STTR Phase I: An Artificial Intelligence (AI)-based algorithm using nanosensor-based salivary analytics to predict clinical outcomes in symptomatic COVID-19 patients

这项小企业技术转移（STTR）第一阶段项目的更广泛影响/商业潜力是为美国医院提供一种工具，帮助在初步诊断时准确预测COVID-19感染患者的预期疾病严重程度。一个简单的界面使用基于人工智能的预测算法，帮助医院做出明智和准确的决定，确定哪些患者需要特定的护理和治疗干预措施。这一改进的流程使医院能够更好、更快地对患者护理和治疗做出决策，重新分配医院资源（包括员工、医院床位、ICU），以实现最大效益。从长期来看，该平台可用于预测与其他传染病有关的疾病，也可用于医院基础设施有限的国家。这个小型企业技术（STTR）一期项目将通过一种新的方法开发一种全新的医疗诊断和预后工具，该方法依赖于分析复杂的多变量信号，反映患者的整个唾液代谢组和蛋白质组。人工智能工具将用于查看是否可以识别与患者结果相关的信号簇。这是对传统医学诊断的彻底背离，传统医学诊断通过评估个体生物标志物来进行临床诊断。这种方法不适合预测病人未来的结果。该试点阶段工作的范围是开发一种有效的算法，并了解算法在预测和分类COVID-19患者结局方面的有效性和可靠性。试点项目的目标是：(i)获取COVID-19患者的生物液样本，以及（ii）开发用于有效预测算法的机器学习技术。多种机器学习技术和比较策略将用于算法开发和功效测试。该奖项反映了美国国家科学基金会的法定使命，并通过使用基金会的知识价值和更广泛的影响审查标准进行评估，被认为值得支持。

The broader impact/commercial potential of this Small Business Technology Transfer (STTR) Phase I project is to provide U.S. hospitals with a tool to help accurately predict the expected severity of illness for COVID-19 infected patients at the time of initial diagnosis. A simple interface uses Artificial Intelligence-based predictive algorithms to help hospitals make informed and accurate decisions about which patients require specific care and treatment interventions. This enhanced process allows hospitals better and faster decision-making on patient care and treatments, redirecting hospital resources (including staff, hospital beds, ICU) for maximum effectiveness. In the longer term, the platform can be adapted to predict illnesses related to other infectious diseases, and also scaled for countries where availability of hospital infrastructure is limited.This Small Business Technology (STTR) Phase I project will develop a completely new category of medical diagnostic and prognostic tools via a novel approach that relies on analysis of a complex multi-variate signal, reflective of the patient’s entire salivary metabolome and proteome. Artificial intelligence tools will be used to see if signal clusters correlating with patient outcomes can be identified. This is a radical departure from traditional medical diagnostics which evaluate individual biomarkers for a clinical diagnosis. Such approaches are ill suited to the task of predicting future patient outcomes. The scope of this pilot phase work is the development of an effective algorithm and understanding algorithm efficacy and reliability in prediction and classification of outcomes for COVID-19 patients. The goals of the pilot project are to: (i) obtain COVID-19 patient bio-fluid samples, and (ii) develop machine learning techniques for an effective predictive algorithm. Multiple machine learning techniques and comparison strategies will be used for algorithm development and efficacy testing.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.