Early detection and diagnosis of age-related macular degeneration AMD using machine learning

AMD是导致老年人失明的最常见原因。到2020年，预计将有2亿人受到AMD的影响，到2040年将增加到近3亿人[1]。这是一种复杂的、可遗传的和异质性的疾病，影响黄斑，黄斑是中央视网膜，负责详细的中央视力。AMD的发病机制尚不清楚。玻璃斑疹是该病的早期临床特征，可见主要位于整个黄斑的黄色沉淀物。目前，AMD的严重程度被分为早、中和晚期AMD，根据黄斑玻璃疣的大小和形态以及色素变化。然而，基于这些特征的AMD的个体间进展率差异极大。最近成像技术的进步使AMD的其他成像生物标记物得以识别，如网状假性玻璃体(视网膜下沉积)，其发生和发展独立于玻璃体[2]。光学相干断层扫描(OCT)还表明，患者体内可能同时存在38种不同类型的玻璃疣，它们在AMD进展中的意义尚不清楚[3]。此外，并不是所有视网膜的高危特征都进展到个体中的晚期AMD，这突出了计算分析的必要性，以更好地了解这些标记物的轨迹。同样，晚期AMD可能表现为黄斑或脉络膜新生血管(CNV)的地理性萎缩(GA)和视网膜色素上皮(RPE)萎缩，脉络膜血管迁移到视网膜，目前尚不清楚为什么、何时以及哪个标志物/事件先于或预测每种形式的晚期AMD。因此，有一个尚未得到满足的需求，以更好地了解AMD的病理生理学，准确地对AMD进行分类，并开发个性化的风险预测模型，以进入该疾病的预防试验。PHD项目将与由惠康信托资助的顶峰财团合作进行。参考：[1]Wong，W.L.，X.Su，X.Li，C.M.Cheung，R.Klein，C.Y.Cheng和T.Y.Wong，2020和2040年老年性黄斑变性的全球流行和疾病负担预测：系统回顾和荟萃分析。《柳叶刀全球健康》杂志，2014。2(2)：E106-E116页。[2]Sivolrasad，S.，A.Bird，R.NitiahPapand，L.Nicholson，P.Hykin，I.Chatziralli，等人，《老年性黄斑变性中网状假性黄斑病变的前景》。《冲浪眼科》，2016。61(5)：第521-37页。[3]Khanifar，A.A.，A.F.Korea ishi，J.A.Izatt，和C.A.Toth，光谱域光学相干断层扫描在老年性黄斑变性中的玻璃体超微结构成像。眼科，2008年。115(11)：1883-90页。

AMD is the commonest cause of blindness in the elderly. By 2020, 200 million people are expected to be affected with AMD, increasing to nearly 300 million by 2040 [1]. It is a complex, heritable and heterogeneous disease that affects the macula, the central retina that is responsible for detailed central vision. The pathogenesis of AMD remains unclear. Drusen are an early clinical feature of the disease and are visualised as yellow deposits predominantly located throughout the macula. Currently, severity of AMD is classified as early, intermediate and late AMD based on the size and morphology of drusen and pigmentary changes at the macula. However, the inter-individual progression rates of AMD based on these characteristics are extremely variable.Recent advances in imaging technologies have enabled identification of other imaging biomarkers of AMD such as reticular pseudodrusen (subretinal deposits) that develop and progress independent of drusen [2]. Optical coherence tomography (OCT) also indicates that there are 38 different types of drusen that may co-exist in a patient and their significance in AMD progression is unknown [3]. Furthermore, not all high-risk features in the retina progress to late AMD within an individual highlighting the need for computational analysis to better understand the trajectories of these markers. Similarly, late AMD may present as geographic atrophy (GA) with atrophy of photoreceptors and retinal pigment epithelium (RPE) in the macula or choroidal neovascularization (CNV) where choroidal blood vessels migrate into the retina and it remains unknown why, when and which marker/event precede or predict each form of late AMD. Therefore, there is an unmet need to better understand the pathophysiology of AMD, accurately classify AMD and develop personalised risk prediction models to progress into prevention trials for this condition.The PhD project will be carried in collaboration with the PINNACLE consortium funded by the Wellcome Trust.References:[1] Wong, W.L., X. Su, X. Li, C.M. Cheung, R. Klein, C.Y. Cheng, and T.Y. Wong, Global prevalence of age-related macular degeneration and disease burden projection for 2020 and 2040: a systematic review and meta-analysis. The Lancet Global Health, 2014. 2(2): p. e106-e116.[2] Sivaprasad, S., A. Bird, R. Nitiahpapand, L. Nicholson, P. Hykin, I. Chatziralli, et al., Perspectives on reticular pseudodrusen in age-related macular degeneration. Surv Ophthalmol, 2016. 61(5): p. 521-37.[3] Khanifar, A.A., A.F. Koreishi, J.A. Izatt, and C.A. Toth, Drusen ultrastructure imaging with spectral domain optical coherence tomography in age-related macular degeneration. Ophthalmology, 2008. 115(11): p. 1883-90.