Developing next-generation, AI-enabled, medical image processing for multiple sclerosis clinical trials and routine care.

项目背景：多发性硬化症(MS)是一种导致年轻人残疾的致残性疾病，每年给英国造成超过30亿英镑的损失。从大脑和脊髓拍摄的图像可以告诉我们患者的当前状态以及它们可能会如何演变。目前用于测量大脑和脊髓变化的图像处理工具非常耗时，需要昂贵的多步骤管道，并且依赖于高质量数据的可用性，而这些数据在现实世界的临床实践中经常缺失。候选人将利用计算机视觉领域的最新进展开发一套图像处理工具，以分割脑MRI扫描和脊髓MRI，并使用来自英国各地的临床试验和医院的数据来检测这些结构随时间的变化。该项目将使用皇后广场多发性硬化症中心独特的临床试验数据集、来自英国8家以上医院的MRI数据，并能够与我们的合作者(IXICO)一起测试它们在行业临床试验中的应用潜力。这是学术界和产业界之间开发下一代人工智能算法以影响药物开发和患者护理的独特机会。研究目标：-开发基于深度学习的图像模拟器，以实现脑和脊髓MRI的高效和半监督图像处理(第1年和第2年)-开发用于检测多发性硬化症的疾病活动和疾病进展的超高效图像分割模型(第2年和第3年)-在临床试验和来自我们的合作医院(UCLH Trust/国家神经学和神经外科医院)和合作医院的真实世界数据中验证所开发的模型在整个英国，该项目的结果将是一个独特的模型，将接受大脑或脊髓NHS核磁共振，并将输出大量的相关结构。

Project Background:Multiple sclerosis (MS) is a disabling disease that causes disability in young persons costing the UK more than £3 billion per year. Images taken from brain and spinal cord can tell us what the current status of a patient is and how they may evolve. Current image processing tools to measure changes on brain and spinal cord are time consuming, require expensive and multi-step pipelines, and rely on the availability of high quality data which are often missing in the real world clinical practice. The candidate will develop a set of image processing tools using the latest advances in computer vision field to segment brain MRI scans, spinal cord MRI, and detect changes in these structures over time using data from both clinical trials and hospitals from across the UK. This project will use unique clinical trial datasets at the Queen Square Multiple Sclerosis Centre, MRI data from more than 8 hospitals from across the UK, and be able to test their potential for their applications in clinical trials in the industry with our collaborator (IXICO). This is a unique opportunity in between the academia and industry to develop the next generation of AI algorithms to impact drug development and patient care.Research aims:- Develop deep-learning based image simulators to enable efficient and semi-supervised image processing for the brain and spinal cord MRI (year 1 and year2)- Develop ultra-efficient image segmentation models for detecting disease activity and disease progression in multiple sclerosis using longitudinal image processing (year 2 and year 3)- Validate the developed models in clinical trials and real world data from our partner hospital (UCLH Trust / National Hospital for Neurology and Neurosurgery) and collaborating hospitals from across the UKThe outcome of the project will be a unique model that will receive either brain or spinal cord NHS MRI and will output volumes of relevant structures.