Final Report Summary - INSERT (Development of an integrated SPECT/MRI system for enhanced stratification of brain tumour patients prior to patient-specific radio-chemo therapy and early assessment of treatment efficacy)
Executive Summary:
The motivation of the INSERT project laid in the high mortality rate of patients with recurrent glioma. Unfortunately, notwithstanding the innovations brought by the scientific research, the management of these patients, both adults and in paediatric age, is still a challenge, since the life expectancy of these patients is very low and the surgery does not represent a definitive solution in all cases, given to the nature of these tumours.
INSERT purpose was to provide improved personalized radio-chemo therapies and clinically relevant stratification for brain tumour patients with potentially better outcome in survival and quality-of-life using a specifically developed multi-modality imaging tool.
The specific project objectives were related to:
1) the development of a SPECT (Single Photon Emission Computed Tomography) system to be used as an insert to an MRI (Magnetic Resonance Imaging) gantry, thus enabling the simultaneous acquisition of images resulting from the two systems. Two types of integrated SPECT/MRI imager were planned to be produced: one specifically dedicated to preclinical imaging, the second one designed for clinical imaging.
2) the use of preclinical tumour models and in vivo molecular imaging strategies with potential for significant impact in the framework of neuro oncology, to explore the characterization of tumour biology and its progression as well as treatment efficacy.
3) the evaluation of treatments based on multi-parametric imaging in preclinical trials and, upon approval from regulatory authorities and institutional ethical committee, in selected clinical evaluation studies.
At the end of the project, the consortium has completed the development of the gamma-ray detection modules, which are key components that compose the SPECT systems. The gamma-ray detection module is based on the well-established Anger architecture, with an array of Silicon PhotoMultipliers to readout the light emitted by a single and continuous CsI:Tl scintillator. The module is equipped with a readout ASIC to readout the signals provided by the SiPMs. Moreover, the module is designed in order to be compatible for both preclinical and clinical SPECTs, in order to simplify the development effort and to have a production of a volume of modules that can be employed for both the preclinical and clinical systems. The production of detection modules to equip both instruments has been successfully completed. The components of the detection module have being produced with MRI-compatible materials and they should also operate without artifacts when MRI is active. One main goal was therefore to carry compatibility tests of the components of the SPECT prototypes. After several campaigns on tests in laboratory, with a custom developed gradients emulator, and in MRI a satisfactory, mutual compatibility of the detection modules with the MRI has been achieved.
At the end of the period, the preclinical SPECT has been completed and tested also in a MRI. The preclinical SPECT has been assembled with all related components, including DAQ system, reconstruction software, high-resolution collimators, animal MRI coil, cooling systems and several other important components here not listed. The preclinical SPECT have been tested in MRI for what concerns compatibility validation and also tests on radioactive phantoms and on one animal have been performed.
At the end of the project, the clinical SPECT has been assembled with all its components (detectors, collimator, MRI coil, ...) and is available for tests and qualification in lab. and in MRI.
As far as preclinical activities are concerned, the project aimed to validate the biomarker array identified during the project duration, to set up strategies for RT monitoring, to compare INSERT system features to the already available techniques and to understand the molecular mechanisms underlying HIF-1a activity modulation in glioma in relation to responsiveness.
To reach these goals, the following activities were performed:
1. Analyze by different techniques (biochemical assays, FACS, IHC, Real Time PCR) the reliability of identified biomarkers;
2. Compare different tracers for glioma imaging;
3. Verify the possibility to perform multi tracer imaging by SPECT;
4. Compare different strategies for RT monitoring;
5. Assess the performance of the integrated system compared to already available techniques;
6. Exploit the cell/animal model to assess the reliability of new biomarkers and the efficacy of new treatment options;
7. Analyze the role of HIF-1a as a theranostic biomarker
For what concerns the clinical activity, the unavoidable technological delays in developing the clinical SPECT system resulted in a major deviation in the objectives of the clinical work package. The modified objectives were therefore: 1) to prepare a strong case for the clinical utility of SPECT/MRI, b) to review the current status of SPECT/MRI imaging in glioma and 3) to demonstrate the feasibility of the designed system to produce tomographic images of the brain.
In conclusion, thanks to the possibility of obtaining multiple parameters, once clinically validated, the INSERT system will enable to better define the tumour biology and to give relevant information for a personalized treatment, with a considerable impact on the efficacy of the treatment itself.
INSERT will lay the foundation for long-term innovation in the field of treatment planning and response monitoring. Many of its results will be translatable to clinical trials in different fields of oncology, ultimately facilitating treatment planning in individual patients. Therefore, the uptake of the proposed technology in the European and potentially global markets would further enhance the development of personalized therapies in the oncological setting.
Project Context and Objectives:
INSERT – “Development of an integrated SPECT/MRI system for enhanced stratification of brain tumour patients prior to patient specific radio-chemo therapy and early assessment of treatment efficacy” – is a new collaborative research project founded by the European Union within the HEALTH Theme of its Seventh Research Framework Program (FP7).
INSERT, launched on March 1st 2013, was granted 4,600,000 EUR of funding over a 48 month period. The project is coordinated by Politecnico di Milano and includes nine partners whose expertise covers the whole verticality of medical device discovery, from sensor and instrument development to pre-clinical testing and clinical application development and evaluation.
INSERT purpose is to provide improved personalized radio-chemo therapies and clinically relevant stratification for brain tumour patients with potentially better outcome in survival and quality-of-life using a specifically developed multi-modality imaging tool. The proposed tool is based on the development of a SPECT (Single Photon Emission Computed Tomography) system to be used as an insert to an MRI (Magnetic Resonance Imaging) gantry, thus enabling the simultaneous acquisition of images resulting from the two systems.
The specific objectives of this project are:
1) To develop an integrated SPECT/MRI system for simultaneous dual modality imaging. The SPECT system is designed to be combined with a general-purpose MRI system. Two types of integrated SPECT/MRI imager will be designed and produced: one will be specifically dedicated to preclinical imaging, the second one will be designed for clinical imaging
2) To use preclinical tumour models and in vivo molecular imaging strategies with potential for significant impact in the framework of neuro oncology, to explore the characterization of tumour biology and its progression as well as treatment efficacy.
3) To plan and evaluate treatments based on multi-parametric imaging in preclinical trials and, upon approval from regulatory authorities and institutional ethical committee, in selected clinical evaluation studies. This will facilitate the optimization of treatment scheduling using a personalized approach.
Therefore, INSERT will develop an integrated SPECT/MRI system for molecular imaging which will give the unique opportunity to non-invasively and dynamically describe biological and functional tumour features by permitting SPECT/MRI measurement of multiple imaging biomarkers within preclinical research protocols and their translation into the clinics thanks to the creation of animal models (preclinical level) and a pilot study (clinical level). It will develop a novel multimodality imaging technique, which will be validated and exploited by:
• monitoring the tumour growth, biological features and tumour progression in experimental mouse and rat models of glioma;
• improving clinical protocol design for the treatment of these tumours by using biological information obtained by non-invasive multimodal acquisitions.
Project Results:
INSERT purpose was to develop innovative instruments to provide improved personalized radio-chemo therapies and clinically relevant stratification for brain tumour patients with potentially better outcome in survival and quality-of-life using a specifically developed multi-modality imaging tool. The proposed tool is based on the development of a SPECT (Single Photon Emission Computed Tomography) system to be used as an insert to an MRI (Magnetic Resonance Imaging) gantry, thus enabling the simultaneous acquisition of images resulting from the two systems. The development of such instrument was very challenging, as the simultaneous use of SPECT inside a MRI requires mutual compatibility for these two kinds of scanners. A very specific study of the components of the SPECT, with also several compatibility tests, has been done during the project, leading finally to the production of the two foreseen prototypes: a preclinical SPECT and a clinical SPECT. The preclinical SPECT have been developed and used in first experimentation, the clinical SPECT has been assembled with all components, including an innovative collimator and MRI coil and is now available for qualification and tests.
Overall the project met significant achievements in terms of development of unique instruments that the Consortium hopes could open new opportunities of preclinical experiments and clinical trials in personalized medicine.
In the following, the main achievements of the project workpackages are summarized.
The aim of WP1 (Specifications) was to provide all specifications useful for the design of the SPECT system to be integrated in the MRI and used in the applications foreseen in the project. The specifications have been derived from the assessment of the requirements of the experiments (preclinical and clinical) to be carried out during the project. Specifications on the single subsystems of the SPECT apparatus have been defined accordingly in order to achieve the requested performance in terms of spatial resolution, energy resolution, sensitivity and others. The approach to investigate compatibility issues was defined together with plan of validation of components and of the whole SPECT apparatus when integrated in the MRI.
All the goals of WP2 (Gamma detection modules and electronics), focusing on the development of MRI-compatible gamma-ray detection modules for simultaneous SPECT imaging, have been successfully achieved. The module is based on the Anger architecture and features a CsI(Tl) scintillator crystal coupled to tiles of silicon photo-detectors. The results of the WP2 activity can be classified in three majors groups: (1) realization of state-of-the-art silicon photomultipliers (SiPM), (2) realization of the readout electronics completely compatible with the intense (3 T) and time-varying magnetic field inside the MRI bore, (3) realization of complete detection modules offering very good spatial (1 mm) and energetic (14% at 662 keV) resolution, for both the preclinical and clinical detection rings. In fact, for the preclinical instrument, 10 modules (of sensing area 5 cm × 5 cm, covered by 4 SiPM tiles) have been fabricated and tested, while for the clinical prototype, 20 modules (5 cm × 10 cm covered by 8 tiles) have been realized and qualified.
After the initial evaluation of silicon drift detector, an important technological milestone has been the optimization of the cleanroom fabrication process of SiPMs by Fondazione Bruno Kessler (FBK - Trento, Italy). The last fabrication batch (with yield higher than 80% over 17 wafers and deeper insight into the role of the initial substrate type) of RGB-HD SiPM (with cell pitch of 25µm and enhanced PDE/ENF ratio) offers reliable low-noise performance (dark count rate on average below 200 kHz/mm2).
Every part of the acquisition chain is optimized in order to be MRI-compatible, while preserving the best possible performance, spanning from SiPM biasing, to the readout ASIC, the analog board as well as the digital and power supply sections. ANGUS is a 36-channel integrated (CMOS AMS 0.35um technology) current amplifier and programmable shaper (RC filtering) expressively designed for this module. Given the large capacitance of the SiPM detectors (about 1 nF per SiPM, four of which are merged together), a feedback topology is not the preferred solution, while a current conveyor input stage is selected. The acquisition architecture is modular in order to adapt to a variable number of acquisition nodes. Each DAQ, containing an ADC and FPGA, communicates in daisy-chain through optical fibers for better signal immunity (along a 15-meter distance from the MRI scanner to the control room).
Reciprocal MRI-SPECT compatibility has been assessed and finally achieved. The compatibility of the module towards MRI is obtained by (i) minimizing the amount of metal (the presence of the collimator affecting the spatial field homogeneity is corrected by means of field shimming), (ii) by choosing frequencies of the digital clocks far from the resonance imaging frequency (123 MHz) and (iii) by using a filter plate on the power supply cables entering the MRI shielded room through a waveguide. The impact of MRI on SPECT modules has been analyzed in depth and led to non-standard modifications of the layout of the ASIC boards in order to reduce the impact of switching gradients (while the impact of RF is minimized by using a custom and shielded TX-RX RF coil placed inside the SPECT ring). Pick-up is reduced by: (i) two-pole filtering of the HV lines (biasing the SiPM and identified as critical lines by means of a laboratory mapping tool developed on purpose) close to the SiPM connector, (ii) use of meshed ground and power planes to reduce the width of current loops (Eddy currents), (iii) minimization of the number of vias connecting planes on different layers to avoid vertical loops in the PCB (orthogonal to the MRI z axis). These solutions allowed achieving the final immunity condition in which, the FWHM of the 99mTc peak (14%) is not broadened when the detector ring is operated inside the MRI scanner performing standard Gradient Echo sequences.
The choice of non-standard non-metallic (especially non ferromagnetic) materials required particular efforts and satisfactory solutions have been identified. The cooling block (a heat-sink sandwiched between the SiPM tiles and the ASIC board required for detector operation around 0°C granting suitable spectroscopic resolution) is made of a thermally conductive plastic (Coolpoly), whose porosity is corrected by spray painting. Since the ASIC carriers contains nickel, they are avoided and replaced by direct wire-bonding of the ASICs on the PCB, allowing a yield of 100%.
Important results have been also achieved regarding the reconstruction algorithms used for image processing at module level. Acquired planar images are processed with novel statistical reconstruction techniques (such as maximum likelihood) based on pixels light response functions which are iteratively identified from flood homogeneous irradiation. Wider field of view, better noise filtering and better uniformity are achieved with respect to the centroid method.
Goal of WP3 (SPECT/MRI systems development, integration and characterization) was the development of the SPECT prototypes, preclinical and clinical.
During the design of the pre-clinical system we performed several experiments and simulations. We used Monte-Carlo photon transportation simulations to design a multi-pinhole collimator for mouse and rat brain imaging. For the mouse brain collimator this complex optimization problem was addressed by evaluating 10 designs with different number of regularly placed pinholes. The size of the holes was fixed at 0.8 mm and for each configuration, and we tested several field-of-view (FOV) sizes. We calculated resolution and sensitivity for these configurations, and simulated imaging of a mathematical phantom. As the zoom factor decreases with increasing number of holes, the achievable reconstructed resolution also decreases, while sensitivity increases with increasing number of holes. Based on parameters and imaging artefacts we selected 5 designs for further refinement. Next we calculated hole-size vs. peak sensitivity curves. For each design we selected a hole size to obtain 1.5 kcps/MBq peak sensitivity and simulated imaging of a hot-rod phantom. Based on the quality of reconstructed images we selected the 3x3 pinhole design, and refined it to improve angular sampling density. Next we optimized the hole size performing a series of hot-rod phantom simulations. Using this optimal design, we investigated effects of the finite penetration of gamma photons into the collimator. Decision was that for Tc99m imaging Tungsten-polyamide composite material would be optimal, with pure Tungsten pinhole-insets.
As we needed a non-conductive radiation shielding material at several places in the instrument, we developed the technology of manufacturing Tungsten-polyamide composite structures. We experimented with Tungsten powders of different grain size and different powder glue mixtures to find the material with highest density. The optimal mixture reached a density of over 11g/ml.
We have put significant effort into designing the cooling for the system. From a system-design perspective, air-cooling was very desirable, but its effectivity was questionable. We performed simulations and several test measurements with various cooling blocks and both air and liquid cooling, to evaluate efficiency of the various alternatives. We selected liquid cooling and a thermally conductive plastic material. Unfortunately, though the initial samples of this thermally conductive plastic material were perfect the larger blocks we obtained for manufacturing the prototype systems were porous, that we realized only during pressure testing of the cooling blocks. Sealing the pores caused several weeks delay.
MRI compatibility was a main design requirement, and we performed several material and component level tests throughout the development process. To avoid eddy currents we could not use bulky metal cooling blocks, and had to find and evaluate thermally conductive plastics and ceramics instead. MRI compatibility of the 3D printed Tungsten pinhole insets was tested in simulations and tests. All electronic component designs were tested separately in the MRI.
Designing MRI compatible electronics was a major challenge of the project. We introduced several special designs, like hatching the power distribution and reference plane layers of the PCB, to reduce eddy currents. Radiated RF noise is a critical issue in MRI environment. This was regularly checked during the development using a shielded box, near field antennas and a spectrum analyzer. We payed special attention to use non-magnetic electronic components, to avoid susceptibility artefacts in the MRI.
To improve MRI compatibility and also to facilitate both the pre-clinical and the clinical system with the same hardware, we developed a distributed data acquisition strategy. Each detector module contains a small FPGA, and data is processed locally. The modules are daisy chained with an optical bus, greatly reducing the size of conducting loops in the system and enabling very flexible design.
Due to delays in the project the software was simplified compared to the original plans. Never the less, we developed a graphical data acquisition tool and a separate program for automatic calibration of the system and for application of corrections to the data. Images from the pre-clinical prototype confirmed the original design specifications.
Finally, WP3 has lead to the construction of the following prototypes:
Pre-clinical system (assembly completed; acquisition and correction software implemented and tested; test of the full prototype in MRI).
Clinical system (design completed, assembly completed with detection modules, clinical RF coil, collimator and mechanical frame).
In the framework of WP4 (Reconstruction software development), software for tomographic image reconstruction and for system calibration was developed for the pre-clinical scanner equipped with multi-pinhole collimators and for the clinical scanner equipped with multi-mini-slit-slat collimators. The iterative reconstruction algorithms ML-EM and OS-EM were used, which are based on forward and back-projection routines to transform data between the object domain and the projection domain. For the pre-clinical system, these procedures were based on Monte Carlo simulation, incorporating a 3D model of the imaging apparatus and the most important physical effects that affect the acquisition. For the clinical system, a flexible angular blurring algorithm was developed, allowing easy modification of the system parameters.
The calibrating procedure consists of estimating a specific set of parameters, describing the acquisition geometry, by comparing measured and calculated projection data from a set of line or point sources. Attenuation correction is performed using a pseudo-CT generated from standard T1-weighted or T2-weighted MRI images by an algorithm developed previously for PET/MRI, and scatter correction is done with an energy-window based approach.
Advanced software algorithms were developed in order to take advantage of simultaneous SPECT/MRI dual modality acquisition. A MAP-EM reconstruction algorithm was implemented, which incorporates high resolution anatomical MRI images in order to improve the quality of the reconstructed SPECT images. Thereby noise amplification can be reduced while preserving anatomic features in the image.
Motion-correction routines were implemented that can utilize high temporal-resolution information, obtained from fast MRI sequences, to reduce motion artefacts in the SPECT images. For the INSERT system, some degree of patient motion may actually be advantageous, as long as it can be monitored accurately and corrected for, as this improves the spatial sampling of the data.
We have implemented a range of methods for partial volume correction, which are post-reconstruction image processing procedures to improve both the qualitative and quantitative aspects of the reconstructed SPECT images by utilizing high-resolution anatomical MRI images. These methods do not suffer from the excessive noise-amplification and ringing artefacts associated with standard de-convolution algorithms. The package includes a novel, practical method, which only requires segmentation of one single region, such as a tumour, rather than the whole image.
Since the INSERT scanner is a stationary system, as opposed to practically all commercial SPECT scanners, it can be used for acquisition of dynamic data, which can provide additional information of clinical relevance. We have developed a series of software tools for performing kinetic analysis in order to take advantage of this type of studies.
In the framework of WP5 (Preclinical evaluation), we have characterized a glioma murine model for the assessment of the INSERT new multimodal system. Moreover this work package has identified important biomarkers for the early estimation of glioma response to TMZ treatment. These biomarkers have been validated in several glioma cell lines, animal models and by different non-invasive imaging techniques.
In detail, UMIL dealt with cell characterization, animal model production, identification of molecular pathways involved in glioma TMZ response, supporting also HSR and CROMed in their activities. HSR identified translational biomarkers able to report about variation in HIF-1a activity with translational imaging techniques (MRI, PET and SPET). Cromed tested the possibility to perform non-invasive imaging with the identified SPECT tracer, set up RT protocol and an ischemia model for further application of the INSERT instrument.
UMIL activities: U251 cells engineered to express the reporter gene luciferase under control of either a constitutive (U251-pGL3) or a HIF-1a inducible promoter (U251-HRE) were in vitro validated. U251-HRE cells were also engineered to express a constitutive fluorescent reporter (mCherry) to in vivo assess their viability by fluorescence imaging and HIF-1a activity by Bioluminescence imaging using an IVIS Spectrum/CT system by Perkin Elmer.
The effect of TMZ treatment was assessed in these cells allowing us to conclude that TMZ treatment is able to reduce HIF-1a activity and that this reduction precedes cell death. Moreover we showed that by inhibiting HIF-1a expression in TMZ resistant cells, glioma sensitivity to this drug can be restored.
HIF-1a activity has been validated as response biomarker in several other cell lines both responsive and resistant to TMZ (U87, LN229, U130, T98G, GLI36s, GLI36r), and after different treatments (MG132, Bafilomycin A1, Geldanamycin, Metformin, miR675-5p) allowing to consider HIF-1a activity reduction as an early biomarker of glioma response to treatment.
An orthotopic glioma model has been set up and characterized both for the growth kinetic, and for its responsiveness to TMZ. HIF-1a activity was confirmed as early biomarker of tumour response to treatment, also in vivo by non-invasive analyses (non invasive reporter gene acitivity by BLI and FLI, uptake assessment of specific commercially available fluorescent probes able to report upon HIF-1a activity (Hypoxisense) and neo angiogenesis (Integrisense)). Changes in HIF-1a activity was also able to induce changes in different processes that can be studied by translatable imaging techniques. Within the project DWI and DCE MRI studies were carried out demonstrating the possibility to early estimate response to TMZ by this procedure. In the same animals PET imaging allowed us to monitor both hypoxia induction and cell proliferation changes after treatment using 18F-FAZA and 18F-FLT tracers, respectively. A number of PSECT tracers were evaluated as well as radiotherapeutic protocol have been set up.MIBI uptake showed a significant reduction after treatment with TMZ and it was chosen for further clinical evaluations.
All these results have been confirmed and validated by immunohistochemical analysis.
Due to technical problems occurred during INSERT instrument delivery in Milan we could not perform the integrated imaging experiments we set up, but biological results obtained with already available instruments allowed us to confirm previous data and the features of INSERT SPECT was evaluated in another experimental set.
The comparison of the the NanoSPECT system and the Consortium-built INSERT system has been performed. While the conditions of preparedness of the INSERT SPECT`s reconstruction interface were suboptimal, the imaging performance measured by contrast means and practicalities of imaging time does support an advantageous judgement of future biological measurements with the pre-clinical INSERT SPECT in mouse brain glioma models.
The highest advantage over state-of-the art is that INSERT SPECT applies all-round detectors and list mode data acquisition, rendering short acquisitions and time-activity curve generation using motionless data acquisition possible.
Moreover, for the general application of the INSERT SPECT it can be inferred that the use of the system for multiple small volume definitions to describe glioma model heterogeneity and its changes is preferred to any other existing SPECT device.
The overall imaging performance of the INSERT is remarkably good even when compared to NanoSPECT.
In the framework of WP6 (Clinical evaluation), we have conducted an up to date audit of patients with glioblastoma multiforme (GBM) treated at UCLH and these data confirm the clinical utility of multi-modality imaging to define early response to standard of care.
We have therefore designed and activated a feasibility protocol in which patients undergoing standard of care chemo-radiotherapy at UCLH will be scanned using sequential MRI and MIBI-SPECT. This protocol will recruit patients over the next 6-12 months at UCLH and provide proof of principle for the use of combined MIBI-SPECT MRI in glioma. The important outputs will include an assessment of utility for radiotherapy planning and as a novel early response scan during chemo-radiotherapy.
We have undertaken an updated review of the literature and re-assessed the potential demand for multi-modality imaging in the context of brain cancer, with reference to on-going studies in the consortium. There remains a clear unmet need for imaging that more closely represents biology in high-grade glioma. A recent published review confirms this perspective.
99mTc- MIBI SPECT is a promising molecular imaging technique based on published literature and data from INSERT WP5 and we have developed a clinical protocol to allow further investigation and provide clinical data on GBM.
Due to delays in the technical development of the system, no clinical studies have been performed with INSERT in the time frame of the funding. We have therefore focused on gaining relevant experience within the consortium on using imaging end-points that will be relevant when the system becomes operational.
Specific on-going work that will support the clinical implementation of this work package beyond the end of the funding period are discussed above and include:
• Clinical protocol developed at UCLH and to be activated at UCLH and University of Leeds
• Phase IIa trial is currently investigating Lutetium DOTATATE in children with primary refractory or relapsed high risk neuroblastoma. Patients are imaged with 123I-mIBG, 18F-FDG PET/CT, Gallium DOTATAE PET/CT plus conventional MRI or CT, and SPECT is used for dosimetry purposes.
• A phase I/II trial is comparing [124I]mIBG PET/CT to[123I]mIBG scintigraphy in patients with metastatic neuroblastoma, with the aim of detecting metastatic neuroblastoma deposits and establish SUV values (Standardized Uptake Value) for [124I]mIBG PET/CT that define positive vs negative findings.
• Functional MRI imaging in patients undergoing standard of care chemo-radiotherapy
• White cell tracking study in patients with solid tumours (University of Leeds)
Due to delays in the project we were not able to perform any experiments on the final system within the time-frame of the project. Instead we used a test-system with a single detector and a rotating platform. We used this system to develop a practical calibration procedure for the final system equipped with multi-mini-slit-slat (MSS) collimators. The procedure consists of a sequence of scans using plane- and line sources for estimation of different calibration parameters. We also used the test-system to perform a number of phantom experiments, which confirmed results we had obtained previously with simulations, showing that the final INSERT system will be able to provide images adequate for the intended purpose – i.e. SPECT imaging of brain tumours.
Potential Impact:
The motivation of the INSERT project lays in the high mortality rate of patients with recurrent glioma. Unfortunately, notwithstanding the innovations brought by the scientific research, the management of these patients, both adults and in paediatric age, is still a challenge, since the life expectancy of these patients is very low and the surgery does not represent a definitive solution in all cases, given to the nature of these tumours.
Thanks to the possibility of obtaining multiple parameters, the INSERT system will enable to better define the tumour biology and to give relevant information for a personalized treatment, with a considerable impact on the efficacy of the treatment itself.
The system will be validated at the pre-clinical level thanks to the creation of animal models, and at the clinical level. The initial focus will be on patients with glioma, but there is future potential to target a range of tumours in the head and neck region.
INSERT will lay the foundation for long-term innovation in the field of treatment planning and response monitoring. Many of its results will be translatable to clinical trials in different fields of oncology, ultimately facilitating treatment planning in individual patients. Therefore, the uptake of the proposed technology in the European and potentially global markets would further enhance the development of personalized therapies in the oncological setting.
List of Websites:
http://www.insert-project.eu
The motivation of the INSERT project laid in the high mortality rate of patients with recurrent glioma. Unfortunately, notwithstanding the innovations brought by the scientific research, the management of these patients, both adults and in paediatric age, is still a challenge, since the life expectancy of these patients is very low and the surgery does not represent a definitive solution in all cases, given to the nature of these tumours.
INSERT purpose was to provide improved personalized radio-chemo therapies and clinically relevant stratification for brain tumour patients with potentially better outcome in survival and quality-of-life using a specifically developed multi-modality imaging tool.
The specific project objectives were related to:
1) the development of a SPECT (Single Photon Emission Computed Tomography) system to be used as an insert to an MRI (Magnetic Resonance Imaging) gantry, thus enabling the simultaneous acquisition of images resulting from the two systems. Two types of integrated SPECT/MRI imager were planned to be produced: one specifically dedicated to preclinical imaging, the second one designed for clinical imaging.
2) the use of preclinical tumour models and in vivo molecular imaging strategies with potential for significant impact in the framework of neuro oncology, to explore the characterization of tumour biology and its progression as well as treatment efficacy.
3) the evaluation of treatments based on multi-parametric imaging in preclinical trials and, upon approval from regulatory authorities and institutional ethical committee, in selected clinical evaluation studies.
At the end of the project, the consortium has completed the development of the gamma-ray detection modules, which are key components that compose the SPECT systems. The gamma-ray detection module is based on the well-established Anger architecture, with an array of Silicon PhotoMultipliers to readout the light emitted by a single and continuous CsI:Tl scintillator. The module is equipped with a readout ASIC to readout the signals provided by the SiPMs. Moreover, the module is designed in order to be compatible for both preclinical and clinical SPECTs, in order to simplify the development effort and to have a production of a volume of modules that can be employed for both the preclinical and clinical systems. The production of detection modules to equip both instruments has been successfully completed. The components of the detection module have being produced with MRI-compatible materials and they should also operate without artifacts when MRI is active. One main goal was therefore to carry compatibility tests of the components of the SPECT prototypes. After several campaigns on tests in laboratory, with a custom developed gradients emulator, and in MRI a satisfactory, mutual compatibility of the detection modules with the MRI has been achieved.
At the end of the period, the preclinical SPECT has been completed and tested also in a MRI. The preclinical SPECT has been assembled with all related components, including DAQ system, reconstruction software, high-resolution collimators, animal MRI coil, cooling systems and several other important components here not listed. The preclinical SPECT have been tested in MRI for what concerns compatibility validation and also tests on radioactive phantoms and on one animal have been performed.
At the end of the project, the clinical SPECT has been assembled with all its components (detectors, collimator, MRI coil, ...) and is available for tests and qualification in lab. and in MRI.
As far as preclinical activities are concerned, the project aimed to validate the biomarker array identified during the project duration, to set up strategies for RT monitoring, to compare INSERT system features to the already available techniques and to understand the molecular mechanisms underlying HIF-1a activity modulation in glioma in relation to responsiveness.
To reach these goals, the following activities were performed:
1. Analyze by different techniques (biochemical assays, FACS, IHC, Real Time PCR) the reliability of identified biomarkers;
2. Compare different tracers for glioma imaging;
3. Verify the possibility to perform multi tracer imaging by SPECT;
4. Compare different strategies for RT monitoring;
5. Assess the performance of the integrated system compared to already available techniques;
6. Exploit the cell/animal model to assess the reliability of new biomarkers and the efficacy of new treatment options;
7. Analyze the role of HIF-1a as a theranostic biomarker
For what concerns the clinical activity, the unavoidable technological delays in developing the clinical SPECT system resulted in a major deviation in the objectives of the clinical work package. The modified objectives were therefore: 1) to prepare a strong case for the clinical utility of SPECT/MRI, b) to review the current status of SPECT/MRI imaging in glioma and 3) to demonstrate the feasibility of the designed system to produce tomographic images of the brain.
In conclusion, thanks to the possibility of obtaining multiple parameters, once clinically validated, the INSERT system will enable to better define the tumour biology and to give relevant information for a personalized treatment, with a considerable impact on the efficacy of the treatment itself.
INSERT will lay the foundation for long-term innovation in the field of treatment planning and response monitoring. Many of its results will be translatable to clinical trials in different fields of oncology, ultimately facilitating treatment planning in individual patients. Therefore, the uptake of the proposed technology in the European and potentially global markets would further enhance the development of personalized therapies in the oncological setting.
Project Context and Objectives:
INSERT – “Development of an integrated SPECT/MRI system for enhanced stratification of brain tumour patients prior to patient specific radio-chemo therapy and early assessment of treatment efficacy” – is a new collaborative research project founded by the European Union within the HEALTH Theme of its Seventh Research Framework Program (FP7).
INSERT, launched on March 1st 2013, was granted 4,600,000 EUR of funding over a 48 month period. The project is coordinated by Politecnico di Milano and includes nine partners whose expertise covers the whole verticality of medical device discovery, from sensor and instrument development to pre-clinical testing and clinical application development and evaluation.
INSERT purpose is to provide improved personalized radio-chemo therapies and clinically relevant stratification for brain tumour patients with potentially better outcome in survival and quality-of-life using a specifically developed multi-modality imaging tool. The proposed tool is based on the development of a SPECT (Single Photon Emission Computed Tomography) system to be used as an insert to an MRI (Magnetic Resonance Imaging) gantry, thus enabling the simultaneous acquisition of images resulting from the two systems.
The specific objectives of this project are:
1) To develop an integrated SPECT/MRI system for simultaneous dual modality imaging. The SPECT system is designed to be combined with a general-purpose MRI system. Two types of integrated SPECT/MRI imager will be designed and produced: one will be specifically dedicated to preclinical imaging, the second one will be designed for clinical imaging
2) To use preclinical tumour models and in vivo molecular imaging strategies with potential for significant impact in the framework of neuro oncology, to explore the characterization of tumour biology and its progression as well as treatment efficacy.
3) To plan and evaluate treatments based on multi-parametric imaging in preclinical trials and, upon approval from regulatory authorities and institutional ethical committee, in selected clinical evaluation studies. This will facilitate the optimization of treatment scheduling using a personalized approach.
Therefore, INSERT will develop an integrated SPECT/MRI system for molecular imaging which will give the unique opportunity to non-invasively and dynamically describe biological and functional tumour features by permitting SPECT/MRI measurement of multiple imaging biomarkers within preclinical research protocols and their translation into the clinics thanks to the creation of animal models (preclinical level) and a pilot study (clinical level). It will develop a novel multimodality imaging technique, which will be validated and exploited by:
• monitoring the tumour growth, biological features and tumour progression in experimental mouse and rat models of glioma;
• improving clinical protocol design for the treatment of these tumours by using biological information obtained by non-invasive multimodal acquisitions.
Project Results:
INSERT purpose was to develop innovative instruments to provide improved personalized radio-chemo therapies and clinically relevant stratification for brain tumour patients with potentially better outcome in survival and quality-of-life using a specifically developed multi-modality imaging tool. The proposed tool is based on the development of a SPECT (Single Photon Emission Computed Tomography) system to be used as an insert to an MRI (Magnetic Resonance Imaging) gantry, thus enabling the simultaneous acquisition of images resulting from the two systems. The development of such instrument was very challenging, as the simultaneous use of SPECT inside a MRI requires mutual compatibility for these two kinds of scanners. A very specific study of the components of the SPECT, with also several compatibility tests, has been done during the project, leading finally to the production of the two foreseen prototypes: a preclinical SPECT and a clinical SPECT. The preclinical SPECT have been developed and used in first experimentation, the clinical SPECT has been assembled with all components, including an innovative collimator and MRI coil and is now available for qualification and tests.
Overall the project met significant achievements in terms of development of unique instruments that the Consortium hopes could open new opportunities of preclinical experiments and clinical trials in personalized medicine.
In the following, the main achievements of the project workpackages are summarized.
The aim of WP1 (Specifications) was to provide all specifications useful for the design of the SPECT system to be integrated in the MRI and used in the applications foreseen in the project. The specifications have been derived from the assessment of the requirements of the experiments (preclinical and clinical) to be carried out during the project. Specifications on the single subsystems of the SPECT apparatus have been defined accordingly in order to achieve the requested performance in terms of spatial resolution, energy resolution, sensitivity and others. The approach to investigate compatibility issues was defined together with plan of validation of components and of the whole SPECT apparatus when integrated in the MRI.
All the goals of WP2 (Gamma detection modules and electronics), focusing on the development of MRI-compatible gamma-ray detection modules for simultaneous SPECT imaging, have been successfully achieved. The module is based on the Anger architecture and features a CsI(Tl) scintillator crystal coupled to tiles of silicon photo-detectors. The results of the WP2 activity can be classified in three majors groups: (1) realization of state-of-the-art silicon photomultipliers (SiPM), (2) realization of the readout electronics completely compatible with the intense (3 T) and time-varying magnetic field inside the MRI bore, (3) realization of complete detection modules offering very good spatial (1 mm) and energetic (14% at 662 keV) resolution, for both the preclinical and clinical detection rings. In fact, for the preclinical instrument, 10 modules (of sensing area 5 cm × 5 cm, covered by 4 SiPM tiles) have been fabricated and tested, while for the clinical prototype, 20 modules (5 cm × 10 cm covered by 8 tiles) have been realized and qualified.
After the initial evaluation of silicon drift detector, an important technological milestone has been the optimization of the cleanroom fabrication process of SiPMs by Fondazione Bruno Kessler (FBK - Trento, Italy). The last fabrication batch (with yield higher than 80% over 17 wafers and deeper insight into the role of the initial substrate type) of RGB-HD SiPM (with cell pitch of 25µm and enhanced PDE/ENF ratio) offers reliable low-noise performance (dark count rate on average below 200 kHz/mm2).
Every part of the acquisition chain is optimized in order to be MRI-compatible, while preserving the best possible performance, spanning from SiPM biasing, to the readout ASIC, the analog board as well as the digital and power supply sections. ANGUS is a 36-channel integrated (CMOS AMS 0.35um technology) current amplifier and programmable shaper (RC filtering) expressively designed for this module. Given the large capacitance of the SiPM detectors (about 1 nF per SiPM, four of which are merged together), a feedback topology is not the preferred solution, while a current conveyor input stage is selected. The acquisition architecture is modular in order to adapt to a variable number of acquisition nodes. Each DAQ, containing an ADC and FPGA, communicates in daisy-chain through optical fibers for better signal immunity (along a 15-meter distance from the MRI scanner to the control room).
Reciprocal MRI-SPECT compatibility has been assessed and finally achieved. The compatibility of the module towards MRI is obtained by (i) minimizing the amount of metal (the presence of the collimator affecting the spatial field homogeneity is corrected by means of field shimming), (ii) by choosing frequencies of the digital clocks far from the resonance imaging frequency (123 MHz) and (iii) by using a filter plate on the power supply cables entering the MRI shielded room through a waveguide. The impact of MRI on SPECT modules has been analyzed in depth and led to non-standard modifications of the layout of the ASIC boards in order to reduce the impact of switching gradients (while the impact of RF is minimized by using a custom and shielded TX-RX RF coil placed inside the SPECT ring). Pick-up is reduced by: (i) two-pole filtering of the HV lines (biasing the SiPM and identified as critical lines by means of a laboratory mapping tool developed on purpose) close to the SiPM connector, (ii) use of meshed ground and power planes to reduce the width of current loops (Eddy currents), (iii) minimization of the number of vias connecting planes on different layers to avoid vertical loops in the PCB (orthogonal to the MRI z axis). These solutions allowed achieving the final immunity condition in which, the FWHM of the 99mTc peak (14%) is not broadened when the detector ring is operated inside the MRI scanner performing standard Gradient Echo sequences.
The choice of non-standard non-metallic (especially non ferromagnetic) materials required particular efforts and satisfactory solutions have been identified. The cooling block (a heat-sink sandwiched between the SiPM tiles and the ASIC board required for detector operation around 0°C granting suitable spectroscopic resolution) is made of a thermally conductive plastic (Coolpoly), whose porosity is corrected by spray painting. Since the ASIC carriers contains nickel, they are avoided and replaced by direct wire-bonding of the ASICs on the PCB, allowing a yield of 100%.
Important results have been also achieved regarding the reconstruction algorithms used for image processing at module level. Acquired planar images are processed with novel statistical reconstruction techniques (such as maximum likelihood) based on pixels light response functions which are iteratively identified from flood homogeneous irradiation. Wider field of view, better noise filtering and better uniformity are achieved with respect to the centroid method.
Goal of WP3 (SPECT/MRI systems development, integration and characterization) was the development of the SPECT prototypes, preclinical and clinical.
During the design of the pre-clinical system we performed several experiments and simulations. We used Monte-Carlo photon transportation simulations to design a multi-pinhole collimator for mouse and rat brain imaging. For the mouse brain collimator this complex optimization problem was addressed by evaluating 10 designs with different number of regularly placed pinholes. The size of the holes was fixed at 0.8 mm and for each configuration, and we tested several field-of-view (FOV) sizes. We calculated resolution and sensitivity for these configurations, and simulated imaging of a mathematical phantom. As the zoom factor decreases with increasing number of holes, the achievable reconstructed resolution also decreases, while sensitivity increases with increasing number of holes. Based on parameters and imaging artefacts we selected 5 designs for further refinement. Next we calculated hole-size vs. peak sensitivity curves. For each design we selected a hole size to obtain 1.5 kcps/MBq peak sensitivity and simulated imaging of a hot-rod phantom. Based on the quality of reconstructed images we selected the 3x3 pinhole design, and refined it to improve angular sampling density. Next we optimized the hole size performing a series of hot-rod phantom simulations. Using this optimal design, we investigated effects of the finite penetration of gamma photons into the collimator. Decision was that for Tc99m imaging Tungsten-polyamide composite material would be optimal, with pure Tungsten pinhole-insets.
As we needed a non-conductive radiation shielding material at several places in the instrument, we developed the technology of manufacturing Tungsten-polyamide composite structures. We experimented with Tungsten powders of different grain size and different powder glue mixtures to find the material with highest density. The optimal mixture reached a density of over 11g/ml.
We have put significant effort into designing the cooling for the system. From a system-design perspective, air-cooling was very desirable, but its effectivity was questionable. We performed simulations and several test measurements with various cooling blocks and both air and liquid cooling, to evaluate efficiency of the various alternatives. We selected liquid cooling and a thermally conductive plastic material. Unfortunately, though the initial samples of this thermally conductive plastic material were perfect the larger blocks we obtained for manufacturing the prototype systems were porous, that we realized only during pressure testing of the cooling blocks. Sealing the pores caused several weeks delay.
MRI compatibility was a main design requirement, and we performed several material and component level tests throughout the development process. To avoid eddy currents we could not use bulky metal cooling blocks, and had to find and evaluate thermally conductive plastics and ceramics instead. MRI compatibility of the 3D printed Tungsten pinhole insets was tested in simulations and tests. All electronic component designs were tested separately in the MRI.
Designing MRI compatible electronics was a major challenge of the project. We introduced several special designs, like hatching the power distribution and reference plane layers of the PCB, to reduce eddy currents. Radiated RF noise is a critical issue in MRI environment. This was regularly checked during the development using a shielded box, near field antennas and a spectrum analyzer. We payed special attention to use non-magnetic electronic components, to avoid susceptibility artefacts in the MRI.
To improve MRI compatibility and also to facilitate both the pre-clinical and the clinical system with the same hardware, we developed a distributed data acquisition strategy. Each detector module contains a small FPGA, and data is processed locally. The modules are daisy chained with an optical bus, greatly reducing the size of conducting loops in the system and enabling very flexible design.
Due to delays in the project the software was simplified compared to the original plans. Never the less, we developed a graphical data acquisition tool and a separate program for automatic calibration of the system and for application of corrections to the data. Images from the pre-clinical prototype confirmed the original design specifications.
Finally, WP3 has lead to the construction of the following prototypes:
Pre-clinical system (assembly completed; acquisition and correction software implemented and tested; test of the full prototype in MRI).
Clinical system (design completed, assembly completed with detection modules, clinical RF coil, collimator and mechanical frame).
In the framework of WP4 (Reconstruction software development), software for tomographic image reconstruction and for system calibration was developed for the pre-clinical scanner equipped with multi-pinhole collimators and for the clinical scanner equipped with multi-mini-slit-slat collimators. The iterative reconstruction algorithms ML-EM and OS-EM were used, which are based on forward and back-projection routines to transform data between the object domain and the projection domain. For the pre-clinical system, these procedures were based on Monte Carlo simulation, incorporating a 3D model of the imaging apparatus and the most important physical effects that affect the acquisition. For the clinical system, a flexible angular blurring algorithm was developed, allowing easy modification of the system parameters.
The calibrating procedure consists of estimating a specific set of parameters, describing the acquisition geometry, by comparing measured and calculated projection data from a set of line or point sources. Attenuation correction is performed using a pseudo-CT generated from standard T1-weighted or T2-weighted MRI images by an algorithm developed previously for PET/MRI, and scatter correction is done with an energy-window based approach.
Advanced software algorithms were developed in order to take advantage of simultaneous SPECT/MRI dual modality acquisition. A MAP-EM reconstruction algorithm was implemented, which incorporates high resolution anatomical MRI images in order to improve the quality of the reconstructed SPECT images. Thereby noise amplification can be reduced while preserving anatomic features in the image.
Motion-correction routines were implemented that can utilize high temporal-resolution information, obtained from fast MRI sequences, to reduce motion artefacts in the SPECT images. For the INSERT system, some degree of patient motion may actually be advantageous, as long as it can be monitored accurately and corrected for, as this improves the spatial sampling of the data.
We have implemented a range of methods for partial volume correction, which are post-reconstruction image processing procedures to improve both the qualitative and quantitative aspects of the reconstructed SPECT images by utilizing high-resolution anatomical MRI images. These methods do not suffer from the excessive noise-amplification and ringing artefacts associated with standard de-convolution algorithms. The package includes a novel, practical method, which only requires segmentation of one single region, such as a tumour, rather than the whole image.
Since the INSERT scanner is a stationary system, as opposed to practically all commercial SPECT scanners, it can be used for acquisition of dynamic data, which can provide additional information of clinical relevance. We have developed a series of software tools for performing kinetic analysis in order to take advantage of this type of studies.
In the framework of WP5 (Preclinical evaluation), we have characterized a glioma murine model for the assessment of the INSERT new multimodal system. Moreover this work package has identified important biomarkers for the early estimation of glioma response to TMZ treatment. These biomarkers have been validated in several glioma cell lines, animal models and by different non-invasive imaging techniques.
In detail, UMIL dealt with cell characterization, animal model production, identification of molecular pathways involved in glioma TMZ response, supporting also HSR and CROMed in their activities. HSR identified translational biomarkers able to report about variation in HIF-1a activity with translational imaging techniques (MRI, PET and SPET). Cromed tested the possibility to perform non-invasive imaging with the identified SPECT tracer, set up RT protocol and an ischemia model for further application of the INSERT instrument.
UMIL activities: U251 cells engineered to express the reporter gene luciferase under control of either a constitutive (U251-pGL3) or a HIF-1a inducible promoter (U251-HRE) were in vitro validated. U251-HRE cells were also engineered to express a constitutive fluorescent reporter (mCherry) to in vivo assess their viability by fluorescence imaging and HIF-1a activity by Bioluminescence imaging using an IVIS Spectrum/CT system by Perkin Elmer.
The effect of TMZ treatment was assessed in these cells allowing us to conclude that TMZ treatment is able to reduce HIF-1a activity and that this reduction precedes cell death. Moreover we showed that by inhibiting HIF-1a expression in TMZ resistant cells, glioma sensitivity to this drug can be restored.
HIF-1a activity has been validated as response biomarker in several other cell lines both responsive and resistant to TMZ (U87, LN229, U130, T98G, GLI36s, GLI36r), and after different treatments (MG132, Bafilomycin A1, Geldanamycin, Metformin, miR675-5p) allowing to consider HIF-1a activity reduction as an early biomarker of glioma response to treatment.
An orthotopic glioma model has been set up and characterized both for the growth kinetic, and for its responsiveness to TMZ. HIF-1a activity was confirmed as early biomarker of tumour response to treatment, also in vivo by non-invasive analyses (non invasive reporter gene acitivity by BLI and FLI, uptake assessment of specific commercially available fluorescent probes able to report upon HIF-1a activity (Hypoxisense) and neo angiogenesis (Integrisense)). Changes in HIF-1a activity was also able to induce changes in different processes that can be studied by translatable imaging techniques. Within the project DWI and DCE MRI studies were carried out demonstrating the possibility to early estimate response to TMZ by this procedure. In the same animals PET imaging allowed us to monitor both hypoxia induction and cell proliferation changes after treatment using 18F-FAZA and 18F-FLT tracers, respectively. A number of PSECT tracers were evaluated as well as radiotherapeutic protocol have been set up.MIBI uptake showed a significant reduction after treatment with TMZ and it was chosen for further clinical evaluations.
All these results have been confirmed and validated by immunohistochemical analysis.
Due to technical problems occurred during INSERT instrument delivery in Milan we could not perform the integrated imaging experiments we set up, but biological results obtained with already available instruments allowed us to confirm previous data and the features of INSERT SPECT was evaluated in another experimental set.
The comparison of the the NanoSPECT system and the Consortium-built INSERT system has been performed. While the conditions of preparedness of the INSERT SPECT`s reconstruction interface were suboptimal, the imaging performance measured by contrast means and practicalities of imaging time does support an advantageous judgement of future biological measurements with the pre-clinical INSERT SPECT in mouse brain glioma models.
The highest advantage over state-of-the art is that INSERT SPECT applies all-round detectors and list mode data acquisition, rendering short acquisitions and time-activity curve generation using motionless data acquisition possible.
Moreover, for the general application of the INSERT SPECT it can be inferred that the use of the system for multiple small volume definitions to describe glioma model heterogeneity and its changes is preferred to any other existing SPECT device.
The overall imaging performance of the INSERT is remarkably good even when compared to NanoSPECT.
In the framework of WP6 (Clinical evaluation), we have conducted an up to date audit of patients with glioblastoma multiforme (GBM) treated at UCLH and these data confirm the clinical utility of multi-modality imaging to define early response to standard of care.
We have therefore designed and activated a feasibility protocol in which patients undergoing standard of care chemo-radiotherapy at UCLH will be scanned using sequential MRI and MIBI-SPECT. This protocol will recruit patients over the next 6-12 months at UCLH and provide proof of principle for the use of combined MIBI-SPECT MRI in glioma. The important outputs will include an assessment of utility for radiotherapy planning and as a novel early response scan during chemo-radiotherapy.
We have undertaken an updated review of the literature and re-assessed the potential demand for multi-modality imaging in the context of brain cancer, with reference to on-going studies in the consortium. There remains a clear unmet need for imaging that more closely represents biology in high-grade glioma. A recent published review confirms this perspective.
99mTc- MIBI SPECT is a promising molecular imaging technique based on published literature and data from INSERT WP5 and we have developed a clinical protocol to allow further investigation and provide clinical data on GBM.
Due to delays in the technical development of the system, no clinical studies have been performed with INSERT in the time frame of the funding. We have therefore focused on gaining relevant experience within the consortium on using imaging end-points that will be relevant when the system becomes operational.
Specific on-going work that will support the clinical implementation of this work package beyond the end of the funding period are discussed above and include:
• Clinical protocol developed at UCLH and to be activated at UCLH and University of Leeds
• Phase IIa trial is currently investigating Lutetium DOTATATE in children with primary refractory or relapsed high risk neuroblastoma. Patients are imaged with 123I-mIBG, 18F-FDG PET/CT, Gallium DOTATAE PET/CT plus conventional MRI or CT, and SPECT is used for dosimetry purposes.
• A phase I/II trial is comparing [124I]mIBG PET/CT to[123I]mIBG scintigraphy in patients with metastatic neuroblastoma, with the aim of detecting metastatic neuroblastoma deposits and establish SUV values (Standardized Uptake Value) for [124I]mIBG PET/CT that define positive vs negative findings.
• Functional MRI imaging in patients undergoing standard of care chemo-radiotherapy
• White cell tracking study in patients with solid tumours (University of Leeds)
Due to delays in the project we were not able to perform any experiments on the final system within the time-frame of the project. Instead we used a test-system with a single detector and a rotating platform. We used this system to develop a practical calibration procedure for the final system equipped with multi-mini-slit-slat (MSS) collimators. The procedure consists of a sequence of scans using plane- and line sources for estimation of different calibration parameters. We also used the test-system to perform a number of phantom experiments, which confirmed results we had obtained previously with simulations, showing that the final INSERT system will be able to provide images adequate for the intended purpose – i.e. SPECT imaging of brain tumours.
Potential Impact:
The motivation of the INSERT project lays in the high mortality rate of patients with recurrent glioma. Unfortunately, notwithstanding the innovations brought by the scientific research, the management of these patients, both adults and in paediatric age, is still a challenge, since the life expectancy of these patients is very low and the surgery does not represent a definitive solution in all cases, given to the nature of these tumours.
Thanks to the possibility of obtaining multiple parameters, the INSERT system will enable to better define the tumour biology and to give relevant information for a personalized treatment, with a considerable impact on the efficacy of the treatment itself.
The system will be validated at the pre-clinical level thanks to the creation of animal models, and at the clinical level. The initial focus will be on patients with glioma, but there is future potential to target a range of tumours in the head and neck region.
INSERT will lay the foundation for long-term innovation in the field of treatment planning and response monitoring. Many of its results will be translatable to clinical trials in different fields of oncology, ultimately facilitating treatment planning in individual patients. Therefore, the uptake of the proposed technology in the European and potentially global markets would further enhance the development of personalized therapies in the oncological setting.
List of Websites:
http://www.insert-project.eu