Molecular imaging of pulmonary diseases

Lung cancer is the leading cause of cancer-related death in men worldwide, and the second leading cause in women [3]. In the United States, lung cancer is the leading cause of cancer death in both men and women [3]. It is estimated that greater than 25% of cancer-related deaths in the United States in 2017 will be caused by lung cancer, and that it will kill more people than breast, colon, and prostate cancer combined [4]. Lung cancer has three main histologic categories: non-small cell lung cancer (NSCLC), small-cell lung cancer (SCLC), and neuroendocrine tumors (NET) [5]. NSCLC is most common, accounting for 85–90% of all lung cancers, and is made up of squamous cell carcinoma, adenocarcinoma, and large cell carcinoma. The main risk factor for the development of lung cancer is smoking. Other risk factors include exposure to air pollution, second-hand smoke, radiation (either radon decay or from medical X-rays), and asbestos [5]. Independent of type, early and accurate diagnosis of lung cancer is important to increase the likelihood of survival. Chest radiographs are the first line of conventional imaging used to detect abnormal areas in the lung but are limited in their capacity to detect cancer due to low-resolution and the two-dimensional nature of the image. Low-dose spiral CT is therefore the gold standard for lung cancer diagnosis and staging [5]. The National Lung Screening Trial found that patients at high-risk for lung cancer randomized to yearly CT scans had a 20% reduction in lung cancer mortality compared with patients randomized to yearly chest radiographs [6]. However, low-dose CT does have limitations, including a higher dose of radiation, more frequent detection of questionable lesions leading to potentially unnecessary biopsies or surgical intervention, and inability to detect chest wall invasion [7, 8]. Due to the limitations of low-dose CT and the fact that lung nodules are often found incidentally, patients with lung cancer would benefit from improved imaging modalities that can provide early detection, diagnosis, and appropriate staging.

Solitary pulmonary nodule (SPN) is the name given to a small isolated dense region observed on chest X-ray and CT scans. SPNs are sources of preoccupation since they can be of benign or malignant causes such as inflammation, infection or lung cancer [9]. PET/CT and its high detection sensitivity associated with molecularly targeted probes can permit further distinction. In this regard, [¹⁸F]-fluorodeoxyglucose ([¹⁸F]-FDG) is the most commonly use radiotracer in oncology. Its ability to mimic glucose reveals information about cellular glucose metabolism and allows discrimination of cell populations or tissues with high glucose metabolic rate, usually active proliferation, from a quiescent cell population. However, not all cancers are [¹⁸F]-FDG avid, leading to false negative interpretation. In addition, metastatic spread can be misinterpreted if the primary tumor showed initially low [¹⁸F]-FDG uptake. On the other hand, false positive scans can occur as inflammatory or infectious condition protagonists tend to have high metabolic rate, thus exhibiting high [¹⁸F]-FDG uptake. A more accurate differential diagnosis is necessary to improve the outcome of lung cancer patients via appropriate staging, treatment, and assessment of the lung disease.

A hexosamine pathway tracer, [^99mTc]-ethylenedicysteine-glucosamine ([^99mTc]-ECG), was evaluated in phase 1 and phase 2 clinical trials where investigators were able to discriminate inflammation from tumors in NSCLC patients [10]. Further studies have been performed in mesothelioma-bearing rats [11]. Here, [^99mTc]-ECG and [⁶⁸Ga]-ECG were taken up by tumors at a lower extent than [¹⁸F]-FDG, but still enough to exhibit a positive signal in comparison to muscle on in vivo images 30 min post-injection, thus providing discriminatory information on the hexosamine pathway in such tumors.

3′-deoxy-3′-[¹⁸F]-fluorothymidine ([¹⁸F]-FLT), an analogue of thymidine that is incorporated in DNA during proliferation, has been developed as an imaging marker of proliferation. Recently a meta-analysis study comparing Ki67 index from biopsies of 1213 lung cancer patients to [¹⁸F]-FDG or [¹⁸F]-FLT standard uptake values revealed respective pooled Rho values of 0.45 and 0.65, indicating higher reliability for [¹⁸F]-FLT to assess lung cancer cell proliferation [12]. Artificial amino acid tracers such as O-[¹⁸F]-fluoromethyl-L-tyrosine (FMT) [13], O-[¹⁸F]-fluoropropyl-L-tyrosine (FPT), and O-(2-[¹⁸F]-fluoroethyl)-L-tyrosine (FET), have also been evaluated as amino acid metabolism and protein synthesis is enhanced in tumors [14].

Angiogenesis is a phenomenon of great interest in oncology. The more a tumor is vascularized, the higher is the risk of metastasis [15‐17]. Expression of vascular endothelial growth factor receptor type 2 (VEGFR-2) in tumors has been proportionally correlated to poor patient prognosis. Luo et al. developed a PET [⁶⁴Cu]-labeled monoclonal antibody targeted to VEGFR-2 [18]. In this study, the investigators were able to successfully discriminate high versus low VEGFR-2 expression levels in a mouse model bearing human NSCLC tumors using [⁶⁴Cu]-NOTA-ramucirumab. Such capability would allow effective monitoring of the efficacy of therapeutic treatments with angiogenesis inhibitors in lung cancer.

Hypoxia has been observed and characterized in many tumors. Hypoxic cells are phenotypically different compared to the non-hypoxic cells that might have been originally used to validate the efficacy of the treatment. Hypoxia is one of the main players that will determine therapy efficacy. Nitroimidazole derivatives have been used in the development of molecular imaging tracers. [¹⁸F]-fluoromisonidazole-1-(2-nitroimidazolyl)-2-hydroxy-3-fluoropropane (FMISO) and [¹⁸F]-fluoroazomycin-arabinofuranoside (FAZA) showed strong correlation with measurements of tumor oxygenation [19]. However, FAZA exhibited a better signal-to-background ratio than FMISO due to FAZA hydrophilicity inducing fast clearance in preclinical studies [20]. Clinically, little impact has been made by hypoxia imaging tracers in lung cancer mostly due to the lack of understanding of the mechanistic actions of these imaging agents [19].

Matrix metalloproteinases (MMPs) are key players in cancer progression and are known to be overexpressed in lung tumor tissue [21]. Fluorescent bioactivable probes such as MMPSense 680 showed potential to delineate tumor mass in mutant K-ras mice [22]. This study demonstrated the ability of MMPSense 680 to discriminate normal tissue from adenoma or adenocarcinoma via fluorescent molecular tomography. Unfortunately, due to the lack of tissue penetration depth, the translation of fluorescent probes to a non-invasive clinical application is not yet feasible. However, intraoperative molecular imaging using such probes can be of tremendous help for surgeons as it facilitates tumor delineation, thus confirming that complete removal of tumor residue during surgical resection is possible [23].

In lung cancer as in other cancers, conventional imaging provides evidence for the existence of small tissue abnormalities or lesions to help guide decision making regarding whether to further investigate for malignancy. Among the various imaging modalities used in nuclear medicine, PET is the closest modality to CT imaging in terms of size limits of detection. The primary clinical role of PET imaging is the detection of anomalous regions of [¹⁸F]-FDG uptake, which are often indicative of malignant lesions. The limits of detectability of tumors depend on many parameters including location of the tumor and imaging isotope. Modern clinical PET scanners have a resolution limit of 4 mm, corresponding to the detection of tumors with a volume of 0.2 ml (7 mm diameter) in 5:1 tumor-to-background ratio; whereas the minimum lesion size that can be measured with CT is about 3 mm [24]. Significant research efforts are underway to develop new detector materials, improved camera design, and new reconstructive algorithms in order to improve the limits of tumor detectability.

ARDS is an acute, progressive inflammatory lung disease that affects 10% of all intensive care unit (ICU) patients worldwide [25, 26]. The pathophysiology of ARDS includes lung inflammation leading to increased permeability of the pulmonary vasculature, accumulation of pulmonary edema, increased lung weight, and a reduction in lung surface area participating in gas exchange [27]. ARDS is diagnosed clinically using the Berlin definition, which includes timing of onset, appearance of chest imaging, origin of edema, and oxygenation capacity. Timing of onset for ARDS is within 1 week of new or worsening respiratory symptoms or of a known respiratory insult. Common risk factors for the development of ARDS include trauma caused by aspiration, burns, blood transfusion and noncardiogenic shock. However, most ARDS cases are mainly caused by pneumonia and sepsis [25, 28].

Many different viral, bacterial, and fungal organisms cause respiratory infections [29]. Collectively, these infections are one of the most prevalent worldwide public health problems [30]. Although many community-acquired respiratory infections are short-lived and are successfully managed as an outpatient with little to no treatment, respiratory infections cause more morbidity and mortality in the United States than any other infection [29]. The mortality rate for respiratory infections has remained rather stagnant for the past 50 years [31]. Patients who acquire a respiratory infection and are elderly, immunocompromised, or have multiple comorbidities or baseline lung dysfunction, may require inpatient hospitalization, ICU admission, mechanical ventilation, and broad-spectrum antibiotics. Respiratory infections are diagnosed through a combination of a thorough history and physical exam, clinical indicators (leukocytosis, presence of microorganisms in sputum culture), and imaging.

Common imaging modalities useful in the diagnosis of respiratory infections are chest X-ray and CT [32]. Both chest X-rays and CT scans provide nonspecific, supporting evidence for the presence of a respiratory infection, but they do not allow for identification of the etiology of a specific infection or the involved immune cell types. This lack of clinically available imaging techniques that could help identify the specifics of an infection impacts the clinician’s ability to provide timely, targeted therapeutic interventions. The use of molecular imaging may allow for earlier identification of patients who are progressing towards ARDS. Additionally, the ability to image respiratory infections using molecular markers may provide important information regarding the likely etiology of the infection (i.e. bacterial versus viral). This improvement in diagnosis may allow for earlier, targeted antibiotic treatment and protection from overtreatment when it is not likely to be beneficial.

Ex vivo administration of [¹¹¹In]-labeled leukocytes are useful tools to detect infection [33]. In a cohort of 145 patients, administration of [¹¹¹In]-WBCs showed a 94% sensitivity and 64% specificity for detecting pulmonary infections [34]. In addition, a 99% negative predictive value has been observed in 137 immuno-deficient patients without focal pulmonary activity [35]. The detection of bacterial load by targeting specific prokaryotic signatures is an alternative method that can be achieved using molecularly targeted probes. For instance, a radiolabeled nucleoside analog that is an exclusive substrate of bacterial thymidine kinase (TK), has been studied for this purpose. 2′-fluoro-2′-deoxy-1β-D-arabinofuranosyl-5-[^124/5I]iodouracil ([^124/5I]-FIAU) was injected in a mouse model of lipopolysaccharide (LPS)-induced inflammation as well as an E. coli RS218 TK-positive infected mouse lung model (Fig. 1) [36]. The investigators were able to discriminate the inflammation state from infection state in the mice, reporting an [^124/5I]-FIAU uptake 4 times greater (1.69 ± 0.40 vs. 7.14 ± 1.09 %ID/g, respectively). In vivo SPECT/CT images confirmed these results. However, the use of [¹²⁵I] is usually not recommended for SPECT imaging purposes as the emitting energy is very low. Switching to PET would improve the quality of the images as well as quantitative analysis. Unfortunately, [¹²⁵I]-FIAU PET has only been tested in humans for infection condition that is not lung related and did not show suitable results [37].

Bacterial ribosomal RNAs have also been studied as a bacterial marker. Chen et al. generated a [^99mTc]-radiolabeled 18-mer oligomer phosphorodiamidate morpholino sequence ([^99mTc]-MORF) that is complementary to a 16S rRNA exhibited by most bacteria [38]. The researchers found that [^99mTc]-MORF uptake was higher in mice with Klebsiella pneumoniae infection versus mice with sterile inflammation within 90 min. Qualitatively, radiotracer uptake was observed on SPECT/CT whole body images. However, the background appeared elevated.

Another interesting study by Jørgensen et al. evaluated inflammation and infection processes by comparing [¹⁸F]-FDG PET to uptake of cholinergic PET imaging agents such as [¹⁸F]-fluoroethoxybenzovesamicol ([¹⁸F]-FEOBV), a marker of vesicle cholinergic transporter, and [¹¹C]-donepezil, a marker of acetylcholinesterase [39]. Using a murine model of Staphylococcus aureus infection as well as pigs displaying post-operative abscesses, the authors demonstrated a different accumulation kinetics compared to [¹⁸F]-FDG, suggesting that these two radiotracers image different aspects of immune response. Human data showed higher [¹¹C]-donepezil activity at sites of chronic inflammation when compared to [¹⁸F]-FDG.

Radiolabeled antibiotics have been thought to be promising as they would be targeting specific bacteria items. Many antibiotics have been tested to this purpose [40]. Among them, ciprofloxacin, a member of the fluoroquinolone family, binds to gram-negative and gram-positive bacterial DNA gyrase. When radiolabeled with [^99mTc], ciprofloxacin penetrates mammalian cells and tissues but with low affinity and quick wash out. However, after a prolonged time, [^99mTc]-ciprofloxacin is retained at sites of infection allowing a target-to-background ratio favorable for imaging. High sensitivity (>90%) has been found in tuberculosis and other conditions reported from a large multicenter study [41]. However, a PET version of ciprofloxacin radiolabeled with [¹⁸F] showed rapid wash-in, wash-out and non-specific binding in the infected area, which depreciate the suitability of this radiotracer for infection sites [42]. Another fluoroquinolone-complex named norfloxacin dithiocarbamate (NFXDTC) has been well-characterized [43]. Zhang et al. radiolabeled NFXDTC with [^99mTc] and showed a 3.43±0.53% injected dose per gram (%ID/g) of infected area compared to 1.76±0.20%ID/g obtained with [^99mTc]-ciprofloxacin. However, lung and liver uptake are very elevated which prevent its use in related conditions. As many radiolabeled antibiotics have been or still are under investigation, none has yet demonstrated significant capacity to differentiate sterile inflammation from infection [40].

Chronic respiratory diseases, including chronic obstructive pulmonary disease (COPD) and asthma, cause significant morbidity and mortality worldwide, accounting for 4 million deaths annually [44, 45]. COPD, which includes emphysema and chronic bronchitis, can be suspected based on the presence of dyspnea, chronic cough, chronic sputum production, recurrent respiratory infections, and are associated with known risk factors such as exposure to tobacco smoke and occupational vapors [46, 47]. The gold standard for diagnosing COPD is the pulmonary function test (PFT), which uses spirometry to measure lung volumes. A post-bronchodilator ratio of the fraction of expired volume in one second to the forced vital capacity (FEV1/FVC) of less than 0.70 is indicative of persistent airflow limitation [46]. The mainstay of treatment for COPD is bronchodilation, using either short- or long-acting β2-adrenergic agonists or anticholinergic agents, depending on the severity of symptoms [46].

The role for conventional imaging modalities (chest x-rays and CT scans) in the diagnosis of COPD and other chronic lung diseases is to identify structural abnormalities (air trapping, airway septal thickening) and to rule out concomitant diseases that may be responsible for the patient’s respiratory symptoms. Unfortunately, these conventional modalities are limited in their ability to diagnose chronic lung diseases. Chest x-rays, while inexpensive and associated with low radiation exposure, are two-dimensional and have low resolution (Table 1). CT scans are significantly more valuable due to the increased spatial resolution and improved depth of penetration; however, they are expensive and require higher radiation doses (Table 1). Additionally, research has shown that the visual evaluation of CT scans has low reproducibility [48]. However, the use of densitometric evaluation improves the utility of CT scans in diagnosing chronic lung disease [49].

Asthma is another common chronic respiratory disease that can cause airflow obstruction and symptoms similar to COPD but differs in that asthma is reversible with proper treatment. Asthma is characterized by intermittent dyspnea, cough, and wheezing. Intermittent asthma is managed with short-acting β2-adrenergic agonists such as albuterol, while persistent asthma is treated with inhaled corticosteroids [50]. PFTs are used during the diagnostic workup of asthma, however results are often normal considering that asthma symptoms occur intermittently. Additionally, the use of bronchodilators during PFTs will demonstrate the reversible nature of the airway obstruction associated with asthma, which differs from the irreversible nature of COPD. Chest x-rays and CT scans have little role in the diagnosis of asthma and will usually not demonstrate any structural abnormalities. Asthma affects 7.5% of the population in the United States, with onset usually in childhood or as a young adult [50]. A family history of asthma, exposure to environmental toxins such as tobacco smoke, and low socioeconomic status are risk factors for the development of asthma [50]. The hallmark of airway inflammation in patients with asthma is allergen-specific IgE- and eosinophil-mediated neutrophil activation. Additional physiologic changes include mast cell degranulation and infiltration of T-helper lymphocytes, and is associated with elevated levels of IL-4, IL-5, and IL-13 [51‐54].

Currently, planar lung scintigraphy is usually prescribed to visualize distribution of ventilation and perfusion in the lungs of patients with pulmonary embolism but also in cases of emphysema patients that undergo lung volume reduction surgery. However, use of 3-dimensional ventilation/perfusion (v/q) SPECT imaging methods have increased the diagnostic accuracy and expanded the application range to various pulmonary diseases. Technegas, ^99mTc-labelled carbon particles, has been tested as an aerosol in COPD patients [55]. With a moderate to strong correlation with spirometric lung function, Technegas has potential to provide valuable information for the diagnosis of COPD. However, more studies that link v/q SPECT to lung function need to be performed to fully validate the method.

Considering that conventional imaging modalities play a limited role in the diagnosis and management of chronic respiratory diseases, there exists a significant opportunity to develop cell type-specific molecular imaging strategies than can improve diagnosis, provide monitoring of treatment responsiveness, and limit invasive and often traumatic tissue biopsies. In addition, molecular imaging could have a role in quantifying with accuracy lung involvement in COPD, asthma and other chronic pulmonary diseases. The involvement of immune cells in the development of chronic lung diseases is an important topic, which was addressed in a study by Jones et al. that assessed neutrophilic infiltration and presence of macrophages in a small group of COPD and asthmatic patients [56]. Using PET imaging associated with [¹⁸F]-FDG to assess metabolism and [¹¹C]-PK11195 as a marker of macrophages, the investigators found a 2.7- and 2.5-fold increase in [¹⁸F]-FDG uptake in COPD patients versus, respectively, asthmatic and healthy patients whereas [¹¹C]-PK11195 tissue-to-plasma ratios were equally higher in asthma and COPD conditions compared to normal patients. The difference in [¹⁸F]-FDG uptake between COPD and asthma might come from a different state of neutrophil activation between the two conditions. [¹⁸F]-FDG avidity to neutrophil activation has also been demonstrated and quantified by PET imaging methods (Fig. 2) [57, 58].

MMPs are known to play an important role in lung inflammation and remodeling. Using a transgenic mouse model that exclusively expresses IL-13 in lungs and an RP805 precursor [^99mTc]-labeled macrocyclic MMP-targeted tracer, Golestani et al. obtained a significant 2.5-fold increase in tracer uptake in these mice compared to wild-type animals via SPECT/CT imaging (Fig. 3) [59]. RP805 exhibited good correlation with CD68 expression (r = 0.70, P < 0.01). Other radiotracers such as [⁶⁸Ga]-DOTANOC and [¹¹¹In]-DTPA-A20FMDV2 may be useful for detection of fibroblast activity and AvB6-integrin expression in pulmonary fibrosis, respectively [60, 61].

In addition to IRI and PGD, allograft rejection is another serious complication following lung transplantation that contributes to poor outcomes. Cell-mediated (recognition of foreign human leukocyte antigens by recipient T lymphocytes) and antibody-mediated (binding of recipient antibodies to donor cell antigens) rejection can occur both in the short- and long-term following lung transplantation [78]. Diagnosing allograft rejection can be difficult due to variations in clinical presentation, with some patients being completely asymptomatic. Transbronchial biopsies are therefore required and are the gold standard for diagnosing rejection [78]. Given the inherent risks associated with transbronchial biopsies (bleeding, infection, pneumothorax), a noninvasive molecular imaging strategy using cell type- or antibody-specific labeling could provide a significant advancement in the diagnosis and surveillance of allograft rejection.

As previously mentioned for PGD, molecular imaging for acute or chronic lung rejection has not been widely studied. However, methods used to visualize cellular and molecular components of rejection such as cell membrane disruption, apoptosis and monitoring of immune cell activation in heart rejection may be applied to the lung [79]. In addition, a study by Chen et al. demonstrated that [¹⁸F]-FDG uptake was significantly increased in a mouse model of orthotopic lung transplantation and rejection [80]. These authors demonstrated that [¹⁸F]-FDG uptake occurred largely in CD8(+) T cells and was markedly attenuated following T cell depletion therapy.