Optimising a microfluidic assay for sepsis diagnostics by combining numerical simulations with machine learning

改善医疗保健是我们这个时代的全球性挑战之一。脓毒症（由宿主对感染的免疫反应失调引起的危及生命的器官功能障碍）与世界上20%的死亡有关。快速诊断败血症对患者的生存至关重要，因为每延迟治疗一小时，败血症的死亡率就会增加8%。微流体的出现为疾病诊断提供了令人着迷的机会。美国公司Cytovale （https://cytovale.com/）正在开发用于快速（少于10分钟）败血症诊断的微流体技术。尽管最近取得了进展，但在这个博士项目中仍有几个悬而未决的问题需要解决：血细胞的特性（如粘弹性、大小）如何与Cytovale用于败血症诊断的微流体装置中观察到的物理细胞行为联系起来？如何使用机器学习和数据挖掘来构建细胞行为的预测模型？如何优化微流控装置以获得最大的诊断性能？除了在微流体和细胞力学方面产生新的基础知识外，该项目的最终成果是一个用于优化微流体装置的软件工具，用于探测生物细胞的力学特性。

Improving healthcare is one of the global challenges of our time. Sepsis (a life-threatening organ dysfunction caused by a dysregulated host immune response to infection) is linked to 20% of all deaths in the world. Diagnosing sepsis quickly is of utmost importance to the survival of a patient, as mortality from sepsis increases as much as 8% for every hour that treatment is delayed. The emergence of microfluidics has enabled fascinating opportunities for disease diagnostics. The US-based company Cytovale (https://cytovale.com/) is developing microfluidic technologies for rapid (less than 10 min) sepsis diagnostics. Despite recent progress, there are several open questions that will be addressed in this PhD project:How are properties of blood cells (e.g. viscoelasticity, size) linked to the observed physical cell behaviour in Cytovale's microfluidic devices for sepsis diagnostics?;How can machine learning and data mining be used to build a predictive model of the cell behaviour?;How can the microfluidic device be optimised for maximal diagnostic performance?Beside the generation of new fundamental knowledge in microfluidics and cell mechanics, the ultimate outcome of the project is a software tool for the optimisation of microfluidic devices that probe mechanical properties of biological cells.