Current status of targeted microbubbles in diagnostic molecular imaging of pancreatic cancer

Abstract Pancreatic ductal adenocarcinoma (PDAC) is often associated with a poor prognosis due to silent onset, resistance to therapies, and rapid spreading. Most patients are ineligible for curable surgery as they present with advanced disease at the time of diagnosis. Present diagnostic methods relying on anatomical changes have various limitations including difficulty to discriminate between benign and malignant conditions, invasiveness, the ambiguity of imaging results, or the inability to detect molecular biomarkers of PDAC initiation and progression. Therefore, new imaging technologies with high sensitivity and specificity are critically needed for accurately detecting PDAC and noninvasively characterizing molecular features driving its pathogenesis. Contrast enhanced targeted ultrasound (CETUS) is an upcoming molecular imaging modality that specifically addresses these issues. Unlike anatomical imaging modalities such as CT and MRI, molecular imaging using CETUS is promising for early and accurate detection of PDAC. The use of molecularly targeted microbubbles that bind to neovascular targets can enhance the ultrasound signal specifically from malignant PDAC tissues. This review discusses the current state of diagnostic imaging modalities for pancreatic cancer and places a special focus on ultrasound targeted‐microbubble technology together with its clinical translatability for PDAC detection.


| INTRODUCTION
Pancreatic ductal adenocarcinoma (PDAC) has one of the worst prognoses of all types of cancer, and its rising incidence projects the disease to become the second deadliest cancer by 2030 after lung cancer. 1 Approximately 80% of patients with PDAC present with locally advanced or metastatic disease at their initial diagnosis, or do not qualify for a complete tumor resection due to its diffused nature.
All tumor stages combined, PDAC has a dismal prognosis with a 5-year survival rate of <9%. 2 Even for the small percentage of patients diagnosed with localized disease, the 5-year survival rate is only of 37%. The need for effective screening methods is globally recognized since diagnosing pancreatic diseases at an early stage (for example by serendipitous discovery in asymptomatic patients when evaluating unrelated disease) can drastically improve outcomes by providing patients opportunities for effective treatments with fewer complications. 3 The detection of pancreatic cancer biomarkers in plasma or serum (e.g., proteins, circulating tumor cells, circulating nucleic acids, aberrantly expressed cancer-associated antigens, metabolites, small molecules, and exosomes) appears promising. [4][5][6] Nevertheless, serum carbohydrate antigen  is, for now, the only serum biomarker that has been routinely used in clinical practice to monitor PDAC progression, recurrence, and therapy response. 7 Of high interest, serum CA 19-9 level has been shown to be significantly upregulated as early as 2 years before pancreatic cancer diagnosis. 8 Unfortunately, the reliability of CA 19-9 as a biomarker has been compromised by false negative results in patients lacking fucosyltransferase activity. 9 Moreover, false positive results are also observed in benign pancreaticobiliary diseases such as hepatic cyst, 10 obstructive jaundice, 11 cholangitis, 12 and pancreatitis. 13 Therefore, the accuracy of CA19-9 is debatable, and it is recommended to diagnose patients using imaging techniques. 14 16 This review brings attention to the current status of the targeted microbubble technology for PDAC diagnosis via ultrasound molecular imaging. We specifically discuss the details of this technology and compare it to alternative molecular imaging strategies, with a special emphasis on the sensitivity and specificity of these imaging modalities in differentiating PDAC from benign conditions. Finally, we culminate this review by providing an outlook toward clinical translation of ultrasound targeted PDAC contrast agents.

| CURRENT STATUS OF PDAC DIAGNOSTIC IMAGING IN THE CLINIC
Over the past decade, multiple publications have reviewed the accuracy of the current clinical imaging modalities (CT, MRI, US, and PET) in detecting pancreatic cancers. 15,[17][18][19] These modalities are briefly presented below under two different categories such as, anatomical imaging modalities, and molecular/metabolic imaging modalities. Main characteristics of those imaging modalities are summarized in Table 1.

| Anatomical imaging modalities
Anatomical imaging modalities provide useful information by detecting primary tumor or metastatic foci and aid in determining tumor resectability. Such imaging modalities exploit the differences between normal and abnormal tissue properties to create contrast.
Transabdominal ultrasound, endoscopic ultrasound, CT and MRI, are some primary imaging modalities falling under this category.

| Transabdominal ultrasound
Transabdominal ultrasound imaging uses sound waves to produce pictures of the structures within the upper abdomen. It involves an ultrasound transducer pressed against the skin of the abdomen. Sound waves delivered from the transducer at a specific frequency bounce off tissues and create echoes. A computer, connected to the transducer, utilizes those sound waves to create an image (sonogram). For patients presenting with jaundice or abdominal pain, noninvasive transabdominal ultrasound is a frequently recommended initial imaging modality. Transabdominal ultrasound imaging is convenient, devoid of ionizing radiation, portable, widely available and economical.
On the contrary, ultrasound imaging is a highly operator-dependent modality, that sometimes cannot allow a panoramic view of the pancreas due to the presence of gas in the stomach or duodenum. As a direct consequence, the sensitivity of transabdominal ultrasound for detecting pancreatic cancer varies widely, from 50% (tumor <1 cm in diameter) to 90% (tumor >3 cm in diameter). 20,21 Moreover, differentiating pancreatic cancer from other focal lesions, such as chronic pancreatitis, is very challenging due to similarities in their superficial imaging features. 21

| Endoscopic ultrasound
In EUS, ultrasound transducer is placed much closer to the pancreas than in transabdominal ultrasound. The ultrasound transducer is inserted into the mouth and positioned down into the first part of the small intestine. EUS imaging represents one of the most sensitive methods for detecting PDAC. It is especially useful to detect small tumors that are not visualized by other imaging modalities (typically <2 cm). 22,23 EUS also provides the opportunity to collect tissue samples by fine needle aspiration (FNA) of suspected pancreatic lesion during imaging session. Samples can be used for biochemical, cytologic, and/or DNA analysis, thus helping to confirm the diagnosis or to further characterize the tumor. However, the performance of traditional EUS can be limited, especially when imaging patients with symptomatic or asymptomatic inflammatory status, the inflammation can interfere with pancreatic cancer diagnosis. 24 EUS interpretation relies on the normal pancreatic parenchyma as point of distinction to pathological lesions, typically hypoechoic. Acute pancreatitis can mask significant findings or be misinterpreted as a mass, and chronic pancreatitis can result in an extensive loss of normal pancreatic parenchyma. For those patients, EUS has a sensitivity and a specificity of 63.6 and 75.9%, respectively. 25 EUS-FNA constitutes a more accurate diagnostic tool. A meta-analysis evaluates the accuracy of EUS-FNA with 86.9% sensitivity and 95.9% specificity. 26 Nevertheless, to reach high sensitivity in patients with chronic pancreatitis is still a challenging task. 17,27,28 Furthermore, EUS-FNA is an operator dependent technique that requires sedation or sometimes general anesthesia, which may induce risks in some patients (e.g., cognitive dysfunction, malignant hyperthermia, breathing problems). 29,30 Although endoscopic ultrasound has high accuracy for the detection of small pancreatic tumors, it is invasive and considered as a complementary tool in current clinical practice.

| Computed tomography
Multi-detector computed tomography (MDCT) is the most readily available imaging modality and the most frequently used technique in the diagnosis and staging of PDAC. It is also the gold standard to evaluate pancreatic resectability. For optimal performance, CT examination is performed with an intravenous iodinated contrast agent injection. CT has an overall sensitivity between 76 and 94% for diagnosing PDAC with a specificity between 70 and 96%. [31][32][33] However, the sensitivity can decrease to 70% for smaller tumors (<2 cm), and can be as low as 67% for lesions <1.5 cm. 31,34 Thus, MDCT is inefficient in detecting small pancreatic lesions (<1 cm). Yet, as pancreatic cancer can start metastasize when tumors are <1 cm, detecting lesions at the earliest stage is of great importance. 35  lesions. 17,19 Furthermore, CT is accompanied by the risk of nephrotoxicity from the injected iodine contrast agent, and capable of causing DNA damage from the ionizing radiations.

| Magnetic resonance imaging
MRI is another important imaging modality that can help in diagnosing patients at initial presentation. Usually, magnetic resonance is reserved as a second-line imaging modality when suspected pancreatic tumors are not visible on CT or in case of equivocal CT findings.
Because of its high soft tissue contrast, MRI has been preferred over CT for assessing small tumors in the pancreas, and for precisely detecting enlarged lymph nodes and distant metastases. 36,37 Magnetic resonance cholangiopancreatography (MRCP) is often included in the MRI examination to improve evaluation of the biliary and pancreatic ductal system. The lack of ionizing radiation for image acquisition makes MRI an ideal tool for follow-up examinations. The sensitivity of MRI is low but can be improved by contrast agents modifying the T1 or T2 relaxation time constants of tissues (small gadoliniumcontaining contrast agents and superparamagnetic iron oxide nanoparticles, respectively). MRI has a reported sensitivity and specificity for pancreatic cancer diagnosis ranging from 85 to 93% and from 72 to 79%, respectively. 38,39 Nevertheless, chronic pancreatitis and autoimmune pancreatitis can appear as focal mass mimicking pancreatic adenocarcinoma. 31 Drawbacks of MRI include the susceptibility of the image quality to internal and external motions, long acquisition times, relatively high costs, lower availability relative to ultrasound and CT imaging, and potential contraindications (e.g., cardiac implantable electronic devices, metallic foreign bodies, implantable neurostimulation systems, drug infusion pumps, magnetic dental implants).

| Metabolic imaging modalities
Compared to anatomical imaging modalities that are offering morphological information, PET provides exclusive information about the molecular and metabolic changes associated with the disease. This modality uses the metabolic radiotracer 18 F-fluoro-2-deoxy-D-glucose ( 18 F-FDG) that accumulates in cells with increased glycolytic metabolism. The high glycolytic rate in malignant cells forms the basis of PET imaging. Studies have reported its relatively high sensitivity and specificity in the detection of pancreatic malignancy, ranging from 85 to 96% and 61 to 94%, respectively. 38,[40][41][42] Several studies have evaluated the ability to differentiate chronic pancreatitis from pancreatic cancer based on accumulation patterns of 18 F-FDG within the pancreas. [43][44][45] Those studies have shown that the accumulation of 18 F-FDG in chronic pancreatitis generally demonstrated to be at low level with diffused 18 F-FDG uptake pattern compared to pancreatic cancer.
However, 18 F-FDG lacks target-related specificity, and active metabolic uptake, that is, positive results, can also be found in patients with benign conditions. 46 New PET radiotracers with high specificity to differentiate pancreatic lesions with different level of malignancies are under development. 47,48 Globally, MDCT constitutes the first imaging modality in suspected pancreatic cancer patients. Alternatively, MRCP may be used in centers where this facility is readily available. PET-CT could serve as a useful functional imaging approach. Finally, ultrasonography is an easy and fast way to image the pancreas area of patients with upper gastrointestinal complaints. EUS appears complementary to CT and MRI in the assessment of lesions not clearly detected, but suspected, on CT/MRI. 22

| Clinically approved ultrasound contrast agents
Ultrasound imaging could be a valuable diagnostic tool when a wholebody imaging is not required. To improve the quality and the reliability of ultrasound scans, various methods are currently available. Among which contrast-enhanced endoscopic ultrasound (CE-EUS), endoscopic ultrasound elastography, and CEUS are being frequently employed in the clinic. Microbubbles are gas-filled microparticles of approximately 2-6 μm in diameter, and are pure vascular contrast agents. Their inert gas core can be air but most frequently heavier gases such as perfluorocarbon with low diffusion constants and low solubility in blood have been used. The outer shell of the contrast agents can consist of a variety of materials including albumin, polymers or phospholipids. 54 Additionally, microbubbles can also be exploited to carry molecules, and have important applications as transporters of therapeutic drugs or genes. 55 Microbubbles have the possibility to be conjugated to targeting ligands, referred as targeted-microbubbles. By contrast, microbubbles without molecular targeting capacities are referred as nontargeted microbubbles. Microbubbles are highly echogenic in response to an incident ultrasound beam due to the impedance mismatch against the surrounding tissue. 56 They can oscillate nonlinearly upon interacting with the ultrasound wave and provide greater tissue contrast in relation to the background signal. 57 The desmoplastic reac- Such treatment approaches may transform PDAC vascularity in humans, which in turn will aid in increasing not only the delivery of therapeutics but also the sensitivity of CETUS imaging. However, increasing angiogenesis is a double-edged sword and must be carefully controlled since it promotes tumor growth and metastasis. 61 This is the reason why many therapeutic strategies either involve the inhibition of pro-angiogenic factors (e.g., vascular endothelial growth factor [VEGF]), either use endogenous angiogenesis inhibitors such as endostatin or angiostatin. A meta-analysis has shown that CE-EUS has a sensitivity of 92% and a specificity of 86% in the diagnosis of pancreatic tumors. 62 On the other hand, CEUS has proved to provide tissue information with both high specificity and sensitivity. Two meta-analysis have been published with pooled sensitivity and specificity of 93-94% and 88-89%, respectively. 63,64 Compared to CE-EUS, CEUS has the main advantage to work without endoscopy for similar sensitivity and specificity, and thus is often preferred. Details on microbubbles used in clinics for CEUS imaging of pancreatic cancer are presented below. Biology. 66,67 In the United States, the Food and Drug Administration (FDA) has approved several nontargeted microbubbles to improve organ imaging and assess tissue perfusion. 68  and Sonazoid (GE Healthcare, Olso, Norway) ( Table 2). The gas contained in the first-generation microbubbles, such as Albunex, Levovist and Echovist, included air with poor duration of ultrasound enhancement due to rapid diffusion out of the microbubbles into the blood.

| FDA-approved ultrasound contrast agents
The second-generation utilized inert and heavy gases with low diffusion coefficient such as sulfur hexafluoride or perfluorobutane (Imagent, SonoVue, Definity, Optison and Sonazoid). 69 However, currently used microbubbles acknowledge limitation with their wide size distribution range, which needs optimization according to the transducer frequency bandwidth to increase imaging efficiency. Nevertheless, these microbubbles provided the basis for the design and synthesis of molecularly targeted UCAs, ultimately contributing to the development of ultrasound molecular imaging techniques for accurate and early detection of PDAC.

| Molecularly targeted UCAs in PDAC imaging
Molecularly targeted ultrasound imaging relies on systemically delivering contrast agents that bind to biological markers overexpressed in vascular margins of the pathological tissues. Contrary to nontargeted UCAs, targeted UCAs have the great potential to improve pancreatic cancer early detection imaging with greater sensitivity and specificity.    Ongoing research projects for complete clinical application of targeted microbubbles explore other signal quantification methods, which are compatible for human testing, faster to analyze, and more real-time. 77 Contrary to most other imaging techniques like PET/CT that frequently require more than 1 hr between the imaging agent injection and data acquisition, ultrasound imaging using targeted microbubbles could be performed a few minutes following intrave-

| Nanobubbles, alternative to microbubbles?
Nano-size UCAs are nanobubbles small enough (diameter typically ranging from 100 to 600 nm) to pass through the tumor vessels into the interstitial tissues. They are composed of a phospholipid shell and a solid, liquid or gas core.

| Targeting ligands
Ultrasound molecular imaging aims to selectively detect and enhance echo signals after specific binding of targeted-contrast agents to their biological targets. Many targeting ligands have been reported in PDAC molecular imaging research (Table 3). These consist in full-length antibodies, antibody fragments, oligonucleotides, peptides, proteins, and small molecule ligands. Such ligands may be applied for designing targeted microbubbles (Figure 3). We discuss here the various targeting ligands explored for PDAC.

| Full-length antibodies
Among targeting ligands, antibodies (molecular weight 150 kDa) have been the most widely adopted for active targeting in cancer imaging. 129 Several preclinical and clinical imaging studies have exploited monoclonal antibodies as targeting ligands. 92,95,[97][98][99][100][109][110][111]114,115,[122][123][124][125]130 There are many advantages in the use of antibodies as molecular imaging probes. First, a wide range of humanized antibodies targeting endothelial molecules are commercially available. Antibodies are bivalent and have the inherent ability to bind to antigens with high affinity and high specificity. However, their utilization is limited due to timeconsuming and expensive development processes. 131 Importantly, their long blood circulation times (ranging from days to weeks) and slow background clearance rate (optimal tumor uptake between 2and 5-days postinjection) further limits the application of full-length antibodies as ideal imaging agents for clinical practice. Considering the relatively short half-life of most microbubbles (<10 min), it is important to select ligands with compatible clearance time to minimize unnecessary side-effects and to achieve high imaging signal in a relatively shorter timeframe. Therefore, lower molecular weight ligands are being preferred for the development of targeted UCAs.

| Antibody fragments
Imaging probes using antibody fragments (e.g., Fab or Fab 0 (50/55 kDa), scFv (26 kDa) and F(ab 0 ) 2 (110 kDa)) or a combination of antibody fragments (e.g., diabodies (55 kDa), minibodies (75 kDa), triabodies (90 kDa), and tetrabodies (120 kDa)) are being frequently used. 132 It has been recognized that blood clearance rate is inversely related to the size of the protein that is, clearance rate of scFV > Fab or Fab 0 > diabody > minibody > triabody > F (ab 0 ) 2 > tetrabody > IgG. 133 For PDAC ultrasound imaging with targeted microbubbles, Bam et al. used a scFv ligand identified and engineered from a yeast-surface-display technique panned against human Thy1 protein. 72 Zou and co-workers expoited a trio of two single chain antibodies and one scFv conjugated to magnetic iron nanoparticles targeted three over- window. To prolong their mean retention time, molecules can be conjugated to polyethylene glycol (PEG) polymers. 134 On the other hand, numerous approaches to genetically engineer multivalent fragments have been pursued. 135 Using scFvs as building blocks, fragments such as scFv polymers and scFv-fusion proteins have been generated. As examples, diabodies and minibodies form stable molecules that have been shown to have a longer blood clearance and improved tumor uptake due to their larger size and multivalence properties. 136 As the smallest protein scaffolds, the preclinical imaging use of nanobodies (12 kDa) and affibodies (7 kDa [98] (Continues)

| Oligonucleotides
Aptamers are a class of short single-stranded oligonucleotides (DNA or RNA) that can bind to cellular targets with high affinity and specificity, and can inhibit protein function. When chemically modified, aptamers show enhanced stability, tissue permeability, and lower immunogenicity compared to antibodies. 139

| Peptides
Small peptides, usually less than 50 amino acids, are advantageous over other classes of targeting ligands due to their good tissue and cell penetration, easy production with lower cost, compared to the expensive production of monoclonal antibodies in hybridoma cell cultures. 141 Peptide ligands have high potential for clinical applications in a wide range of pathologies. 142 Microbubbles have been reported to be capable of bearing a variety of peptide ligands targeted to endothelial cell adhesion molecules F I G U R E 3 Molecularlytargeted microbubble strategies. (a), Native microbubble consists in a gas core and a shell (lipids, polymers or proteins) and (b), can be converted into molecularlytargeted microbubble by functionalization with various targeting ligands (antibody, fragment of antibody, protein, peptide, oligonucleotide, small molecule) (i.e., P-selectin, vascular cell adhesion molecule-1 [VCAM-1]), von Willebrand factor (VWF), and oxidized low-density lipoprotein cholesterol receptor-1 (LOX-1)) for imaging the recruitment of inflammatory cells and/or platelets. 143 In the context of pancreatic cancer imaging, a wide range of biomarkers have been imaged through peptides as targeting ligands. Those targets are proteins mainly located on plasma membrane such as neuropilin-1, 89 EGFR, 116,127 uPAR, 96 and integrin αvβ6, 47,48,104,112,121 and several others have tested theranostic application of peptides in various models. 48,90,112,116 Moreover, the peptide transporter 1 (PEPT1) is also investigated as another relevant biomarker of pancreatic cancer. The Ser-Glu dipeptide was used as a PEPT1 ligand, 144 later functionalized with nanoparticles and evaluated in pancreatic cancer using FI 127 . The probe successfully accumulated in the tumor with a signal intensity about 3.6-fold higher than in the normal tissue. Of high interest, plectin-1 and integrin β4, have been detected through the utilization of a bispecific molecular probe and allowed imaging of pancreatic neoplasms and angiogenesis simultaneously. 101 Such methodologies can greatly increase the targeting efficiency compared to that of either single peptide. Additionally, an in vitro proof-of-principle study performed for plectin-1 receptor imaging using peptide-labeled microbubbles using multiphoton microscopy. 87 Such imaging modality has the ability to detect directly the microbubbles, without the need of contact medium (ultrasound gel) between the transducer and the zone being imaged. The motivation here has been to develop a technology using targeted microbubbles to ensure that all pancreatic cancerous tissues has been removed during resection procedures.
Peptides also have the capacity to target markers located intracellularly or within the extracellular matrix. This property was exploited to visualize fibronectin in pancreatic cancer environment using PET imaging. 117 Preclinical quantitative analysis revealed that the radio- higher tumor uptake than the peptide. 120 Other researchers posted that the mutagenesis of key amino acid residues could optimize the binding ability of MT1-MMP-peptide ligand. 146 In breast cancer mouse model, Ren and co-workers enhanced the fluorescence signal in tumor by 3.2-fold compared to the unmodified peptide.

| Proteins
Recognition and targeting of pancreatic cancer biomarkers can also be done by proteins. Those proteins have the capacity to bind receptors as illustrated by the serine protease urokinase-type plasminogen activator (uPA) and its receptor (uPAR) used in preclinical research for PDAC imaging. 126 Additionally, integrin αvβ3 is a key regulator of adhesion and signaling in numerous biological processes, including tumor cell migration, metastasis, and angiogenesis.
Integrin αvβ3 appears to be functionally coupled with syndecan-1 molecule that regulates its activity during carcinoma cell spreading and migration. Using a syndecan-1 based probe, integrin αvβ3 expression was assessed by optoacoustic tomography in PDAC mouse models. 128  shell, but more frequently a polyethylene glycol spacer arm is added to display the ligands away from the microbubble surface. 152 The most common conjugation strategy is the biotin-(strept)avidin linkage. 74,106,112 Each streptavidin molecule consists of four binding pockets for biotin with noncovalent interaction of dissociation constant in the femtomolar range. However, this approach is restrained to proof of concept in preclinical studies due to the potential immunogenicity associated with (strept)avidin, and the cross reactivity of unconjugated (strept)avidin to endogenous biotin in human body. 153  shell forms a bond with amine groups in protein ligands but is less controllable than maleimide chemistry because of the presence of multiple amine groups in peptides and proteins. 154 Maleimide-thiol conjugation chemistry has proven to be clinically translatable as evidenced by its adoption in many FDA approved antibody-drug conjugates. 155 The maleimide on the extremity of the PEG spacer arm allows covalent conjugation to a thiol group on the targeting agent, (e.g., a terminal cysteine residue). This strategy permits site-specific conjugation of targeting molecules to microbubbles with possibly lower target-ligand steric hindrances.

| PRECLINICAL MODELS FOR PDAC AND PANCREATITIS DIFFERENTIATION BY ULTRASOUND IMAGING
The ability of imaging tools to differentiate between PDAC and pancreatitis tissue is crucial. Numerous rodent models have contributed to the understanding of PDAC pathogenesis and thereby, provided opportunities to characterize and detect disease progression from benign to malignant stages with imaging tools.

| PDAC mouse models
With the aim of producing high fidelity, preclinical models representative of the architectural and functional complexity of human PDAC, various mouse models including carcinogen-induced and genetically engineered murine pancreatic tumor models have been currently employed for preclinical ultrasound imaging researches.  Last mentioned model has been utilized in molecular ultrasound imaging due to the expression of Thy1 or VEGFR2 on its tumor neovasculature. 16,73 Mice lose weight followed by development of ascites, jaundice and then pancreatic tumors with frequent involvement in the duodenum, stomach and/or spleen. The model demonstrates an early and rapid appearance of PanIN lesions and a tumor development time typically three-fold faster than LSL-Kras G12D/+ ; Pdx-1-Cre alone (2 vs. 6 months). 157 GEMMs provide a highly heterogeneous tumor microenvironment and contrary to xenografts, tumors exist in the presence of a competent immune system thus improving the ability to examine therapeutic monitoring to immune-therapies. Moreover, these animal models provide the possibility to follow the disease formation from early stages of PanIN to primary and metastatic tumors (commonly within the abdomen and to the liver, lung, and brain). However, tumor development is slow and colony maintenance is costly.

| Carcinogen-induced PDAC in mice
The administration of carcinogens via a combination of DNAdamaging agents (7, per genome); locally rearranged genomes (>200 structural variants clustered on <3 chromosomes); or unstable genomes (>200 structural variants distributed across the genome). 166 Thus, given the heterogeneous nature of PDAC in patients, use of an universal biomarker may not be realistic, and therefore the potential to perform multiplexing is crucial.
In PDAC, a large number of biomarkers are known to be overexpressed.
However, a limited number of these markers are eligible candidates for vascular targeted imaging. Biomarker expression must be exclusive of chronic pancreatitis, homogeneous through tumor tissue, and have significant upregulation relative to normal and surrounding tissue. The ideal early detection diagnostic strategy will likely use multiple biomarkers with confirmatory real-time molecular imaging.

| Clinical trials using targeted UCAs
Transabdominal ultrasound is the only imaging technology to be at the same time relatively nonexpensive, available, noninvasive and without ionizing radiation. Moreover, some patients cannot undergo imaging using microbubbles, we refer our readers to the publication by Abou-Elkacem et al. 77 Although targeted CEUS exhibits high sensitivity, molecular targets are currently constrained to the vascular lumen as the microbubbles are unable to extravasate from the vessels into the interstitium owing to their larger size. The use of other imaging modalities, like PET/CT, can thus bring additional, rather than competitive, information by using probes able to extravasate the vasculature. Importantly, its ability to discriminate between pancreatitis and PDAC in mouse models should certainly provide opportunities for improved PDAC prognosis.

ACKNOWLEDGMENTS
Part of this work was funded by Phase I and Phase II SBIR Grant R41 CA203090 and R44 CA203090 to Stanford and NuvOx Pharma.