15th European Molecular Imaging Meeting
supported by:

To search for a specific ID please enter the hash sign followed by the ID number (e.g. #123).

Cancer Detection

Session chair: Twan Lammers (Aachen, Germany); David Lewis (Glasgow, UK)
Shortcut: PS 21
Date: Friday, 28 August, 2020, 10:00 a.m. - 11:30 a.m.
Session type: Parallel Session


Abstract/Video opens by clicking at the talk title.

10:00 a.m. PS 21-01

Introductory Lecture

Danielle J. Vugts1

1 Amsterdam UMC, Amsterdam, Germany

10:18 a.m. PS 21-02

USPIO-enhanced MRI for lymph node staging after neoadjuvant chemoradiotherapy in esophageal cancer: a single-centre feasibility study

Didi de Gouw1, Bastiaan Klarenbeek1, Marnix Maas2, Atsushi Nakamoto3, Tom Scheenen2, John Hermans2, Camiel Rosman1

1 Radboudumc, department of surgery, Nijmegen, Netherlands
2 Radboudumc, department of radiology and nuclear medicine, Nijmegen, Netherlands
3 Osaka University Graduate School of Medicine, department of radiology, Suita, Japan


Lymph node dissections during esophagectomy may be omitted or minimized in patients with esophageal cancer without or with limited lymph node metastases (LNM), thereby reducing associated morbidity. A promising technique to detect LNM is T2*-weighted MRI after the injection of ultrasmall superparamagnetic iron oxide nanoparticles (USPIO, Ferumoxtran-10)[1]. The aim of this study is to evaluate the feasibility of USPIO-enhanced MRI in the detection of locoregional LNM in patients with esophageal cancer whom underwent neoadjuvant chemoradiotherapy (nCRT).


We performed a prospective, single centre, feasibility study in patients undergoing minimally invasive esophagectomy with suspected LNM. USPIO-nanoparticles (2.6 mg Fe/kg body weight) were infused intravenously, followed by MRI after 24 to 36 hours. All patients underwent a USPIO-enhanced MRI examination with multiple breathholds before nCRT (3T Magnetom PrismaFit, Siemens Healthcare). After nCRT (~3 months), USPIO-enhanced MRI was repeated on the day of surgery with the patient under controlled prolonged apnea of four minutes in a hybrid operation room with 3T MRI (Magnetom Skyra, Siemens Healthcare). The MRI images from both pre-nCRT and post-nCRT examinations were analyzed independently by two experienced radiologists using diagnostic guidelines proposed by Anzai et al[2].


15 patients were enrolled into the study. Five patients developed distant metastases after nCRT, and did therefore not undergo a second USPIO-enhanced MRI scan or surgery. In 15 patients, some 435 lymph nodes were evaluated by the first radiologist with a mean short axis length of 5 mm (range 2-18) and 233 lymph nodes by the second radiologist with a mean short axis length of 6 mm (range 2-25) (figure 1). 193 lymph nodes could be matched and were evaluated by both radiologists. 75% of the matched nodes received the identical scores by both radiologists, with corresponding kappa value for inter-observer agreement of 0.57 (P = 0.044). In 10 patients USPIO-enhanced MRI was performed before and after nCRT and 72 nodes were analyzed by both radiologists. From these matched nodes, 5 (observer 1) and 3 (observer 2) out of 22 suspicious nodes before nCRT received a non-suspicious score after nCRT (figure 2).


MRI with controlled mechanical ventilation was safe and feasible and USPIO uptake in regional lymph nodes was seen on T2*-weighted sequences in all patients. Radiological evaluation of lymph nodes on USPIO-enhanced MRI was performed with high interobserver agreement. This technique might be of added value in the indication for nCRT, in monitoring the response of nCRT, and in modifying the surgical treatment plan.

[1] Fortuin, A.S., et al., Ultra-small superparamagnetic iron oxides for metastatic lymph node detection: back on the block. Wiley Interdiscip Rev Nanomed Nanobiotechnol, 2018. 10(1).
[2] Anzai, Y., et al., Evaluation of neck and body metastases to nodes with ferumoxtran 10-enhanced MR imaging: phase III safety and efficacy study. Radiology, 2003. 228(3): p. 777-88.
Figure 1. Example of USPIO-enhanced MRI scans before and after nCRT

Example of USPIO-enhanced MRI scans before and after nCRT in the axial (left) and coronal plane (right) one day after the administration of Ferumoxtran-10. The T1-weighted in-phase Dixon images (a+c) are used to identify lymph nodes. In the water-selective iron sensitive T2* weighted MRI (b+d, root-mean-square addition of 5 gradient echoes) the signal of healthy lymph nodes disappears, whereas suspicious lymph nodes, in which the USPIO-particles do not accumulate, retain MRI signal. The blue circles point out one suspicious lymph node without USPIO contrast.

Figure 2. Difference in USPIO score before and after nCRT evaluated by two radiologists.
A horizontal line indicates the same result before and after nCRT. Lymph nodes with a full or partial high signal intensity were considered suspicious (score 1-4) and lymph nodes with overall dark intensity or lymph nodes with fatty hilum were considered not suspicious for metastases (score 5-7). From these matched nodes, 5 (observer 1) and 3 (observer 2) out of 22 suspicious nodes before nCRT received a non-suspicious score after nCRT, and 66 (observer 1) and 67 (observer 2) out of the 72 matched lymph nodes had the same USPIO score before and after nCRT.
Keywords: Esophageal carcinoma, MRI, lymph node staging, neoadjuvant therapy, USPIO
10:30 a.m. PS 21-03

68Ga-NODAGA-exendin-4 PET/CT for the localization of insulinomas

Marti Boss1, Kirsi Mikkola2, Maarten Brom1, Annemarie Eek1, Mijke Buitinga1, Olof Eriksson3, Damian Wild4, Vikas Prasad5, Adrienne Brouwers6, Francois Pattou7, Hans Hofland8, Pirjo Nuutila2, Martin Gotthardt1

1 Radboud University Medical Center, Nuclear Medicine, Nijmegen, Netherlands
2 University of Turku, Turku PET Center, Turku, Finland
3 Uppsala University, Nuclear Medicine, Uppsala, Sweden
4 University of Basel, Nuclear Medicine, Basel, Switzerland
5 Charite University Hospital, Nuclear Medicine, Berlin, Germany
6 University Medical Center Groningen, Nuclear Medicine, Groningen, Netherlands
7 University Hospital Lille, General Surgery and Endocrine Surgery, Lille, France
8 Erasmus Medical Center, Internal Medicine, Rotterdam, Netherlands


Insulinomas are usually small, benign pancreatic neuroendocrine tumors. Precise anatomical localization is crucial for surgical treatment. The current standard imaging techniques CT, MRI and somatostatin receptor (SSTR) PET have limited sensitivity. The glucagon like peptide-1 (GLP-1) analog exendin specifically binds to the GLP-1 receptor, which is overexpressed in most insulinomas. We have performed a prospective multicenter imaging study to compare the effectiveness of [68Ga]Ga-NODAGA-exendin-4 with all current standard non-invasive imaging procedures for the localization of insulinomas.


42 adults aged 24-62 with biochemically proven hyperinsulinemic hypoglycemia were included. PET/CT images were obtained one and two hours after injection of 95-105 MBq 68Ga-NODAGA-exendin-4 (5-7 µg). Current standard imaging, consisting of CT or MRI and SSTR PET, was performed within 8 weeks of 68Ga-NODAGA-exendin-PET in all patients. A patient-based analysis was performed with histopathology as a reference standard.


Lesions were identified in 33 patients. 31 patients underwent surgery and presence of an insulinoma was confirmed histopathologically. Analysis showed that [68Ga]Ga-NODAGA-exendin-4 PET localized insulinomas with a higher accuracy and sensitivity (90.6% and 93.5% respectively) than conventional imaging (78.1% and 80.6% respectively) and SSTR PET (59.4% and 61.3% respectively). In 12.5% of patients, a correct diagnosis and decision to perform surgery was only reached after [68Ga]Ga-NODAGA-exendin-4 PET. This novel technique therefore significantly influenced the clinical management of the patients in this population. The median size of the lesions identified by GLP-1R PET was 12 (10 – 18) mm, including 5 lesions smaller than 10 mm, showing the excellent sensitivity of the technique. The lower peptide dose used in this study compared to previous studies with [68Ga]Ga-DOTA-exendin-4 (4-7 µg vs. 12-24 µg respectively) resulted in fewer occurrences of nausea (5% vs. 27% of patients).


This study demonstrates the superior performance of [68Ga]Ga-NODAGA-exendin-4 PET/CT compared to current standard non-invasive imaging modalities for pre-operative localization of benign insulinomas. Because of its high sensitivity and excellent imaging quality, [68Ga]Ga-NODAGA-exendin-4 PET/CT could have the potential to become the primary diagnostic imaging modality in patients with AHH.

AcknowledgmentThis work is supported by BetaCure (FP7/2014–2018, grant agreement 602812)
PET images
GLP-1R PET/CT and SSTR PET/CT images of patient 4 (A) and patient 9 (B). Location of tumors is indicated with green arrows.
Keywords: Insulinoma, Exendin, PET imaging, GLP-1 receptor
10:42 a.m. PS 21-04

89Zr-DFO-Dinutuximab-Beta to assess differential expression of surface GD2 in neuroblastoma.

Stephen Turnock1, Chiara Da Pieve1, Graham Smith1, Gabriela Kramer-Marek1

1 The Institute of Cancer Research, SUTTON, United Kingdom


GD2 is a highly restricted surface ganglioside that is a key target for immunotherapy in high-risk neuroblastoma (NB) patients. Over 95% of primary NBs express GD2, yet target expression levels are heterogeneous both inter- and intra-tumourally [1]. Disease relapse has been associated with a positive, yet low GD2 expression in tumour biopsies. Currently, immunotherapy is given regardless of the degree of positivity in tumour GD2 expression. We hypothesise that patient stratification using PET imaging and quantification of whole-tumour GD2 expression will improve treatment planning and outcome.


GD2 expression was characterised by flow cytometry (FC) and confocal microscopy in Kelly, SK-N-BE(2)C and SK-N-AS NB cells using  GD2-PE antibody and an in-house prepared Dinutuximab-Beta-IR700 conjugate (GD2-IR700). For PET studies, DFO was attached to Dinutuximab-Beta and the immunoconjugate was subsequently radiolabelled with zirconium-89 (89Zr-DFO-GD2). The specificity of 89Zr-DFO-GD2 binding was performed on the panel of cell lines (4oC, 1 hr). Mice bearing SK-N-AS (GD2high) and SK-N-BE(2)C (GD2low) tumours were injected with 89Zr-DFO-GD2 (10 µg, 1.8 MBq) and imaged daily up to 96 hr post-injection. Biodistribution studies, autoradiography (AR), FC and immunofluorescence (IF) analyses of the resected tumours were performed and correlated with PET imaging data.


FC and confocal analysis showed that SK-N-AS cells express a high level of GD2, whereas SK-N-BE(2)C expression is relatively low. Kelly cells had the greatest GD2 expression out of the three analysed cell lines, but distribution of GD2 was heterogeneous. Therefore, the more homogenous Kelly-GD2high population was successfully isolated and used for further studies. The 89Zr-DFO-GD2 radioconjugate was prepared and isolated with a 99% radiochemical purity, and a specific activity of 0.19-0.25 MBq/µg. Cell uptake of 89Zr-DFO-GD2 correlated to the GD2 expression levels measured via FC, and was successfully blocked with a 50-fold excess of non-labelled antibody. Analysis of PET data (24-96 hr post-injection) confirmed tumour specific uptake of 89Zr-DFO-GD2 and clear differences in GD2 expression between SK-N-AS and SK-N-BE(2)C tumours. This was corroborated by ex vivo AR and IF of tumour sections. GD2-IR700 FC on fresh tumour samples correlated with 89Zr-DFO-GD2 binding characteristics.


Labelled GD2 antibody can determine surface ganglioside expression levels in vitro and in vivo. 89Zr–DFO-GD2 accurately delineates between low and high GD2 expression in tumour models of NB as early as 48 hours post-injection. Further characterisation using other mouse models and patient samples could provide insight in to the benefits and limitations of 89Zr-DFO-GD2 PET imaging to guide GD2 immunotherapy.

[1] T. Terzic et al., Expression of Disialoganglioside (GD2) in Neuroblastic Tumors. Pediatric and Developmental Pathology. vol. 21, 355-362, 2017.
Keywords: Immunotherapy, PET, In Vivo, Neuroblastoma
10:54 a.m. PS 21-05

In vivo head to head comparison of [89Zr]Zr-DFO-NCS-trastuzumab and [89Zr]Zr-DFO*-NCS-trastuzumab in a breast bone model of metastasis

Marion Chomet1, Maxime Schreurs1, Mariska Verlaan1, Wissam Beaino1, Guus A. M. S. van Dongen1, Danielle J. Vugts1

1 Amsterdam UMC, VU University, Radiology and Nuclear Medicine, Radionuclide center, Amsterdam, Netherlands


Labeling of monoclonal antibodies (mAbs) with 89Zr is done using the clinically approved chelator desferrioxamine (DFO). However, this hexadentate chelator shows some instability of the complex, likely due to incomplete coordination, resulting in unwanted bone uptake of 89Zr. This can lead to false-positive observations in the detection of bone metastasis. Therefore, DFO*, an octadentate analogue was synthesized and this chelator demonstrated improved stability with the anti-HER2-mAb trastuzumab [1-2]. We aim to confirm the superiority of DFO* over DFO in a breast bone model of metastasis.


A bone metastasis model was developed by injecting 1.5x10e6 luciferase transfected BT-474 (HER2+) cells in left tibias of nu/nu mice while PBS was injected as negative control in the other tibia. Disease progression was followed by bioluminescence and CT until tracer injection. Trastuzumab and the non-binding control mAb B12 were labeled with 89Zr using either DFO-pPhe-NCS or DFO*-pPhe-NCS and 100 µg (2-3 MBq) of each construct were injected IV in 6 mice per group. PET imaging (n=4/group) was performed at 24, 72 and 144h p.i. followed by tissue distribution (TD). To assess uptake “free 89Zr” in tumor and nontumor involved bones, 0.5 MBq of either [89Zr]-oxalate, [89Zr]-citrate or [89Zr]-chloride was injected in 10 mice (n=3-4/group) followed by PET imaging and tissue distribution 24h p.i.


89Zr uptake with the nonbinding mAb in tumor involved tibiae was significantly lower for [89Zr]Zr-DFO*-B12 than [89Zr]Zr-DFO-B12: 1.7±0.3 and 7.1±2.3%ID/g respectively. In unaffected tibiae this was 1.4±0.3 and 5.8±1.3 %ID/g respectively. High uptake in case of 89Zr in case of [89Zr]Zr-DFO-B12 might be due to nonspecific uptake of released 89Zr, as confirmed by PET imaging and TD of “free 89Zr” (Fig.1). [89Zr]Zr-oxalate, citrate and chloride showed similar uptake especially in shoulders, spinal cord and tibiae with nonsignificant difference between tumor involved and noninvolved tibiae. In mice injected with [89Zr]Zr -DFO*-trastuzumab, 89Zr uptake was much higher in tumor involved tibiae than in noninvolved tibiae (7.3±5.6 and 1.5±0.3%ID/g respectively). In contrast, [89Zr]Zr -DFO-trastuzumab was not able to discriminate metastatic involvement with 13.2±6.0 %ID/g in metastases involved tibiae and 4.4±1.3%ID/g in non-involved tibiae. PET imaging confirmed those results (Fig.2).


This study indicates the release of 89Zr and its nonspecific uptake in bone containing organs in case DFO is used for 89Zr-immuno-PET. This hampers specific detection of breast cancer bone metastases with [89Zr]Zr-DFO-trastuzumab. DFO* seems better capable for stable binding of 89Zr. As a result [89Zr]Zr-DFO*-trastuzumab appeared much better qualified for specific detection of bone metastases than [89Zr]Zr-DFO-trastuzumab.


This project has received funding from the European Union’s Horizon 2020 research and innovation programme under the Marie Sklodowska–Curie grant agreement No 675417

[1] Patra M, Bauman A, Mari C, Fischer CA, Blacque O, Haussinger D,Gasser G, Mindt TM, “An octadentate bifunctional chelating agent for the development of stable zirconium-89 based molecular imaging probes,” Chem. Commun., vol. 50, no. 78, pp. 11523–11525, 2014.
[2] Vugts DJ, Klaver C, Sewing C, Poot AJ, Adamzek K, Huegli S, Mari C, Visser GWM, Valverde IE, Gasser G, Mindt TM, van Dongen GAMS, “Comparison of the octadentate bifunctional chelator DFO*-pPhe-NCS and the clinically used hexadentate bifunctional chelator DFO-pPhe-NCS for 89Zr-immuno-PET,” Eur. J. Nucl. Med. Mol. Imaging, pp. 1–10, Aug. 2016.

Biodistribution of [89Zr]Zr-DFO-NCS-trastuzumab, [89Zr]Zr-DFO*-NCS-trastuzumab, [89Zr]Zr-DFO-NCS-B12 and [89Zr]Zr-DFO*-NCS-B12 in collected bones 144 h p.i. 144 p.i. of 100 µg per construct (n=6 mice per group).


PET images of mice injected with 110 µg of either [89Zr]Zr-DFO-NCS-trastuzumab (A), [89Zr]Zr-DFO*-NCS-trastuzumab (B) , [89Zr]Zr-DFO-NCS-B12 (C)  or [89Zr]Zr-DFO*-NCS-B12 (D) and scanned 144 h p.i. All mice had received an intratibial injection of BT-474 cells in the left leg and PBS in the right leg.

Keywords: Immuno-PET, 89Zr, DFO*, bone model
11:06 a.m. PS 21-06

Multispectral endoscopy for early detection of dysplasia in Barrett’s oesophagus (MuSE): a first-in-human pilot study

Dale J. Waterhouse1, 2, Massimiliano di Pietro3, Wladyslaw Januszewicz3, Tara Evans3, Rebecca C. Fitzgerald3, Sarah E. Bohndiek1, 2

1 University of Cambridge, Department of Physics, Cambridge, United Kingdom
2 University of Cambridge, Cancer Research UK Cambridge Institute, Cambridge, United Kingdom
3 University of Cambridge, Hutchison MRC Research Centre, Cambridge, United Kingdom


Barrett’s oesophagus is an acquired condition that predisposes patients to the development of oesophageal adenocarcinoma through intermediate stages of dysplasia. Early detection of dysplasia allows curative endoscopic therapy, but current standard of care surveillance achieves only around 40% sensitivity for dysplasia. Multispectral imaging (MSI) allows simultaneous collection of morphological (spatial) and biochemical (spectral) information from tissue, which we hypothesised could help to more effectively delineate disease.


To test our hypothesis, we constructed a clinically translatable multispectral endoscope (MuSE) using an imaging fibre bundle (PolyScope) that can be introduced through the accessory channel of a standard gastroscope to collect spectral information in vivo. The device (Fig 1A) includes both a spectrometer, to capture high spectral resolution information from a single point, and a spectrally resolved detector array (SRDA), to capture spatially-resolved spectral information across 9 spectral filters.
Subjects (n=20) due to undergo clinically indicated endoscopy with a previous diagnosis of Barrett’s were enrolled in a pilot clinical study (approved by research ethics committee, NCT03388047). MSI was performed by introducing the MuSE through the working channel of the SOC endoscope.


In each patient, two regions of interest, one with a suspicious lesion and one with Barrett’s oesophagus were identified in standard-of-care (SOC) endoscopy, marked, and then interrogated using the MuSE before finally being biopsied to provide gold standard histopathological diagnosis (Fig 1B).
Study recruitment has now completed and preliminary analysis of the available data yields apparent differences between spectra from different disease classes across multiple patients (Figs 2A-C). Interestingly, dysplasia and cancer seem to be intermediate in spectral profile between squamous and Barrett’s. Most of the observed contrast appears to come from the diffuse reflection spectrum of haemoglobin and distinct haemoglobin absorption is apparent at ~540 and ~570 nm.


Preliminary analysis suggests promising spectral contrast between dysplastic and non-dysplastic regions. With further development of both the hardware, potentially using miniaturised SRDAs on tip, and analysis algorithms, potentially with neural networks for classification, MSI may have the potential to provide high contrast for dysplasia and cancer in surveillance of Barrett’s oesophagus, thus providing a much-needed improvement to the SOC.

AcknowledgmentThis work was funded by CRUK (C14303/A17197, C9545/A29580, C47594/A16267, C47594/A21102, C55962/A24669), EPSRC (EP/N014588/1, EP/R003599/1) and the EU FP7 agreement FP7-PEOPLE-2013-CIG-630729.
Figure 1. Multispectral endoscopy (MuSE): trial overview

A. A clinically translatable multispectral endoscope (MuSE) constructed around an imaging fibre bundle (PolyScope) that can be introduced through the accessory channel of a standard gastroscope to collect spectral information in vivo. Multispectral imaging is achieved using a spectrally resolved detector array (SRDA) with 9 spectral filters. B. Schematic of the trial protocol.

Figure 2. Spectra measured in the MuSE trial
Spectra measured using the multispectral endoscope in the MuSE trial. A. Patient 11. B. Patient 12. C. Patient 15. Shaded areas represent the standard deviation of spectra collected within each patient.
Keywords: endoscopy, Barrett's oesophagus, early detection, multispectral, clinical trials