Final Report Summary - MITIGATE (Closed-loop Molecular Environment for Minimally Invasive Treatment of Patients with metastatic Gastrointestinal Stromal Tumours)
Executive Summary:
Gastrointestinal stromal tumour (GIST) is a rare disease with high potential for metastasising that frequently affects young patients and results in a short life expectancy. Currently there is only one class of effective medication and often the tumours develop drug resistance after a few years. No molecular imaging technologies indicating drug resistance, therapeutic response or disease progression are clinically available. Furthermore, effective targeted agents such as endoradiopharmaceuticals are not designed for GIST and alternative minimally-invasive treatment options are not well explored. The overall objective of the MITIGATE project was to develop and validate an integrated closed-loop process to effectively treat metastatic GIST patients resistant to the currently available class of medication.
MITIGATE’s objectives were:
1) Optimisation of biopsy and tissue preparation in GIST for analysis and subtyping of molecular markers.
2) Provision of a tumour bank of TKI-responsive and -resistant GIST
3) Application of molecular imaging probes indicating resistance, early therapeutic response or progression.
4) Development of endoradiopharmaceuticals for treatment in patients.
5) Combination of the therapeutic approach with alternative, minimally-invasive treatment options.
6) Non-invasive monitoring of treatment effects in clinical routine.
After 4 years the highly complex research lead to a variety of project results that will strongly impact GIST patients and improve their personalised diagnosis and treatment. Our established GIST-biopsy process involves a novel endoscopic biopsy device, biological upstream processing systems, a bespoke immunomagnetic single cell separation procedure and innovative mass spectrometry approaches. Through the development of the GIST subtype database and GIST subtype tumour animal models, combined with the possibility to predict TKI-resistant GIST behaviours, personalised selection of an appropriate PET probe or endoradiotherapeutic tumour treatment was achieved. A SOP for radiolabelling of peptides with 68Gallium using a kit approach that is successfully applied to the NeoBOMB1 peptide was prepared. The establishment of 68Ga-NeoBOMB1 PET-CT for molecular phenotyping in GIST tumours now enables to develop a novel diagnostic approach for GIST patients. The therapeutic efficacy of the 177Lu-NeoBOMB1 was confirmed in pre-clinical GIST models, making it a new potential treatment option for GIST patients. A dosimetry study using an already validated 68Ga-NeoBOMB1 PBPK model indicated a prediction of the compound’s behaviour for individualised targeted therapy approaches. A first in human clinical trial with 68Ga-NeoBOMB1 focussing on safety and pharmacokinetics was successfully completed. Initial results indicate an excellent safety profile of 68Ga-NeoBOMB1 in patients and targeting visualisation by PET/CT of GIST tumours was achieved. Novel, PET-CT-based target volume definition for external beam radiotherapy based on 68Ga-NeoBOMB1 were established. The high-risk areas within the tumour volume can be used for an intensified treatment. The evaluation of preliminary animal PET-CT data for EBRT therapy planning indicate that the new PET-tracer allows escalation of the dose to vital tumour cells, thereby further increasing tumour control probability. A multimodal protocol for treatment monitoring shows that several functional imaging modalities can assess the treatment response to imatinib. A new multimodal coil (23Na/1H) for assessing anatomical and metabolic information at clinical level was validated and provides a new technology based on 23Na imaging for GIST treatment in clinical routine.
The close interdisciplinary collaboration led to an integrated, closed-loop treatment platform that induces a paradigm shift for individually tailored therapy of patients with TKI-resistant disease. MITIGATE findings may also be transferable to other cancer types (e.g. lung cancer) for even higher socio-economic impact.
Project Context and Objectives:
Gastrointestinal stromal tumour (GIST) is a rare disease with an annual incidence of about 1.5 /100,000 persons. GIST frequently affects young patients with high potential for metastasising, which often results in a short life expectancy of less than 3 years on average. Currently there is only one class of effective medications for systemic GIST therapy and often the tumours develop drug resistance after a few years. No molecular imaging technologies indicating drug resistance, early therapeutic response or disease progression are clinically available. Furthermore, effective targeted agents with other mechanisms of action such as endoradiopharmaceuticals are not designed for treatment of GIST and alternative minimally-invasive treatment options for metastatic disease are not well explored nor has the clinical value of their combined use been established. This leaves GIST patients with very limited treatment options.
The overall objective of the MITIGATE project is to develop and validate an integrated closed-loop process to effectively treat metastatic GIST patients resistant to the currently available class of medication, the tyrosine kinase inhibitors (TKI).
This envisaged personalised treatment concept combines innovative strategies for biopsy, inline tissue analysis, molecular tumour characterisation, theranostics by imaging technologies (PET and MRI) and companion radiopharmaceuticals followed by the assessment of biodistribution, dose calculation and measurement of therapeutic effectiveness
MITIGATE`s ultimate goal is to develop new protocols and guidelines to effectively diagnose and treat patients with metastatic GIST resistant to current treatment. The project will develop new targeted personalised treatment for GIST metastases which can be then used as a role model for other molecularly characterised cancer subtypes no-longer responsive to targeted therapy.
The overall objectives of MITGATE are to
1) Optimise biopsy and tissue preparation in suspected GIST for integrated analysis and subtyping of molecular markers.
2) Provide a large and broadly characterised tumour bank of TKI-responsive and –resistant GIST for generating novel pre-clinical models envisioned to reveal novel markers for TKI-resistance.
3) Apply molecular imaging probes for specific GIST features indicating drug resistance, early therapeutic response or progression.
4) Develop effective targeted agents with other mechanisms of therapeutic action such as endoradiopharmaceuticals for treatment in patients.
5) Combine this therapeutic approach with alternative, minimally-invasive treatment options for a personalised treatment in the metastatic situation with fewer side effects.
6) Non-invasively monitor treatment effects in clinical routine.
Project Results:
Biopsy and Preparation for Molecular Analysis
Novel Endoscopic Biopsy System
A novel endoscopic biopsy system was developed. First, a demonstrator of a biopsy needle taking samples for molecular analysis through a flexible endoscope was designed. Requirements for cutting and transport of the sample were specified to safeguard the sample quality. Several cutting mechanisms were designed and evaluated. A cutting mechanism based on a sharp blade at the tip of the insertion tube was chosen as the appropriate mechanism in combination with vacuum transportation. Several transport methods were conceptualised and evaluated and an experimental set up was designed to test the simplest method of applying of vacuum. Hence it was concluded that a continuously applied vacuum in combination with a uniformly cut sample presented a valid method for biopsy transport. The optimised entire endoscopic biopsy system, as shown in Figure 1 was validated by showing that the cutting mechanism and vacuum transport work as a functional unit in a phantom model, shown in Figure 2. As model, an endoscopy simulator with an artificial stomach was chosen. The resulting pictures and videos show that the developed system is completely functional.
Biopsy Needle interface
This task aimed to connect the biopsy system with the subsequent tissue disruption module. The tissue disruption module is based on grinding of the biopsy between rows of teeth. Therefore, no standard container had to be used as an interface and the grinder was directly connected to the tubing of the biopsy system.
Tumour Cell Isolation
The aim of this task was to isolate specific tumour cells from a single cell suspension obtained from a GIST tumour tissue biopsy. A first objective of this task was the development of a tissue disruption device which ensures the disruption of tissue into single cells without destroying the material. After researching many different cell dissociation methods and the performance of several experiments, a grinding device inspired by a commercially available herb grinder was developed. The mechanism is based on the movement of rows of teeth grinding against each other. Experiments were mainly performed with chicken liver, of which single cells could be obtained. Due to the limited availability of tumour material (GIST tumours are rare and frozen tissue was not suitable) only few experiments were performed with fresh GIST samples. Since consistency of GIST can vary a lot depending on the tumour site, inserts varying in design and material were manufactured to adapt the grinder to every consistency of the material. Pre-treatment with enzymes might be necessary for certain GIST consistencies. The grinder was directly connected with the biopsy system and integrated into an automated process of single cell generation. The description of work intended to design a device with different stations for mechanic and enzymatic treatment of the tissue. The automated grinder setup allows for the addition of buffer and tissue dissociation enzyme, thus mechanical and enzymatic disruption can happen simultaneously.
Isolation of tumour cells with magnetic bead technology using specific antibodies turned out to be challenging. The commercially available DOG1 antibodies, often used in immunohistochemical staining, were not suitable for positive GIST cell isolation as they exclusively bind to intracellular domains of the surface protein. Initially, three GIST cell culture lines (TKI resistant and non-resistant) were assessed phenotypically for their defined GIST tumour markers. An identification of GIST cell surface markers allowed for the design of a selection cocktail which would theoretically immunomagnetically deplete all non-GIST cells in a single cell suspension thereby enriching for GIST cells. In a mock tumour design, initial experiments combined peripheral blood mononuclear cells (PBMCs) endothelial, epithelial and mesenchymal cells with GIST cells from cell lines in specific ratios. Non-GIST cells where immunomagnetically depleted, enriching for GIST cells with purity in excess of 98% and an average recovery of 65%. To provide a proof of principle, an alternative cell line was used. The results show that cells can not only be distracted but also be distinguished by their magnetic charge in a microfluidic device. The technology is applicable for every kind of tumour cells, and as soon as a suitable antibody will be available for the positive isolation of GIST cells experiments will be adapted and performed. A negative selection of GIST cells is conceivable as well, by simply running several positive isolations of non-desired cells in a row. However, this will probably result in a loss of a large fraction of material. Due to limited availability of GIST tissue not many experiments could be performed with patient-derived tumour material. A total of eight experiments were performed with patient-derived tumour tissue. The quality of the extracted tissue was extremely variable with respect to cell viability, organ/tissue of origin and expression levels of known GIST surface phenotype markers (CD34, CD117 and CD90). Much of this variability in viability is likely due to patient treatment regimens with TKIs, such as imatinib, which are known to induce large-scale apoptosis in tumour cells. Furthermore, reduction in CD34 levels after imatinib treatment is known and found to be related to cystic degeneration. Some of the variability in surface marker expression is due to patient’s GIST phenotype prior to treatment and the tissue of origin. CD34 and CD117 are expressed on average in 60-70% and 85-95%, respectively but these expression levels vary dramatically depending on GIST tissue location. Tissue with poor initial viability resulted in very few viable cells remaining following tissue dissociation and GIST cell isolation. It was found that tissue excised from patients undergoing treatment was very dense and rubbery in consistency, with high collagen contents resulting in difficult tissue dissociation. Conversely, GIST tissue excised from stomach or the peritoneal cavity was much easier to dissociate and resulted in higher viability. On average, the GIST cell isolation cocktail developed for the immunomagnetic separation of GIST from non-GIST cells resulted in a complete removal of all CD45 positive cells from the single cell suspensions and a 3 to 6 fold enrichment of GIST cells depending on the quality and initial phenotype of the GIST tissue.
GIST Substype Classification and GIST Subtype Database
Potential relevant masses for a robust classification of different tumour samples were identified. After a successful setup of a suitable data acquisition workflow based on MALDI MS spectrometry measurements, a standard procedure for pre-processing the measurement data and a graphical user interface for MATLAB tools was developed. Because of the “fresh GIST sample” bottleneck studies were based on three distinct GIST cell lines, namely GIST-T1, GIST-882 and GIST-430, with different response rates towards the tyrosine kinase inhibitor imatinib. We used Fisher Score Estimation methods to identify the relevant features for a following classification model which could possibly also be used for sorted cells obtained.
The developed methods are not limited to the processing and analysis of GIST, but can also be applied to other cancer cells and possibly to those derived from other solid tumours like mamma- or prostate carcinoma. To validate the model, in addition to Hierarchical Cluster Analysis (HCA), blinded data was classified by using Principle Component Analysis (PCA) and k-nearest neighbours (kNN) algorithms. In this way, a cross validation was performed. Since the real assignments between GIST type and the correspondent spectra are known, the performance of the blinded experiments can be expressed by calculation of corresponding sensitivity and specificity. In a final step, we applied the complete processing pipeline to an independent set of breast cancer cell line data to demonstrate the transferability of this approach. The suitability of determined fingerprints to predict the response upon TKI was examined. Since human tissue samples of GIST are rare, a different model was used for initial verification of the predictive power given by the identified MS features. Xenograft tissue from highly immune-deficient NSG mice that were engrafted with GIST-T1, GIST-882 and GIST-430 cells was used. The NSG mouse model lacks mature T cells, B cells as well as natural killer cells and holds deficiencies in multiple cytokine signalling pathways. Therefore, it has proven to be well suited for GIST xenografts development. The predictive power of compressed MS fingerprints derived from GIST cell lines was successfully determined on a model closer to GIST biopsy material. Compressed Data of TKI-resistant and TKI-responsive NSG xenografts was separable with 100% accuracy by Principle Component and Hierarchical Cluster Analysis.For future applications and an iterative improvement of the classification quality it's essential to create a knowledge base containing all historical measurement and data analysis. This database must not only store the corresponding device measurements but needs also to register all relevant properties of the biological sample and potential parameters influencing the measurement result. In this way correlations between features in the high dimensional feature space can be found and incorporated to optimise the classification result. A suitable database system for managing MS fingerprinting data was designed, acquired and installed. With further data input an iterative improvement of classification quality is expected.
Summary of significant results
▪ Novel endoscopic biopsy system
▪ Novel tissue dissociation device directly coupled with biopsy system
▪ Proof of principle of specific tumour isolation via magnetic distraction
▪ GIST subtype classification system
▪ GIST subtype database (based on cell lines)
Target Analysis and Molecular Probe Design
Analysing biopsies of GIST with respect to expression of a series of potential targets and identification of candidate targeting vectors
A characterisation of Imatinib resistant GIST animal models by providing spatially resolved molecular mass spectrometry (MS) data was performed. Based on lipid fingerprints, differentiating mass features between resistant and susceptive animals were identified to confirm previously observed drug effects. In conjunction with MS results obtained from cell culture experiments discrepancies between the molecular profiles of tumour cells cultivated in laboratory flasks and mice were observed. However, the established methodological workflow for MS data acquisition was proven to be stable and reproducible for a transfer to monitor human tissue samples. MS imaging was used to detect potential radiodiagnostics, but was proven incapable of monitoring small dosages in tissues. For that reason, MS image triplicates of 31 human patient GIST resectates were recorded and molecular lipid signatures were linked to histopathological and immunohistochemistry data. For that purpose, a sophisticated sample logistics, measurement and data processing workflow was established. This allowed for the MS imaging analysis of high quality cryo-preserved GIST tissue at HM. As the instrumental equipment highly improved during the project period a faster MS analysis with high spatial and mass resolution for highly specific molecular information was enabled. In this regard staining workflows to determine the presence/absence of new marker candidates were developed. A supervised classification pipeline was successfully established based on an efficient “big-data” analytics framework to identify molecular signatures characterizing specific targets. Also the identification of potential candidate molecules by high resolution FTICR-MS was pursued. Proof of concept for the characterisation of lipid fingerprints by means of MS imaging have been accomplished to evaluate the pharmacological properties of potential targeting vectors in humans.
Establishment of an animal model of imatinib resistant GIST
This task focused on the establishment of a mouse model of imatinib-resistant GIST using highly immunocompromised mice. To allow a personalised approach to cancer therapy, patient-derived xenografts (PDX) based on the transfer of primary tumours directly from the patient into immunodeficient mice represent a suitable model. Tumours can be engrafted heterotopically (subcutaneously at the flanks of mouse) or orthotopically (implantation into specific organs). To find the best model for GIST implants, several mouse models with different levels of immunosuppression were developed. NSG mice were reported to be the most appropriate immunodeficient host animal for direct xenografting of fresh tumour tissues. In total 30 GIST tissue samples from Germany and Italy were collected. Pieces of tumour tissue were subcutaneously and orthotopically (renal capsule and peritoneum) implanted into NSG mice. To evaluate tumour engraftment, growth and dissemination animals were monitored by digital palpation and MRI analysis. After 6-9 months, no tumour growth was detected and the autoptical examinations confirmed the absence of either tumour masses and/or metastases. Parts of GIST fresh samples were subjected to mechanical and enzymatic dissociation to obtain primary cells, which, after molecular and genetic characterisation, have been inoculated subcutaneously or orthotopically into spleen and liver. Cell viability was higher than 50% and KIT expression was variable among tumours. Primary GIST cells inoculated into NSG mice were monitored for tumour growth by MRI analysis for histological injection and by digital palpation for subcutaneous inoculation. However, no tumour growth was discovered after 3 months. To understand engraftment failure of in vivo experiments, primary cells isolated from GIST were characterised at different culture passages. We performed real time PCR to evaluate c-KIT and DOG1 expression in primary tumours and in cells at passages 1, 2 and 6. We observed that KIT and DOG1 expressions progressively decreased with culture passages and lastly disappeared. Accordingly, direct DNA sequencing analysis revealed the loss of the original KIT mutations with culture passages indicating the proliferation of fibroblasts instead of tumour patient isolated cells. Although we have performed 230 implants using 140 NSG mice, no tumour mass has been observed in any of the injection sites (orthotopical, subcutaneous or peritoneal). Several reasons could explain the difficulties we and others faced in generating GIST PDX models. First, GIST is a low grade tumour and most likely hard to engraft and grow by xenotransplant. In fact, a limited number of successful studies displaying GIST-PDX model has been published. Second, GIST PDX models are not available in public repositories such as the European PDX consortium and US National Cancer Institute repository. As an alternative strategy, we decided to develop a new approach based on the use of Imatinib-sensitive and -resistant GIST cell lines (GIST-T1, GIST-882 and GIST-430) to generate GIST animal models suitable to investigate biological mechanisms related to tumour resistance to individuate new molecular markers and to design effective targeted treatments.
We performed orthotopic, subcutaneous and intravenous injection into NSG mice. Subcutaneous inoculation developed localised non-invasive tumours. Orthotopic injection, in particular intrahepatic inoculation, revealed diffuse tumour masses and numerous metastases with GIST-T1 and GIST-43 cell lines. On the contrary, inoculation of the Imatinib sensitive cells GIST-882 led to tumour masses with slower growth rate without formation of metastases. No tumour masses were detected by i.v. injection of all GIST cell lines. In conclusion, we established that in vivo GIST cell lines growth is dependent on the site of injection and that the most suitable protocol to generate metastatic GIST mouse models by mimicking the tumour-stromal microenvironment is to inject GIST-T1 (Imatinib-sensitive) and GIST430 (Imatinib-resistant) cell lines orthotopically in the liver of NSG mice.
Chemical development of new radiopharmaceuticals for molecular imaging of resistant GIST with a focus on PET-Tracers and potential endoradiopharmaceuticals and biological evaluation of the potential new contrast agents/radiotracers
DOTA-NT20.3 was successfully synthesised in high purity (>98%) via solid phase peptide synthesis using the Fmoc-strategy. DOTA was chosen as chelating agent due to the easy introduction of several radionuclides (e.g. 68Ga, 177Lu) allowing a theranostic approach. After establishing the radiolabelling protocol with 68Ga, DOTA-NT20.3 and the commercially available Neurotensin-1 receptor ligand DOTA-NT(8-13) were evaluated in vitro in an internalisation assay using three GIST cell lines, as well as HT-29 cells as a positive control. Both NT-1 ligands did not show any internalisation or membrane binding on GIST cell lines (HT-29 tested positive), indicating an insufficient number of NT-1 receptors on the cell surface of GIST cells. To image angiogenesis in GIST tumours, the VEGF analogue peptide HPLW was selected as targeting ligand. The prosthetic group N-succinimidyl-4-[18F]fluorobenzoate ([18F]SFB) was linked to the targeting vector through its terminal lysine residue. The radiolabelling of [18F]SFB was established and optimised on a AiO (Trasis) synthesis module in a three-step procedure with final SPE purification. This way, [18F]SFB could be synthesised in a radiochemical yield (n.d.c.) of 44±6% and a radiochemical purity of 98±2% within 45 minutes. [18F]SFB could be successfully linked to the targeting vector HPLW within 10 minutes and after semipreparative HPLC purification the radiotracer could be obtained in a radiochemical yield of 1.54±0.04% and a radiochemical purity of 94±6% after 100 min of synthesis in total. Following appropriate formulation, the radiotracer [18F]SFB-HPLW proved to be suitable for in vivo evaluation.
The Glucagon like receptor 2 (GLP-2R) has been found to be expressed in 68% of gastrointestinal stromal tumours (GIST). Consequently, GLP-2R could be considered a potential target for the visualisation of GIST by both optical imaging and PET. Several GLP-2R ligands have been synthesised via solid phase peptide synthesis, comprising 4 DOTA-derivatised Teduglutide analogues (24-34 amino acids) for radiolabelling, as well as short peptides coupled with fluorescent dyes (carboxyfluorescein, Cy3) and AAZTA. The synthesis of the non-truncated Teduglutide derivatives was performed via fragment condensation to obtain the 34 amino acid sequence in a high purity of >95%. The radiolabelling protocol using 68Ga was established for the DOTA-derivatised Teduglutide analogues ([68Ga]DOTA-Lys10)-Teduglutide, ([68Ga]DOTA-His1)-Teduglutide, ([68Ga]DOTA-Lys30)-Teduglutide and the truncated ([68Ga]DOTA-Lys34)-(Trp10, Ala12)-GLP-2(10-33) with high 68Ga-incorporation of >94% for all compounds with varying reaction times depending on the presence of oxidation-sensitive methionine in the amino acid sequence. The truncated radiotracer showed the most promising results in vitro in an internalisation assay and flow cytometric analyses of all three GIST cell lines revealed comparable mean fluorescence intensities as the positive control cell line A549. In vivo evaluation of the truncated derivative in tumour-bearing mice (GIST-T1 and GIST882) showed a favourable renal excretion, but a slow blood clearance as well as a low uptake into the tumour.
Real-Time PCR (RT-PCR) was performed to evaluate GLP-2R expression in GIST cell lines (430, -T1 and 882) in addition to MDA-MB-231 and HeLa cell lines, indicated by the human transcriptome database (http://ist.medisapiens.com/) as negative and positive control cell lines, respectively. Imatinib-resistant GIST430 show the highest expression, about 20 times higher than the positive control, whereas GIST-T1 and GIST882 show intermediate expression (16.2 and 46.7% expression, respectively). Therefore, the generation of a peptide-targeting GLP-2R is crucial for improving GIST diagnosis, detection and seeking for alternative therapy. In order to correctly design a peptide able to interact with GLP-2R, homology modelling of GLP-2 receptor was firstly performed and the GLP-2 sequence of 20 amino acids was derived and modelled on GLP-2R. Consequently, a first batch of GLP-2 derivative was synthesised and conjugated with carboxyfluorescein for optical imaging and chelated with AAZTA for PET/MRI applications. However, due to the high impurity of both derivatives, a new synthetic strategy was exploited. GLP-2 conjugated with Cy3 fluorescent dye was obtained with 90% of purity, whereas GLP-2 AAZTA reached a purity of 80%. GLP-2 Cy3 demonstrated its ability to bind on cells expressing GLP-2R by flow cytometry. However, the competition assay with a competitor (unlabelled GLP-2) was unable to provide information as a control. For this purpose, the synthesis of a scramble peptide labelled to a fluorescent dye for improving the control assays was done. Cy5 dye was chosen to avoid background fluorescence coming from tissue for subsequent in vivo applications. Moreover, both GLP-2 targeting and scramble peptides were synthesised adding a PEGylate linker to improve binding affinity to the target on cell lines. GLP-2 and Scramble-Cy5 were then in vitro tested on GIST cell lines for evaluating their binding affinity. Confocal microscopy images showed that GLP2-Cy5 appeared to localise on the cell membrane and in the cytosol region of GIST430 cells. In particular, the signal diffuses from the membrane to the cytosol. By contrast, the scramble peptide shows a weak fluorescence signal. Further in vivo studies will validate the binding affinity of GLP2-Cy5 in GIST tumour murine models.
The radiolabelling protocol for the commercially available polyclonal DOG1 antibody using 124I was established and the radioiodinated antibody was evaluated in vitro in a binding assay. The control antibodies (e.g. 124I-Cetuximab, 124I-Rituximab) for radiolabelling and in vitro evaluation showed binding to the respective control cell lines, but nut to GIST cell lines. The polyclonal 124I-DOG1 antibody, however, showed no binding to GIST cell lines and was therefore further evaluated. FACS analysis revealed that the native polyclonal DOG1 antibody is not an appropriate agent for GIST imaging.
After the successful 6-step and 7-step organic syntheses of the nitropyridine precursor and the fluoro reference compound, the radiolabelling of [18F]fluoronorimatinib (FNI) has been established. This two-step reaction consisting of radiofluorination and deprotection yielded the radiotracer in a good radiochemical yield (RCY) of 20±3%. After ensuring high stability in buffered solution (pH 7.4) as well as human and murine serum, the in vitro evaluation of [18F]FNI was performed. Cell permeability proved to be good despite the increased hydrophilicity, but competitive binding assays with viable cells gave no satisfactory result due to a high background accumulation. Therefore, the binding affinity of [18F]FNI towards 10 different KIT mutants has been evaluated in a kinase assay confirming a similar binding profile as Norimatinib and thus a distinction between Imatinib sensitive and resistant KIT mutants in vitro. The in vivo evaluation was performed in Balb/c SCID mice which were inoculated with GIST-T1, GIST882 and GIST430 tumours. Despite the promising in vitro results, the radiotracer was not able to distinguish between sensitive and resistant tumours in vivo due to the vastly different vascularisation. The tumour-to-background ratios were in the range of similar radiotracers (SKI696), however, the pharmacokinetic behaviour could be improved by the reduction of lipophilicity leading to an enhanced, more favourable renal excretion. The co-administration of ketoconazole led to an increased metabolic stability of the tracer as well as an increased tumour uptake in first experiments resulting in a better demarcation of tumours in PET imaging.
Several precursor strategies for the small molecule radiotracer [18F]fluoro-DOG1 have been pursued and the respective precursor molecules have been synthesised including nitro, triazene and boronic acid ester leaving groups. Thereof, the latter was successfully radiolabelled under copper-catalysis and obtained in a good RCY of 33±11%. Although [18F]fluoro-DOG1 proved to be stable in buffered solution (pH 7.4) and human serum, a fast degradation was observed in murine serum. Consequently, PET imaging of GIST430 tumour-bearing mice showed only marginal uptake in the tumour and strong accumulation in the liver and the gastrointestinal tract.
Calculation and optimisation of biodistribution and dosimetry for diagnostic purposes
PET biokinetic data from experiments in 5 tumour-bearing Balb/c SCID mice injected with [18F]FNI was used to develop a mouse [18F]FNI PBPK model. The PBPK model bases on anatomical and physiological information obtained from the mouse experiments and from literature. In the model three kinds of parameters were used: fixed, Bayes (based on a known distribution), and fitted (without previous information). Data from the experiments (total mouse weight and organ weights) were used as fixed parameters. Data from the literature (organ vascular fractions, organ interstitial fractions, organ permeability surface area products and organ plasma flows) were used as Bayes parameters. Unknown parameters (metabolic rates, internalisation rates, tumour plasma flow and tumour permeability surface are product) were fitted. The developed mouse [18F]FNI PBPK model individually considers the most relevant organs and the organs with high accumulation (liver, kidneys, spleen, intestines, pancreas, tumour, muscle, bones, brain, lungs and bladder). The other organs and tissues were grouped together. The developed model was well supported by the data in terms of visual inspection, coefficients of variation of the fitted parameters and correlation among the model parameters. In the [18F]FNI PBPK model is considered both the whole [18F]FNI molecule and its metabolites. Additionally, a human PBPK model for [18F]FNI was created based on the fitted mouse [18F]FNI PBPK model by scaling the physiological and anatomical parameters from mice to humans using the allometric equation. Volumes and flows were scaled in a different way while some rates such as the internalisation rate were kept equal. This model will be further used to perform dosimetric estimations in humans. Dosimetric estimations in humans using the scaled human [18F]FNI PBPK model were performed and compared with traditional dosimetric scaling methods.
Summary of significant results
▪ MALDI MSI was applied to detect fluoronorimatinib and NeoBOMB1 in tissue; however, drug concentrations were too low for complete distribution studies
▪ MALDI MS imaging study of 31 human GIST samples was performed and is still ongoing. Supervised classifications pipeline was established based on an efficient “big data” analytics framework to identify molecular signatures
▪ GRPR staining of cryosectioned tissues was established and will be coupled with MALDI-MSI
▪ Initial results indicate observable separation between GRPR-positive and –negative MS spectra based on lipidomic foot print in tissues
▪ proof of concept study was conducted to detect gene mutations in KIT and PDGFRA based on minute disturbances in the lipidomic foot print with MALDI-ToF MSI
▪ establishment of immunodeficient xenograft GIST animal models in NSG and Balb/c SCID mice with GIST-T1, GIST882 and GIST430
▪ patient-derived xenografts could not be established in NSG mice
▪ subcutaneous injection of GIST cell lines in NSG mice induced localised tumours without metastazation
▪ orthotopic injection of GIST cell lines in NSG mice induced tumour growth with all cell lines and metastases with GIST-T1 and GIST430 within 7 weeks
▪ intravenous injection of GIST cell lines in NSG mice did not result in tumour growth
▪ synthesis of several precursors for potential radiotracers targeting KIT, NTSR1, GLP-2R, VEGFR and DOG1
▪ syntheses of GLP-2 derivatives labelled with fluorescent dyes for optical imaging
▪ establishment of radiolabelling protocols for 68Ga (DOTA) and 124I
▪ establishment of radiolabelling protocol for [18F]SFB-HPLW
▪ establishment of radiolabelling protocol for [18F]fluoronorimatinib and [18F]fluoro-DOG1
▪ establishment of in vitro assays to determine target affinities and specific binding
▪ [18F]fluoronorimatinib showed similar binding profile as Norimatinib in a kinase assay with 10 KIT mutants (Imatinib sensitive and resistant)
▪ flow cytometry analyses showed presence of GLP-2R on the cell surface of all GIST cell lines
▪ RT-PCR showed highest GLP-2R expression in GIST430; intermediate expression levels were found for GIST-T1 and GIST882
▪ GLP2-Cy5 showed specific binding to GLP-2R positive GIST430 cells in an internalization assay
▪ Fluorescence microscopy showed localization of GLP2-Cy5 on the cell membrane and in the cytosol of GIST430
▪ in vivo evaluation of several radiotracers targeting KIT, GLP-2R and DOG1
▪ ([68Ga]DOTA-Lys34)-(Trp10, Ala12)-GLP-2 (10-33) showed high renal excretion, but slow blood clearance and low tumour uptake
▪ [18F]fluoronorimatinib showed reduced liver uptake and enhanced renal excretion compared to similar radiotracers and similar tumour uptake, but distinction between Imatinib-resistant and –sensitive tumours was not possible
▪ [18F]fluoro-DOG1 showed unfavourable biodistribution in mice due to the lack of stability in murine blood serum
▪ general PBPK model (peptides) for mice and humans was developed
▪ mouse PBPK model for [18F]fluoronorimatinib (small molecule) was developed and fitted to data from 5 xenograft bearing mice
▪ human PBPK model for [18F]fluoronorimatinib was created by allometrically scaling the physiological and anatomical parameters from mice to humans
Preclinical Theranostics
Development of a standard procedure for the radiolabelling of peptides with Ga-68
A standard procedure for 68Ga radiolabelling of peptides with high affinity for somatostatin (SST) and gastrin-releasing peptide (GRP) receptors was developed. In particular, a kit approach for direct radiolabelling with 68Ga acid solution coming directly from the 68Ge-68Ga generator was set up and validated initially for the DOTA-TATE peptide (targeting SST receptors). The same kit approach was then successfully applied also to the novel bombesin analogue NeoBOMB1 (targeting the GRP receptors). This procedure avoids the laborious post-processing of the eluate and should represent a considerable improvement in the labelling of DOTA-peptide with 68Ga both in terms of synthesis time, as well as for the quality of the radiopharmaceuticals (specific activity, radionuclidic and radiochemical purities). By testing different conditions of reaction (pH of complexation, volume of buffer, type of buffer, amount of precursor and final purification procedure), a standard procedure for 68Ga -labelling was set up, validated and applied in GMP conditions.
The main steps of the standard operating procedure for 68Ga -labelling of the NeoBOMB1 peptide are:
• Elution of 68Ga from the 68Ge/68Ga Generator into the vial containing the lyophilised peptide and selected excipients;
• Addition of the reaction buffer solution (provided in the kit), to reach the optimal pH for the coordination of the 68Ga with the DOTA-chelator;
• Reaction for 7 minutes at 95°C;
• Quality control and radioactivity measurement of the 68Ga-NeoBOMB1.
The radiopharmaceutical 68Ga-NeoBOMB1 obtained following this procedure was then characterised and tested both in vitro and in vivo in relevant animal models bearing the GIST tumour. The first GMP batch of the kit for radiopharmaceutical preparation of 68Ga -NeoBOMB1 was manufactured in early 2016, and was used exclusively for assessing the long-term stability of the product at the intended storage conditions. Subsequent GMP-grade batches were manufactured to guarantee supply of the product for clinical use.
In vitro and in vivo evaluation of specificity of synthesised radiopharmaceuticals in GIST animal models
Using relevant GIST tumour models the in vitro and in vivo characterisation of the 68Ga -labelled DOTA-peptides synthesised as described in Task 5.1 was performed. Binding and internalisation studies were conducted using three GIST cell lines (GIST-T1, GIST-882 and GIST-430) characterised by different degrees of sensitivity to the tyrosine kinase inhibitor imatinib. Several peptides (targeting different receptors, all known to be overexpressed in GIST) were initially screened in vitro, however very low or no specific binding to GIST tumour cells was found for all 68Ga-DOTA labelled peptides except for bombesin derivatives. The NeoBOMB1 peptide was therefore characterised further and confirmed to specifically bind the GRPR, shown to be expressed at similar levels in all three GIST cell lines. In vitro displacement assays were also performed, resulting in high affinity of the NeoBOMB1 peptide to the GRPR; additionally, no differences in terms of peptide-receptor binding potency were observed among the three models.
The internalisation properties of 68Ga-NeoBOMB1 were also studied in vitro and the well-known bombesin agonist analogue AMBA was compared with the NeoBOMB1. A clear difference in binding behaviour was found between the two bombesin analogues, reflecting the agonist and antagonist properties, respectively. In the case of 68Ga-AMBA activity was mainly internalised, whereas for 68Ga-NeoBOMB1 membrane bound activity was predominant, confirming the antagonistic nature of the latter bombesin analogue.
Biodistribution properties of 68Ga-NeoBOMB1 were assessed in vivo, intravenously injecting the tracer in healthy mice and following its distribution in main organs over time. The radiolabelled NeoBOMB1 was rapidly cleared from the blood, showing no retention in kidneys. Background radioactivity was observed in GRPR-expressing tissues (mostly pancreas), which however decreased over time, consistently with a GRPR antagonist profile. The ability of the 68Ga-radiolabelled NeoBOMB1 peptide to display the GRPR expressing GIST tumour has been confirmed in in vivo imaging PET studies. The whole-body picture of a representative GIST-882 bearing animal showed good imaging performances, allowing a clear visualisation of the tumour. Uptake in the liver and intestinal area was also observed, due to the partial hepatobiliary excretion of the peptide and to a specific uptake in these areas. Finally, as for many other peptides of the same class/type, renal excretion is responsible for the high uptake in the bladder area. Based on the excellent in vitro and in vivo results obtained with 68Ga-NeoBOMB1 in GRPR-expressing GIST models and on the availability of the GMP-grade pharmaceutical product (kit for the radiopharmaceutical preparation of 68Ga-NeoBOMB1), 68Ga-NeoBOMB1 was selected as the optimal candidate for clinical translation. In the last period of the MITIGATE project, efforts have been focused on the assessment of the 177Lu-labelled NeoBOMB1. Both biodistribution and efficacy experiments were conducted in mice bearing GIST tumour, which confirmed the therapeutic potential of this radiopharmaceutical in GRPR-expressing GIST patients.
Database of mass spectrometry signatures as indicators as acute markers of response to radiotherapy in GIST animal models
The work of HM in WP5 focused on the application of MALDI mass spectrometry (MS) signatures to investigate the response to novel radiotherapeutics in preclinical GIST animal models. Acquisition methods for protein and metabolite MS fingerprints were successfully established and provided in form of standard operating procedures (SOPs). Due to the successful acquisition of new MALDI TOF mass spectrometers during the period of the MITIGATE project which allow for higher acquisition speed and a superior mass resolution, it was possible to measure spatially resolved molecular data using MS Imaging. A preclinical NeoBOMB1 study in a GIST animal model was carried out to assess MS signatures as acute markers of response. Remarkably, all mice responded to the radiotherapy with 177Lu-NeoBOMB1. Therefore, no tumour tissue of treated animals was available for analysis. Consequently, the database to explore therapeutic efficacy of radiotherapy in preclinical studies correlating treatment regimes, MS fingerprints and outcomes was established but not implemented due to the fact that fingerprints of responders and non-responders could not be compared. Instead, MALDI MS imaging was used to analyse potential adverse effects in other tissues such as pancreas. Additionally the spatial distribution of the MITIGATE radiodiagnostics Fluoronorimatinib and NeoBOMB1 by means of MALDI MS imaging was analysed. Due to the low doses applied, detection of the drugs masses was not possible. Instead, spatially resolved molecular MS data was acquired for the first line treatment Imatinib in human GIST samples. Finally, the MALDI MS imaging approach for analysis of response signatures is being translated to resectates from GIST patients.
Modelling, simulation and prediction of human biodistribution and absorbed doses
PET biokinetic data from experiments in healthy Balb/c SCID mice injected with 68Ga-NeoBOMB1 were used to perform dosimetric estimations for humans. For this, the time integrated activity coefficients (TIACs) obtained from the mouse biokinetics were scaled-up to humans based on the mass of the organs and the total body mass, according to the AAPM Report No. 71. The scaled TIACs were then used as an input in OLINDA/EXM software, a program with which the dosimetric calculations were performed. The results of the dosimetric calculations for 68Ga-NeoBOMB1 were comparable to dosimetric estimations reported for other bombesin antagonist substances labelled with 68Ga. The biokinetic data from the experiments in healthy mice were also used to develop a mouse 68Ga-NeoBOMB1 PBPK model. The PBPK model is based on anatomical and physiological information obtained from the mouse experiments and from the literature. In the model three kinds of parameters were used: fixed, Bayes (based on a known distribution), and fitted (without previous information). Data from the experiments (total mouse weight and organ weights) were used as fixed parameters. Data from the literature (organ vascular fractions, organ interstitial fractions, organ permeability surface area products and organ plasma flows) were used as Bayes parameters. Unknown parameters (metabolic rates, internalisation rates and receptor densities) were fitted. The developed mouse 68Ga-NeoBOMB1 PBPK model individually considers the most relevant organs and the organs with high accumulation (liver, kidneys, spleen, intestines, pancreas, muscle, bones, brain, lungs and bladder). The other organs and tissues were grouped together. The model was well supported by the biokinetic data gathered from the experiments with 68Ga-NeoBOMB1 in healthy mice.
Moreover, a human 68Ga-NeoBOMB1 PBPK model was created by scaling-up the mouse 68Ga-NeoBOMB1 PBPK model using allometry. Some mouse parameters were used in the translated human PBPK model such as receptor densities and transfer rates (i.e. internalisation, metabolic rate, etc.). Volumes and flows were considered human specific parameters and were taken from the literature for an average healthy man of 71 kg. Permeability surface area product values per organ mass were scaled from mice to humans using allometry based on the organ mass.
In addition, experiments in tumour-bearing mice intravenously injected with 68Ga-NeoBOMB1 were performed and biokinetic data were obtained. Pancreas biokinetic data from these experiments (not considered in the healthy mice 68Ga-NeoBOMB1 PBPK model) were successfully retrieved using an automated biokinetic-based clustering segmentation method combined with previous knowledge of the kinetics and anatomy. The segmentation process was performed using PMOD v3.8 software. A special strategy had to be developed and applied for the pancreas (limiting organ at risk) since identifying the pancreas from mouse PET/CT images is a complex task due to its location, variable form and similarity to the surrounding organs in a CT or PET image. The developed mouse 68Ga-NeoBOMB1 PBPK model was adapted and fitted to the data from tumour-bearing mice (including biokinetic data of the pancreas) and scaled-up to humans using allometry.
Summary of significant results
▪ The standard radiolabelling procedure based on the kit approach was set-up and validated for 68Ga-DOTA-TATE by exploiting AAA Patent
▪ The standard procedure for 68Ga-labelling developed by AAA with the DOTATATE peptide was successfully applied to the novel GRPR antagonist, NeoBOMB1.
▪ In vitro experiments demonstrated the high affinity of 68Ga-NeoBOMB1 to the GRPR expressed by GIST cell lines, as well its low internalisation properties, which is consistent with the antagonist nature of this peptide.
▪ Suitable pharmacokinetics and imaging properties by the 68Ga-NeoBOMB1 were confirmed in GIST animal models.
▪ Confirmation of the therapeutic potential of the 177Lu-NeoBOMB1 in GRPR-expressing, GIST-bearing animal models.
▪ A database structure for the therapeutic effect of radiopharmaceuticals in non-tumour tissues correlating treatment regime, MS fingerprints and outcomes is being developed.
▪ Dosimetric calculations for humans were performed based on healthy mouse biokinetic data
▪ A mouse 68Ga-NeoBOMB1 PBPK model was developed and fitted to experimental data from healthy mice
▪ A scaled human 68Ga-NeoBOMB1 PBPK model was created
▪ Mouse pancreas biokinetics was assessed by using a developed automated biokinetic-based clustering segmentation approach.
Translational Studies
Planning of study, initiating study, monitoring of study, data evaluation
In September 2015 an Interim Meeting of the involved MITIGATE partners was initiated to decide on the selection of the compound to be translated into the clinic for molecular imaging in patients with advanced GIST tumours. In the meeting all radiopharmaceuticals developed so far or in development were reviewed and there was a uniform consensus that by far the best candidate for clinical translation was 68Ga-NeoBOMB1 with excellent GIST targeting properties in vitro and in vivo and the basis of pharmaceutical development by one of the involved partners.
The aim of the clinical trial with 68Ga-NeoBOMB1 is to assess the safety, tolerability and tumour targeting properties of 68Ga-NeoBOMB1 in patients with GIST, as a combined phase I/IIa first-in-human study. The study was designed to minimise patient burden caused by the assessment procedures. Only patients with advanced GIST – meaning metastatic disease in the liver or other intestinal organs – with progressive disease (i.e. failure of 1st, 2nd or 3rd-line tyrosine-kinase-inhibitor [TKI] therapy) were included. This population is also the most likely to benefit from future development of therapeutic concepts revolving around NeoBOMB1. The final study design was aimed to keep the radiation dose for patients as low as possible to achieve the study aims. Two subgroups of patients were defined: For the first six patients, safety and tolerability was the main objective, with measurements of physiological parameters, and repeated PET-scans to obtain data on the biodistribution of the radiopharmaceutical as basis for the absorbed dose (exposure) in the patients. The second phase – involving another 5 to 6 patients – was focussed on the targeting properties of 68Ga-NeoBOMB1 in GIST tumours to further demonstrate the benefit of 68Ga-NeoBOMB1 as a novel diagnostic tool. Besides the clinical investigation plan, also a number of other documents was prepared for the submission to both the Ethical Committee and the Austrian drug authority. This included patient informed consent forms, Investigator Brochure for the physician conducting the trial, standardised operating procedures for various procedures such as radiolabelling of the drug, plans for monitoring of the study and many more. Furthermore, insurance coverage was obtained for all participants.
The clinical trial MITIGATE-NeoBOMB1 “A Phase I/IIa study to evaluate safety, biodistribution, dosimetry and preliminary diagnostic performance of 68Ga-NeoBOMB1 in patients with advanced TKI-treated GIST using positron-emission tomography/computer tomography (PET/CT)” was registered in the EU database for clinical trials (EudraCT 2016-002053-38). Submission of the Clinical Investigation Plan and other relevant files to the Ethics Committee was performed in May 2016 and the final positive ethics vote was issued on the 5th of August 2016. In parallel, in July 2016, the official submission was sent to the competent authority, the Austrian Federal Office for safety in healthcare (BASG)/Agency for Food Safety (AGES). End of November 2016, the final approval of the authority was finally granted.
Preclinical studies for translation of radiopharmaceutical development into the clinic
The developed kit formulation and lyophilisation process for the industrial manufacturing of a GMP kit for the radiopharmaceutical preparation of 68Ga-NeoBOMB1 was used. The kit is composed of 2 vials, one containing the active substance plus excipients, to be reconstituted with 68Ga eluted from a 68Ge/68Ga generator and one vial containing the buffer solution to be added to achieve radiolabelling. The radiolabelling procedure was qualified and the formulation was successfully tested. All preclinical data of 68Ga-NeoBOMB1 required for conducting a clinical trial were developed.
First in human clinical trial
The Clinical Trial was formally initiated in December 2016 and the first patient was successfully recruited in January 2017. In the initial part of the study dynamic PET scans in three patients with advanced GIST – most of them resistant to TKI treatment – were included in the study. All study related investigations could be performed as planned. In particular safety monitoring parameters were included, whereby no severe adverse events due to the administration of 68Ga-NeoBOMB1 were found. Physiological accumulation of 68Ga-NeoBOMB1 was observed most strongly in the pancreas. Rapid renal clearance was observed. Tumour enhancement increased over the time course of the study giving high contrast images at later time points. Pharmacokinetics` investigations revealed a high metabolic stability of 68Ga-NeoBOMB1. A report detailing these initial results was again evaluated by the Ethical Committee of The Medical University Innsbruck. The Ethical Committee had no objections to continue the clinical trial. Based on this vote, patient recruitment went on and three more patients could be included until September 2017 confirming the findings in the first three patients.
All these first data indicate a very good safety profile of 68Ga-NeoBOMB1 showing excellent image quality in PET. The first subgroup of patients was completed in September 2017 and the study is currently moving into its second phase (Phase IIa) to provide additional data on specific receptor targeting in GIST patients. Confirmation of GRP receptor status on histological samples is ongoing, required for inclusion of patients in this Phase. Study completion is planned in the first quarter of 2018. Overall the study is expected to provide the basis for further development of the promising novel imaging agent for the benefit of GIST patients in an advanced stage of disease, but also to open other applications in oncological patients.
Proof-of-concept for multimodal treatment
The potential inclusion of patients for a minimally-invasive therapeutic approach, based on loco-regional stereotactic radiofrequency ablation of tumour lesions of patients with advanced GIST was attempted in parallel with the clinical trial. This approach could offer patients with localised, yet advanced disease a therapeutic option in the form of radiofrequency ablation based on multimodality imaging approach. So far none of the patients was clinically suited to benefit from such a loco-regional therapy. This multidisciplinary approach is continuing to select patients for this novel multi-modality treatment approach in GIST.
PBPK Optimisation of Individual Patient Biokinetics and Dosimetry
Biokinetic data from animal experiments in mice were used to develop a PBPK (Physiologically based pharmacokinetics) mouse model. Based on this a scaled human [68Ga]NeoBOMB1 PBPK model was created taking into account anatomical and physiological parameters from the literature.
High accumulation was seen in tumour and pancreas, as expected, whereas liver and kidneys did not present considerable accumulation. The scaled human [68Ga]NeoBOMB1 PBPK model is currently being validated and, if necessary, adapted using biokinetic data retrieved from PET/CT images from 5-6 patients injected with [68Ga]NeoBOMB1. The validated human PBPK model will be used to optimise the therapeutic index for each patient for potential translation of the imaging data towards a therapeutic option for patients, in which NeoBOMB1 labelled with a therapeutic radionuclide such as Lutetium-177 will be used. To optimise the therapeutic index, different scenarios (varying peptide amount, activity and administration schedule) will be simulated in silico using the PBPK model. The results of these simulations will be used to perform dosimetric calculations to define conditions for the highest therapeutic index in patients.
Summary of significant results
▪ 68Ga-NeoBOMB1 was selected for clinical translation and molecular imaging in GIST tumour patients and a pharmaceutical formulation for the radiopharmaceutical developed.
▪ A Clinical Trial Application for a first in human clinical trial with 68Ga-NeoBOMB1 was prepared according to EU regulations and approved by ethics committee and competent authority according to EU regulations in 2016.
▪ A clinical trial with this novel molecular imaging approach for GIST patients was initialised end of 2016 and the first patient was enrolled in January 2017.
▪ The first part of the clinical trial focussing on safety and pharmacokinetics was successfully completed. Initial results indicate an excellent safety profile of 68Ga-NeoBOMB1 in patients.
▪ Tumour visualisation by PET/CT of GIST tumours was achieved in these patients.
▪ Data for potential proof-of-concept of multimodal treatment were generated
▪ Biokinetic data for PBPK modelling and dosimetry were obtained, dosimetric predictions for [68Ga]NeoBOMB1 based on mouse and patient data were performed.
Minimally Invasive Therapy
Protocol and treatments design
The objective of this task was to design a concept to investigate different combinations of treatment and have the following protocol approved: patients with GIST receive multiple radiologic examinations to detect and follow up the disease status. In case progressive disease is seen on conventional imaging techniques (CT, MRI, 18FDG PET/CT), patients will receive a PET-CT examination with the new tracer NeoBOMB1. Unclear findings will be re-evaluated using MRI imaging and if necessary, an image guided biopsy (CT or ultrasound) will be performed. This approach helps to avoid under- and over-treatment. Patients will be treated by current state of the art endo-radiotherapy or image guided minimally invasive treatment options.
Therapy planning
Tumour volume definition based on new tracer: The main objective of this task is to develop a strategy and the tools to integrate the new PET-CT tracer NeoBOMB1 into the combination of treatments. Therefore preliminary animal PET CT was used to evaluate tumour volume definition based on the new tracer NeoBOMB1. An additional objective is the development of an assisting device for minimally invasive treatments. The purpose of the assisting technology is to support physicians performing minimally invasive treatments in planning, execution, and monitoring of the intervention. The assisting manipulator consists of a robotic unit with modifications tailored to the interventional task. Preliminary animal data suggest that the new PET tracer NeoBOMB1 enables us to escalate the dose to vital tumour cells even more. We expect increased tumour control and less damage to risk structures. Improved tumour volume definition can be used to adapt target volume for other treatments like minimally invasive ablations or radioactive seed implantation.
Evaluation of the manipulator system: The purpose of the assisting technology is to support physicians performing minimally invasive treatments in planning, execution, and monitoring of the intervention. To meet the requirements of a complex minimally invasive treatment the following concept was developed.
▪ A software tool allowing the user to combine and view medical imaging data and to select target points and entry points for the needle-based treatment.
▪ A manipulator or robot to provide positioning and orientation of a needle guide. The KUKA LBR iiwa MED is selected for its design purpose: a component for medical products.
▪ A registration method to determine the robot position and orientation in image coordinates.
▪ A communication protocol to send target and entry points from the planning software tool to the robot.
The evaluation of the manipulator systems shows promising results in terms of precision and intervention time. These characteristics have the potential to render it a useful tool in minimally invasive treatments.
Summary of significant results
▪ Positive vote of ethics committee for minimally invasive treatment of GIST metastases based on Ga-NeoBOMB1 PET-CT
▪ Promising preliminary animal results concerning tumour volume definition for therapy planning
▪ The evaluation of the manipulator systems shows promising results in terms of precision and intervention time. These characteristics have the potential to render it a useful tool in minimally invasive treatments.
▪ Improved tumour volume definition can be used to adapt target volume for other treatments like minimally invasive ablations or radioactive seed implantation
Multimodal therapy control
Work package 8 aims at developing multi-modality imaging protocols, MRI sequences and hardware for the MRI-based functional assessment of GIST tumour microenvironment and for evaluating the therapeutic effect on GIST tumour murine models upon imatinib treatment. Several academic institutions (UNITO, UHEI, CKM) and commercial partners (RAPID, CAGE, AAA) were involved in this work-package. Imaging-based approaches are promising approaches to detect early response to Tyrosin Kinase Inhibitors (TKI) in GIST patients. Tomographic imaging modalities, such as MRI and CT, provide anatomical volumetric information that can report on tumor volume changes upon treatment. These metrics are commonly exploited for assessing treatment response in longitudinal studies. However, assessment of therapy response that occurs at 3-month intervals is hampered by null or minimal decrease in tumor size, thus not providing useful clinical information. Therefore, alternative imaging approaches as developed within WP8 should improve the diagnosis and assessment of GIST therapy, hence their development is of relevant importance.
At preclinical level several MRI-based approaches and multi-modality imaging techniques (MRI and PET) were investigated for assessing GIST tumor microenvironment in murine tumor models, whereas at clinical level dedicated multimodal coil for 23Na/1H imaging was developed and tested in human volunteers and GIST patients. At preclinical level, one of the objectives of WP8 was to evaluate the ability of functional MRI-based approaches to highlight differences in tumor microenvironment properties related to imatinib resistance in GIST murine models. We investigated GIST tumor vascularization using a Dynamic Contrast Enhanced-MRI approach in combination with a blood pool contrast agent at low-magnetic field.Functional MRI imaging was performed on immundeficient mice bearing three different GIST tumors (imatinib-sensitive GIST-T1 and GIST882; imatinib-resistant GIST430). By applying a pharmacokinetic model, functional derived parameters informative of vessel permeability (Ktrans) and plasmatic volumes (vp) were obtained. Our findings demonstrated that DCE-MRI can detect differences in plasmatic volume and vessel permeability among the investigated GIST tumor cell lines. In particular, GIST430 tumors display more than twofold higher Ktrans and vp values compared to imatinib-sensitive tumors. Increased Ktrans and vp suggested a more unstructured and deregulated vasculature in terms of blood flow in imatinib-resistant tumors. In addition, our in vivofunctional findings were confirmed ex-vivo by histological quantifications of endothelial vessels and permeability (by observing the extravasation of labeled-dextran). This study indicates that characterizing the tumor microenvironment and vasculature of GIST tumors using a functional MRI-based approach can allow imatinib-responsive tumors to be discriminated from imatinib-resistant ones. Moreover, assessing angiogenesis by functional MRI approaches may provide alternatives to conventional imaging modalities for the early detection of tumor response.
A second objective of WP8 was to investigate whether changes in tumor microenvironment properties as measured by MRI could assess imatinib response in comparison to standard approaches in GIST models. MRI offers the unique possibility to in vivo characterize vascularization, acidosis and cellularity within the same modality. For this purpose, imaging-based protocols were established, in addition to anatomical information (T2w-MRI), based on functional MRI procedures to characterize vascularization (DCE-MRI), acidosis (pH-CEST imaging) and cellularity (DWI) in combination with PET for assessing metabolism (18FDG-PET) in imatinib-sensitive (GIST-T1 and GIST882) and imatinib resistant (GIST430) preclinical murine models upon imatinib treatment. Mice were treated with imatinib for 14 days and imaging acquisitions were performed before and after 1 week and 2 weeks of treatment. Our results showed that all the proposed functional imaging approaches were able to detect changes between treated and untreated tumors. Briefly, T2w-MRI images showed that imatinib-sensitive GIST882 and GIST-T1 tumors respond to imatinib treatment by showing a decrease in tumor size in comparison to the untreated group, whereas tumor volume increases for both treated and untreated imatinib-resistant (GIST430). All treated GIST tumors show decreased FDG uptake (FDG-PET) and increased vessel permeability (DCE-MRI), reflecting metabolic changes and normalization of tumor vasculature upon imatinib treatment, respectively. In addition, ADC values (DWI) increased in imatinib-sensitive GIST-T1 and GIST882 tumors due to cytotoxic effect of imatinib. Finally, imatinib-sensitive tumors showed a decrease of tumor acidosis (pH-CEST). Summarizing the obtained results at preclinical level, we can state that investigating GIST tumor microenvironment will be of relevant importance for improving GIST diagnosis and treatment. In addition, the proposed MRI functional imaging approaches can be considered promising alternatives to standard clinical criteria.
Additionally, within WP8 a dedicated molecular diagnostic agent capable of detecting cell-induced apoptosis was assessed for its potentialto be used in GIST patients to monitor effectiveness of the employed treatments.The tracer is based on a recombinant version of wild-type Annexin V, called rhAnnexin V-128. This recombinant protein contains an endogenous chelation site for technetium (Tc-99m) and it can be used to detect apoptosis and cell death by SPECT imaging.The product is a single vial, lyophilized cold kit, which is stored at refrigerated temperature.The procedure for the kit radiolabeling with technetium is user-friendly and standardized. Briefly, the lyophilized powder is reconstituted directly with the eluate from a technetium generator. Reaction takes 1 and an half hours, leading to a final injectable product with considerably highradiochemical purity (above 95%) and stability (up to 4-6 hours after end of radiolabeling).As a proof of concept, the 99mTc-rhAnnexin V-128 was successfully used in a pancreactic tumour model (Panc-1 xenografted in nude mice), as tracer to assess the efficacy of the therapeutic treatment.
A third objective of WP8 was to develop a multimodal MRI coil for 23Na imaging. A 23Na coil array setup for investigation of the human abdomen was designed and tested on phantoms. It yields good 23Na Tx efficiency and homogeneity, good 23Na Rx sensitivity and allows for 1H imaging by providing 6 1H Rx elements while allowing the 1H body coil being used for 1H Tx. The coil was, even if being a one-of-a-kind prototype, CE certified as a medical device for allowing patient scanning. With the development of RF arrays for X nuclei (i.e. other nuclei than 1H) RAPID is now capable to provide a technology which is currently not utilized in high numbers. However, there are promising applications coming up, including 23Na imaging for GIST treatment as well as others which were not subject to this project (e.g. hyperpolarized 13C Metabolic Imaging for tumor detection and treatment). RAPID is able to provide RF coil arrays for X nuclei even on a level of very small sales numbers (one-of-a-kind coils) with medical device quality which enables research groups all over the world to further develop these technologies on volunteers and in patient studies and, such, bring them to the state where they will enable new treatment solutions in clinics. With this project, RAPID has strengthened its presence and position in the field of dedicated RF coils for MRI. In the international competition, RAPID is and stays a well established provider of such coils and is visible to both, researchers and MR system vendors as the first address for finding customized or low number RF coil solutions.
A fourth objective of WP8 is the translation of 23Na imaging in clinical studies.
During the Mitigate project, CKM (UHEI) has optimized a multimodal 23Na/1H MR imaging protocol for abdominal and body stem applications in GIST patients, which is specially adjusted to the new built double resonant coil by RAPID and was evaluated in volunteers. One GIST patient was scanned with another previously designed sodium coil by RAPID and the new imaging protocol. A second GIST patient received multimodal measurements without sodium imaging. Several steps have been conducted until designing the GIST protocol:
- Correction methods for the inhomogeneous transmit field (separate scans) have been evaluated, optimized and included into the final protocol (double angle method and phase sensitivity approach).
- A correction for the inhomogeneous receiving field due to the sensitivity profile of the phased array coils used in the newly built coil was proposed, which is completely done in post-processing and is based on ensemble empirical mode decomposition.
- The post-processing of the images was improved:
o Gridding and filtering of data was adapted.
o A post-processing pipeline designed for the GIST protocol was developed.
o Denoising of images based on total variation was implemented.
o Receiving field inhomogeneities were corrected in post-processing.
o Inhomogeneous transmission fields were corrected using the additional correction scans.
o Sodium images were translated to DICOM format.
o Spatial combination of proton and sodium images was conducted for easier incorporation of anatomical data.
The protocol and hardware were tested stepwise on 16 volunteers. Ten volunteers were scanned in the region of lower body stem, four in abdominal region and two in upper body stem region. Sodium MR images provided good quality but signal decreased too fast towards body center. This leads to high standard deviations of tissue sodium concentration in deep located tissue, where distances from the coil surfaces are larger than 10 cm. Therefore, coil design was further discussed with RAPID and promising changes for improvements were set. Hardware changes were started but could not be finished by now. We expect the upgraded coil design to enable us to improve signal to noise ratio especially from deeper located tissue and regions with a large distance from the sodium coils. Two GIST patients received a multimodal imaging. One patient received a dual CT, MR morphological and functional sequences as well as sodium imaging. Patient 2 obtained dual CT and proton MR imaging which was functional as well as morphological. Tumors could be displayed and evaluated.
Summary of significant results
• MRI-based approaches and multi-modality imaging techniques (MRI and PET) were investigated for assessing GIST tumour microenvironment in murine tumour models
o With a functional MRI-based approach imatinib-responsive tumours can be discriminated from imatinib-resistant ones
o Assessing angiogenesis by functional MRI approaches may provide alternatives to conventional imaging modalities for the early detection of tumour response
• Dedicated multimodal coil for 23Na/1H imaging was developed and tested in human volunteers and GIST patients
Potential Impact:
Description of the potential impacts
The highly complex research carried out within MITIGATE lead to a variety of project results and outputs. Each of these achievements will strongly impact GIST patients, improving their personalised diagnosis and treatment. MITIGATE findings may in many cases be transferable to other cancer types (e.g. lung cancer) for even higher socio-economic impact. Our established GIST-biopsy process involves a novel endoscopic biopsy device, biological upstream processing systems, a bespoke immunomagnetic single cell separation procedure and innovative mass spectrometry approaches. It will ensure high-quality molecular analysis of the tumour samples. Through the development and application of the GIST subtype database and GIST subtype tumour animal models, combined with the possibility to predict TKI-resistant GIST behaviours, personalised selection of an appropriate PET probe or endoradiotherapeutic tumour treatment was achieved. The whole procedure will result in an accelerated decision making process for the treatment of individual patients. AAA has developed an SOP for the radiolabelling of peptides with 68Gallium using a proprietary kit approach that is successfully applied to the NeoBOMB1 peptide, the selected probe for the clinical trial. The establishment of 68Ga-NeoBOMB1 PET-CT for molecular phenotyping in GIST tumours enables AAA to develop a novel diagnostic approach for GIST patients. The therapeutic efficacy of the 177Lu-NeoBOMB1 was confirmed in pre-clinical GIST models, making it a new potential treatment option for GIST patients. A dosimetry study was performed on GIST patients receiving the diagnostic agent 68Ga-NeoBOMB1. Using an already validated 68Ga-NeoBOMB1 PBPK model, these results will also allow a prediction of the compound’s behaviour for individualised targeted therapy approaches. MALDI MS imaging was found to differentiate between GRPR-positive and –negative GIST based on the lipidomic footprint which also allows further characterisation of gene mutations of human GIST resectates. To generate valid animal models for metastasizing GIST, orthotopic inoculation proved to be the most reliable method. The evaluation of radiotracers targeting NTSR1, GLP-2R, KIT and DOG1 showed that these were not suitable for imaging of GIST in mice, but a PBPK model for [18F]FNI was developed and is currently transferred to a human model to estimate dosimetry. Also, a fluorescent probe targeting GLP-2R for a multimodal approach is currently under investigation for enhanced visualisation of GIST. Results of the clinical study showed a high safety profile and highly promising results in specific molecular targeting of GIST tumours in PET/CT using 68Ga-NeoBOMB1. With this the initial phase providing safety and pharmacokinetic data was completed. Our patient data are the basis for further pharmacokinetic modelling and dosimetry calculations and to select patients benefitting from multimodal treatment. Endoradiotherapeutic treatment can be supported by minimally-invasive interventions guided by molecular PET imaging. Accordingly, novel, PET-CT-based target volume definition for external beam radiotherapy based on 68Ga-NeoBOMB1 were established. The high-risk areas within the tumour volume can be used for an intensified treatment. The evaluation of preliminary animal PET-CT data for EBRT therapy planning indicate that the new PET-tracer allows escalation of the dose to vital tumour cells, thereby further increasing tumour control probability and proves its significance for the minimally invasive interventions. A multimodal protocol for treatment monitoring has been investigated by CAGE and UNITO, showing that several functional imaging modalities, besides clinical ones, can assess the treatment response to imatinib. Rapid and UHEI have validated a new multimodal coil (23Na/1H) for assessing anatomical and metabolic information at clinical level, thus providing a new technology based on 23Na imaging for GIST treatment in clinical routine. By the electronic website enrolment, access to potentially more precise and less toxic treatment is provided for oligometastatic GIST patients throughout Europe
Main dissemination activities and exploitation of results
MITIGATE featured a separate Work Package dedicated to the effective dissemination and exploitation of the Foreground. The dissemination activities were based on the Dissemination plan. Exploitation activities were and are based on the intermediate and final exploitation plans. These plans set out the strategy and activities that were undertaken during the project’s lifetime, as well as exploitation activities that will continue beyond the lifetime of MITIGATE.
Communication and dissemination
The project first developed the overall dissemination strategy. The dissemination plan lists target groups and contacts from the scientific community, specific clinical disciplines, medical centres, healthcare provides, patient organisations and the general public. For each group it was specified how it will be addressed, for instance through attendance of conferences, the MITIGATE project website, mailing lists, workshops etc. It also included a list of main conferences of interest for presenting MITIGATE results. Given that the plan was developed and submitted at the beginning of the project, it remained a flexible, living document that was continuously adapted to the demands of the project and the evolving demands and interests of the stakeholders.
The overall dissemination objectives were to
• share new insights in GIST treatment with patients, patient’s advocacy groups and the general public.
• ensure project conclusions are considered by the International Rare Disease Research Consortium (IRDiRC) for fostering international collaboration in rare diseases research.
• validate that project developments are designed and adapted to better respond to patients’ needs.
• increase awareness among healthcare providers to facilitate rapid update of project outcomes into patient care, realising the value of advancements made during MITIGATE.
• inform the international scientific community active in the field of GIST about the objectives of the project and subsequent advancements made to the state of the art in GIST diagnosis and treatment through MITIGATE.
To realise the project objectives, the dissemination of information about MITIGATE was directed to the appropriate channels, from scientific publications to general media. Execution of the dissemination strategy reflected the following questions:
1. What information do we have to disseminate? (e.g. general project information, novelties in technology and tools, on-going project work and interim results, state-of-the-art of scientific questions related to GIST, etc).
2. To whom do we need to disseminate the information? (EIBIR members, partners’ contacts and networks, national and internal authorities and associations, industries, etc).
3. How will we disseminate the information? (printing articles in media, brochure and attendance at conferences, online articles to be published on websites and external web forums, media briefing, position papers, press campaign, etc).
4. When do we disseminate the information? Each dissemination activity should be followed by a time estimate if no exact dissemination schedule is provided.
5. What do we hope to achieve by disseminating this information? (what is the intended impact, what reaction or change is expected from the target audience?).
The dissemination strategy identified the unique needs of each stakeholder group, and the activities employed corresponded to these needs, ensuring effective and efficient distribution of project information.
A corporate identity was created for the MITIGATE project to facilitate clear and consistent communications. It included a logo, colour scheme, presentation slides and the overall look for the website. The project aimed to generate a brand image that could be used to promote awareness of the project and its activities. The logo and the MITIGATE layout were used on all documentation, promotional/publicity literature, exhibition boards and on the project website, etc.
The website (www.mitigate-project.eu) presents the project and its results to a worldwide audience. The website has been online since December 2013. It provides an overview of the project and the objectives (also per Work Package), information on partners and about the project progress. Non-restricted project deliverables, press items and articles have been made available on the website in the publications and media section. The website summarises the main impacts of MITIGATE for patients, researchers and clinicians separately. Scientific and technical achievements, return on investment, i.e. demonstrating the importance of research spending at European level, are also highlighted. Updates to the MITIGATE website were then promoted via social media through EIBIR’s Twitter account.
The following dissemination material was produced:
• Folders
o MITIGATE folder to introduce and promote the project
o Updated MITIGATE folder to summarise and highlight achievements of the project
o Information folder for patients (in English, German, Bulgarian)
• Four newsletters
o 1st MITIGATE newsletter was distributed in September 2014 to MITIGATE stakeholders by email and published on www.mitigate-project.eu as well as on websites of participating organisations (D2.2). It was also distributed by EIBIR at the European Congress of Radiology 2015.
o 2nd MITIGATE newsletter, distributed in September 2015 to MITIGATE stakeholders via email and published on the project website. It was also distributed by EIBIR at the European Cancer Congress 2015 and at the European Congress of Radiology 2016.
o 3rd MITIGATE newsletter was, distributed in September 2016 to MITIGATE stakeholders via email and published on the project website. The 3rd Newsletter was also distributed by EIBIR at the European Congress of Radiology (March 2017), the German Röntgen Congress 2017 (May 2017) and the Congress of the European Society for Radiotherapy and Oncology (May 2017).
o 4th MITIGATE newsletter, distributed in September 2017 to MITIGATE stakeholders via email and published on the project website.
• Informational poster to display at conferences and congresses
• Introductory roll-up to display at conferences and congresses
MITIGATE held an innovation workshop, titled “Innovations for diagnostic imaging and minimally invasive treatment: ready for the market” at the 98th German Röntgen Congress in Leipzig, Germany on May 26th 2017 (approx. 6500 attendees from academia, hospitals and industry). MITIGATE also had a booth during the whole congress duration for appropriate on-site advertisement. The MITIGATE Innovation Workshop was attended by representatives from universities, hospitals and industry. Each talk was followed by an engaging discussion, that was also continued at the MITIGATE booth. Attendees highly appreciated the presented innovations made by the project and endorsed their translation to the market.
Overview of dissemination activities
• 22 scientific publications in high-ranking, relevant, peer-reviewed journals.
• 18 articles in popular press
• Sessions and representation at the European Congress of Radiology (ECR)
o ECR 2014: Molecular imaging and targeted image-guided therapy in gastrointestinal stromal tumours
o ECR 2015: MITIGATE Consortium: State of the art imaging and therapy in GIST
o ECR 2016: MITIGATE: What does it take to perform clinical trials in interventional radiology?
o ECR 2017: Innovative solutions for diagnosis and treatment concepts for GIST patients from the MITIGATE project
o Articles in the congress newspapers “ECR Today” (print of 26,000)
▪ ECR 2014
▪ ECR 2015
▪ ECR 2016
▪ ECR 2017
o Representation and distribution of dissemination material through the EIBIR booth
• MITIGATE was presented and promoted at networking events (e-rare workshop; CommHERE workshop)
• MITIGATE Joint Scientific Symposium: ADVANCES in THERANOSTICS of RARE CANCER, at the Medical University Innsbruck, Austria
• Promotion and dissemination through EIBIR social media
• Workshop “Innovations for diagnostic imaging and minimally invasive treatment: ready for the market” at the German Röntgen Congress 2017, 26 May 2017
• Presentations at scientific events
o 54th Annual Meeting of the German Society for Nuclear Medicine
o 5th Annual Symposium of the Oncology Network of West Austria and South Tyrol, 27 May 2015
o Multiple presentation at the 23rd ISMRM annual meeting, 3-4 June 2015
o Deutscher Krebskongress, 24-27 February 2016
o German Sarkomkonferenz, 17 March 2016
o Multiple presentations at events of the German Society for Medical Physics
o Four presentations at the EANM Annual Meeting 2016, 15 October 2016
o OurCon IV conference in imaging mass spectrometry, 17 October 2016
o Mass Spectrometry and Proteomics Congress 2016, 15 November 2016
o Invited talk about 0D- to 4D-Mass Spectrometry in or near the Operating Room at University of California, Los Angeles, 18 November 2016
o Presentation about MALDI MS imaging in a clinical context at the BW-CAR Kickoff-meeting, 30 November 2016
o EMBL target validation conference, 6 December 2016
o Medizintechnologie made in Mannheim event, 8 December 2016
o Invited talk about at MALDI mass spectrometry fingerprinting and imaging in pharma industry and in the clinic at Leipzig University, 9 December 2016
o 16th meeting of the Norwegian Society for Mass Spectrometry, 23 January 2017
o Deutsche Gesellschaft für Endoskopie und Bildgebende Verfahren Kongress 2017, 7 April 2017
o Annual Meeting of the Canadian Association of Nuclear Medicine 2017, 21 April 2017
o Annual Meeting of the ISMRM 2017, 27 April 2017
o International Symposium on Radiopharmaceutical Sciences 2017, 19 May 2017
o SNMMI Annual Meeting 2017, 10 June 2017
o DGN Summer School 2017, 1 September 2017
o 14th UHEI Summer Workshop, 2 September 2017
• 17 scientific posters
o European Molecular Imaging Meeting 2015, 2016
o World Molecular Imaging Congress
o Deutscher Krebskongress 2016
o Jahrestagung der Deutschen Geselleschaft für Medizinische Physik 2015, 2016
o Jahrestagung der Deutschen Gesellschaft für Nuklearmedizin 2015, 2016
o 13th South East Asia Congress of Medical Physics
o EANM Annual Meeting 2016
o ISMRM 2017
o SNMMI Annual Meeting 2017
o EMBL 18th PhD symposium
• Two exhibitions
o Medica 2016, 14 November 2016
o RSNA 2016, 26 November 2016
• Distribution of the MITIGATE dissemination material and representation at other scientific events
o European Cancer Congress, 25-29 September 2015
o Congress of the European Society for Radiotherapy and Oncology, May 2017
• Engagement with patient advocacy groups (PAGs)
o Judith Robinson, GIST Support UK
o Amy Bruno-Lindner, GIST Support Austria
o Feedback about the MITIGATE folder for patients, distribution though PAGs at PAG events.
o Feedback on the MITIGATE Clinical Study Patient Recruitment Form
o Presentations at PAG events
▪ Patient Forum on GIST, Das Lebenshaus, Mannheim, Oct 8, 2015
▪ NEW HORIZONS GIST Conference (Sitges/ES, May 2016)
• Promotion and dissemination through EIBIR social media
Exploitation
In the first project period the project developed an intermediate exploitation plan. This plan detailed how partners intend to exploit the results during the project and how exploitation will continue at the end and shortly after the project. The project’s Clinical Innovation Advisory Board provided expert advice on exploitation planning throughout the lifetime of the project and highlighted the importance of analysing and specifying the market as far and detailed as possible, in quantitative and qualitative aspects. To this end, MITIGATE launched an end user survey to assess the market potential, market access and market environment from MITIGATE’s exploitable results. The survey was promoted and distributed through the MITIGATE newsletter, at the European Congress of Radiology 2016, and through social media. A final exploitation plan consisting of individual exploitation plans and activities for the key exploitable results was developed in the final period of the project.
Overview of the project’s key exploitable results and the exploiting partners:
RF Coil Arrays for X Nuclei
RAPID and UHEI developed an RF coil package for doing MR imaging and spectroscopy on the human body. The RF coil package consists of a 23Na transmit coil for exciting the abdominal region, a 23Na receive array for getting the best possible signal-to-noise ratio and an an integrated 1H receive array for achieving clinical 1H imaging performance. The RF coil technology is not limited to 23Na but works with other nuclei as well, such as 13C. The RF coil array shall be exploited by manufacturing and selling the coil through RAPID. RAPID will use its own sales force and sell the RF coil package to researchers directly.
Tissue Grinder
FHI has developed an innovative tissue dissociation module (tissue grinder) which allows the dissociation of tissue samples into single-cell suspensions with lower enzyme concentrations. This device enables the isolation of a higher number of intact cell surface markers and subsequent analysis. After securing the intellectual propetry, the technology will be licensed to an industry partner who will manufacture and sell the product. To achieve this, FHI will have to contact relevant industrial parties. FHI has successfully applied for other grants to further develop the tissue grinder technology and has started the FHI internal project “MAVO LyDia HD”. FHI has also applied for a patent for the tissue grinder under application number 10 2016 216 345.0 which is expected to be granted shortly.
Magnetic Separation Technology
SCT developed a methodology to generate a single cell suspension from three main heterogeneous GIST enriched tissues and to enrich GIST cells from these tissue types using an immunomagnetic isolation system. The isolation kit is based on two main components: one vial of isolation cocktail and one vial of paramagnetic particles. SCT expanded this technology to isolate other cancer cells. SCT owns to intellectual property, but also assesses OEM (original equipment manufacturer) products and licensing opportunities identified by its business development department for the exploitation of the magnetic separation technology. SCT will market the GIST and non-GIST cancer cell isolation reagents, and intends to establish collaborations with academia for the further development and exploitation of these products. A targeted marketing effort will have to be made towards cancer related institutions and laboratories.
Robotic Assistant for needle placement
FHI has, together with UHEI, developed and successfully tested in a preclinical setting, a robotic assistance system for interventional needle placement. The system consists of a navigation software and a robotic guidance device, which together enable the fast and precise placement of needles under image guidance. The system is slightly more precise than the manual needle placement and significantly reduces the procedural time. FHI has presented the technology at RSNA 2015, Medica 2016 and RSNA 2016 in order to convince industry partners to join FHI’s efforts to make the product ready for the market. FHI has found a partner, who is willing to bring the assistance device to market. Together with an industry partner, the technology will be developed further and then brought to market by the industrial partner, licensed by FHI.
Kit for radiopharmaceutical preparation of 68Ga-NeoBOMB1 for diagnosis and/or tumour assessment in patients with GRPR-positive malignancies, including GIST
AAA, with the support of MUI, developed a GMP-compliant kit for radiopharmaceutical preparation of 68Ga-NeoBOMB1 to be used for diagnosis and/or tumour assessment in patients with GRPR-positive malignancies, including GIST. AAA is also pursuing opportunities to develop a kit for the therapeutic counterpart (177Lu-NeobOMB1) as a theragnostic option. AAA has extensive expertise and knowledge in bringing new radiopharmaceuticals to the market and aims at pursuing marketing authorisation for NeoBOMB1 in GIST as well as in other potentially interesting oncological indications where GRPR expression is confirmed. On 12 December 2016, orphan designation (EU/3/16/1794) was granted by the European Commission to AAA for 68Ga-NeoBOMB1 for the diagnosis of gastrointestinal stromal tumours. Orphan designation for the therapeutic counterpart, 177Lu-NeoBOMB1, will be pursued by AAA as well. MUI is currently using the AAA kit in a phase I/IIa clinical trial in TKI-resistant GIST patients to evaluate safety, biodistribution, dosimetry and preliminary diagnostic performance of 68Ga-NeoBOMB1.
Mouse models for GIST implants
UNITO developed imatinib-resistant and sensitive GIST mouse models. These models recapitulate typical GIST features by developing a primary tumour and subsequent liver metastasis that were easily detected by MRI. The animal models can be used to test new diagnostic tools and pharmaceutical treatments on imatinib-resistant tumours. UNITO will license the animal models to dedicated companies (e.g. Charles River, Jackson, Oncodesign) that can ensure efficient animal generation by guaranteed manufacturing procedures in order to be competitive on the market.
New imaging probes able to distinguish imatinib-resistant and imatinib-sensitive GIST tumours
CAGE designed new imaging probes targeting the GLP-2 receptor to distinguish imatinib-resistant and sensitive GIST tumours. A GLP-2-like peptide conjugated with a fluorescent or a MRI/PET probe was synthesised. In addition, a novel class of chelates, able to bind both the Gadolinium cation for MRI applications or radionuclide metals for PET/SPECT, was developed to extend the use of the GLP-2 probe. The GLP-2 imaging probes will be patented and then licensed to pharmaceutical companies for planning the required clinical trials. A patent application will be submitted after further validation (estimated in spring 2018), after which CAGE will actively seek for a pharmaceutical company interested in a product license. In parallel, CAGE will maintain production of small quantities of the product to satisfy potential requests for preclinical studies from research centres.
Summary of dissemination and exploitation activities
• MITIGATE website
• Dissemination plan
• Dissemination material
o Flyers, folders, posters, roll-up, newsletters
• 22 papers published in peer-reviewed journals
• 48 presentations at various scientific meetings
• 17 posters at various scientific meetings
• 18 articles in popular press
• Intermediate Exploitation Plan
• End User Survey
• 4 exhibitions
• 2 workshops
• Final Exploitation Plan
List of Websites:
http://www.mitigate-project.eu/
Gastrointestinal stromal tumour (GIST) is a rare disease with high potential for metastasising that frequently affects young patients and results in a short life expectancy. Currently there is only one class of effective medication and often the tumours develop drug resistance after a few years. No molecular imaging technologies indicating drug resistance, therapeutic response or disease progression are clinically available. Furthermore, effective targeted agents such as endoradiopharmaceuticals are not designed for GIST and alternative minimally-invasive treatment options are not well explored. The overall objective of the MITIGATE project was to develop and validate an integrated closed-loop process to effectively treat metastatic GIST patients resistant to the currently available class of medication.
MITIGATE’s objectives were:
1) Optimisation of biopsy and tissue preparation in GIST for analysis and subtyping of molecular markers.
2) Provision of a tumour bank of TKI-responsive and -resistant GIST
3) Application of molecular imaging probes indicating resistance, early therapeutic response or progression.
4) Development of endoradiopharmaceuticals for treatment in patients.
5) Combination of the therapeutic approach with alternative, minimally-invasive treatment options.
6) Non-invasive monitoring of treatment effects in clinical routine.
After 4 years the highly complex research lead to a variety of project results that will strongly impact GIST patients and improve their personalised diagnosis and treatment. Our established GIST-biopsy process involves a novel endoscopic biopsy device, biological upstream processing systems, a bespoke immunomagnetic single cell separation procedure and innovative mass spectrometry approaches. Through the development of the GIST subtype database and GIST subtype tumour animal models, combined with the possibility to predict TKI-resistant GIST behaviours, personalised selection of an appropriate PET probe or endoradiotherapeutic tumour treatment was achieved. A SOP for radiolabelling of peptides with 68Gallium using a kit approach that is successfully applied to the NeoBOMB1 peptide was prepared. The establishment of 68Ga-NeoBOMB1 PET-CT for molecular phenotyping in GIST tumours now enables to develop a novel diagnostic approach for GIST patients. The therapeutic efficacy of the 177Lu-NeoBOMB1 was confirmed in pre-clinical GIST models, making it a new potential treatment option for GIST patients. A dosimetry study using an already validated 68Ga-NeoBOMB1 PBPK model indicated a prediction of the compound’s behaviour for individualised targeted therapy approaches. A first in human clinical trial with 68Ga-NeoBOMB1 focussing on safety and pharmacokinetics was successfully completed. Initial results indicate an excellent safety profile of 68Ga-NeoBOMB1 in patients and targeting visualisation by PET/CT of GIST tumours was achieved. Novel, PET-CT-based target volume definition for external beam radiotherapy based on 68Ga-NeoBOMB1 were established. The high-risk areas within the tumour volume can be used for an intensified treatment. The evaluation of preliminary animal PET-CT data for EBRT therapy planning indicate that the new PET-tracer allows escalation of the dose to vital tumour cells, thereby further increasing tumour control probability. A multimodal protocol for treatment monitoring shows that several functional imaging modalities can assess the treatment response to imatinib. A new multimodal coil (23Na/1H) for assessing anatomical and metabolic information at clinical level was validated and provides a new technology based on 23Na imaging for GIST treatment in clinical routine.
The close interdisciplinary collaboration led to an integrated, closed-loop treatment platform that induces a paradigm shift for individually tailored therapy of patients with TKI-resistant disease. MITIGATE findings may also be transferable to other cancer types (e.g. lung cancer) for even higher socio-economic impact.
Project Context and Objectives:
Gastrointestinal stromal tumour (GIST) is a rare disease with an annual incidence of about 1.5 /100,000 persons. GIST frequently affects young patients with high potential for metastasising, which often results in a short life expectancy of less than 3 years on average. Currently there is only one class of effective medications for systemic GIST therapy and often the tumours develop drug resistance after a few years. No molecular imaging technologies indicating drug resistance, early therapeutic response or disease progression are clinically available. Furthermore, effective targeted agents with other mechanisms of action such as endoradiopharmaceuticals are not designed for treatment of GIST and alternative minimally-invasive treatment options for metastatic disease are not well explored nor has the clinical value of their combined use been established. This leaves GIST patients with very limited treatment options.
The overall objective of the MITIGATE project is to develop and validate an integrated closed-loop process to effectively treat metastatic GIST patients resistant to the currently available class of medication, the tyrosine kinase inhibitors (TKI).
This envisaged personalised treatment concept combines innovative strategies for biopsy, inline tissue analysis, molecular tumour characterisation, theranostics by imaging technologies (PET and MRI) and companion radiopharmaceuticals followed by the assessment of biodistribution, dose calculation and measurement of therapeutic effectiveness
MITIGATE`s ultimate goal is to develop new protocols and guidelines to effectively diagnose and treat patients with metastatic GIST resistant to current treatment. The project will develop new targeted personalised treatment for GIST metastases which can be then used as a role model for other molecularly characterised cancer subtypes no-longer responsive to targeted therapy.
The overall objectives of MITGATE are to
1) Optimise biopsy and tissue preparation in suspected GIST for integrated analysis and subtyping of molecular markers.
2) Provide a large and broadly characterised tumour bank of TKI-responsive and –resistant GIST for generating novel pre-clinical models envisioned to reveal novel markers for TKI-resistance.
3) Apply molecular imaging probes for specific GIST features indicating drug resistance, early therapeutic response or progression.
4) Develop effective targeted agents with other mechanisms of therapeutic action such as endoradiopharmaceuticals for treatment in patients.
5) Combine this therapeutic approach with alternative, minimally-invasive treatment options for a personalised treatment in the metastatic situation with fewer side effects.
6) Non-invasively monitor treatment effects in clinical routine.
Project Results:
Biopsy and Preparation for Molecular Analysis
Novel Endoscopic Biopsy System
A novel endoscopic biopsy system was developed. First, a demonstrator of a biopsy needle taking samples for molecular analysis through a flexible endoscope was designed. Requirements for cutting and transport of the sample were specified to safeguard the sample quality. Several cutting mechanisms were designed and evaluated. A cutting mechanism based on a sharp blade at the tip of the insertion tube was chosen as the appropriate mechanism in combination with vacuum transportation. Several transport methods were conceptualised and evaluated and an experimental set up was designed to test the simplest method of applying of vacuum. Hence it was concluded that a continuously applied vacuum in combination with a uniformly cut sample presented a valid method for biopsy transport. The optimised entire endoscopic biopsy system, as shown in Figure 1 was validated by showing that the cutting mechanism and vacuum transport work as a functional unit in a phantom model, shown in Figure 2. As model, an endoscopy simulator with an artificial stomach was chosen. The resulting pictures and videos show that the developed system is completely functional.
Biopsy Needle interface
This task aimed to connect the biopsy system with the subsequent tissue disruption module. The tissue disruption module is based on grinding of the biopsy between rows of teeth. Therefore, no standard container had to be used as an interface and the grinder was directly connected to the tubing of the biopsy system.
Tumour Cell Isolation
The aim of this task was to isolate specific tumour cells from a single cell suspension obtained from a GIST tumour tissue biopsy. A first objective of this task was the development of a tissue disruption device which ensures the disruption of tissue into single cells without destroying the material. After researching many different cell dissociation methods and the performance of several experiments, a grinding device inspired by a commercially available herb grinder was developed. The mechanism is based on the movement of rows of teeth grinding against each other. Experiments were mainly performed with chicken liver, of which single cells could be obtained. Due to the limited availability of tumour material (GIST tumours are rare and frozen tissue was not suitable) only few experiments were performed with fresh GIST samples. Since consistency of GIST can vary a lot depending on the tumour site, inserts varying in design and material were manufactured to adapt the grinder to every consistency of the material. Pre-treatment with enzymes might be necessary for certain GIST consistencies. The grinder was directly connected with the biopsy system and integrated into an automated process of single cell generation. The description of work intended to design a device with different stations for mechanic and enzymatic treatment of the tissue. The automated grinder setup allows for the addition of buffer and tissue dissociation enzyme, thus mechanical and enzymatic disruption can happen simultaneously.
Isolation of tumour cells with magnetic bead technology using specific antibodies turned out to be challenging. The commercially available DOG1 antibodies, often used in immunohistochemical staining, were not suitable for positive GIST cell isolation as they exclusively bind to intracellular domains of the surface protein. Initially, three GIST cell culture lines (TKI resistant and non-resistant) were assessed phenotypically for their defined GIST tumour markers. An identification of GIST cell surface markers allowed for the design of a selection cocktail which would theoretically immunomagnetically deplete all non-GIST cells in a single cell suspension thereby enriching for GIST cells. In a mock tumour design, initial experiments combined peripheral blood mononuclear cells (PBMCs) endothelial, epithelial and mesenchymal cells with GIST cells from cell lines in specific ratios. Non-GIST cells where immunomagnetically depleted, enriching for GIST cells with purity in excess of 98% and an average recovery of 65%. To provide a proof of principle, an alternative cell line was used. The results show that cells can not only be distracted but also be distinguished by their magnetic charge in a microfluidic device. The technology is applicable for every kind of tumour cells, and as soon as a suitable antibody will be available for the positive isolation of GIST cells experiments will be adapted and performed. A negative selection of GIST cells is conceivable as well, by simply running several positive isolations of non-desired cells in a row. However, this will probably result in a loss of a large fraction of material. Due to limited availability of GIST tissue not many experiments could be performed with patient-derived tumour material. A total of eight experiments were performed with patient-derived tumour tissue. The quality of the extracted tissue was extremely variable with respect to cell viability, organ/tissue of origin and expression levels of known GIST surface phenotype markers (CD34, CD117 and CD90). Much of this variability in viability is likely due to patient treatment regimens with TKIs, such as imatinib, which are known to induce large-scale apoptosis in tumour cells. Furthermore, reduction in CD34 levels after imatinib treatment is known and found to be related to cystic degeneration. Some of the variability in surface marker expression is due to patient’s GIST phenotype prior to treatment and the tissue of origin. CD34 and CD117 are expressed on average in 60-70% and 85-95%, respectively but these expression levels vary dramatically depending on GIST tissue location. Tissue with poor initial viability resulted in very few viable cells remaining following tissue dissociation and GIST cell isolation. It was found that tissue excised from patients undergoing treatment was very dense and rubbery in consistency, with high collagen contents resulting in difficult tissue dissociation. Conversely, GIST tissue excised from stomach or the peritoneal cavity was much easier to dissociate and resulted in higher viability. On average, the GIST cell isolation cocktail developed for the immunomagnetic separation of GIST from non-GIST cells resulted in a complete removal of all CD45 positive cells from the single cell suspensions and a 3 to 6 fold enrichment of GIST cells depending on the quality and initial phenotype of the GIST tissue.
GIST Substype Classification and GIST Subtype Database
Potential relevant masses for a robust classification of different tumour samples were identified. After a successful setup of a suitable data acquisition workflow based on MALDI MS spectrometry measurements, a standard procedure for pre-processing the measurement data and a graphical user interface for MATLAB tools was developed. Because of the “fresh GIST sample” bottleneck studies were based on three distinct GIST cell lines, namely GIST-T1, GIST-882 and GIST-430, with different response rates towards the tyrosine kinase inhibitor imatinib. We used Fisher Score Estimation methods to identify the relevant features for a following classification model which could possibly also be used for sorted cells obtained.
The developed methods are not limited to the processing and analysis of GIST, but can also be applied to other cancer cells and possibly to those derived from other solid tumours like mamma- or prostate carcinoma. To validate the model, in addition to Hierarchical Cluster Analysis (HCA), blinded data was classified by using Principle Component Analysis (PCA) and k-nearest neighbours (kNN) algorithms. In this way, a cross validation was performed. Since the real assignments between GIST type and the correspondent spectra are known, the performance of the blinded experiments can be expressed by calculation of corresponding sensitivity and specificity. In a final step, we applied the complete processing pipeline to an independent set of breast cancer cell line data to demonstrate the transferability of this approach. The suitability of determined fingerprints to predict the response upon TKI was examined. Since human tissue samples of GIST are rare, a different model was used for initial verification of the predictive power given by the identified MS features. Xenograft tissue from highly immune-deficient NSG mice that were engrafted with GIST-T1, GIST-882 and GIST-430 cells was used. The NSG mouse model lacks mature T cells, B cells as well as natural killer cells and holds deficiencies in multiple cytokine signalling pathways. Therefore, it has proven to be well suited for GIST xenografts development. The predictive power of compressed MS fingerprints derived from GIST cell lines was successfully determined on a model closer to GIST biopsy material. Compressed Data of TKI-resistant and TKI-responsive NSG xenografts was separable with 100% accuracy by Principle Component and Hierarchical Cluster Analysis.For future applications and an iterative improvement of the classification quality it's essential to create a knowledge base containing all historical measurement and data analysis. This database must not only store the corresponding device measurements but needs also to register all relevant properties of the biological sample and potential parameters influencing the measurement result. In this way correlations between features in the high dimensional feature space can be found and incorporated to optimise the classification result. A suitable database system for managing MS fingerprinting data was designed, acquired and installed. With further data input an iterative improvement of classification quality is expected.
Summary of significant results
▪ Novel endoscopic biopsy system
▪ Novel tissue dissociation device directly coupled with biopsy system
▪ Proof of principle of specific tumour isolation via magnetic distraction
▪ GIST subtype classification system
▪ GIST subtype database (based on cell lines)
Target Analysis and Molecular Probe Design
Analysing biopsies of GIST with respect to expression of a series of potential targets and identification of candidate targeting vectors
A characterisation of Imatinib resistant GIST animal models by providing spatially resolved molecular mass spectrometry (MS) data was performed. Based on lipid fingerprints, differentiating mass features between resistant and susceptive animals were identified to confirm previously observed drug effects. In conjunction with MS results obtained from cell culture experiments discrepancies between the molecular profiles of tumour cells cultivated in laboratory flasks and mice were observed. However, the established methodological workflow for MS data acquisition was proven to be stable and reproducible for a transfer to monitor human tissue samples. MS imaging was used to detect potential radiodiagnostics, but was proven incapable of monitoring small dosages in tissues. For that reason, MS image triplicates of 31 human patient GIST resectates were recorded and molecular lipid signatures were linked to histopathological and immunohistochemistry data. For that purpose, a sophisticated sample logistics, measurement and data processing workflow was established. This allowed for the MS imaging analysis of high quality cryo-preserved GIST tissue at HM. As the instrumental equipment highly improved during the project period a faster MS analysis with high spatial and mass resolution for highly specific molecular information was enabled. In this regard staining workflows to determine the presence/absence of new marker candidates were developed. A supervised classification pipeline was successfully established based on an efficient “big-data” analytics framework to identify molecular signatures characterizing specific targets. Also the identification of potential candidate molecules by high resolution FTICR-MS was pursued. Proof of concept for the characterisation of lipid fingerprints by means of MS imaging have been accomplished to evaluate the pharmacological properties of potential targeting vectors in humans.
Establishment of an animal model of imatinib resistant GIST
This task focused on the establishment of a mouse model of imatinib-resistant GIST using highly immunocompromised mice. To allow a personalised approach to cancer therapy, patient-derived xenografts (PDX) based on the transfer of primary tumours directly from the patient into immunodeficient mice represent a suitable model. Tumours can be engrafted heterotopically (subcutaneously at the flanks of mouse) or orthotopically (implantation into specific organs). To find the best model for GIST implants, several mouse models with different levels of immunosuppression were developed. NSG mice were reported to be the most appropriate immunodeficient host animal for direct xenografting of fresh tumour tissues. In total 30 GIST tissue samples from Germany and Italy were collected. Pieces of tumour tissue were subcutaneously and orthotopically (renal capsule and peritoneum) implanted into NSG mice. To evaluate tumour engraftment, growth and dissemination animals were monitored by digital palpation and MRI analysis. After 6-9 months, no tumour growth was detected and the autoptical examinations confirmed the absence of either tumour masses and/or metastases. Parts of GIST fresh samples were subjected to mechanical and enzymatic dissociation to obtain primary cells, which, after molecular and genetic characterisation, have been inoculated subcutaneously or orthotopically into spleen and liver. Cell viability was higher than 50% and KIT expression was variable among tumours. Primary GIST cells inoculated into NSG mice were monitored for tumour growth by MRI analysis for histological injection and by digital palpation for subcutaneous inoculation. However, no tumour growth was discovered after 3 months. To understand engraftment failure of in vivo experiments, primary cells isolated from GIST were characterised at different culture passages. We performed real time PCR to evaluate c-KIT and DOG1 expression in primary tumours and in cells at passages 1, 2 and 6. We observed that KIT and DOG1 expressions progressively decreased with culture passages and lastly disappeared. Accordingly, direct DNA sequencing analysis revealed the loss of the original KIT mutations with culture passages indicating the proliferation of fibroblasts instead of tumour patient isolated cells. Although we have performed 230 implants using 140 NSG mice, no tumour mass has been observed in any of the injection sites (orthotopical, subcutaneous or peritoneal). Several reasons could explain the difficulties we and others faced in generating GIST PDX models. First, GIST is a low grade tumour and most likely hard to engraft and grow by xenotransplant. In fact, a limited number of successful studies displaying GIST-PDX model has been published. Second, GIST PDX models are not available in public repositories such as the European PDX consortium and US National Cancer Institute repository. As an alternative strategy, we decided to develop a new approach based on the use of Imatinib-sensitive and -resistant GIST cell lines (GIST-T1, GIST-882 and GIST-430) to generate GIST animal models suitable to investigate biological mechanisms related to tumour resistance to individuate new molecular markers and to design effective targeted treatments.
We performed orthotopic, subcutaneous and intravenous injection into NSG mice. Subcutaneous inoculation developed localised non-invasive tumours. Orthotopic injection, in particular intrahepatic inoculation, revealed diffuse tumour masses and numerous metastases with GIST-T1 and GIST-43 cell lines. On the contrary, inoculation of the Imatinib sensitive cells GIST-882 led to tumour masses with slower growth rate without formation of metastases. No tumour masses were detected by i.v. injection of all GIST cell lines. In conclusion, we established that in vivo GIST cell lines growth is dependent on the site of injection and that the most suitable protocol to generate metastatic GIST mouse models by mimicking the tumour-stromal microenvironment is to inject GIST-T1 (Imatinib-sensitive) and GIST430 (Imatinib-resistant) cell lines orthotopically in the liver of NSG mice.
Chemical development of new radiopharmaceuticals for molecular imaging of resistant GIST with a focus on PET-Tracers and potential endoradiopharmaceuticals and biological evaluation of the potential new contrast agents/radiotracers
DOTA-NT20.3 was successfully synthesised in high purity (>98%) via solid phase peptide synthesis using the Fmoc-strategy. DOTA was chosen as chelating agent due to the easy introduction of several radionuclides (e.g. 68Ga, 177Lu) allowing a theranostic approach. After establishing the radiolabelling protocol with 68Ga, DOTA-NT20.3 and the commercially available Neurotensin-1 receptor ligand DOTA-NT(8-13) were evaluated in vitro in an internalisation assay using three GIST cell lines, as well as HT-29 cells as a positive control. Both NT-1 ligands did not show any internalisation or membrane binding on GIST cell lines (HT-29 tested positive), indicating an insufficient number of NT-1 receptors on the cell surface of GIST cells. To image angiogenesis in GIST tumours, the VEGF analogue peptide HPLW was selected as targeting ligand. The prosthetic group N-succinimidyl-4-[18F]fluorobenzoate ([18F]SFB) was linked to the targeting vector through its terminal lysine residue. The radiolabelling of [18F]SFB was established and optimised on a AiO (Trasis) synthesis module in a three-step procedure with final SPE purification. This way, [18F]SFB could be synthesised in a radiochemical yield (n.d.c.) of 44±6% and a radiochemical purity of 98±2% within 45 minutes. [18F]SFB could be successfully linked to the targeting vector HPLW within 10 minutes and after semipreparative HPLC purification the radiotracer could be obtained in a radiochemical yield of 1.54±0.04% and a radiochemical purity of 94±6% after 100 min of synthesis in total. Following appropriate formulation, the radiotracer [18F]SFB-HPLW proved to be suitable for in vivo evaluation.
The Glucagon like receptor 2 (GLP-2R) has been found to be expressed in 68% of gastrointestinal stromal tumours (GIST). Consequently, GLP-2R could be considered a potential target for the visualisation of GIST by both optical imaging and PET. Several GLP-2R ligands have been synthesised via solid phase peptide synthesis, comprising 4 DOTA-derivatised Teduglutide analogues (24-34 amino acids) for radiolabelling, as well as short peptides coupled with fluorescent dyes (carboxyfluorescein, Cy3) and AAZTA. The synthesis of the non-truncated Teduglutide derivatives was performed via fragment condensation to obtain the 34 amino acid sequence in a high purity of >95%. The radiolabelling protocol using 68Ga was established for the DOTA-derivatised Teduglutide analogues ([68Ga]DOTA-Lys10)-Teduglutide, ([68Ga]DOTA-His1)-Teduglutide, ([68Ga]DOTA-Lys30)-Teduglutide and the truncated ([68Ga]DOTA-Lys34)-(Trp10, Ala12)-GLP-2(10-33) with high 68Ga-incorporation of >94% for all compounds with varying reaction times depending on the presence of oxidation-sensitive methionine in the amino acid sequence. The truncated radiotracer showed the most promising results in vitro in an internalisation assay and flow cytometric analyses of all three GIST cell lines revealed comparable mean fluorescence intensities as the positive control cell line A549. In vivo evaluation of the truncated derivative in tumour-bearing mice (GIST-T1 and GIST882) showed a favourable renal excretion, but a slow blood clearance as well as a low uptake into the tumour.
Real-Time PCR (RT-PCR) was performed to evaluate GLP-2R expression in GIST cell lines (430, -T1 and 882) in addition to MDA-MB-231 and HeLa cell lines, indicated by the human transcriptome database (http://ist.medisapiens.com/) as negative and positive control cell lines, respectively. Imatinib-resistant GIST430 show the highest expression, about 20 times higher than the positive control, whereas GIST-T1 and GIST882 show intermediate expression (16.2 and 46.7% expression, respectively). Therefore, the generation of a peptide-targeting GLP-2R is crucial for improving GIST diagnosis, detection and seeking for alternative therapy. In order to correctly design a peptide able to interact with GLP-2R, homology modelling of GLP-2 receptor was firstly performed and the GLP-2 sequence of 20 amino acids was derived and modelled on GLP-2R. Consequently, a first batch of GLP-2 derivative was synthesised and conjugated with carboxyfluorescein for optical imaging and chelated with AAZTA for PET/MRI applications. However, due to the high impurity of both derivatives, a new synthetic strategy was exploited. GLP-2 conjugated with Cy3 fluorescent dye was obtained with 90% of purity, whereas GLP-2 AAZTA reached a purity of 80%. GLP-2 Cy3 demonstrated its ability to bind on cells expressing GLP-2R by flow cytometry. However, the competition assay with a competitor (unlabelled GLP-2) was unable to provide information as a control. For this purpose, the synthesis of a scramble peptide labelled to a fluorescent dye for improving the control assays was done. Cy5 dye was chosen to avoid background fluorescence coming from tissue for subsequent in vivo applications. Moreover, both GLP-2 targeting and scramble peptides were synthesised adding a PEGylate linker to improve binding affinity to the target on cell lines. GLP-2 and Scramble-Cy5 were then in vitro tested on GIST cell lines for evaluating their binding affinity. Confocal microscopy images showed that GLP2-Cy5 appeared to localise on the cell membrane and in the cytosol region of GIST430 cells. In particular, the signal diffuses from the membrane to the cytosol. By contrast, the scramble peptide shows a weak fluorescence signal. Further in vivo studies will validate the binding affinity of GLP2-Cy5 in GIST tumour murine models.
The radiolabelling protocol for the commercially available polyclonal DOG1 antibody using 124I was established and the radioiodinated antibody was evaluated in vitro in a binding assay. The control antibodies (e.g. 124I-Cetuximab, 124I-Rituximab) for radiolabelling and in vitro evaluation showed binding to the respective control cell lines, but nut to GIST cell lines. The polyclonal 124I-DOG1 antibody, however, showed no binding to GIST cell lines and was therefore further evaluated. FACS analysis revealed that the native polyclonal DOG1 antibody is not an appropriate agent for GIST imaging.
After the successful 6-step and 7-step organic syntheses of the nitropyridine precursor and the fluoro reference compound, the radiolabelling of [18F]fluoronorimatinib (FNI) has been established. This two-step reaction consisting of radiofluorination and deprotection yielded the radiotracer in a good radiochemical yield (RCY) of 20±3%. After ensuring high stability in buffered solution (pH 7.4) as well as human and murine serum, the in vitro evaluation of [18F]FNI was performed. Cell permeability proved to be good despite the increased hydrophilicity, but competitive binding assays with viable cells gave no satisfactory result due to a high background accumulation. Therefore, the binding affinity of [18F]FNI towards 10 different KIT mutants has been evaluated in a kinase assay confirming a similar binding profile as Norimatinib and thus a distinction between Imatinib sensitive and resistant KIT mutants in vitro. The in vivo evaluation was performed in Balb/c SCID mice which were inoculated with GIST-T1, GIST882 and GIST430 tumours. Despite the promising in vitro results, the radiotracer was not able to distinguish between sensitive and resistant tumours in vivo due to the vastly different vascularisation. The tumour-to-background ratios were in the range of similar radiotracers (SKI696), however, the pharmacokinetic behaviour could be improved by the reduction of lipophilicity leading to an enhanced, more favourable renal excretion. The co-administration of ketoconazole led to an increased metabolic stability of the tracer as well as an increased tumour uptake in first experiments resulting in a better demarcation of tumours in PET imaging.
Several precursor strategies for the small molecule radiotracer [18F]fluoro-DOG1 have been pursued and the respective precursor molecules have been synthesised including nitro, triazene and boronic acid ester leaving groups. Thereof, the latter was successfully radiolabelled under copper-catalysis and obtained in a good RCY of 33±11%. Although [18F]fluoro-DOG1 proved to be stable in buffered solution (pH 7.4) and human serum, a fast degradation was observed in murine serum. Consequently, PET imaging of GIST430 tumour-bearing mice showed only marginal uptake in the tumour and strong accumulation in the liver and the gastrointestinal tract.
Calculation and optimisation of biodistribution and dosimetry for diagnostic purposes
PET biokinetic data from experiments in 5 tumour-bearing Balb/c SCID mice injected with [18F]FNI was used to develop a mouse [18F]FNI PBPK model. The PBPK model bases on anatomical and physiological information obtained from the mouse experiments and from literature. In the model three kinds of parameters were used: fixed, Bayes (based on a known distribution), and fitted (without previous information). Data from the experiments (total mouse weight and organ weights) were used as fixed parameters. Data from the literature (organ vascular fractions, organ interstitial fractions, organ permeability surface area products and organ plasma flows) were used as Bayes parameters. Unknown parameters (metabolic rates, internalisation rates, tumour plasma flow and tumour permeability surface are product) were fitted. The developed mouse [18F]FNI PBPK model individually considers the most relevant organs and the organs with high accumulation (liver, kidneys, spleen, intestines, pancreas, tumour, muscle, bones, brain, lungs and bladder). The other organs and tissues were grouped together. The developed model was well supported by the data in terms of visual inspection, coefficients of variation of the fitted parameters and correlation among the model parameters. In the [18F]FNI PBPK model is considered both the whole [18F]FNI molecule and its metabolites. Additionally, a human PBPK model for [18F]FNI was created based on the fitted mouse [18F]FNI PBPK model by scaling the physiological and anatomical parameters from mice to humans using the allometric equation. Volumes and flows were scaled in a different way while some rates such as the internalisation rate were kept equal. This model will be further used to perform dosimetric estimations in humans. Dosimetric estimations in humans using the scaled human [18F]FNI PBPK model were performed and compared with traditional dosimetric scaling methods.
Summary of significant results
▪ MALDI MSI was applied to detect fluoronorimatinib and NeoBOMB1 in tissue; however, drug concentrations were too low for complete distribution studies
▪ MALDI MS imaging study of 31 human GIST samples was performed and is still ongoing. Supervised classifications pipeline was established based on an efficient “big data” analytics framework to identify molecular signatures
▪ GRPR staining of cryosectioned tissues was established and will be coupled with MALDI-MSI
▪ Initial results indicate observable separation between GRPR-positive and –negative MS spectra based on lipidomic foot print in tissues
▪ proof of concept study was conducted to detect gene mutations in KIT and PDGFRA based on minute disturbances in the lipidomic foot print with MALDI-ToF MSI
▪ establishment of immunodeficient xenograft GIST animal models in NSG and Balb/c SCID mice with GIST-T1, GIST882 and GIST430
▪ patient-derived xenografts could not be established in NSG mice
▪ subcutaneous injection of GIST cell lines in NSG mice induced localised tumours without metastazation
▪ orthotopic injection of GIST cell lines in NSG mice induced tumour growth with all cell lines and metastases with GIST-T1 and GIST430 within 7 weeks
▪ intravenous injection of GIST cell lines in NSG mice did not result in tumour growth
▪ synthesis of several precursors for potential radiotracers targeting KIT, NTSR1, GLP-2R, VEGFR and DOG1
▪ syntheses of GLP-2 derivatives labelled with fluorescent dyes for optical imaging
▪ establishment of radiolabelling protocols for 68Ga (DOTA) and 124I
▪ establishment of radiolabelling protocol for [18F]SFB-HPLW
▪ establishment of radiolabelling protocol for [18F]fluoronorimatinib and [18F]fluoro-DOG1
▪ establishment of in vitro assays to determine target affinities and specific binding
▪ [18F]fluoronorimatinib showed similar binding profile as Norimatinib in a kinase assay with 10 KIT mutants (Imatinib sensitive and resistant)
▪ flow cytometry analyses showed presence of GLP-2R on the cell surface of all GIST cell lines
▪ RT-PCR showed highest GLP-2R expression in GIST430; intermediate expression levels were found for GIST-T1 and GIST882
▪ GLP2-Cy5 showed specific binding to GLP-2R positive GIST430 cells in an internalization assay
▪ Fluorescence microscopy showed localization of GLP2-Cy5 on the cell membrane and in the cytosol of GIST430
▪ in vivo evaluation of several radiotracers targeting KIT, GLP-2R and DOG1
▪ ([68Ga]DOTA-Lys34)-(Trp10, Ala12)-GLP-2 (10-33) showed high renal excretion, but slow blood clearance and low tumour uptake
▪ [18F]fluoronorimatinib showed reduced liver uptake and enhanced renal excretion compared to similar radiotracers and similar tumour uptake, but distinction between Imatinib-resistant and –sensitive tumours was not possible
▪ [18F]fluoro-DOG1 showed unfavourable biodistribution in mice due to the lack of stability in murine blood serum
▪ general PBPK model (peptides) for mice and humans was developed
▪ mouse PBPK model for [18F]fluoronorimatinib (small molecule) was developed and fitted to data from 5 xenograft bearing mice
▪ human PBPK model for [18F]fluoronorimatinib was created by allometrically scaling the physiological and anatomical parameters from mice to humans
Preclinical Theranostics
Development of a standard procedure for the radiolabelling of peptides with Ga-68
A standard procedure for 68Ga radiolabelling of peptides with high affinity for somatostatin (SST) and gastrin-releasing peptide (GRP) receptors was developed. In particular, a kit approach for direct radiolabelling with 68Ga acid solution coming directly from the 68Ge-68Ga generator was set up and validated initially for the DOTA-TATE peptide (targeting SST receptors). The same kit approach was then successfully applied also to the novel bombesin analogue NeoBOMB1 (targeting the GRP receptors). This procedure avoids the laborious post-processing of the eluate and should represent a considerable improvement in the labelling of DOTA-peptide with 68Ga both in terms of synthesis time, as well as for the quality of the radiopharmaceuticals (specific activity, radionuclidic and radiochemical purities). By testing different conditions of reaction (pH of complexation, volume of buffer, type of buffer, amount of precursor and final purification procedure), a standard procedure for 68Ga -labelling was set up, validated and applied in GMP conditions.
The main steps of the standard operating procedure for 68Ga -labelling of the NeoBOMB1 peptide are:
• Elution of 68Ga from the 68Ge/68Ga Generator into the vial containing the lyophilised peptide and selected excipients;
• Addition of the reaction buffer solution (provided in the kit), to reach the optimal pH for the coordination of the 68Ga with the DOTA-chelator;
• Reaction for 7 minutes at 95°C;
• Quality control and radioactivity measurement of the 68Ga-NeoBOMB1.
The radiopharmaceutical 68Ga-NeoBOMB1 obtained following this procedure was then characterised and tested both in vitro and in vivo in relevant animal models bearing the GIST tumour. The first GMP batch of the kit for radiopharmaceutical preparation of 68Ga -NeoBOMB1 was manufactured in early 2016, and was used exclusively for assessing the long-term stability of the product at the intended storage conditions. Subsequent GMP-grade batches were manufactured to guarantee supply of the product for clinical use.
In vitro and in vivo evaluation of specificity of synthesised radiopharmaceuticals in GIST animal models
Using relevant GIST tumour models the in vitro and in vivo characterisation of the 68Ga -labelled DOTA-peptides synthesised as described in Task 5.1 was performed. Binding and internalisation studies were conducted using three GIST cell lines (GIST-T1, GIST-882 and GIST-430) characterised by different degrees of sensitivity to the tyrosine kinase inhibitor imatinib. Several peptides (targeting different receptors, all known to be overexpressed in GIST) were initially screened in vitro, however very low or no specific binding to GIST tumour cells was found for all 68Ga-DOTA labelled peptides except for bombesin derivatives. The NeoBOMB1 peptide was therefore characterised further and confirmed to specifically bind the GRPR, shown to be expressed at similar levels in all three GIST cell lines. In vitro displacement assays were also performed, resulting in high affinity of the NeoBOMB1 peptide to the GRPR; additionally, no differences in terms of peptide-receptor binding potency were observed among the three models.
The internalisation properties of 68Ga-NeoBOMB1 were also studied in vitro and the well-known bombesin agonist analogue AMBA was compared with the NeoBOMB1. A clear difference in binding behaviour was found between the two bombesin analogues, reflecting the agonist and antagonist properties, respectively. In the case of 68Ga-AMBA activity was mainly internalised, whereas for 68Ga-NeoBOMB1 membrane bound activity was predominant, confirming the antagonistic nature of the latter bombesin analogue.
Biodistribution properties of 68Ga-NeoBOMB1 were assessed in vivo, intravenously injecting the tracer in healthy mice and following its distribution in main organs over time. The radiolabelled NeoBOMB1 was rapidly cleared from the blood, showing no retention in kidneys. Background radioactivity was observed in GRPR-expressing tissues (mostly pancreas), which however decreased over time, consistently with a GRPR antagonist profile. The ability of the 68Ga-radiolabelled NeoBOMB1 peptide to display the GRPR expressing GIST tumour has been confirmed in in vivo imaging PET studies. The whole-body picture of a representative GIST-882 bearing animal showed good imaging performances, allowing a clear visualisation of the tumour. Uptake in the liver and intestinal area was also observed, due to the partial hepatobiliary excretion of the peptide and to a specific uptake in these areas. Finally, as for many other peptides of the same class/type, renal excretion is responsible for the high uptake in the bladder area. Based on the excellent in vitro and in vivo results obtained with 68Ga-NeoBOMB1 in GRPR-expressing GIST models and on the availability of the GMP-grade pharmaceutical product (kit for the radiopharmaceutical preparation of 68Ga-NeoBOMB1), 68Ga-NeoBOMB1 was selected as the optimal candidate for clinical translation. In the last period of the MITIGATE project, efforts have been focused on the assessment of the 177Lu-labelled NeoBOMB1. Both biodistribution and efficacy experiments were conducted in mice bearing GIST tumour, which confirmed the therapeutic potential of this radiopharmaceutical in GRPR-expressing GIST patients.
Database of mass spectrometry signatures as indicators as acute markers of response to radiotherapy in GIST animal models
The work of HM in WP5 focused on the application of MALDI mass spectrometry (MS) signatures to investigate the response to novel radiotherapeutics in preclinical GIST animal models. Acquisition methods for protein and metabolite MS fingerprints were successfully established and provided in form of standard operating procedures (SOPs). Due to the successful acquisition of new MALDI TOF mass spectrometers during the period of the MITIGATE project which allow for higher acquisition speed and a superior mass resolution, it was possible to measure spatially resolved molecular data using MS Imaging. A preclinical NeoBOMB1 study in a GIST animal model was carried out to assess MS signatures as acute markers of response. Remarkably, all mice responded to the radiotherapy with 177Lu-NeoBOMB1. Therefore, no tumour tissue of treated animals was available for analysis. Consequently, the database to explore therapeutic efficacy of radiotherapy in preclinical studies correlating treatment regimes, MS fingerprints and outcomes was established but not implemented due to the fact that fingerprints of responders and non-responders could not be compared. Instead, MALDI MS imaging was used to analyse potential adverse effects in other tissues such as pancreas. Additionally the spatial distribution of the MITIGATE radiodiagnostics Fluoronorimatinib and NeoBOMB1 by means of MALDI MS imaging was analysed. Due to the low doses applied, detection of the drugs masses was not possible. Instead, spatially resolved molecular MS data was acquired for the first line treatment Imatinib in human GIST samples. Finally, the MALDI MS imaging approach for analysis of response signatures is being translated to resectates from GIST patients.
Modelling, simulation and prediction of human biodistribution and absorbed doses
PET biokinetic data from experiments in healthy Balb/c SCID mice injected with 68Ga-NeoBOMB1 were used to perform dosimetric estimations for humans. For this, the time integrated activity coefficients (TIACs) obtained from the mouse biokinetics were scaled-up to humans based on the mass of the organs and the total body mass, according to the AAPM Report No. 71. The scaled TIACs were then used as an input in OLINDA/EXM software, a program with which the dosimetric calculations were performed. The results of the dosimetric calculations for 68Ga-NeoBOMB1 were comparable to dosimetric estimations reported for other bombesin antagonist substances labelled with 68Ga. The biokinetic data from the experiments in healthy mice were also used to develop a mouse 68Ga-NeoBOMB1 PBPK model. The PBPK model is based on anatomical and physiological information obtained from the mouse experiments and from the literature. In the model three kinds of parameters were used: fixed, Bayes (based on a known distribution), and fitted (without previous information). Data from the experiments (total mouse weight and organ weights) were used as fixed parameters. Data from the literature (organ vascular fractions, organ interstitial fractions, organ permeability surface area products and organ plasma flows) were used as Bayes parameters. Unknown parameters (metabolic rates, internalisation rates and receptor densities) were fitted. The developed mouse 68Ga-NeoBOMB1 PBPK model individually considers the most relevant organs and the organs with high accumulation (liver, kidneys, spleen, intestines, pancreas, muscle, bones, brain, lungs and bladder). The other organs and tissues were grouped together. The model was well supported by the biokinetic data gathered from the experiments with 68Ga-NeoBOMB1 in healthy mice.
Moreover, a human 68Ga-NeoBOMB1 PBPK model was created by scaling-up the mouse 68Ga-NeoBOMB1 PBPK model using allometry. Some mouse parameters were used in the translated human PBPK model such as receptor densities and transfer rates (i.e. internalisation, metabolic rate, etc.). Volumes and flows were considered human specific parameters and were taken from the literature for an average healthy man of 71 kg. Permeability surface area product values per organ mass were scaled from mice to humans using allometry based on the organ mass.
In addition, experiments in tumour-bearing mice intravenously injected with 68Ga-NeoBOMB1 were performed and biokinetic data were obtained. Pancreas biokinetic data from these experiments (not considered in the healthy mice 68Ga-NeoBOMB1 PBPK model) were successfully retrieved using an automated biokinetic-based clustering segmentation method combined with previous knowledge of the kinetics and anatomy. The segmentation process was performed using PMOD v3.8 software. A special strategy had to be developed and applied for the pancreas (limiting organ at risk) since identifying the pancreas from mouse PET/CT images is a complex task due to its location, variable form and similarity to the surrounding organs in a CT or PET image. The developed mouse 68Ga-NeoBOMB1 PBPK model was adapted and fitted to the data from tumour-bearing mice (including biokinetic data of the pancreas) and scaled-up to humans using allometry.
Summary of significant results
▪ The standard radiolabelling procedure based on the kit approach was set-up and validated for 68Ga-DOTA-TATE by exploiting AAA Patent
▪ The standard procedure for 68Ga-labelling developed by AAA with the DOTATATE peptide was successfully applied to the novel GRPR antagonist, NeoBOMB1.
▪ In vitro experiments demonstrated the high affinity of 68Ga-NeoBOMB1 to the GRPR expressed by GIST cell lines, as well its low internalisation properties, which is consistent with the antagonist nature of this peptide.
▪ Suitable pharmacokinetics and imaging properties by the 68Ga-NeoBOMB1 were confirmed in GIST animal models.
▪ Confirmation of the therapeutic potential of the 177Lu-NeoBOMB1 in GRPR-expressing, GIST-bearing animal models.
▪ A database structure for the therapeutic effect of radiopharmaceuticals in non-tumour tissues correlating treatment regime, MS fingerprints and outcomes is being developed.
▪ Dosimetric calculations for humans were performed based on healthy mouse biokinetic data
▪ A mouse 68Ga-NeoBOMB1 PBPK model was developed and fitted to experimental data from healthy mice
▪ A scaled human 68Ga-NeoBOMB1 PBPK model was created
▪ Mouse pancreas biokinetics was assessed by using a developed automated biokinetic-based clustering segmentation approach.
Translational Studies
Planning of study, initiating study, monitoring of study, data evaluation
In September 2015 an Interim Meeting of the involved MITIGATE partners was initiated to decide on the selection of the compound to be translated into the clinic for molecular imaging in patients with advanced GIST tumours. In the meeting all radiopharmaceuticals developed so far or in development were reviewed and there was a uniform consensus that by far the best candidate for clinical translation was 68Ga-NeoBOMB1 with excellent GIST targeting properties in vitro and in vivo and the basis of pharmaceutical development by one of the involved partners.
The aim of the clinical trial with 68Ga-NeoBOMB1 is to assess the safety, tolerability and tumour targeting properties of 68Ga-NeoBOMB1 in patients with GIST, as a combined phase I/IIa first-in-human study. The study was designed to minimise patient burden caused by the assessment procedures. Only patients with advanced GIST – meaning metastatic disease in the liver or other intestinal organs – with progressive disease (i.e. failure of 1st, 2nd or 3rd-line tyrosine-kinase-inhibitor [TKI] therapy) were included. This population is also the most likely to benefit from future development of therapeutic concepts revolving around NeoBOMB1. The final study design was aimed to keep the radiation dose for patients as low as possible to achieve the study aims. Two subgroups of patients were defined: For the first six patients, safety and tolerability was the main objective, with measurements of physiological parameters, and repeated PET-scans to obtain data on the biodistribution of the radiopharmaceutical as basis for the absorbed dose (exposure) in the patients. The second phase – involving another 5 to 6 patients – was focussed on the targeting properties of 68Ga-NeoBOMB1 in GIST tumours to further demonstrate the benefit of 68Ga-NeoBOMB1 as a novel diagnostic tool. Besides the clinical investigation plan, also a number of other documents was prepared for the submission to both the Ethical Committee and the Austrian drug authority. This included patient informed consent forms, Investigator Brochure for the physician conducting the trial, standardised operating procedures for various procedures such as radiolabelling of the drug, plans for monitoring of the study and many more. Furthermore, insurance coverage was obtained for all participants.
The clinical trial MITIGATE-NeoBOMB1 “A Phase I/IIa study to evaluate safety, biodistribution, dosimetry and preliminary diagnostic performance of 68Ga-NeoBOMB1 in patients with advanced TKI-treated GIST using positron-emission tomography/computer tomography (PET/CT)” was registered in the EU database for clinical trials (EudraCT 2016-002053-38). Submission of the Clinical Investigation Plan and other relevant files to the Ethics Committee was performed in May 2016 and the final positive ethics vote was issued on the 5th of August 2016. In parallel, in July 2016, the official submission was sent to the competent authority, the Austrian Federal Office for safety in healthcare (BASG)/Agency for Food Safety (AGES). End of November 2016, the final approval of the authority was finally granted.
Preclinical studies for translation of radiopharmaceutical development into the clinic
The developed kit formulation and lyophilisation process for the industrial manufacturing of a GMP kit for the radiopharmaceutical preparation of 68Ga-NeoBOMB1 was used. The kit is composed of 2 vials, one containing the active substance plus excipients, to be reconstituted with 68Ga eluted from a 68Ge/68Ga generator and one vial containing the buffer solution to be added to achieve radiolabelling. The radiolabelling procedure was qualified and the formulation was successfully tested. All preclinical data of 68Ga-NeoBOMB1 required for conducting a clinical trial were developed.
First in human clinical trial
The Clinical Trial was formally initiated in December 2016 and the first patient was successfully recruited in January 2017. In the initial part of the study dynamic PET scans in three patients with advanced GIST – most of them resistant to TKI treatment – were included in the study. All study related investigations could be performed as planned. In particular safety monitoring parameters were included, whereby no severe adverse events due to the administration of 68Ga-NeoBOMB1 were found. Physiological accumulation of 68Ga-NeoBOMB1 was observed most strongly in the pancreas. Rapid renal clearance was observed. Tumour enhancement increased over the time course of the study giving high contrast images at later time points. Pharmacokinetics` investigations revealed a high metabolic stability of 68Ga-NeoBOMB1. A report detailing these initial results was again evaluated by the Ethical Committee of The Medical University Innsbruck. The Ethical Committee had no objections to continue the clinical trial. Based on this vote, patient recruitment went on and three more patients could be included until September 2017 confirming the findings in the first three patients.
All these first data indicate a very good safety profile of 68Ga-NeoBOMB1 showing excellent image quality in PET. The first subgroup of patients was completed in September 2017 and the study is currently moving into its second phase (Phase IIa) to provide additional data on specific receptor targeting in GIST patients. Confirmation of GRP receptor status on histological samples is ongoing, required for inclusion of patients in this Phase. Study completion is planned in the first quarter of 2018. Overall the study is expected to provide the basis for further development of the promising novel imaging agent for the benefit of GIST patients in an advanced stage of disease, but also to open other applications in oncological patients.
Proof-of-concept for multimodal treatment
The potential inclusion of patients for a minimally-invasive therapeutic approach, based on loco-regional stereotactic radiofrequency ablation of tumour lesions of patients with advanced GIST was attempted in parallel with the clinical trial. This approach could offer patients with localised, yet advanced disease a therapeutic option in the form of radiofrequency ablation based on multimodality imaging approach. So far none of the patients was clinically suited to benefit from such a loco-regional therapy. This multidisciplinary approach is continuing to select patients for this novel multi-modality treatment approach in GIST.
PBPK Optimisation of Individual Patient Biokinetics and Dosimetry
Biokinetic data from animal experiments in mice were used to develop a PBPK (Physiologically based pharmacokinetics) mouse model. Based on this a scaled human [68Ga]NeoBOMB1 PBPK model was created taking into account anatomical and physiological parameters from the literature.
High accumulation was seen in tumour and pancreas, as expected, whereas liver and kidneys did not present considerable accumulation. The scaled human [68Ga]NeoBOMB1 PBPK model is currently being validated and, if necessary, adapted using biokinetic data retrieved from PET/CT images from 5-6 patients injected with [68Ga]NeoBOMB1. The validated human PBPK model will be used to optimise the therapeutic index for each patient for potential translation of the imaging data towards a therapeutic option for patients, in which NeoBOMB1 labelled with a therapeutic radionuclide such as Lutetium-177 will be used. To optimise the therapeutic index, different scenarios (varying peptide amount, activity and administration schedule) will be simulated in silico using the PBPK model. The results of these simulations will be used to perform dosimetric calculations to define conditions for the highest therapeutic index in patients.
Summary of significant results
▪ 68Ga-NeoBOMB1 was selected for clinical translation and molecular imaging in GIST tumour patients and a pharmaceutical formulation for the radiopharmaceutical developed.
▪ A Clinical Trial Application for a first in human clinical trial with 68Ga-NeoBOMB1 was prepared according to EU regulations and approved by ethics committee and competent authority according to EU regulations in 2016.
▪ A clinical trial with this novel molecular imaging approach for GIST patients was initialised end of 2016 and the first patient was enrolled in January 2017.
▪ The first part of the clinical trial focussing on safety and pharmacokinetics was successfully completed. Initial results indicate an excellent safety profile of 68Ga-NeoBOMB1 in patients.
▪ Tumour visualisation by PET/CT of GIST tumours was achieved in these patients.
▪ Data for potential proof-of-concept of multimodal treatment were generated
▪ Biokinetic data for PBPK modelling and dosimetry were obtained, dosimetric predictions for [68Ga]NeoBOMB1 based on mouse and patient data were performed.
Minimally Invasive Therapy
Protocol and treatments design
The objective of this task was to design a concept to investigate different combinations of treatment and have the following protocol approved: patients with GIST receive multiple radiologic examinations to detect and follow up the disease status. In case progressive disease is seen on conventional imaging techniques (CT, MRI, 18FDG PET/CT), patients will receive a PET-CT examination with the new tracer NeoBOMB1. Unclear findings will be re-evaluated using MRI imaging and if necessary, an image guided biopsy (CT or ultrasound) will be performed. This approach helps to avoid under- and over-treatment. Patients will be treated by current state of the art endo-radiotherapy or image guided minimally invasive treatment options.
Therapy planning
Tumour volume definition based on new tracer: The main objective of this task is to develop a strategy and the tools to integrate the new PET-CT tracer NeoBOMB1 into the combination of treatments. Therefore preliminary animal PET CT was used to evaluate tumour volume definition based on the new tracer NeoBOMB1. An additional objective is the development of an assisting device for minimally invasive treatments. The purpose of the assisting technology is to support physicians performing minimally invasive treatments in planning, execution, and monitoring of the intervention. The assisting manipulator consists of a robotic unit with modifications tailored to the interventional task. Preliminary animal data suggest that the new PET tracer NeoBOMB1 enables us to escalate the dose to vital tumour cells even more. We expect increased tumour control and less damage to risk structures. Improved tumour volume definition can be used to adapt target volume for other treatments like minimally invasive ablations or radioactive seed implantation.
Evaluation of the manipulator system: The purpose of the assisting technology is to support physicians performing minimally invasive treatments in planning, execution, and monitoring of the intervention. To meet the requirements of a complex minimally invasive treatment the following concept was developed.
▪ A software tool allowing the user to combine and view medical imaging data and to select target points and entry points for the needle-based treatment.
▪ A manipulator or robot to provide positioning and orientation of a needle guide. The KUKA LBR iiwa MED is selected for its design purpose: a component for medical products.
▪ A registration method to determine the robot position and orientation in image coordinates.
▪ A communication protocol to send target and entry points from the planning software tool to the robot.
The evaluation of the manipulator systems shows promising results in terms of precision and intervention time. These characteristics have the potential to render it a useful tool in minimally invasive treatments.
Summary of significant results
▪ Positive vote of ethics committee for minimally invasive treatment of GIST metastases based on Ga-NeoBOMB1 PET-CT
▪ Promising preliminary animal results concerning tumour volume definition for therapy planning
▪ The evaluation of the manipulator systems shows promising results in terms of precision and intervention time. These characteristics have the potential to render it a useful tool in minimally invasive treatments.
▪ Improved tumour volume definition can be used to adapt target volume for other treatments like minimally invasive ablations or radioactive seed implantation
Multimodal therapy control
Work package 8 aims at developing multi-modality imaging protocols, MRI sequences and hardware for the MRI-based functional assessment of GIST tumour microenvironment and for evaluating the therapeutic effect on GIST tumour murine models upon imatinib treatment. Several academic institutions (UNITO, UHEI, CKM) and commercial partners (RAPID, CAGE, AAA) were involved in this work-package. Imaging-based approaches are promising approaches to detect early response to Tyrosin Kinase Inhibitors (TKI) in GIST patients. Tomographic imaging modalities, such as MRI and CT, provide anatomical volumetric information that can report on tumor volume changes upon treatment. These metrics are commonly exploited for assessing treatment response in longitudinal studies. However, assessment of therapy response that occurs at 3-month intervals is hampered by null or minimal decrease in tumor size, thus not providing useful clinical information. Therefore, alternative imaging approaches as developed within WP8 should improve the diagnosis and assessment of GIST therapy, hence their development is of relevant importance.
At preclinical level several MRI-based approaches and multi-modality imaging techniques (MRI and PET) were investigated for assessing GIST tumor microenvironment in murine tumor models, whereas at clinical level dedicated multimodal coil for 23Na/1H imaging was developed and tested in human volunteers and GIST patients. At preclinical level, one of the objectives of WP8 was to evaluate the ability of functional MRI-based approaches to highlight differences in tumor microenvironment properties related to imatinib resistance in GIST murine models. We investigated GIST tumor vascularization using a Dynamic Contrast Enhanced-MRI approach in combination with a blood pool contrast agent at low-magnetic field.Functional MRI imaging was performed on immundeficient mice bearing three different GIST tumors (imatinib-sensitive GIST-T1 and GIST882; imatinib-resistant GIST430). By applying a pharmacokinetic model, functional derived parameters informative of vessel permeability (Ktrans) and plasmatic volumes (vp) were obtained. Our findings demonstrated that DCE-MRI can detect differences in plasmatic volume and vessel permeability among the investigated GIST tumor cell lines. In particular, GIST430 tumors display more than twofold higher Ktrans and vp values compared to imatinib-sensitive tumors. Increased Ktrans and vp suggested a more unstructured and deregulated vasculature in terms of blood flow in imatinib-resistant tumors. In addition, our in vivofunctional findings were confirmed ex-vivo by histological quantifications of endothelial vessels and permeability (by observing the extravasation of labeled-dextran). This study indicates that characterizing the tumor microenvironment and vasculature of GIST tumors using a functional MRI-based approach can allow imatinib-responsive tumors to be discriminated from imatinib-resistant ones. Moreover, assessing angiogenesis by functional MRI approaches may provide alternatives to conventional imaging modalities for the early detection of tumor response.
A second objective of WP8 was to investigate whether changes in tumor microenvironment properties as measured by MRI could assess imatinib response in comparison to standard approaches in GIST models. MRI offers the unique possibility to in vivo characterize vascularization, acidosis and cellularity within the same modality. For this purpose, imaging-based protocols were established, in addition to anatomical information (T2w-MRI), based on functional MRI procedures to characterize vascularization (DCE-MRI), acidosis (pH-CEST imaging) and cellularity (DWI) in combination with PET for assessing metabolism (18FDG-PET) in imatinib-sensitive (GIST-T1 and GIST882) and imatinib resistant (GIST430) preclinical murine models upon imatinib treatment. Mice were treated with imatinib for 14 days and imaging acquisitions were performed before and after 1 week and 2 weeks of treatment. Our results showed that all the proposed functional imaging approaches were able to detect changes between treated and untreated tumors. Briefly, T2w-MRI images showed that imatinib-sensitive GIST882 and GIST-T1 tumors respond to imatinib treatment by showing a decrease in tumor size in comparison to the untreated group, whereas tumor volume increases for both treated and untreated imatinib-resistant (GIST430). All treated GIST tumors show decreased FDG uptake (FDG-PET) and increased vessel permeability (DCE-MRI), reflecting metabolic changes and normalization of tumor vasculature upon imatinib treatment, respectively. In addition, ADC values (DWI) increased in imatinib-sensitive GIST-T1 and GIST882 tumors due to cytotoxic effect of imatinib. Finally, imatinib-sensitive tumors showed a decrease of tumor acidosis (pH-CEST). Summarizing the obtained results at preclinical level, we can state that investigating GIST tumor microenvironment will be of relevant importance for improving GIST diagnosis and treatment. In addition, the proposed MRI functional imaging approaches can be considered promising alternatives to standard clinical criteria.
Additionally, within WP8 a dedicated molecular diagnostic agent capable of detecting cell-induced apoptosis was assessed for its potentialto be used in GIST patients to monitor effectiveness of the employed treatments.The tracer is based on a recombinant version of wild-type Annexin V, called rhAnnexin V-128. This recombinant protein contains an endogenous chelation site for technetium (Tc-99m) and it can be used to detect apoptosis and cell death by SPECT imaging.The product is a single vial, lyophilized cold kit, which is stored at refrigerated temperature.The procedure for the kit radiolabeling with technetium is user-friendly and standardized. Briefly, the lyophilized powder is reconstituted directly with the eluate from a technetium generator. Reaction takes 1 and an half hours, leading to a final injectable product with considerably highradiochemical purity (above 95%) and stability (up to 4-6 hours after end of radiolabeling).As a proof of concept, the 99mTc-rhAnnexin V-128 was successfully used in a pancreactic tumour model (Panc-1 xenografted in nude mice), as tracer to assess the efficacy of the therapeutic treatment.
A third objective of WP8 was to develop a multimodal MRI coil for 23Na imaging. A 23Na coil array setup for investigation of the human abdomen was designed and tested on phantoms. It yields good 23Na Tx efficiency and homogeneity, good 23Na Rx sensitivity and allows for 1H imaging by providing 6 1H Rx elements while allowing the 1H body coil being used for 1H Tx. The coil was, even if being a one-of-a-kind prototype, CE certified as a medical device for allowing patient scanning. With the development of RF arrays for X nuclei (i.e. other nuclei than 1H) RAPID is now capable to provide a technology which is currently not utilized in high numbers. However, there are promising applications coming up, including 23Na imaging for GIST treatment as well as others which were not subject to this project (e.g. hyperpolarized 13C Metabolic Imaging for tumor detection and treatment). RAPID is able to provide RF coil arrays for X nuclei even on a level of very small sales numbers (one-of-a-kind coils) with medical device quality which enables research groups all over the world to further develop these technologies on volunteers and in patient studies and, such, bring them to the state where they will enable new treatment solutions in clinics. With this project, RAPID has strengthened its presence and position in the field of dedicated RF coils for MRI. In the international competition, RAPID is and stays a well established provider of such coils and is visible to both, researchers and MR system vendors as the first address for finding customized or low number RF coil solutions.
A fourth objective of WP8 is the translation of 23Na imaging in clinical studies.
During the Mitigate project, CKM (UHEI) has optimized a multimodal 23Na/1H MR imaging protocol for abdominal and body stem applications in GIST patients, which is specially adjusted to the new built double resonant coil by RAPID and was evaluated in volunteers. One GIST patient was scanned with another previously designed sodium coil by RAPID and the new imaging protocol. A second GIST patient received multimodal measurements without sodium imaging. Several steps have been conducted until designing the GIST protocol:
- Correction methods for the inhomogeneous transmit field (separate scans) have been evaluated, optimized and included into the final protocol (double angle method and phase sensitivity approach).
- A correction for the inhomogeneous receiving field due to the sensitivity profile of the phased array coils used in the newly built coil was proposed, which is completely done in post-processing and is based on ensemble empirical mode decomposition.
- The post-processing of the images was improved:
o Gridding and filtering of data was adapted.
o A post-processing pipeline designed for the GIST protocol was developed.
o Denoising of images based on total variation was implemented.
o Receiving field inhomogeneities were corrected in post-processing.
o Inhomogeneous transmission fields were corrected using the additional correction scans.
o Sodium images were translated to DICOM format.
o Spatial combination of proton and sodium images was conducted for easier incorporation of anatomical data.
The protocol and hardware were tested stepwise on 16 volunteers. Ten volunteers were scanned in the region of lower body stem, four in abdominal region and two in upper body stem region. Sodium MR images provided good quality but signal decreased too fast towards body center. This leads to high standard deviations of tissue sodium concentration in deep located tissue, where distances from the coil surfaces are larger than 10 cm. Therefore, coil design was further discussed with RAPID and promising changes for improvements were set. Hardware changes were started but could not be finished by now. We expect the upgraded coil design to enable us to improve signal to noise ratio especially from deeper located tissue and regions with a large distance from the sodium coils. Two GIST patients received a multimodal imaging. One patient received a dual CT, MR morphological and functional sequences as well as sodium imaging. Patient 2 obtained dual CT and proton MR imaging which was functional as well as morphological. Tumors could be displayed and evaluated.
Summary of significant results
• MRI-based approaches and multi-modality imaging techniques (MRI and PET) were investigated for assessing GIST tumour microenvironment in murine tumour models
o With a functional MRI-based approach imatinib-responsive tumours can be discriminated from imatinib-resistant ones
o Assessing angiogenesis by functional MRI approaches may provide alternatives to conventional imaging modalities for the early detection of tumour response
• Dedicated multimodal coil for 23Na/1H imaging was developed and tested in human volunteers and GIST patients
Potential Impact:
Description of the potential impacts
The highly complex research carried out within MITIGATE lead to a variety of project results and outputs. Each of these achievements will strongly impact GIST patients, improving their personalised diagnosis and treatment. MITIGATE findings may in many cases be transferable to other cancer types (e.g. lung cancer) for even higher socio-economic impact. Our established GIST-biopsy process involves a novel endoscopic biopsy device, biological upstream processing systems, a bespoke immunomagnetic single cell separation procedure and innovative mass spectrometry approaches. It will ensure high-quality molecular analysis of the tumour samples. Through the development and application of the GIST subtype database and GIST subtype tumour animal models, combined with the possibility to predict TKI-resistant GIST behaviours, personalised selection of an appropriate PET probe or endoradiotherapeutic tumour treatment was achieved. The whole procedure will result in an accelerated decision making process for the treatment of individual patients. AAA has developed an SOP for the radiolabelling of peptides with 68Gallium using a proprietary kit approach that is successfully applied to the NeoBOMB1 peptide, the selected probe for the clinical trial. The establishment of 68Ga-NeoBOMB1 PET-CT for molecular phenotyping in GIST tumours enables AAA to develop a novel diagnostic approach for GIST patients. The therapeutic efficacy of the 177Lu-NeoBOMB1 was confirmed in pre-clinical GIST models, making it a new potential treatment option for GIST patients. A dosimetry study was performed on GIST patients receiving the diagnostic agent 68Ga-NeoBOMB1. Using an already validated 68Ga-NeoBOMB1 PBPK model, these results will also allow a prediction of the compound’s behaviour for individualised targeted therapy approaches. MALDI MS imaging was found to differentiate between GRPR-positive and –negative GIST based on the lipidomic footprint which also allows further characterisation of gene mutations of human GIST resectates. To generate valid animal models for metastasizing GIST, orthotopic inoculation proved to be the most reliable method. The evaluation of radiotracers targeting NTSR1, GLP-2R, KIT and DOG1 showed that these were not suitable for imaging of GIST in mice, but a PBPK model for [18F]FNI was developed and is currently transferred to a human model to estimate dosimetry. Also, a fluorescent probe targeting GLP-2R for a multimodal approach is currently under investigation for enhanced visualisation of GIST. Results of the clinical study showed a high safety profile and highly promising results in specific molecular targeting of GIST tumours in PET/CT using 68Ga-NeoBOMB1. With this the initial phase providing safety and pharmacokinetic data was completed. Our patient data are the basis for further pharmacokinetic modelling and dosimetry calculations and to select patients benefitting from multimodal treatment. Endoradiotherapeutic treatment can be supported by minimally-invasive interventions guided by molecular PET imaging. Accordingly, novel, PET-CT-based target volume definition for external beam radiotherapy based on 68Ga-NeoBOMB1 were established. The high-risk areas within the tumour volume can be used for an intensified treatment. The evaluation of preliminary animal PET-CT data for EBRT therapy planning indicate that the new PET-tracer allows escalation of the dose to vital tumour cells, thereby further increasing tumour control probability and proves its significance for the minimally invasive interventions. A multimodal protocol for treatment monitoring has been investigated by CAGE and UNITO, showing that several functional imaging modalities, besides clinical ones, can assess the treatment response to imatinib. Rapid and UHEI have validated a new multimodal coil (23Na/1H) for assessing anatomical and metabolic information at clinical level, thus providing a new technology based on 23Na imaging for GIST treatment in clinical routine. By the electronic website enrolment, access to potentially more precise and less toxic treatment is provided for oligometastatic GIST patients throughout Europe
Main dissemination activities and exploitation of results
MITIGATE featured a separate Work Package dedicated to the effective dissemination and exploitation of the Foreground. The dissemination activities were based on the Dissemination plan. Exploitation activities were and are based on the intermediate and final exploitation plans. These plans set out the strategy and activities that were undertaken during the project’s lifetime, as well as exploitation activities that will continue beyond the lifetime of MITIGATE.
Communication and dissemination
The project first developed the overall dissemination strategy. The dissemination plan lists target groups and contacts from the scientific community, specific clinical disciplines, medical centres, healthcare provides, patient organisations and the general public. For each group it was specified how it will be addressed, for instance through attendance of conferences, the MITIGATE project website, mailing lists, workshops etc. It also included a list of main conferences of interest for presenting MITIGATE results. Given that the plan was developed and submitted at the beginning of the project, it remained a flexible, living document that was continuously adapted to the demands of the project and the evolving demands and interests of the stakeholders.
The overall dissemination objectives were to
• share new insights in GIST treatment with patients, patient’s advocacy groups and the general public.
• ensure project conclusions are considered by the International Rare Disease Research Consortium (IRDiRC) for fostering international collaboration in rare diseases research.
• validate that project developments are designed and adapted to better respond to patients’ needs.
• increase awareness among healthcare providers to facilitate rapid update of project outcomes into patient care, realising the value of advancements made during MITIGATE.
• inform the international scientific community active in the field of GIST about the objectives of the project and subsequent advancements made to the state of the art in GIST diagnosis and treatment through MITIGATE.
To realise the project objectives, the dissemination of information about MITIGATE was directed to the appropriate channels, from scientific publications to general media. Execution of the dissemination strategy reflected the following questions:
1. What information do we have to disseminate? (e.g. general project information, novelties in technology and tools, on-going project work and interim results, state-of-the-art of scientific questions related to GIST, etc).
2. To whom do we need to disseminate the information? (EIBIR members, partners’ contacts and networks, national and internal authorities and associations, industries, etc).
3. How will we disseminate the information? (printing articles in media, brochure and attendance at conferences, online articles to be published on websites and external web forums, media briefing, position papers, press campaign, etc).
4. When do we disseminate the information? Each dissemination activity should be followed by a time estimate if no exact dissemination schedule is provided.
5. What do we hope to achieve by disseminating this information? (what is the intended impact, what reaction or change is expected from the target audience?).
The dissemination strategy identified the unique needs of each stakeholder group, and the activities employed corresponded to these needs, ensuring effective and efficient distribution of project information.
A corporate identity was created for the MITIGATE project to facilitate clear and consistent communications. It included a logo, colour scheme, presentation slides and the overall look for the website. The project aimed to generate a brand image that could be used to promote awareness of the project and its activities. The logo and the MITIGATE layout were used on all documentation, promotional/publicity literature, exhibition boards and on the project website, etc.
The website (www.mitigate-project.eu) presents the project and its results to a worldwide audience. The website has been online since December 2013. It provides an overview of the project and the objectives (also per Work Package), information on partners and about the project progress. Non-restricted project deliverables, press items and articles have been made available on the website in the publications and media section. The website summarises the main impacts of MITIGATE for patients, researchers and clinicians separately. Scientific and technical achievements, return on investment, i.e. demonstrating the importance of research spending at European level, are also highlighted. Updates to the MITIGATE website were then promoted via social media through EIBIR’s Twitter account.
The following dissemination material was produced:
• Folders
o MITIGATE folder to introduce and promote the project
o Updated MITIGATE folder to summarise and highlight achievements of the project
o Information folder for patients (in English, German, Bulgarian)
• Four newsletters
o 1st MITIGATE newsletter was distributed in September 2014 to MITIGATE stakeholders by email and published on www.mitigate-project.eu as well as on websites of participating organisations (D2.2). It was also distributed by EIBIR at the European Congress of Radiology 2015.
o 2nd MITIGATE newsletter, distributed in September 2015 to MITIGATE stakeholders via email and published on the project website. It was also distributed by EIBIR at the European Cancer Congress 2015 and at the European Congress of Radiology 2016.
o 3rd MITIGATE newsletter was, distributed in September 2016 to MITIGATE stakeholders via email and published on the project website. The 3rd Newsletter was also distributed by EIBIR at the European Congress of Radiology (March 2017), the German Röntgen Congress 2017 (May 2017) and the Congress of the European Society for Radiotherapy and Oncology (May 2017).
o 4th MITIGATE newsletter, distributed in September 2017 to MITIGATE stakeholders via email and published on the project website.
• Informational poster to display at conferences and congresses
• Introductory roll-up to display at conferences and congresses
MITIGATE held an innovation workshop, titled “Innovations for diagnostic imaging and minimally invasive treatment: ready for the market” at the 98th German Röntgen Congress in Leipzig, Germany on May 26th 2017 (approx. 6500 attendees from academia, hospitals and industry). MITIGATE also had a booth during the whole congress duration for appropriate on-site advertisement. The MITIGATE Innovation Workshop was attended by representatives from universities, hospitals and industry. Each talk was followed by an engaging discussion, that was also continued at the MITIGATE booth. Attendees highly appreciated the presented innovations made by the project and endorsed their translation to the market.
Overview of dissemination activities
• 22 scientific publications in high-ranking, relevant, peer-reviewed journals.
• 18 articles in popular press
• Sessions and representation at the European Congress of Radiology (ECR)
o ECR 2014: Molecular imaging and targeted image-guided therapy in gastrointestinal stromal tumours
o ECR 2015: MITIGATE Consortium: State of the art imaging and therapy in GIST
o ECR 2016: MITIGATE: What does it take to perform clinical trials in interventional radiology?
o ECR 2017: Innovative solutions for diagnosis and treatment concepts for GIST patients from the MITIGATE project
o Articles in the congress newspapers “ECR Today” (print of 26,000)
▪ ECR 2014
▪ ECR 2015
▪ ECR 2016
▪ ECR 2017
o Representation and distribution of dissemination material through the EIBIR booth
• MITIGATE was presented and promoted at networking events (e-rare workshop; CommHERE workshop)
• MITIGATE Joint Scientific Symposium: ADVANCES in THERANOSTICS of RARE CANCER, at the Medical University Innsbruck, Austria
• Promotion and dissemination through EIBIR social media
• Workshop “Innovations for diagnostic imaging and minimally invasive treatment: ready for the market” at the German Röntgen Congress 2017, 26 May 2017
• Presentations at scientific events
o 54th Annual Meeting of the German Society for Nuclear Medicine
o 5th Annual Symposium of the Oncology Network of West Austria and South Tyrol, 27 May 2015
o Multiple presentation at the 23rd ISMRM annual meeting, 3-4 June 2015
o Deutscher Krebskongress, 24-27 February 2016
o German Sarkomkonferenz, 17 March 2016
o Multiple presentations at events of the German Society for Medical Physics
o Four presentations at the EANM Annual Meeting 2016, 15 October 2016
o OurCon IV conference in imaging mass spectrometry, 17 October 2016
o Mass Spectrometry and Proteomics Congress 2016, 15 November 2016
o Invited talk about 0D- to 4D-Mass Spectrometry in or near the Operating Room at University of California, Los Angeles, 18 November 2016
o Presentation about MALDI MS imaging in a clinical context at the BW-CAR Kickoff-meeting, 30 November 2016
o EMBL target validation conference, 6 December 2016
o Medizintechnologie made in Mannheim event, 8 December 2016
o Invited talk about at MALDI mass spectrometry fingerprinting and imaging in pharma industry and in the clinic at Leipzig University, 9 December 2016
o 16th meeting of the Norwegian Society for Mass Spectrometry, 23 January 2017
o Deutsche Gesellschaft für Endoskopie und Bildgebende Verfahren Kongress 2017, 7 April 2017
o Annual Meeting of the Canadian Association of Nuclear Medicine 2017, 21 April 2017
o Annual Meeting of the ISMRM 2017, 27 April 2017
o International Symposium on Radiopharmaceutical Sciences 2017, 19 May 2017
o SNMMI Annual Meeting 2017, 10 June 2017
o DGN Summer School 2017, 1 September 2017
o 14th UHEI Summer Workshop, 2 September 2017
• 17 scientific posters
o European Molecular Imaging Meeting 2015, 2016
o World Molecular Imaging Congress
o Deutscher Krebskongress 2016
o Jahrestagung der Deutschen Geselleschaft für Medizinische Physik 2015, 2016
o Jahrestagung der Deutschen Gesellschaft für Nuklearmedizin 2015, 2016
o 13th South East Asia Congress of Medical Physics
o EANM Annual Meeting 2016
o ISMRM 2017
o SNMMI Annual Meeting 2017
o EMBL 18th PhD symposium
• Two exhibitions
o Medica 2016, 14 November 2016
o RSNA 2016, 26 November 2016
• Distribution of the MITIGATE dissemination material and representation at other scientific events
o European Cancer Congress, 25-29 September 2015
o Congress of the European Society for Radiotherapy and Oncology, May 2017
• Engagement with patient advocacy groups (PAGs)
o Judith Robinson, GIST Support UK
o Amy Bruno-Lindner, GIST Support Austria
o Feedback about the MITIGATE folder for patients, distribution though PAGs at PAG events.
o Feedback on the MITIGATE Clinical Study Patient Recruitment Form
o Presentations at PAG events
▪ Patient Forum on GIST, Das Lebenshaus, Mannheim, Oct 8, 2015
▪ NEW HORIZONS GIST Conference (Sitges/ES, May 2016)
• Promotion and dissemination through EIBIR social media
Exploitation
In the first project period the project developed an intermediate exploitation plan. This plan detailed how partners intend to exploit the results during the project and how exploitation will continue at the end and shortly after the project. The project’s Clinical Innovation Advisory Board provided expert advice on exploitation planning throughout the lifetime of the project and highlighted the importance of analysing and specifying the market as far and detailed as possible, in quantitative and qualitative aspects. To this end, MITIGATE launched an end user survey to assess the market potential, market access and market environment from MITIGATE’s exploitable results. The survey was promoted and distributed through the MITIGATE newsletter, at the European Congress of Radiology 2016, and through social media. A final exploitation plan consisting of individual exploitation plans and activities for the key exploitable results was developed in the final period of the project.
Overview of the project’s key exploitable results and the exploiting partners:
RF Coil Arrays for X Nuclei
RAPID and UHEI developed an RF coil package for doing MR imaging and spectroscopy on the human body. The RF coil package consists of a 23Na transmit coil for exciting the abdominal region, a 23Na receive array for getting the best possible signal-to-noise ratio and an an integrated 1H receive array for achieving clinical 1H imaging performance. The RF coil technology is not limited to 23Na but works with other nuclei as well, such as 13C. The RF coil array shall be exploited by manufacturing and selling the coil through RAPID. RAPID will use its own sales force and sell the RF coil package to researchers directly.
Tissue Grinder
FHI has developed an innovative tissue dissociation module (tissue grinder) which allows the dissociation of tissue samples into single-cell suspensions with lower enzyme concentrations. This device enables the isolation of a higher number of intact cell surface markers and subsequent analysis. After securing the intellectual propetry, the technology will be licensed to an industry partner who will manufacture and sell the product. To achieve this, FHI will have to contact relevant industrial parties. FHI has successfully applied for other grants to further develop the tissue grinder technology and has started the FHI internal project “MAVO LyDia HD”. FHI has also applied for a patent for the tissue grinder under application number 10 2016 216 345.0 which is expected to be granted shortly.
Magnetic Separation Technology
SCT developed a methodology to generate a single cell suspension from three main heterogeneous GIST enriched tissues and to enrich GIST cells from these tissue types using an immunomagnetic isolation system. The isolation kit is based on two main components: one vial of isolation cocktail and one vial of paramagnetic particles. SCT expanded this technology to isolate other cancer cells. SCT owns to intellectual property, but also assesses OEM (original equipment manufacturer) products and licensing opportunities identified by its business development department for the exploitation of the magnetic separation technology. SCT will market the GIST and non-GIST cancer cell isolation reagents, and intends to establish collaborations with academia for the further development and exploitation of these products. A targeted marketing effort will have to be made towards cancer related institutions and laboratories.
Robotic Assistant for needle placement
FHI has, together with UHEI, developed and successfully tested in a preclinical setting, a robotic assistance system for interventional needle placement. The system consists of a navigation software and a robotic guidance device, which together enable the fast and precise placement of needles under image guidance. The system is slightly more precise than the manual needle placement and significantly reduces the procedural time. FHI has presented the technology at RSNA 2015, Medica 2016 and RSNA 2016 in order to convince industry partners to join FHI’s efforts to make the product ready for the market. FHI has found a partner, who is willing to bring the assistance device to market. Together with an industry partner, the technology will be developed further and then brought to market by the industrial partner, licensed by FHI.
Kit for radiopharmaceutical preparation of 68Ga-NeoBOMB1 for diagnosis and/or tumour assessment in patients with GRPR-positive malignancies, including GIST
AAA, with the support of MUI, developed a GMP-compliant kit for radiopharmaceutical preparation of 68Ga-NeoBOMB1 to be used for diagnosis and/or tumour assessment in patients with GRPR-positive malignancies, including GIST. AAA is also pursuing opportunities to develop a kit for the therapeutic counterpart (177Lu-NeobOMB1) as a theragnostic option. AAA has extensive expertise and knowledge in bringing new radiopharmaceuticals to the market and aims at pursuing marketing authorisation for NeoBOMB1 in GIST as well as in other potentially interesting oncological indications where GRPR expression is confirmed. On 12 December 2016, orphan designation (EU/3/16/1794) was granted by the European Commission to AAA for 68Ga-NeoBOMB1 for the diagnosis of gastrointestinal stromal tumours. Orphan designation for the therapeutic counterpart, 177Lu-NeoBOMB1, will be pursued by AAA as well. MUI is currently using the AAA kit in a phase I/IIa clinical trial in TKI-resistant GIST patients to evaluate safety, biodistribution, dosimetry and preliminary diagnostic performance of 68Ga-NeoBOMB1.
Mouse models for GIST implants
UNITO developed imatinib-resistant and sensitive GIST mouse models. These models recapitulate typical GIST features by developing a primary tumour and subsequent liver metastasis that were easily detected by MRI. The animal models can be used to test new diagnostic tools and pharmaceutical treatments on imatinib-resistant tumours. UNITO will license the animal models to dedicated companies (e.g. Charles River, Jackson, Oncodesign) that can ensure efficient animal generation by guaranteed manufacturing procedures in order to be competitive on the market.
New imaging probes able to distinguish imatinib-resistant and imatinib-sensitive GIST tumours
CAGE designed new imaging probes targeting the GLP-2 receptor to distinguish imatinib-resistant and sensitive GIST tumours. A GLP-2-like peptide conjugated with a fluorescent or a MRI/PET probe was synthesised. In addition, a novel class of chelates, able to bind both the Gadolinium cation for MRI applications or radionuclide metals for PET/SPECT, was developed to extend the use of the GLP-2 probe. The GLP-2 imaging probes will be patented and then licensed to pharmaceutical companies for planning the required clinical trials. A patent application will be submitted after further validation (estimated in spring 2018), after which CAGE will actively seek for a pharmaceutical company interested in a product license. In parallel, CAGE will maintain production of small quantities of the product to satisfy potential requests for preclinical studies from research centres.
Summary of dissemination and exploitation activities
• MITIGATE website
• Dissemination plan
• Dissemination material
o Flyers, folders, posters, roll-up, newsletters
• 22 papers published in peer-reviewed journals
• 48 presentations at various scientific meetings
• 17 posters at various scientific meetings
• 18 articles in popular press
• Intermediate Exploitation Plan
• End User Survey
• 4 exhibitions
• 2 workshops
• Final Exploitation Plan
List of Websites:
http://www.mitigate-project.eu/
