Robust High-throughput Mechanical Phenotyping Platform for Clinical Diagnostics

描述（由申请人提供）：人们对使用细胞的机械特性作为无标记生物标志物在基础生物学和临床诊断中的应用越来越感兴趣。例如，单细胞变形能力的增加与细胞系和患者样本中的恶性肿瘤有关。与传统的分子生物标志物相比，机械特性是感兴趣的细胞固有的，不需要大量的样品制备或昂贵的标签。尽管如此，该领域的临床应用仅限于研究设置，因为以前的工具缺乏有效检查复杂临床样品异质性的吞吐量。提出的工作建立在可变形性细胞仪平台上，该平台采用惯性微流体原理创建流体剪切拉伸流结，每秒可以分析超过1000个细胞的力学特性。在研究平台阶段，可变形性细胞术系统在识别胸腔液中的恶性细胞方面显示出非常有前景的临床诊断应用，其灵敏度远高于金标准细胞学分析。在这里，我们建立了这一初步成功，并解决了两个主要挑战，以创建一个商业上可行的高通量机械表型平台。我们的目标是降低与高速相机读出相关的成本，减少实现近实时快速决策分析所需的计算资源，更快的数据周转时间和下游排序能力。具体来说，本提案的目的是探索低成本的单点询问模式和高速成像替代方案的可行性，同时保持类似的诊断性能。单点询问方法包括变形前细胞在机械施加应力之前与变形后细胞的传递时间。或者，变形事件期间散射光的单点测量，类似于侧散射流式细胞术测量，将与变形的细胞形状相关。由于输出数据是一维的，这两种方法都有望减少计算资源。三种可供选择的高速成像技术也将被评估1)位置敏感探测器（psd）是一轴成像传感器-通过放置两个psd正交二维空间数据可以在变形事件中收集，2)商用OEM高速图像传感器。可定制，其中板载FPGA集成可用于预筛选数据，以减少离线图像分析工作量，3)适应低帧率相机（10,000 fps）与500ns频闪源耦合。任何提出的替代获取方法都将降低仪器的成本，缩短分析时间，从而创建临床相关的无标签机械生物标志物诊断产品。

DESCRIPTION (provided by applicant): There has been increased interest in using mechanical properties of cells as label-free biomarkers for applications in basic biology and clinical diagnostics. Increased single-cell deformability has been correlated with malignancy in cell lines and patient samples for example. In contrast to conventional molecular biomarkers a mechanical property is intrinsic to the cells of interest and does not require extensive sample preparation or costly labels. Still, clinical applications in this field have been limited to the research setting as previous tools have lacked the throughput to examine the heterogeneity of complex clinical samples effectively. The proposed work builds upon the deformability cytometry platform, which employs inertial microfluidic principles to create a fluid shearing extensional flo junction that can analyze mechanical properties of over 1000 cells per second. At the research platform stage, the deformability cytometry system has shown very promising clinical diagnostic utility in identifying malignant cells in pleural fluid with much higher sensitivity than the gold standard cytological analysis. Here we build off this initial success and address two major challenges to create a commercially viable high- throughput mechanical phenotyping platform. We aim to reduce costs associated with high-speed camera readout, and decrease computational resources required to achieve near-real-time-rapid-decision analysis, faster data turnaround time and downstream sorting capabilities. Specifically, the aims of this proposal are to explore the feasibility of lower cost single-point interrogation modalities and high-speed imaging alternatives while retaining similar diagnostic performance. Single-point interrogation methods include transit-time of a pre- deformed cell prior to mechanically applied stress compared to post-deformed cells. Alternatively, single-point measurements of scattered light during the deformation events, similar to side-scatter flow cytometry measurements will be correlated to deformed cell shape. Both these approaches are expected to reduce computational resources as output data is 1D. Three alternative high-speed imaging techniques will also be evaluated 1) Position sensitive detectors (PSDs) are one axis imaging sensors - by placing two PSDs orthogonally 2D spatial data can be collected during deformation events, 2) Commercially available OEM high- speed image sensors. are customizable where onboard FPGA integration can be used to prescreen data in order to reduce offline image analysis workload, 3) adapted lower frame rate camera (10,000 fps) coupled to 500ns strobe source. Any of the proposed alternative acquisition methods will reduce the cost of the instrument and lower analysis time to create a clinically relevant label-free mechanical biomarker-based diagnostic product.