Comprehensive Assessment of Retinal Vasculature (CARV)

 描述(申请人提供)：在美国，每年有72万例心脏病发作，其中51.5万例是首次心脏病发作。心血管疾病(CVD)的危险因素被认为是明确的，但它们并不能识别所有的高危个体。因此，有必要检测其他因素，为心血管疾病提供更多的近期风险评估。视网膜中发现的危险因素的早期表现可以检测到即将发生的心血管事件的威胁，监测这些前兆可以早期洞察心血管疾病的风险。为此，我们开发了“视网膜血管综合评估”(CARV)，这是一种全自动筛查设备，用于检测视网膜异常并评估相关的CVD风险。在第一阶段，Vision Quest Biomedical的研究团队展示了CARV第二阅读器系统的可行性，该系统可以提供客观测量并检测潜在的血管异常，以帮助视网膜阅读器，从而在减少阅读时间的同时提高疾病检测率。在第二阶段，第二阅读器系统将演变为全自动第一阅读器系统，根据视网膜异常量化心血管疾病的风险。我们将通过两个具体目标来实现这一目标。在第一个目标中，我们将增加检测视网膜异常和病变的其他方法。然后，我们将开发一个分类器，自动为每个病例CAV过程分配低、中或严重的心血管疾病风险。在第二个目标中，我们将结合这些软件工具来产生一个近乎实时的系统，以自动检测高血压视网膜病变的迹象，并根据结果分配CVD风险等级。在第二阶段结束时，我们将演示经过验证的CAV第一读卡器和第二读卡器系统。CAV第二阅读器很可能会得到FDA的I类设备，我们预计一年内就会获得批准。其预期用途将是在远程医疗环境中帮助读者。在CAV第二阅读器作为判定设备的情况下，CAV第一阅读器系统可以在成功完成临床试验后作为510k设备提交给FDA。

 DESCRIPTION (provided by applicant): In the US, there are 720,000 annual incidences of heart attack, out of which 515,000 are first time heart attacks. Risk factors for cardiovascular disease (CVD) are thought to be well-defined, but they do not identify all at-risk individuals. Therefore, there is a need to detect other factors that provide more near-term risk assessment for CVD. The early manifestations of risk factors found in the retina can detect a threat of an impending cardiovascular event, and monitoring these precursors provides an early insight into risk of CVD. For this reason, we have developed the "Comprehensive Assessment of Retinal Vasculature" (CARV), a fully automated screening device for the detection of retinal abnormalities and assessment of associated CVD risk. In Phase I, the research team at Vision Quest Biomedical demonstrated the feasibility of the CARV second reader system to provide objective measurements and detect potential vascular abnormalities to assist a retinal reader, resulting in increased disease detection at a reduced reading time. In Phase II, the second reader system will evolve into a fully automatic first reader system that quantifies the risk of CVD based on retinal abnormalities. We will accomplish this through two specific aims. In the first aim, we will add additional methods for detecting retinal abnormalities and lesions. We will then develop a classifier to automatically assign a CVD risk of low, moderate, or severe to each case CARV processes. In the second aim, we will combine these software tools to produce a near real-time system to automatically detect signs of hypertensive retinopathy and assign a CVD risk grade based on the results. At the end of Phase II, we will demonstrate a validated CARV first reader and second reader system. CARV second reader will likely be given a Class I device by the FDA, and we expect clearance within a year. The intended use will be that of an aid to readers in a telemedicine environment. With CARV second reader serving as a predicate device, the CARV first reader system can be submitted as a 510k device to the FDA, following the successful completion of a clinical trial.