CAREER: Reconfigurable Electro-Fluidic Prescriptions (REFRx): Data-Driven Biosensors for Detection and Treatment of Multidrug-Resistant Cancers

这项提议是开发一种工具，可以快速识别肿瘤中的耐药癌细胞，并为患者开出将癌症复发几率降至最低的疗程。耐药性是治疗癌症和传染病的最大障碍之一，已被确定为未来几十年最大的公共卫生威胁之一。所提出的微型仪器可以用于快速筛查癌症患者的耐药性，识别涉及的关键分子分子，并选择最佳的癌症治疗药物。在这项工作中，提出了一种微流体/电子/数据驱动的横切方法，以实现一种快速技术，该技术可以使用机器学习来识别耐药细胞，并使用无标记传感阵列来检测导致耐药的关键蛋白质通路。建议的平台是自适应的，并重新配置以按需分析相关蛋白质，避免了耗费资源的暴力方法。这一跨学科项目将在不同领域吸引和培训研究生和本科生。PI还将通过外展讲习班吸引K-12学生，通过教育讲座吸引当地行业，通过开发在线课程吸引普通公众，从而广泛传播知识。一种新型的数据驱动的生物传感器将被开发出来，这种传感器可以根据需要进行自我调整，以检测和治疗耐多药癌症。由于肿瘤亚克隆的快速突变和对化疗药物的不敏感，使用静态分析平台很难治疗癌症中的多药耐药，因此适应性方法可能更有效。将开发一个全电子平台，用于快速分选耐药细胞，并使用可重新配置的传感器阵列自适应地分析涉及耐药的分子途径。该分析仪将使用可重构的阻抗细胞仪结合机器学习反复检测耐药细胞，然后使用介电泳法(DEP)对它们进行分类，使用可重构的无标记蛋白质传感器阵列进行分析，然后使用机器学习分类器进一步对耐药细胞进行亚型划分，以选择要在后续迭代中测试的药物组合。将构建和表征一个集成的闭环反馈系统，用于对药物敏感和耐药细胞进行分选，并反复分析针对关键候选药物的潜在耐药途径的差异表达。这项研究的结果将是一种新型的数据驱动生物传感器，可以从乳腺肿瘤组织中检测耐药癌细胞并对其进行亚型划分，可广泛应用于包括抗微生物耐药性在内的多种生物医学应用。这一奖项反映了NSF的法定使命，并通过使用基金会的智力优势和更广泛的影响审查标准进行评估，被认为值得支持。

This proposal is to develop an instrument that can rapidly identify drug resistant cancer cells in tumors and prescribe a course of treatment for the patient that minimizes chance of cancer recurrence. Drug resistance is one of the greatest impediments to treating both cancer and infectious disease and has been identified as one of the greatest public health threats of the next several decades. The proposed miniaturized instrument can be utilized for rapidly screening cancer patients for drug resistance and identifying the key molecular players involved and selecting optimal cancer treatment drugs. In this work, a microfluidics/electronic/data-driven crosscut approach is proposed to enable a rapid technology that can identify drug resistant cells using machine learning and examine the key protein pathways resulting in resistance using a label-free sensing array. The proposed platform is adaptive and reconfigures itself to assay the relevant proteins on-demand and avoids a resource-hungry brute force approach. This interdisciplinary project will engage and train both graduate and undergraduate students in various areas. The PI will also engage K-12 students through outreach workshops, local industry through educational lectures, and the general public through development of an online course, resulting in broad dissemination of knowledge. A new class of data-driven biosensors will be developed that can adapt themselves on-demand to detect and treat multidrug resistant cancers. Treatment of multi-drug resistance in cancer is difficult using static analysis platforms because of the rapid ability of tumor sub-clones to mutate and become insusceptible to a chemotherapeutic drug, thus an adaptive approach can be more efficient. An all-electronic platform will be developed for rapidly sorting drug resistant cells and adaptively analyzing the molecular pathways involved in resistance using a reconfigurable array of sensors. The analyzer will iteratively detect drug resistant cells using reconfigurable impedance cytometry in conjunction with machine learning, then sort them using dielectrophoresis (DEP), analyze them using a reconfigurable array of label-free protein sensors, and then use a machine learning classifier to further sub-type the drug resistant cells to select a drug combination to test in the subsequent iteration. An integrated closed-loop feedback system will be fabricated and characterized for sorting drug-sensitive and drug-resistant cells, and analyze differential expression of potential drug-resistance pathways iteratively against key drug candidates. The result of this research will be a new class of data-driven biosensors that can detect and sub-type drug resistant cancer cells from a breast tumor tissue that can be broadly applied to a multitude of biomedical applications including anti-microbial resistance.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.