Exendin Peptide Imaging for Insulinoma Detection
Peptide-Based Molecular Imaging
94% sensitivity
68Ga-NOTA-Exendin-4 PET/CT achieved 94.11% sensitivity for insulinoma localization in 47 patients, outperforming 68Ga-DOTATATE and conventional imaging.
Yu et al., European Journal of Nuclear Medicine, 2025
Yu et al., European Journal of Nuclear Medicine, 2025
View as imageInsulinomas are the most common functioning pancreatic neuroendocrine tumors, but they are among the hardest to find. These small, insulin-secreting tumors cause dangerous hypoglycemia, and their treatment is surgical removal. The problem is locating them first. Conventional imaging with CT and MRI misses a meaningful percentage of insulinomas because the tumors are often less than 2 cm in diameter. This is where exendin-4, a 39-amino acid peptide originally isolated from Gila monster venom, enters the picture. When radiolabeled with gallium-68 or fluorine-18 and injected as a PET tracer, exendin-4 binds to glucagon-like peptide-1 receptors (GLP-1R) that are massively overexpressed on insulinoma cells, lighting up tumors that other scans cannot see. For a broader look at how peptides are used in molecular imaging, see the pillar article on RGD peptide imaging and PET.
Key Takeaways
- Exendin-4, a 39-amino acid GLP-1 receptor agonist from Gila monster venom, serves as a highly specific imaging probe for GLP-1R-expressing tumors when radiolabeled with 68Ga or 18F.[1]
- In 47 patients, 68Ga-NOTA-Exendin-4 PET/CT achieved 94.11% sensitivity and 95.74% accuracy for insulinoma localization, outperforming 68Ga-DOTATATE PET/CT (47.06% sensitivity).[2]
- 18F-exendin-4 PET/CT achieved 100% sensitivity in 12 patients, exceeding CT (83%), MRI (63%), and endoscopic ultrasonography (90%).[3]
- A meta-analysis of 179 cases found exendin-4 PET/CT achieved 94% sensitivity and 94% positive predictive value for insulinoma localization.[4]
- Kidney accumulation remains the primary limitation; succinylated gelatin (Gelofusine) reduced renal uptake by 18.1% without affecting pancreatic signal.[5]
- Exendin-4 PET can distinguish insulinomas from nesidioblastosis and detect occult tumors missed by all other imaging modalities.[6]
Why insulinomas need a dedicated imaging approach
Insulinomas present a unique diagnostic challenge. They are small (typically 1-2 cm), well-vascularized, and located within the pancreas, an organ where small lesions are difficult to distinguish from surrounding tissue on CT or MRI. Unlike most neuroendocrine tumors, insulinomas are predominantly benign (over 90%), so the clinical goal is precise localization for curative surgical resection rather than staging of metastatic disease.
The standard neuroendocrine tumor imaging agent, 68Ga-DOTATATE, targets somatostatin receptors (SSTR). While SSTR imaging works well for most neuroendocrine tumors, insulinomas frequently have low somatostatin receptor density. Yu et al. (2025) directly compared the two approaches in 47 patients with biochemically confirmed hyperinsulinemic hypoglycemia. 68Ga-NOTA-Exendin-4 PET/CT detected insulinomas with 94.11% sensitivity and 95.74% accuracy. 68Ga-DOTATATE PET/CT achieved only 47.06% sensitivity in the same patients.[2] This twofold difference in detection rate explains why exendin-based imaging has emerged as the preferred molecular imaging approach for insulinoma localization.
The molecular basis: GLP-1 receptor overexpression
The biological foundation of exendin-4 imaging is GLP-1 receptor overexpression. Benign insulinomas express GLP-1R at densities approximately six times higher than normal pancreatic beta cells. This receptor density differential creates the contrast that makes imaging work: the radiolabeled peptide accumulates preferentially in tumor tissue, producing a hot spot on PET that stands out against the lower background signal from normal pancreas.
Yap and Misuan (2019) reviewed the discovery and biology of exendin-4, tracing it from its isolation from Heloderma suspectum (Gila monster) venom to its clinical applications. Exendin-4 is a full agonist of the GLP-1 receptor with a longer half-life than endogenous GLP-1 due to resistance to dipeptidyl peptidase-4 (DPP-4) enzymatic degradation. Its helical region interacts with the extracellular domain of GLP-1R, and this stable, high-affinity binding is what makes it effective both as a diabetes drug (exenatide/Byetta) and as an imaging probe when conjugated to a chelator and radionuclide.[1]
The specificity of GLP-1R targeting is what distinguishes exendin-4 imaging from somatostatin receptor imaging. While somatostatin receptor subtypes 2 and 5 are expressed at variable levels on insulinomas, GLP-1R expression is consistently high across benign insulinomas. This consistent expression is the reason exendin-4 imaging achieves higher sensitivity than DOTATATE in this specific tumor type, even though DOTATATE remains superior for most other neuroendocrine tumors.
Clinical evidence: head-to-head comparisons
The clinical evidence for exendin-4 insulinoma imaging has matured rapidly, with multiple prospective studies and a meta-analysis establishing its superiority over conventional approaches.
The meta-analysis
Shah et al. (2021) conducted a systematic review and meta-analysis of exendin-4-based imaging for insulinoma localization, pooling data from 179 cases across 16 publications. Exendin-4 PET/CT achieved 94% sensitivity and 94% positive predictive value. Exendin-4 SPECT/CT, the earlier nuclear medicine approach using indium-111 or technetium-99m labels, achieved lower sensitivity at 63% but maintained 94% positive predictive value. The analysis confirmed that PET offers a clear advantage over SPECT for this application, driven by PET's superior spatial resolution and sensitivity.[4]
68Ga-labeled exendin-4
Gallium-68-labeled exendin-4 variants (68Ga-NOTA-Exendin-4 and 68Ga-NODAGA-Exendin-4) are the most widely studied formulations. Yu et al. (2025) provided one of the largest prospective head-to-head comparisons to date: in 47 patients, 68Ga-NOTA-Exendin-4 PET/CT achieved 94.11% sensitivity, significantly outperforming not only 68Ga-DOTATATE (47.06%) but also 18F-FDG PET/CT (29.41%) and conventional imaging with CT and MRI.[2]
Zhang et al. (2025) reported a real-world, single-center experience with 68Ga-exendin-4 PET/CT for insulinoma localization, adding to the growing body of evidence from routine clinical use rather than controlled research settings.[7]
18F-labeled exendin-4
Fluorine-18-labeled exendin-4 offers a practical advantage: 18F has a longer half-life (110 minutes vs. 68 minutes for 68Ga), which allows for centralized production and distribution to imaging centers without on-site cyclotrons or generators.
Murakami et al. (2025) reported a phase 2 clinical trial of [18F]FB(ePEG12)12-exendin-4 PET/CT in 12 patients with biochemically confirmed insulinoma. The 18F tracer achieved 100% sensitivity, outperforming CT (83%), MRI (63%), endoscopic ultrasonography (90%), and selective arterial calcium stimulation (89%). The peptide-based probe showed high tumor-to-background contrast, enabling clear visualization of sub-centimeter lesions.[3]
Sakaki et al. (2025) described a case where [18F]FB(ePEG12)12-exendin-4 PET detected an occult insulinoma that CT, MRI, endoscopic ultrasound, and the selective arterial calcium stimulation test had all failed to identify. The patient underwent curative surgery based solely on the exendin-4 PET finding.[6]
Beyond insulinomas: other GLP-1R imaging applications
Exendin-4 imaging is not limited to insulinomas. The same GLP-1R targeting principle extends to other clinical scenarios where beta-cell visualization is valuable.
Congenital hyperinsulinism
Congenital hyperinsulinism (CHI) in neonates can result from either focal or diffuse beta-cell lesions. The treatment differs dramatically: focal lesions require targeted resection, while diffuse disease may require near-total pancreatectomy. Distinguishing focal from diffuse CHI is therefore critical for surgical planning, and incorrect classification can lead to either unnecessary radical surgery or failed targeted resection. GLP-1R imaging with exendin-4 shows promise for identifying focal lesions in CHI because the affected beta cells overexpress GLP-1R in a pattern similar to insulinomas. The evidence base in neonates is smaller than for adult insulinomas, and the smaller body size introduces additional technical challenges for spatial resolution.
Beta-cell mass quantification
Lindheimer et al. (2023) demonstrated noninvasive in vivo imaging of porcine islet xenografts using [68Ga]Ga-exendin-4, establishing a proof of concept for tracking transplanted beta cells over time. This application could eventually allow monitoring of islet transplant survival without invasive biopsies.[8]
Inflammatory bowel disease
Li et al. (2025) explored exendin-4 imaging based on gastrointestinal GLP-1R targets for inflammatory bowel disease (IBD) diagnosis and treatment efficacy assessment, extending the application of this peptide tracer beyond endocrine tumors into the gastrointestinal inflammatory space.[9]
Technical challenges and solutions
Kidney accumulation
The primary technical limitation of exendin-4 imaging is renal accumulation. The radiolabeled peptide is filtered by the kidneys and reabsorbed in proximal tubular cells, creating high kidney uptake that can interfere with imaging of lesions near the left kidney or upper pole of the kidney, and that limits the radiation dose that can be administered.
Buitinga et al. (2019) tested succinylated gelatin (Gelofusine) as a strategy to reduce kidney accumulation in patients undergoing 111In-exendin-4 SPECT/CT. Gelofusine reduced kidney uptake by 18.1% without affecting pancreatic uptake. In 3 of 10 volunteers, this reduction was sufficient to improve the distinction between pancreatic lesions and kidney background signal.[5]
Kaeppeli et al. (2019) took a different approach, engineering exendin-4 derivatives with albumin-binding moieties. These modified peptides showed decreased renal retention and improved GLP-1 receptor targeting, suggesting that molecular engineering of the peptide itself, rather than co-administration of renal protection agents, may eventually solve the kidney accumulation problem.[10]
Probe optimization
Kondo et al. (2025) systematically compared exendin-4 and single amino acid substitution variants as parent peptides for GLP-1R imaging probes. The Ex-D3 variant (Glu3Asp substitution) maintained GLP-1R binding affinity while offering improved properties for radiolabeling, suggesting that the exendin-4 scaffold can be further optimized for imaging applications.[11]
Where this fits in the peptide imaging landscape
Exendin-4 imaging is one of several peptide-based molecular imaging approaches, each targeting a different receptor system to visualize a specific tumor type. RGD peptides target integrin receptors on tumor vasculature. Somatostatin analogs like DOTATATE target somatostatin receptors on neuroendocrine tumors broadly. Bombesin analogs target gastrin-releasing peptide receptors on prostate and breast cancers.
The principle across all these agents is the same: a peptide with high receptor affinity is conjugated to a chelator (NOTA, NODAGA, DOTA) that holds a radioactive isotope (68Ga, 18F, 111In). The peptide provides targeting specificity; the isotope provides the imaging signal. What makes exendin-4 distinctive within this landscape is its receptor target: GLP-1R is abundantly expressed on beta cells and insulinomas but not on most other tumor types, giving it a narrow but exceptionally high-value clinical niche.
What remains uncertain
Exendin-4 imaging is not yet authorized for routine clinical use in most jurisdictions, limiting its availability to specialized research centers. The evidence base, while strong, consists primarily of single-center studies and one meta-analysis. Multicenter randomized trials comparing exendin-4 PET directly to surgical outcomes are needed to establish its role in standard diagnostic pathways.
The optimal radiolabel remains an open question. 68Ga variants have the most clinical data, but 18F variants offer logistical advantages for wider distribution. Head-to-head comparisons between the two isotope platforms in the same patient populations are limited.
For malignant insulinomas (approximately 5-10% of cases), GLP-1R expression may be lower or absent, potentially reducing sensitivity. The meta-analysis by Shah et al. (2021) noted that most studied cases involved benign insulinomas, and sensitivity in the malignant subgroup requires further characterization.[4]
The kidney accumulation issue, while partially addressable with Gelofusine or albumin-binding modifications, has not been fully solved. This limits both imaging quality for peri-renal lesions and the potential transition from diagnostic imaging to peptide receptor radionuclide therapy (PRRT), which would require much higher radiation doses. Whether engineering solutions like albumin-binding moieties or novel chelator-peptide conjugates can reduce renal retention enough to enable therapeutic applications remains an active area of investigation.
The Bottom Line
Radiolabeled exendin-4 targets GLP-1 receptors overexpressed on insulinomas to detect these tumors with 94-100% sensitivity in clinical studies, substantially outperforming CT, MRI, endoscopic ultrasound, and somatostatin receptor imaging. The approach exploits the same peptide-receptor interaction used in GLP-1 diabetes drugs, repurposed as a molecular imaging probe. Clinical adoption remains limited by availability and regulatory status, but the evidence for its superiority in insulinoma localization is consistent across studies.