COmpton CAmera for Low Irradiation Gamma Hyperspectral Tomography – COCALIGHT-3D

Vectorized radiotherapies using alpha radionuclides (e.g. At211) are strongly developing as they allow to overcome hypoxic tumors radio-resistance on metastatic cancers. But, conventional gamma imaging methods (SPECT & PET) don’t allow to get a good image of dose deposition for those therapies. This leads to systematic undertreatment and thus a loss of chance for the patient.
Those “alpha” nuclides, or their parents have emission lines in gamma, but they are mainly observable above 400 keV and with low event rates. Compton camera technology (CC) seems to fit for this task, but today CC image resolution is poor and its sensitivity is still marginal for this task. Recently, CocaLIGHT project’s partners have obtained promising 2D and 3D images both in term of spatial resolution and in term of image quality with a new and disruptive CC technology: Temporal imaging. This technology gives an excellent signal/noise ratio for the detection of Compton events that could allow great progress in image quality.

The aim of this project is to develop a novel 4 heads Compton camera system (COSPECT) with a resolution and a sensitivity adapted to medical imaging requirements. To reach this objective, we need to design new higher performance 3D reconstruction algorithms well suited to CC. Today, this field is poorly documented and existing algorithms are not suited for the topology of the objects used for reconstruction (thick walls cones, widely scattered distribution probabilities, etc).
The challenging aspect of this project is to obtain high quality only based on a limited number of views (3-6) using temporal imaging CC cameras. In the experimental part of the project, we will use mostly PET emitters (511 keV) in Compton Single Photon Emission Tomography mode. We will compare COSPECT images with PET images on pigs (CERMEP) and human (CUSL) using PET as a gold standard.

Project partners have complementary know-how: UTT has co-developed existing 1-head CC algorithms with DI and has long experience with Bayesian regularization approaches. UCL has a large experience in SPECT & PET image reconstruction, while CERMEP is expert of in vivo imaging including pigs.

CocaLIGHT long term objective is to propose a new nuclear medicine modality above 400 keV dedicated to radiotherapy imaging.