as the ideal outcome. On the other hand, acute inflammation can alternatively progress to abscess formation, scarring associated with wound healing, organ fibrosis (diabetes, cardiovascular diseases [CVDs], and asthma), or advance into chronic inflammation. If genetic damage is the “match that lights the fire” of cancer, inflammation can provide the “fuel that feeds the flames.” Dr. Serhan stressed that the resolution of acute inflammation is not a passive process involving the dissipation of inflammatory mediators through time. Instead, inflammation resolution is an active process involving specialized proresolving mediators (SPMs) derived from ω-3 polyunsaturated fatty acids. SPMs are endogenously bioactive metabolome; each stimulates self-limited innate immune responses, kills and clears microbes, and protects organs. Compared with the majority of anti-inflammatory agents, SPMs show potent proresolving actions in several animal disease models (1). Proresolving agents, such as aspirin, enhance cancer therapy by countering debris-stimulated tumor growth (2). Electrical stimulation of the vagus nerve has also been shown to increase endogenous production of SPMs and reduce proinflammatory molecules (3). Dr. Serhan concluded that SPMs and their receptors provide new opportunities for controlling excessive inflammation and infection via resolution physiology and pharmacology. Current tools for identifying and profiling SPMs include liquid chromatography-tandem mass spectrometry–based lipid mediator metabololipidomics in serum or biologic samples. In vivo image-based capabilities to interrogate molecular dynamics in the resolution process could further our understanding of disease-promoting unresolved inflammation-induced disease progression and accelerate drug development to combat inflammation and promote resolution.

CLINICAL CHALLENGES AND NEEDS IN EVALUATING THE INFLAMMATORY STATUS IN PATIENTS

In this session on the clinicians’ needs to monitor inflammatory status in patients for decision making, the clinicians who specialized in different diseases described the critical role of inflammation and its resolution in affecting clinical presentations, treatment responses, treatment-induced adverse effects, and how these processes act as confounders in their clinical decision making. They also discussed their needs for imaging capabilities in evaluating these processes to establish the causal relationship between inflammation and disease progression and for patient stratification and intervention based on the inflammatory status.

Clinical management for patients treated with immune checkpoint inhibition immunotherapy

Cancer immunotherapy enhances a patient’s immune system to fight cancer and has recently shown remarkable benefit in treating many cancers. The most popular immunotherapeutic strategy, immune checkpoint blockade (ICB), increases the host’s antitumor immunity by blocking the negative regulators of T-cell function and enhancing T cells’ ability to attack cancer cells without relying on chemotherapy. The U.S. Food and Drug Administration (FDA) has recently approved a combination of checkpoint inhibitor drugs (e.g., ipilimumab and nivolumab), with which studies have shown a dramatic (53%) increase in survivability for patients with metastatic melanoma. However, autoimmune side effects resulting from disinhibiting the immune system can limit the use of these therapies (4, 5). Dr. Michael Postow (Memorial Sloan Kettering Cancer Center, New York, NY, USA) discussed the clinical challenges in managing patients receiving ICB drugs for advanced cancers. As a clinical oncologist, Dr. Postow indicated that ICB only works in some patients of which a fraction of them experienced significant toxicity. Presently, specific biomarkers or parameters that can be used to monitor treatment response and immune-related adverse events are unknown. Dr. Postow also described the dilemma for clinicians, who need to decide whether to discontinue treatment because of an unconventional clinical response. For example, cancer pseudoprogression, whereby some patients exhibit an initial increase in tumor lesion sizes or new lesions, appears ultimately to have a delayed but positive response to treatment. Dr. Postow suggested that imaging techniques to visualize immune responses and elucidate the mechanism of action during ICB could help explain why checkpoint inhibitors only work in some patients, why only some patients develop significant toxicities, whereas others do not, and help differentiate true tumor progression from pseudoprogression. Following the discussion about cancer pseudoprogression, Dr. Victoria Chiou (FDA, Silver Spring, MD, USA) described cancer imaging and clinical decisions in general, with a focus on regulatory and drug development considerations in cases of pseudoprogression and immuno-oncologic therapy (6–8).

Surveillance in patients with nonalcoholic fatty liver disease

Nonalcoholic fatty liver disease (NAFLD) affects about one-third of the adult population in the United States and is projected to become the leading cause of end-stage liver disease and liver transplantation in the next 10 yr (9). NAFLD is considered a hepatic manifestation of metabolic syndrome and is closely associated with obesity, insulin resistance, and type 2 diabetes. NAFLD can be broadly divided into 2 subtypes: NAFL, the nonprogressive phenotype with minimal risk of progression to cirrhosis, and nonalcoholic steatohepatitis (NASH), the progressive phenotype that has a substantial risk of progression to cirrhosis. NASH is a clinicopathologic entity characterized by steatosis, ballooning, and lobular inflammation with or without perisinusoidal fibrosis on liver histology among individuals who consume little or no alcohol and do not have a secondary cause for hepatic steatosis. It increases the risk of advanced fibrosis, the development of hepatocellular carcinoma (HCC) and other complications, such as cirrhosis. Dr. Rohit Loomba (University of California–San Diego, La Jolla, CA, USA) explained that the current gold standard for the diagnosis for NASH is a liver biopsy, but it is invasive, risky, and prone to sampling errors. At least 1 billion people worldwide and 100 million Americans are affected by NAFLD. Alternate methods for diagnosis and improved risk stratification among the high-risk population are urgently needed (10). Dr. Loomba described 2 MRI-based techniques that are used in his clinics. He discussed studies showing that liver fat content can be quantitatively evaluated using MRI–proton density fat fraction. The other imaging technique he discussed was MRI elastography, which can be used to determine the stage of liver fibrosis in patients with NAFLD (11). Although these MRI techniques have demonstrated efficiency in patient surveillance and in evaluating treatment response in NASH, Dr. Loomba and Dr. Yaron Rotman (National Institute of Diabetes and Digestive and Kidney Diseases, NIH, Bethesda, MD, USA) indicated that, so far, there are no tools available for routine clinical use that are sensitive enough to differentiate NAFL from NASH and to identify individual stages of fibrosis. Studies are underway to assess whether clinical prediction rules [e.g., the Fibrosis-4 (FIB-4) index that includes age, aspartate transaminase, alanine transaminase, and platelets], ultrasound (US)-based elastography methods (e.g., vibration-controlled transient elastography), or serum-fibrosis tests (e.g., enhanced liver fibrosis panel) may be utilized along with advanced MRI methods to cost-effectively risk stratify the at-risk population. Because not all patients with NAFLD will progress to NASH, this presents a challenge for primary care providers in determining when to refer a patient to a hepatologist. A hepatologist wants to know which patient to treat, whom to refer to a clinical trial, and how to best monitor treatment response. As such, it is crucial in the surveillance of patients with NAFLD that physicians be able to identify patients with a high probability of progressing to NASH as early as possible because those at highest risk of HCC have NAFLD with advanced fibrosis and need treatment before it progresses to later stages. Imaging tools, in conjunction with liquid biomarkers, could be very helpful for clinicians to stratify patients and provide appropriate treatment.

Understanding and predicting chronic pain progression

Neuroinflammation involving the infiltration of immune cells in the central nervous system (CNS), activation of glial cells, and production of inflammatory mediators in the peripheral and CNS all play an important role in chronic pain (12). Dr. Ru-Rong Ji (Duke University, Durham, NC, USA) explained that the most popular (and nonaddictive) pain killers address 2 systems: local anesthetic to modulate the nervous system via ion channel blocking and nonsteroidal anti-inflammatory drugs to modulate the immune system. The ideal way to treat pain is to target both systems together. He described how SPMs could modulate these systems in alleviating neuroinflammation-directed chronic pain. Fish oil [containing the ω-3 fatty acids eicosapentaenoic acid (EPA) and docosahexaenoic acid (DHA)] has been touted to have several health benefits, including pain relief. However, clinical trials have provided mixed results regarding pain mediation effects. DHA and EPA may have weak but direct anti-inflammatory and analgesic effects by binding to G-protein–coupled receptor 40/long-chain free fatty acid receptor 1. Dr. Ji also noted that metabolic products of DHA and EPA, such as resolvins and protectins (neuroprotectin D1), have indirect but potent effects on anti-inflammation, proresolution, and analgesia (13). G-protein–coupled receptor 37 is a receptor for neuroprotectin D1 in macrophages, which regulates the resolution of inflammatory pain by inducing phagocytosis of macrophages (14). In the CNS, such as in the spinal cord, microglia activation can drive the pathogenesis of pain by producing proinflammatory cytokines and modulating synaptic transmission leading to increase in neuroexcitation associated with pathologic pain. On the other hand, microglia also play a role in the resolution of pain by releasing SPMs and endocannabinoids, among other factors. Failure to tip the effect of microglial activation toward the resolution phase during acute pain leads to the transition to chronic pain. Dr. Ji discussed a few imaging-based studies for detecting inflammatory pain. There is clinical research suggesting possible increased activation of microglia in patients with chronic low back pain as determined by increased radiolabeled PBR28 (a marker for glial activation) as seen in positron emission tomography (PET) scans (15). Other imaging biomarkers might be useful in identifying neuroinflammation and pain related to it.

Detecting injurious progression in diseases of the veins and arteries

Deep vein thrombosis (DVT) often progresses to symptomatic postthrombotic disease (PTD) and venous ulcers. The majority of studies to date have addressed postthrombotic syndrome, the diagnosis of which is based on a combination of nonspecific signs and symptoms that are often misclassified as PTD. As such, differentiating or accurately diagnosing PTD vs. postthrombotic syndrome is problematic. Dr. Fedor Lurie (Jobst Vascular Institute, Toledo, OH, USA) described that the clinical approach to PTD prevention includes the individualization of DVT treatment plan based on the patient’s risk factors, inflammatory biomarkers, and imaging. He explained that because inflammation is involved in the resolution of acute venous thrombosis and the transition to chronic venous disease, the location and timing of the inflammatory processes could dictate the disease progression in either direction. Clinical studies that have used serum inflammatory biomarkers [e.g., C-reactive protein (CRP), IL- 6, -8, -10] and animal models suggest that an insufficient early inflammatory response negatively affects thrombus resolution, whereas the postthrombotic remodeling of the venous wall is associated with increased inflammation (16). For these reasons, a combination of imaging-based hemodynamic assessment coupled with inflammatory markers and risk factors would help identify risk levels for both insufficient thrombus resolution and vein wall fibrosis. Dr. Lurie further described the current standard of practice in imaging tools, including duplex US, computed tomographic venography, and magnetic resonance venography, used to locate the site of inflammation based on structural and functional changes of the veins. Other techniques, including infrared thermal imaging, Doppler vessel pressurization ultrasonography, and contrast-enhanced US (CEUS) have been used for detecting lower-extremity DVT (17). Although [¹⁸F]-fluorodeoxyglucose (FDG)-PET has been used to evaluate the inflammatory process associated with venous thromboembolic disorders (18) and the acuity of DVT (19), the FDG signals only reveal the site of increased glucose metabolism, which is not necessarily indicative of inflammation. Dr. Lurie concluded that DVT treatment should be based on the differentiation of inflammatory response to thrombosis (magnitude, time, and location); image capabilities that can separate the inflammatory conditions in the vein wall vs. intraluminal thrombus [e.g., matrix metalloproteinases (MMP) and macrophage activity] are needed.

There is a significantly higher incidence of coronary artery disease (CAD) in persons living with human immunodeficiency virus (HIV) (PLWH) receiving antiretroviral therapy (ART). Multiple factors, including chronic inflammation in addition to ART medications, may have a direct effect on lipid metabolism, vascular, and endothelial dysfunction (20). CAD results in a higher rate of acute myocardial infarction in PLWH and the gap between the general population and patients infected with HIV increases with age (21). Dr. Ahmed M Gharib (National Institute of Diabetes and Digestive and Kidney Diseases, NIH, Bethesda, MD, USA) pointed out that a younger patient population should be targeted to better understand the mechanism of CAD. He described MRI studies that revealed thickened coronary vessel walls in younger PLWH (averaged 22 yr old) who were infected at an early stage in life (22); the coronary vessel thickness increases were directly related to the number of years ART is given, as were the elevated inflammatory signatures in serum. Image-based monitoring could improve long-term treatment strategies and prevent acute myocardial infarction in PLWH.

Understanding the dietary effects on inflammation in health and disease

Many dietary constituents affect inflammatory markers, and immune response and individual diets vary in their potential to modulate inflammation. Dr. James R Hebert (University of South Carolina, Columbia, SC, USA) discussed the Dietary Inflammatory Index (DII), which describes the inflammatory potential of diet in health conditions involving inflammation. The DII is based on peer-reviewed literature that identifies 45 dietary constituents known to increase 4 proinflammatory markers (i.e., IL-1β, IL-6, TNF-α, and CRP) or decrease 2 anti-inflammatory markers (i.e., IL-4 and IL-10) (23). The DII provides a score that can be used for assessing or monitoring diet and its effects on inflammation and health outcomes. For example, in a meta-analysis, higher DII was associated with higher colorectal cancer risk in both case-control and prospective studies (24). The DII has also been shown to affect other chronic diseases ranging from CVDs (25) to depression (26). In construct validations, changes in DII score has also been shown to mirror changes in various inflammatory biomarkers, including cytokines (27). Most epidemiologic and clinical studies are limited to measuring diet just once or a few times as a way to describe long-term, habitual intake. Imaging is rarely used beyond simple description and is virtually never focused directly on inflammation. The use of the DII, in combination with careful imaging, for example, in longitudinal studies that monitor inflammatory processes, has the potential to pinpoint the effect of diet as a cause of inflammation-related pathologic conditions.

STATE-OF-THE-ART IN VIVO IMAGING APPROACHES IN DETECTING INFLAMMATION

The presentations in this session focused on in vivo technologies for detecting and measuring inflammation, including the latest preclinical in vivo imaging techniques for detecting a variety of signatures ranging from molecules, cellular trafficking, metabolic changes, structural damage to functional alterations, which are associated with inflammation, and its resolution within the diseased organs.

Optical imaging of immune cell trafficking

Recent advances in intravital microscopy (IVM) have provided insight into dynamic biologic events at the cellular level in both healthy and diseased tissue. The ability to follow the temporal changes in single cells within an organ in live animals has contributed to breakthroughs in many fields, including microbiology, immunology, and cancer. Dr. Stefan Uderhardt (National Institute of Allergy and Infectious Diseases, NIH, Bethesda, MD, USA) explained how IVM technology has helped uncover the dual role of the innate immune cells as an effector (inflammation) or a protector (anti-inflammation) of homeostasis in response to pathogenic infection and tissue injury. IVM studies show that neutrophils undergo a swarming-like migration pattern directly at the site of infection as part of the host defense mechanism. The extent of neutrophil swarming seems to be related to the size of the lesion, with heightened swarming and collateral damage being more prevalent in larger lesions. In addition to neutrophil activation, the resident tissue macrophages play a key role in preventing neutrophil-mediated inflammatory damage, a protective process later dubbed as “cloaking” the microlesion (28). This cloaking mechanism is important for preventing tissue damage and minor local cell injury that occurs regularly to maintain tissue homeostasis. Dr. Uderhardt then introduced 2 emerging microscopic-based technologies: multiplexed immunohistochemistry and histocytometry, which are applicable to monitoring immune system function and maintenance. Immunohistochemistry multiplexing integrates multicolor multispectral imaging with whole-slide scanning to support quantification and analysis of tissue sections, allowing improved characterization of histology. Histocytometry is a multispectral quantitative imaging technique that applies image-derived statistics in flow cytometry analysis software to aid visualization and quantification of complex cell populations directly in tissue sections (29). When coupled with a clearing-enhanced 3-dimension method that improves organ tissue transparency, histocytometry permits simultaneous quantitative analysis of 3-dimensional organ structure and detailed cellular composition as well as identifies subtle differences in cell phenotype and cell-cell interactions.

Optical imaging for detecting metabolic perturbations in tissue

Cellular metabolism is a key regulator of cellular functions, and understanding of cellular metabolic pathways is critical for development and assessment of novel therapies and diagnostics. Dr. Irene Georgakoudi (Tufts University, Medford, MA, USA) unveiled label-free 2-photon excited fluorescence (TPEF) microscopy as a technology to assess metabolic tissue function in live specimens. She presented a quantitative TPEF method as capable of detecting both functional and structural metabolic changes in individual cells. The unique combinations of changes in 3 optical metabolic readouts (i.e., the optical redox ratio, the NADH fluorescence lifetime, and mitochondrial clustering) can elucidate the nature of the metabolic pathway perturbations, leading to the detected optical changes, including glycolysis, glutaminolysis, mitochondrial uncoupling, and fatty acid oxidation and synthesis (30). Such quantitative optical metabolic assessments can identify significant differences in mitochondrial organization patterns in normal and cancerous skin tissues (31) as well as redox and mitochondrial networking changes present in cervical precancers and distinct adipose depots (32).

Macrophages play an important role in the inflammatory process. TPEF detects enhanced glutaminolysis during pro- or anti-inflammatory states that might be dominating optical metabolic readouts in macrophage polarization (33). Dr. Georgakoudi concluded that the typical penetration depth of TPEF could reach hundreds of microns, making it highly relevant for noninvasive approaches to studying epithelial tissues and could be adapted in endoscopic applications. Development of cheaper and more compact fiber lasers, more compact 2-photon probes for efficient light delivery and collection, and machine learning–assisted image analysis will enable functional, label-free high-resolution imaging for identifying new targets to modulate inflammatory responses.

Photoacoustic imaging for detecting joint inflammation

Photoacoustic imaging (PAI) is a nonionizing, noninvasive, low-cost technology that combines the contrast advantage inherent in optical measurements, and the advantage in the depth penetration of US imaging for studying biologic tissue (34). Because the optical absorption contrast of PAI in the visible to near-infrared spectrum is intrinsically sensitive to the contents of oxygenated and deoxygenated hemoglobin, PAI can detect the functional changes in inflamed joints of patients with subclinically active arthritis. Using a US-PA dual-modality imaging system (35), Dr. Xueding Wang (University of Michigan, Ann Arbor, MI, USA) showed that PAI has the capability of generating functional information on physiologic biomarkers such as hyperemia (e.g., enhanced blood volume in response to elevated metabolic demand) and hypoxia (e.g., reduced oxygen saturation due to inadequate oxygen delivery) of the inflamed tissue in patients with arthritis (36). Dr. Wang indicated that the most obvious limitation for clinical applications of PAI remains its dependency on the limited penetration of light waves and propagation of US waves in tissue. As such, not all organs are amenable for PAI. More recently, working with General Electric Global Research (Niskayuna, NY, USA) and Cyberdyne (Tsukuba, Japan), Dr. Wang’s team has been developing a low-profile PAI system for rheumatology clinics by using light-emitting diodes as the light source. Light-emitting diodes enable a smaller footprint for PAI systems when compared with laser sources, thereby improving system portability and accessibility for clinical applications (37).

MRI sensors for biomarkers of inflammation

Molecular sensors that become imageable in response to the inflammation-associated enzymatic activities can locate the sites of inflammation with minimal background signal. Dr. Alexei Bogdanov Jr. (University of Massachusetts Medical School, Worcester, MA, USA) described enzyme-sensing MR contrast probes that can be polymerized because of the presence of myeloperoxidase associated with active inflammation and result in signal amplification; this is an MR technique called MRamp (38). Using an MR imaging probe (gadolinium-5HT-DOTAGA) and micro-MRI on human surgical histology sections, Dr. Bogdanov showed that increased myeloperoxidase content correlates with an increased risk for rupture of intracranial aneurysms (39). This imaging technique is effective for preclinical evaluation of anti-inflammatory therapy after reperfusion therapy in an animal stroke model (40).

Hyperpolarized MRI for detecting oxidative stress associated with inflammation in vivo

Traditional MRI and MR spectroscopy (MRS) detect proton (¹H) signals because of their abundance in the living systems. However, the MR signals of other endogenous molecules (i.e., carbon and nitrogen) are almost undetectable because most are at nanomolar concentration ranges. Dynamic nuclear polarization (DNP) techniques can enhance these signals over 100,000 times, making small-molecule metabolites detectable using hyperpolarized MRI (HP-MRI). HP-MRI can be used to image metabolism using nonradioactive probes (41) as an alternative to PET. Dr. Kayvan R. Keshari (Memorial Sloan Kettering Cancer Center, New York, NY, USA) described the use of HP [¹³C]-pyruvate to investigate changes in metabolic fluxes based on the conversion of HP pyruvate to other HP metabolites for a wide range of disease states in humans, which is currently in multiple human trials (42, 43). Dr. Keshari showed that an endogenous redox sensor, HP [¹³C]-DHA (the oxidized form of vitamin C), can be used to interrogate the redox pair in vivo and to evaluate the level of oxidative stress associated with pathologic conditions where high HP vitamin C was detected, such as in prostate cancer (44), nephropathy (45), NASH (46), and common disorders of the brain (47). Dr. Keshari concluded that characterizing oxidative stress in vivo using HP-DHA provides a means for studying inflammation and can be readily applicable to multiple organs.

Comparative PET imaging of cancer and inflammation in companion animals

Comparative oncology research conducted in companion animals provides opportunities to include naturally occurring cancer models in the study of cancer biology and therapy. Companion animal cancer models are large outbred animals, typically dogs, with strong genetic similarities to humans. Unlike most small animal cancer models, companion animals are immune-competent and develop naturally occurring cancers with relevant tumor histology/genetics (48). They also have relevant responses to chemotherapy as to humans but relatively compressed tumor progression. Thus, study results can typically be obtained within 2 yr (49). Dr. Amy LeBlanc (National Cancer Institute, NIH, Bethesda, MD, USA) first described molecular imaging capabilities that have been applied to companion animals, such as dogs, cats, birds, and fish. Several imaging modalities have been used to study biodistribution of drugs and imaging contrast agents as well as cell proliferation, glucose metabolism, and hypoxia. Imaging modalities [e.g., digital computed tomography (CT), MRI, US] are accessible to veterinary academic institutions and large private veterinary referral hospitals. Hence, magnetic imaging can be utilized to monitor response to clinically relevant therapies and bridge xenograft models to clinical translation. There are, however, logistic challenges when studying companion animals. For example, animals require cage confinement and general anesthesia to facilitate positioning, and disposal of radioactive waste during and after scan also need to be considered. Dr. LeBlanc explained that [¹⁸F]-FDG-PET alone or [¹⁸F]-FDG-PET/CT are emerging diagnostic imaging modalities in detecting the source underlying the fever of unknown origin (50). However, she cautioned potential pitfalls with [¹⁸F]-FDG-PET imaging interpretation being that [¹⁸F]-FDG cannot discriminate heavily inflamed regions vs. neoplastic regions. Also, normalized standardized uptake value and physiologic uptake of tracer need to be taken into consideration when interpreting PET results.

BENEFITS OF INTEGRATING IMAGING AND FLUID-BASED BIOMARKERS OF INFLAMMATION

The speakers in this session focused on the benefits of integrating imaging and biofluid-based inflammatory, and resolution biomarkers emphasized the role of systemic inflammation in affecting localized disease progression. Their presentations showcased the power of establishing qualifiable imaging biomarkers of inflammation and resolution in multiple human diseases, either through combining imaging of the distributions of biomolecules in different pathways and structural changes during the inflammatory process or correlating imaging signals with new or known serum-based inflammatory biomarkers.

The Foundation for the NIH (FNIH) was founded by Congress to support the mission of the NIH by bringing academic, industry, and government entities together to advance biomedical discoveries and improve the quality of people’s lives. The FNIH Biomarkers Consortium undertakes projects to create and lead cross-sector efforts that validate and qualify biomarkers and other drug development tools to accelerate better decision making for the development of new therapeutics and health technologies. Dr. Stephanie Cush (FNIH, Bethesda, MD, USA) described one of the projects in the Biomarker Consortium Inflammation and Immunity Portfolio: Treatments Against RA and Effect on FDG PET-CT (TARGET Biomarkers Study) (https://fnih.org/what-we-do/biomarkers-consortium/programs/treatments-against-rheumatoid-arthritis). The study design includes a comparison of the efficacy of 2 rheumatoid arthritis (RA) treatment regimens in reducing CVD and joint damage, based on [¹⁸F]-FDG-PET/CT imaging, to evaluate vascular and joint inflammation. In parallel, protein serum biomarkers are assessed to analyze key pathways associated with CVD and RA disease to improve cardiovascular risk prediction in patients with RA. The combination of imaging and serum biomarkers of inflammation will be a powerful tool to dissect the molecular mechanisms of CVD in patients with RA and how RA treatments work in patients with active RA disease. Dr. Cush described another FNIH study called PROGRESS OA: Clinical Evaluation and Qualification of Osteoarthritis (OA) Biomarkers (https://fnih.org/what-we-do/biomarkers-consortium/programs/progress-oa). The PROGRESS OA project utilizes data from previously conducted clinical trials to confirm previously validated imaging and biochemical biomarkers that can be used as prognostic markers of knee OA disease progression, and whether those biomarker measurements have the greater prognostic ability for identifying disease progression than the existing standard of radiographic joint space narrowing. The results from this study will be used in a regulatory qualification submission package to the U.S. FDA and European Medicines Agency (EMA) to qualify the biomarkers for use in osteoarthritis drug development.

PET imaging of inflammatory processes in the brain

Neuroinflammation has been implicated in the pathogenesis of a variety of neuropsychiatric disorders. Currently, PET using a radiolabeled ligand for the 18 kDA translocator protein (TSPO) to locate sites of neuroinflammation or glial cell activation has been widely used in neuropsychiatric research. Dr. Martin Pomper (Johns Hopkins Medical School, Baltimore, MD, USA) explained that relying on TSPO-PET as an indicator of glial activation can be limiting under many pathologic conditions because TSPO is cell type–specific and its affinity depends on the disease state and genotype of the patient (51). Additionally, TSPO imaging alone, without other markers of inflammation, may be insufficient for capturing low-grade inflammatory processes, such as seen in schizophrenia (52). By analyzing information from PET using a second-generation tracer for TSPO ([¹¹C]-DPA-713), MRI of white matter and regional brain volume changes, along with psychologic performance, researchers were able to show that persistent microglial activation due to sports-related traumatic injury in National Football League players could increase their susceptibility to cognitive deficits later in life (53).

Dr. Pomper discussed additional targets of neuroinflammation besides TSPO that could be imaged using PET. For example, the α7 nicotinic acetylcholine receptor (α7-nAChR) is posited to be involved in neuroinflammation, and imaging of this receptor has been used to monitor neuroinflammation after cerebral ischemia (54) in preclinical models. Imaging α7-nAChR also showed a link between diminished availability of α7-nAChR and recent onset of psychosis in patients with schizophrenia (55). A second example is the PET imaging of microglia by targeting macrophage colony-stimulating factor 1 receptor (CSF1R). High and specific brain uptake of [¹¹C]-CPPC was seen in a murine and nonhuman primate LPS model of neuroinflammation (56). Another example is with soluble epoxide hydrolase (sEH) enzyme as another target of neuroinflammation that could be imaged with PET. Regulation and expression of sEH enzyme are altered in many brain disorders and are shown to play a key role in Parkinson’s disease (57). PET imaging of sEH using [¹⁸F]-FNDP to assess sEH regulation could thus be a promising area of research (58).

Combinations of serum biomarkers and CEUS (or MRI) for the early detection of HCC during surveillance

Dr. Laura Beretta (University of Texas M. D. Anderson Cancer Center, Houston, TX, USA) discussed how the current surveillance modality frequently used on patients with cirrhosis is a US scan and serum α-fetoprotein (AFP) every 6 mo; both of these tests have low sensitivity and specificity. Based on a meta-analysis of images (MRI or CT) taken from retrospective studies of high-risk patients with liver lesions ranging from benign (LR-1) to HCC (LR-5), it was recommended that more active management than the current management strategy is needed along with prospective studies (59). Using longitudinal blood sample collections and imaging data in a large cohort prospective study, Dr. Beretta’s group combined a list of potential serum biomarkers with MRI/CT or CEUS for detecting small lesion foci to identify early serum biomarkers that become positive as the lesion turns cancerous. Among other promising biomarkers, her group discovered osteopontin (OPN), which has a serum level marker that is more sensitive than AFP in differentiating patients with cirrhosis from those with HCC. Her group demonstrated that OPN levels could be seen to start increasing in samples collected 1 yr before HCC diagnosis, and a combination of serum AFP and OPN enhances the sensitivity and specificity in detecting HCC (60). Dr. Beretta concluded that such a serum biomarker could spare patients from unnecessary imaging tests, identify at-risk patients, and trigger the decision to perform MRI or CEUS instead of the conventional US in patients under surveillance at an early stage, allowing for curative treatment.

Understanding atherosclerosis immunology using serum biomarkers and PET imaging

In contemporary lipid-lowering trials, there are roughly twice as many individuals with elevated residual inflammatory risk (based on serum CRP level) compared with high residual cholesterol risk (61); among patients undergoing percutaneous coronary intervention with low residual cholesterol risk, persistent high residual inflammatory risk is frequent and associated with increased risk of death, myocardial infarction, or stroke (62). An increased baseline inflammation associated with the innate immune cell activation is detected in patients with symptomatic CAD (63). Dr. Zahi Fayad (Icahn School of Medicine at Mount Sinai, New York, NY, USA) described how PET/MRI of immune cell dynamics could help connect the dots in atherosclerosis immunology (64). [¹⁸F]-FDG-PET has been widely used to reveal increased metabolic activities and has been used as a surrogate marker for vascular inflammation. Macrophage-targeting [⁶⁸Ga]-DOTATATE can detect atherosclerotic inflammation with better power to discriminate high-risk vs. low-risk coronary lesions than [¹⁸F]-FDG-PET (65). Other radiotracers, such as [¹⁸F]-NaF-PET, have also shown promise in detecting high-risk coronary lesions (66). In addition to tracking immune cell dynamics, changes in the perivascular adipose tissue induced by coronary inflammation can be detected as attenuated signals on routine CT imaging and could serve as a surrogate measure of coronary inflammation (67). Dr. Fayad emphasized the importance of probing a specific cellular process across multiple organs, which could improve our understanding of the interplay of inflammation in disease development. [¹⁸F]-FDG-PET showed an increased spleen metabolic activity after acute coronary syndrome, which is associated with the proinflammatory remodeling of circulating leukocytes and therefore predictive of subsequent CVD events; these findings provided evidence of a cardiosplenic axis in human (68). Based on a longitudinal and cohort study of [¹⁸F]-FDG-PET of the brain, bone marrow, and aorta of the same subject, links between chronic stress, amygdala, vascular inflammation, and bone marrow and increased risk of cardiovascular events have been shown (69).

HUMAN IMAGING APPROACHES IN DETECTING INFLAMMATION DURING DISEASE PROGRESSION

The needs for the development of noninvasive in vivo imaging markers have become essential for the assessment of disease progression and treatment response in patients. This session was intended to address how imaging inflammation capabilities can help our understanding in predicting disease onset and progression in human. Multimodal clinical imaging techniques and their cross-validations in evaluating different stages of the inflammatory process for risk assessment and understanding the progression of human diseases were presented.

PET and MR imaging of the progression of neuroinflammatory diseases

Many neurodegenerative disorders are associated with significant neuroinflammation. Microglial and astroglial activations are the brain’s response to the neurotoxic process. Excessive glial responses may become pathologic and may further contribute to brain injury. Amyotrophic lateral sclerosis (ALS) is a rapidly progressive neurodegenerative disease that could be affected by neuroinflammation and glial activation. Such phenomena can be detected by proton MRS (¹H-MRS) of the brain, which showed a decrease in the neuronal markers N-acetylaspartate and increase in the glial marker myoinositol in the motor cortex in patients with ALS (70) and with HIV-associated neurologic disorders (71). Dr. Eva Ratai (Massachusetts General Hospital, Boston, MA, USA) described PET imaging techniques using [¹¹C]-PK11195 (72), [¹⁸F]-DPA-714 (73), and [¹¹C]-PBR28-PET (74) that have been developed to detect elevated glial activation in ALS. Dr. Ratai indicated that simultaneous brain multimodal scans using PET/MR could be used to cross-validate cortical glial activation measured by [¹¹C]-PBR28-PET, and spectroscopic markers (decreased N-acetylaspartate and increased myoinositol) (74) and these multimodal imaging techniques are currently being used in 3 clinical trials to serve as outcome measures. She noted that brain tumor imaging is one of the areas in which MRS has impacted patient management by distinguishing primary CNS neoplasm from metastasis, therapeutic response, and progression, and between real response and pseudo-response with antiangiogenesis. Although providing excellent reproducibility, challenges such as low sensitivity, lack of standardization in data acquisition and analysis, and lack of cross-validation with other imaging modalities remain a hurdle in the application of MRS in clinical settings. However, numerous research studies are targeted toward addressing those hurdles.

CT imaging of coronary inflammation to predict cardiovascular risk

CVD is the leading cause of death in the developed world. Current cardiovascular risk management is focused mainly on statin treatment to lower cholesterol (e.g., LDL) levels; however, emerging recognition of atherogenesis as an active process instead of cholesterol storage or calcium repository has brought attention to some key inflammatory mechanisms involved in CVD (75). Dr. Milind Desai (Cleveland Clinic, Cleveland, OH, USA) noted that direct assessment of coronary plaque is better than measuring the serum inflammatory markers, such as CRP, IL-1, and IL-6 (76). He described recent advances in the understanding of the vascular microenvironment related to contributions of adipose tissue to coronary inflammation; this has led to the development of the perivascular fat attenuation index (FAI) as a measure of inflammation-induced change in arterial blood flow. Routine coronary CT angiography coupled with perivascular FAI is being utilized as a noninvasive prognostic marker of cardiac death that is specific for flow-limiting inflammatory changes surrounding coronary arteries (77). He explained that FAI, in combination with artificial intelligence–based technologies, has the power to improve the value of imaging biomarkers in CVD. Dr. Desai emphasized that correction for comorbid conditions, instrument limitations, and consistent validated image analysis tools need to be addressed before FAI technology can be applied to current clinical practice. He also highlighted the need for validation of other imaging techniques like PET imaging and CT by large-scale outcome studies in ethnically varied populations for imaging coronary inflammation.

PET and MRI to locate the generators of pain

Chronic pain is one of the biggest clinical problems in the world, and yet pinpointing the site of pain generation and inflammation has remained a challenge in pain management, leading to significant misdiagnosis, mismanagement, and rampant use of opioids and unhelpful surgeries. Dr. Sandip Biswal (Stanford University, Stanford, CA, USA) explained that chronic, nociceptive pain involves significant biochemical, molecular, and physiologic changes not only in neural structures but also in other cells and tissues, such as immune/inflammatory cells, muscle, synovium, and vessels; imaging tools can help pinpoint the pain generators and proper diagnosis and treatment for chronic pain. Dr. Biswal described initial studies using a combination of [¹⁸F]-FDG-PET for metabolic interrogation and MRI for anatomic localization, demonstrating the potential for identifying the source of sciatic pain. A follow-up study with a larger patient population is needed to validate the clinical impact (78) of this work. He also presented the use of naturally occurring guanidinium-based toxins in the development of a radiotracer for the voltage-gated sodium channel as well as a highly specific [¹⁸F]-labeled PET radiotracer ([¹⁸F]-FTC-146) for imaging the σ-1 receptor (a pronociceptive receptor that is up-regulated in painful and inflamed tissue). Simultaneous PET/MRI with [¹⁸F]-FTC-146 helped accurate diagnoses and objective localization of inflammatory pain generator in a patient with chronic knee pain (79). This PET tracer is currently being investigated in phase 1 clinical trials. Dr. Biswal made clear that challenges remain in pain imaging, as the structures of interest, such as nerves and spinal cord, are relatively small and subject to motion artifact. Thus, an imaging modality or a combination of imaging modalities that can provide high image resolution and signal specificity and sensitivity are needed. Also, as pain generators and inflammation often do not necessarily colocalize with the site of pain sensation, whole-body imaging capability will be extremely helpful in the diagnosis of pain generators and inflammation.

CONCLUSION AND FUTURE DIRECTIONS

Inflammation and its resolution are complex and dynamic processes. Biofluid assays are currently used to monitor inflammatory activity at a systemic level. In vivo imaging techniques that can track the changes of molecular signatures, immune cell trafficking, metabolic, and functional parameters are essential for obtaining a clear picture of the role of unresolved inflammation in the development and progression of disease. Identifying the time scale of the inflammatory process leading to a disease state is critical for early diagnosis, disease prognosis, and treatment planning. The currently known inflammatory markers and tissue phenotypes discussed during the workshop are listed in Table 1. During the workshop, the group of clinical speakers, which included physicians, clinical scientists, and surgeons, expressed a need for imaging approaches for detecting the inflammatory conditions that lead to clinically significant pathologies such as cancer, vascular disease, and pain syndromes, and as input for treatment planning and disease monitoring (see Fig. 1).

TABLE 1.

Inflammatory markers or phenotypes of interest expressed by the clinical presenters

Molecular (M)	Cellular (C)	Functional (F)	Structural (S)
CRP, IL-6, IL-10, MMP9, TLR9, IL-8, VCAM-1, ICAM-1, TLR4, proresolving fatty acids and receptors	Immune cells	Tissue oxygenation	Lesion
	Macrophages	Metabolites	Tissue integrity
	Stem cells		Tissue stiffness

ICAM, intercellular adhesion molecule; MMP, matrix metalloproteinase.

Figure 1.

Clinical wishlist for in vivo imaging capabilities to assess inflammation.

Following the clinical presentations, researchers with expertise in different imaging modalities showcased state-of-the-art imaging techniques that can detect inflammatory signatures in different organs and diseases. Although most imaging techniques and associated imaging probes of either established or newly discovered targets have enabled visualization of local inflammation in animal disease models and some early clinical trials, these imaging approaches are focused mostly on a single time point assessment. These imaging approaches could be amenable to longitudinal measurements to capture the subtle changes in inflammatory signatures. Table 2 lists the states of the imaging approaches discussed by the workshop presenters being used to detect inflammation. In the table, imaging approaches are divided into human and preclinical imaging and then further divided into the types of inflammatory signatures that are measured [e.g., molecular (M), cellular (C), functional (F), and structural (S)]. Although not comprehensive, Table 2 also includes discussion from the imaging presenters as well as currently available imaging approaches that are used or can be adapted for other organs, diseases, and translated to the clinic.

TABLE 2.

Comparison between organ-specific clinical needs and currently available imaging approaches

Organs	Human imaging				Preclinical
	M	C	F	S	M	C	F	S
Brain/CNSa,b,d	MRS, PET				CE-MRI, PET
Peripheral nervesa,b	PET
Musclesa,b	MRS			MRI
Jointsa,b,c	PET		PAI	MRI
Hearta,c	MRS, PET		MRI	MRI	OI, PET	CE-MRI	OI	MRI
Blood vesselsa,c,e	PET		PET	MRI, CT, US	CE-US, CE-MRI, PET	CE-MRI	US
Bone marrowa,c					PET
Mucous membranea			PAIg, OIg				PAI
Skina			PAIg, OI		OI		PAI
							OI
Livera,f	MRS		MRI, US, HP-MRIg	MRI			HP-MRI
Pancreasa	MRSg			MRI
Kidneya							HP-MRI
Breasta	MRS		HP-MRIg	MRI
Prostatea	MRS		HP-MRI	MRI	HP-MRI
Gut and intestinal tracta			PAIg, OIg		CE-US, OI, PET
Eyea			PAIg
Glandsa
Lymph nodes/lymphaticsa				MRI, PET, OI	CE-US, OI			OI

Imaging targets: C, cellular; F, functional; M, molecular; S, structural. CE, imaging approaches requiring the use of imaging probes; OI, optical imaging. Disease areas: ^aoncological, immunotherapy, epidemiology, nutrition science; ^bpain; ^crheumatoid arthritis (RA), cardiovascular diseases (CVD); ^damyotrophic lateral sclerosis (ALS), Alzheimer’s disease (AD), Parkinson’s disease (PD); ^edeep vein thrombosis (DVT); ^fnon-alcoholic fatty liver disease (NAFLD), nonalcoholic steatohepatitis (NASH); ^gpotential applications.

Speakers in human application sessions discussed current imaging techniques in humans that are available for clinical translation. They emphasized the importance of integrating the inflammatory measures obtained from multimodal imaging approaches and disease activity measures derived from biofluid-based biomarkers of inflammation to improve the robustness of clinical interpretations. Both mechanistic studies and epidemiologic datasets are essential in developing new imageable targets, understanding their disease-target interaction, and interpreting the signals obtained from the image approaches. There are data available for inflammatory biomarker discovery in human, as described in Box 2 of our previous meeting report (80). Based on the previous meeting report, apart from the large consortium studies, there was no consistency in the way inflammation biomarker data were collected or measured. Although some biomarkers for monitoring inflammation are in use, very few have high selectivity or sensitivity to be used in a larger population for screening or identifying inflammatory processes that lead to the diseased state.

Both clinicians and imaging researchers in the workshop agreed that although numerous epidemiologic studies have been performed on systemic inflammatory biomarkers, many imaging approaches for detecting inflammation are available but underutilized. Novel approaches, such as artificial intelligence and machine learning techniques and image analysis algorithms can be used to identify imaging features relevant to tissue inflammatory activity to address gaps in current clinical imaging capabilities. Currently, there is a disparity between the lack of availability of information for highly prevalent inflammatory diseases and the wealth of data in low-frequency diseases; both are deficient in precise characterization for clinical assessment and decision making. Integration of imaging data results with biofluid-based biomarkers using artificial intelligence and machine learning could also be highly effective in screening high-risk patients.

The issues presented in this NIH workshop highlighted the clinical need for imaging approaches for inflammatory status in combination with serum-based immunologic assays for patient disease characterization. The presentations also described the promise and readiness of several state-of-the-art imaging approaches for assessing inflammation. The need for developing fast image acquisition for characterizing rapid inflammation-associated target behavior was also highlighted. The importance of integrating results from multimodal imaging approaches, biofluid assays, and longitudinal acquisition of these indices was emphasized. Probing multiple inflammatory signatures associated with different pathways or 1 inflammatory signature across multiple tissues could improve our understanding of the interplay of inflammation in disease development. A coordinated effort between imaging researchers, imaging probe developers, clinicians, molecular biologists, and drug developers is needed to advance the field. Most importantly, these imaging approaches need to be sensitive, specific, and clinically relevant to treatment decision making to be incorporated into clinical practice or serve as a secondary outcome in clinical trials for successful interventions.

Glossary

α7-nAChR

α7 nicotinic acetylcholine receptor

AFP

α-fetoprotein

ALS

amyotrophic lateral sclerosis

ART

antiretroviral therapy

CAD

coronary artery disease

CEUS

contrast-enhanced US

CNS

central nervous system

CRP

C-reactive protein

CT

computed tomography

CVD

cardiovascular disease

DHA

docosahexaenoic acid

DII

Dietary Inflammatory Index

DVT

deep vein thrombosis

EPA

eicosapentaenoic acid

FAI

fat attenuation index

FDA

U.S. Food and Drug Administration

FDG

fluorodeoxyglucose

FNIH

Foundation for the U.S. National Institutes of Health (NIH)

HCC

hepatocellular carcinoma

HIV

human immunodeficiency virus

HP-MRI

hyperpolarized MRI

ICB

immune checkpoint blockade

IVM

intravital microscopy

MRS

magnetic resonance spectroscopy

NAFLD

nonalcoholic fatty liver disease

NASH

nonalcoholic steatohepatitis

OPN

osteopontin

PAI

photoacoustic imaging

PTD

postthrombotic disease

PET

positron emission tomography

PLWH

persons living with HIV

RA

rheumatoid arthritis

sEH

soluble epoxide hydrolase

SPM

specialized proresolving mediator

TPEF

2-photon excited fluorescence

TSPO

translocator protein

US

ultrasound

AUTHOR CONTRIBUTIONS

C. H. Liu, N. D. Abrams, D. M. Carrick, P. Chander, J. Dwyer, M. R. J. Hamlet, A. L. Kindzelski, M. PrabhuDas, S.-Y. A. Tsai, M. M. Vedamony, C. Wang, and P. Tandon jointly wrote the manuscript; and C. H. Liu compiled and finalized the manuscript.