Malarial Retinopathy Screening System for Improved Diagnosis of Cerebral Malaria

摘要 脑型疟疾(CM)是一种与疟疾感染相关的危及生命的临床综合征。 每年，疟疾影响2亿多人，夺去44万多人的生命 世界各地的人们，主要是非洲儿童。由于CM的高发病率，它是 经常被误诊为其他类似症状的病理，导致高误诊 CM阳性率及治疗不正确。一种确认CM存在的准确方法 或者对非疟疾疾病进行调查是改善结果的关键。自.以来 疟疾视网膜病变(MR)对一次CM的特异性和敏感性大于90% 在临床诊断中，视网膜MR筛查是一种有效的辅助和 提高CM诊断的特异性。 VisionQuest Biomedical及其合作者组建了一个跨学科的团队 在自动化视网膜图像分析方面拥有丰富经验的科学家，临床 眼科，专门研究疟疾视网膜病变(MR)和脑型疟疾 诊断(CM)。该团队将开发和测试ASPIRE，这是一种检测MR的系统 由自动磁共振检测软件与低成本和便携式视网膜相集成 摄影机。我们建议的ASPIRE系统将增强而不是取代当前的CM诊断 标准；提高CM诊断的准确性，从而减少错误数量 结果是积极的。 在第一和第二阶段，VisionQuest Biomedical的研究团队开发了自动磁共振 检测软件，并将其与手持视网膜摄像头接口。由此产生的临床 ASPIRE的原型在非洲的一家医疗诊所进行了现场测试，结果显示出极好的性能 无需眼科专家即可检测MR的性能和可用性。同相 II-B，MR检测系统将被精细化、产品化，并由此产生商业 原型将在预期的数据集上进行验证。我们将通过三个方面来实现这一点 明确的目标。在第一个目标中，将适配用于MR检测的软件系统并 改进后可与低成本、便携的InView摄像头配合使用。在第二个目标中，我们将完善 InView的驱动程序软件，并将摄像头与MR检测软件集成，以产生 第一个商业原型。第三个目标将集中于收集视网膜图像数据，以便 算法测试以及用于在观察性临床中验证商业原型 这项研究将在非洲马拉维、乌干达和赞比亚的9个临床地点进行。

Summary Cerebral malaria (CM) is a life-threatening clinical syndrome associated with malarial infection. Annually, malaria affects more than 200 million people and claims the lives of over 440,000 people worldwide, mostly African children. As a consequence of the high incidence of CM, it is often misdiagnosed for other pathologies with similar symptoms, leading to a high false positive rate for CM and incorrect treatment. An accurate means to confirm the presence of CM or to investigate for a non-malarial illness is critically needed to improve outcomes. Since Malarial retinopathy (MR) is greater than 90% specific and sensitive to the presence of CM once clinically diagnosed, retinal screening for MR represents an effective means to assist in and improve the specificity of CM diagnosis. VisionQuest Biomedical and its collaborators have assembled a team of inter-disciplinary scientists with considerable experience in automated retinal image analysis, clinical ophthalmology with specialized research in malarial retinopathy (MR), and cerebral malaria diagnosis (CM). This team will develop and test ASPIRE, a system for detection of MR consisting of automated MR detection software integrated with a low-cost and portable retinal camera. Our proposed ASPIRE system will augment, not replace, the current CM diagnostic standard; increasing the accuracy of CM diagnoses, leading to a smaller number of false positive outcomes. In Phases I and II, the research team at VisionQuest Biomedical developed the automated MR detection software and interfaced it with a handheld retinal camera. The resulting clinical prototype of ASPIRE was tested onsite in a health-clinic in Africa, which demonstrated excellent performance and usability for detecting MR, without the need of an ophthalmic expert. In Phase II-B, the MR detection system will be refined, productized, and the resulting commercial prototype will be validated on prospective datasets. We will accomplish this through three specific aims. In the first aim, the software system for MR detection will be adapted and improved to work with low-cost and portable iNview camera. In the second aim, we will refine iNview’s driver-software and integrate the camera with MR detection software to produce the first commercial prototype. The third aim will focus on collecting the retinal image data for algorithm testing as well as for validating the commercial prototype in an observational clinical study to be conducted at nine clinical sites in Malawi, Uganda, and Zambia in Africa.