Predictive Modeling of COVID-19 Progression in Older Patients

这项提议的目标是开发一种预测性模型来识别感染了 SARS-CoV-2，并有患严重新冠肺炎的风险。路易斯安那州的人均死亡率在#年排名第五。 截至2020年5月4日，美国。严重疾病见于年长的人和潜在的 条件。新奥尔良人口特别容易患严重的新冠肺炎高血压， 糖尿病和肥胖症非常猖獗。感染后，病毒造成的急性肺损伤必须修复 恢复肺功能，避免急性呼吸窘迫综合征和肺纤维化。越来越多的证据 提示严重新冠肺炎患者存在细胞因子风暴综合征，可能会加剧 多器官损伤和纤维化并发症的风险。缺乏有效的方法来识别和减轻严重的 新冠肺炎进展持续是由于对导致细胞因子的生物途径的了解有限 风暴综合症和老年患者风险增加。因此，需要确定关键的 导致严重新冠肺炎进展的细胞因子谱，以开发有效的治疗方法。更进一步说，它是 必须找到一种方法来分期疾病轨迹，以精确地识别治疗靶点以减轻 疾病进展和发现预防策略。为此，我们寻求利用数学上的 SARS-CoV-2诱导的肺损伤模型预测急性呼吸窘迫综合征的严重程度 通过考虑关键细胞因子-细胞间的相互作用来实现肺纤维化。我们假设该模型将准确地 用不同新冠肺炎预测关键细胞因子组合和基质蓄积的数量变化 进展在10%的准确率之内。为了做到这一点，我们在杜兰组建了一个调查小组 在病毒学、临床传染病研究、生物信息学和预测方面拥有关键专业知识的大学 组织重塑的数学模型。在提案的目标1中，我们将确定关键的细胞因子 与病毒引起的肺损伤和肺纤维化有关的标志物。这将通过利用 机器学习确定表征重症肌萎缩侧索硬化症进展的生物标志和分子途径 新冠肺炎要制定分众模式。在目标2中，我们将预测老年患者新冠肺炎的严重程度。 模型预测将与新冠肺炎病的血液标记物在老年患者队列中进行比较 疾病发展的不同阶段。模型将通过细胞因子数据进行细化和通知，以识别 可以测试和定向的因果生物途径和疾病过程。我们的预期结果是 已经确定了导致肺组织损伤的关键细胞因子相互作用和决定途径 用于老年患者不同的疾病轨迹。这些结果预计将产生重要影响，因为 所提出的预测模型将为合理设计药物开辟新的研究途径。 对重症新冠肺炎患者的干预。具体地说，这项研究将提供一种改变模式的开源 描述目标疗法、评估其疗效并发展针对特定患者的工具 针对老年人的治疗计划。

The objective of this proposal is to develop a predictive model to identify individuals who are infected with SARS-CoV-2 and at risk of developing severe COVID-19. Louisiana has the 5th highest death rate per capita in the United States as of May 4th, 2020. Severe disease is seen in older individuals and those with underlying conditions. The New Orleans population is particularly susceptible to severe COVID-19 as hypertension, diabetes and obesity are rampant. After infection, acute lung injury caused by the virus must be repaired to regain lung function and avoid acute respiratory distress syndrome and pulmonary fibrosis. Mounting evidence suggests that patients with severe COVID-19 have cytokine storm syndrome, which may exacerbate multiorgan injury and risk of fibrotic complications. Lack of effective ways to identify and attenuate severe COVID-19 progression persist due to limited understanding of the biological pathways responsible for cytokine storm syndrome and increased risk in older patients. Therefore, there is a need to determine the critical cytokine profiles responsible for severe COVID-19 progression to develop effective treatments. Further, it is essential to find a way to stage disease trajectory(ies) to identify therapeutic targets with precision to attenuate disease progression and uncover preventive strategies. Towards this end, we seek to leverage a mathematical model of SARS-CoV-2-induced lung damage to predict severity of acute respiratory distress syndrome and pulmonary fibrosis by considering key cytokine-cell interactions. We hypothesize that the model will accurately predict quantitative changes in suites of key cytokines and matrix accumulation with varying COVID-19 progression within 10% accuracy. To accomplish this, we have assembled an investigative team at Tulane University with key expertise in virology, clinical infectious disease research, bioinformatics, and predictive mathematical models of tissue remodeling. In Aim 1 of the proposal, we will identify the critical cytokine markers linked to viral-induced lung damage and pulmonary fibrosis. This will be accomplished by leveraging machine learning to determine the biomarkers and molecular pathways characterizing progression of severe COVID-19 to focus model formulation. In Aim 2, we will predict the severity of COVID-19 in older patients. Model predictions will be compared to blood markers of COVID-19 disease in cohorts of older patients at different stages of disease progression. The model will be refined and informed by cytokine data to discern causal biological pathways and disease processes that can be tested and targeted. Our expected outcome is to have determined the critical cytokine interactions responsible for lung tissue damage and dictating pathways for varying disease trajectories in older patients. These results are expected to have an important impact as the proposed predictive model will open new avenues of research to rationally design pharmaceutical interventions for severe COVID-19 patients. Specifically, the study will provide a paradigm-shifting open-source tool to delineate target therapeutics, estimate their efficacy, and move towards development of patient-specific treatment plans for older individuals.