Development of a software system to automatically detect and quantify foot collapse

夏科氏关节病是继发于周围神经病变的关节塌陷，最常见的继发于糖尿病或酒精中毒。这是一种使人衰弱的疾病，会导致关节的进行性破坏，伴随相关的疼痛和行走力学的改变。足部经常变形，增加关节不稳定、软组织溃疡和最终下肢截肢的风险。及时的诊断使患者有适当的足病评估，矫形，保守支具或手术固定。目前的诊断和严重程度的分析依赖于足部侧位X线片上的放射学几何测量。获得这些测量值非常耗时，并且容易受到观察者间和观察者内变化的影响。该项目将开发和验证一个软件系统，以自动检测和量化足部塌陷。机器学习方法将被应用于自动和可靠地识别足部X光照片上的特征点。然后，自动识别的特征点将用于自动计算与检测和测量足部塌陷相关的几何测量值（例如，Meary角）。该系统在获得后者的性能将进行比较，手动地面实况测量。此外，统计形状模型将用于研究患病病例中足部和踝关节的骨骼形状，旨在改进早期检测和评估进展的方法。待开发的方法和软件有可能节省放射科医生的报告时间，并增加疾病相关的几何测量的再现性。此外，早期发现疾病可能会改善患者的预后，降低NHS的成本。学生将加入一个成熟的研究小组，并将获得在跨学科团队中工作的丰富经验。他们将获得用于临床成像问题的最先进机器视觉算法开发的大量知识。学生将被置于翻译领域，并将有机会更多地了解研究进展对临床实践影响的途径。最后，他们将与来自索尔福德皇家NHS基金会信托的临床合作伙伴合作，深入了解足部和踝关节解剖学和病理学，特别是Charcot关节病的临床成像、诊断和治疗。

Charcot arthropathy is collapse of a joint secondary to peripheral neuropathy, most commonly secondary to diabetes or alcoholism. It is a debilitating condition which causes progressive destruction of the joint with associated pain and alteration in the mechanics of walking. The foot often deforms, increasing the risk of joint instability, soft tissue ulceration, and ultimately lower extremity amputation. Prompt diagnosis allows patients to have appropriate podiatry assessment, orthotics, conservative bracing or surgical fixation. Current diagnosis and analysis of severity relies on radiologically derived geometric measurements on a lateral foot radiograph. These measurements are time consuming to obtain and are susceptible to inter- and intra-observer variability. This project will develop and validate a software system to automatically detect and quantify foot collapse. Machine-learning methods will be applied to automatically and reliably identify feature points on foot radiographs. The automatically identified feature points will then be used to automatically calculate geometric measurements of relevance to detecting and measuring foot collapse (e.g. Meary's angle). The performance of the system in obtaining the latter will be compared to manual ground truth measurements. Furthermore, Statistical Shape Models will be used to study the bone shape of the foot and ankle in diseased cases, aiming to improve methods for early detection and assessment of progression. The methods and software to be developed has the potential to save on reporting time for radiologists and to increase the reproducibility of disease-related geometric measurements. Furthermore, earlier detection of disease is likely to improve patient outcome and reduce costs to the NHS.The student will join a well-established research group, and will gain extensive experience of working in an interdisciplinary team. They will gain vast knowledge of state-of-the-art machine vision algorithm development for clinical imaging problems. The student will be placed in the translational space, and will have the opportunity to learn more about the pathway of progressing research towards impact in clinical practice. Finally, they will work with our clinical partners from Salford Royal NHS Foundation Trust and gain in-depth knowledge of foot and ankle anatomy and pathology, particularly the clinical imaging, diagnosis and treatment of Charcot arthropathy.