Using AI to enable imaging of exotic radionuclides for Molecular Radiotherapy

1)分子放射疗法（MRT）是一种快速发展的癌症治疗方式，其中与癌细胞结合的分子用放射性核素标记并注射到患者体内，用于靶向放射治疗。可以使用CT和核成像（SPECT/PET）进行多模态成像，以个性化治疗剂量，并量化肿瘤和危险器官（OAR）的吸收剂量。人们越来越关注使用外来放射性核素，包括α-发射体，如Actinium-225，它们有可能增加肿瘤剂量，同时最大限度地减少对OAR的影响，特别是当与特异性靶向肿瘤中受体或蛋白质的新型分子结合时。2)目的和ObjectiesThe的目的是开发新的图像重建方法，使定量成像的外来放射性核素与低水平的γ发射，专注于α-发射体，如锕-225。该方法将基于物理的机器学习方法集成到重建过程中，结合使用成像物理学的高级建模，规划扫描的数据和最先进的深度学习。将研究采集协议的优化。测试数据将包括蒙特卡罗模拟、体模扫描（作为本项目的一部分获得）以及作为临床试验的一部分获得的潜在患者数据。3）研究方法学的新奇由于这些新型放射性核素产生的伽马光子丰度低，因此成像非常具有挑战性。机器学习技术正在彻底改变图像重建领域，特别是超低信号数据。4）与EPSRC的战略和研究领域保持一致该项目与EPSRC的战略保持一致，即通过使用MRT优化治疗诊断学的成像和剂量测定，“发现和加速新干预措施的开发”和“改善人群健康”。5）Blue Earth Therapeutics Ltd.国家物理实验室（NPL）

1) Brief description of the context of the research including potential impactMolecular Radiotherapy (MRT) is a rapidly growing cancer treatment modality where molecules that bind to cancerous cells are labelled with a radionuclide and injected into the patient for targeted delivery of radiation. Multimodality imaging using CT and nuclear imaging (SPECT/PET) can be performed to personalise the treatment doses as well as quantify absorbed doses to the tumours and organs at risk (OAR). There is increased interest in the use of exotic radionuclides including alpha-emitters such as Actinium-225 that have the potential to increase dose to the tumours while minimising impact on OARs, especially when combined with novel molecules that specifically target receptors or proteins in the tumours. 2) Aims and ObjectivesThe aim is to develop novel image reconstruction methods to enable quantitative imaging of exotic radionuclides with low levels of gamma emissions, focusing on alpha-emitters such as Actinium-225. The approach will integrate physics-based machine learning methods into the reconstruction process, combining the use of advanced modelling of the imaging physics, data from planning scans with state-of-the-art Deep Learning. Optimisation of acquisition protocols will be investigated. Test data will include Monte Carlo simulations, phantom scans (acquired as part of this project) as well as potentially patient data obtained as part of a clinical trial.3) Novelty of Research Methodology Imaging these novel radionuclides is very challenging due to their low abundance of generated gamma photons. Machine learning techniques are revolutionising the field of image reconstruction in general, and for ultra-low signal data in particular.4) Alignment to EPSRC's strategies and research areasThis project is aligned with the EPSRC strategy on "discovering and accelerating the development of new interventions" and "improving population health" by optimising imaging and dosimetry for theranostics using MRT.5) Any companies or collaborators involvedBlue Earth Therapeutics LtdNational Physical Laboratory (NPL)