Morphological and topological image analysis for clinical diagnostics and monitoring of cornea damage

角膜是眼睛透明的前部，覆盖着虹膜、瞳孔和前房。清澈的角膜对良好的视力至关重要。眼睛的化学和热烧伤是一种严重的伤害，可能给患者带来灾难性的长期后果，包括慢性眼痛、视力低下和经常失明。这些伤害影响了患者及其家属的生活质量，并对社会产生了强烈的负面影响。据报道，英国眼部化学损伤的发生率在所有眼部外伤中为7%至10%。英国东北部每年的化学伤害发生率为每10万人5.6例，高于英国每年每10万人0.02例的总体发生率。在英国，角膜热烧伤的频率在烟花季节（篝火之夜、除夕等）期间和之后达到峰值。根据英国内政部最近的报告，在过去两年中，由于“酸攻击”，角膜化学烧伤的数量翻了一番。评估和诊断角膜损伤包括检查角膜不同深度的显微照片，并依赖于角膜细胞的形状和排列的识别，特别是在其上皮层。健康角膜和严重受损角膜之间的区别通常很容易辨别，但对术后恢复的监测涉及更微妙、渐进的变化，如果没有专门的定量分析，很难识别这些变化。与此同时，为了评估医疗的成功并为其战略提供信息，这些变化对于发现和描述是至关重要的。在这个项目中，一名执业眼科医生和细胞生物学家将与天体物理学家和应用数学家合作，开发定量的、计算机辅助的角膜显微照片分析方法，以促进角膜损伤的诊断和监测。图像分析是天文学的核心，从根本上说，望远镜和显微镜图像的分析有很多共同之处。基于先进的形态学和拓扑统计图像分析方法，我们建议开发用于天文图像的定量临床诊断工具，该工具基于来自维多利亚皇家医院（NHS, Newcastle upon Tyne）的患者健康和受损眼睛的约10万张体内共聚焦显微镜角膜图像。该项目的目的是探索和应用新颖、先进的图像分析数学方法，以识别和量化角膜损伤的性质和程度，并监测其术后恢复，包括一般和患者特异性。我们的目标是最终开发出一套软件，供眼科医生在日常临床工作中使用，以促进快速、客观和可靠的诊断、损害评估、监测疾病进展和对手术干预的反应，并提供适当准确和可重复的诊断和监测方法，以协助日常医疗决策。

The cornea is the transparent front part of the eye that covers the iris, pupil, and anterior chamber. A clear cornea is essential for good vision. Chemical and thermal burns to the eye constitute a serious form of injury with potentially catastrophic long-term consequences to patients including chronic eye pain, poor vision and frequently blindness. These injuries affect the quality of life of the patients and their families and have a strong negative impact on society. The reported incidence of ocular chemical injuries in the UK ranges from 7% to 10% among all ocular traumas. The annual incidence of chemical injuries in the north-eastern UK is 5.6 cases per 100,000 population, higher than the overall UK rate of 0.02 per 100,000 population per annum. The frequency of thermal cornea burns in the UK peaks during and after the fireworks seasons (Bonfire Night, New Year Eve, etc.). Chemical burns of the cornea have doubled in numbers over the last two years due to "acid attacks" as has been recently reported by the Home Office.Assessment and diagnosis of cornea damage involves inspection of microscopic photographs of the cornea at various depths and relies on the identification of the shapes and arrangements of the cornea cells, especially in its epithelial layer. Distinctions between healthy and strongly damaged cornea are often easily discernible, but monitoring of the post-operative recovery involves more subtle, gradual changes that are difficult to identify without dedicated quantitative analysis. Meanwhile, such changes are critical to discover and characterise in order to assess the success of the medical treatments and to inform their strategy.In this project, a practising ophthalmologist and cell biologist will combine their effort with astrophysicists and applied mathematicians to develop quantitative, computer-aided methods of analysis of microscopic cornea photographs in order to facilitate diagnostics and monitoring of the cornea damage.Image analysis is at the heart of astronomy and, fundamentally, analyses of telescopic and microscopic images have much in common. Building upon advanced morphological and topological methods of statistical image analysis developed in the current STFC grant in application to to astronomical images, we propose to develop tools for quantitative clinical diagnostics of cornea cell damage based on approximately 100,000 in vivo confocal microscopic cornea images of healthy and damaged eyes obtained from patients of the Royal Victoria Infirmary (NHS, Newcastle upon Tyne). The aim of this project is to explore and apply novel, advanced mathematical methods of image analysis to identify and quantify the nature and degree of cornea damage and monitor its post-operative recovery, both general and patient-specific. Our ambition is to develop, eventually, a software package that could be used by ophthalmologists in their day-to-day clinical work in order to facilitate a speedy, objective and reliable diagnostics, damage assessment, monitoring progression of disease and response to surgical intervention, and to provide a suitably accurate and reproducible diagnostic and monitoring methods to assist in the day-to-day medical decision making.