Real-time, large-area microbial mapping to prevent the spread of healthcare-associated infections

摘要 医疗相关感染（HAI）是对患者安全的威胁。阻止HAI的传播需要预防， 监测和疫情调查。许多医院正在采用增强的清洁协议，如紫外线消毒， 他们的标准清洁协议，当正确实施时，已被证明可以减少HAI的传播。 实施增强清洁方案的医疗保健提供者面临的一个重大挑战是验证其有效性 这些干预措施。使用常规微生物计数方法（如培养和菌落计数）进行采样。 可以使用计数方法、聚合酶链反应方法和免疫测定方法，但是这些方法是不可行的。 劳动密集型，仅对有限区域进行采样，且不提供实时结果。光谱和光谱成像 由于最少的样品制备和快速的数据采集，这些技术已经变得流行和有吸引力。荧光 光谱学使用紫外线光源来激发微生物中分子中的电子， 发出的光。不同的微生物将基于其组成分子产生不同的荧光特征， 如蛋白质、维生素和辅酶。基于传统光谱仪的荧光光谱工具 目前用于量化制药和食品生产应用中的生物负载。然而，这些系统基于 常规光谱仪上的光谱仪具有有限的采样区域。将荧光光谱法应用于大面积区域需要 高光谱成像仪测量图像中每个像素的空间和光谱信息。在实验室环境中， 这种高光谱成像仪通常是扫描图像光谱仪，其体积大、昂贵、需要样品， 传感器是静止的，可能需要几分钟才能捕捉到一个图像。Nanohohead开发了一种固态芯片级 高光谱成像仪提供实时全帧数据收集和空间配准的光谱数据。 Nanohnik建议开发一种手持式，大面积，实时荧光成像检测器（HEART-FID）， 在医疗保健环境中实现生物负载的映射。HEART-FID系统的关键部件是定制的 具有由嵌入式图像采集和处理控制的激发源的荧光成像器， 指纹和机器学习来区分细菌、真菌和其他有机材料。在I期 该计划，该小组将证明该系统可以测量不同浓度的目标微生物上， 我们将这些结果与医院中常见的相关表面进行比较，并将这些结果与现有的微生物监测方法进行比较。 第二阶段计划的目标是HEART-FID系统的设计、优化和性能演示 这可以纳入已建立的消毒程序。该原型将在2015年期间升级到TRL 5-6。 第二阶段计划的能力图像100厘米x 100厘米在不到10秒，并区分细菌， 真菌和无害的有机物质。该系统将为医疗保健提供者提供新的数据， 轻松评估其清洁程序，并快速跟踪和预防潜在的爆发。

Summary Abstract Healthcare-associated infections (HAIs) are a threat to patient safety. Stopping the spread of HAIs requires prevention, surveillance, and outbreak investigations. Many hospitals are employing enhanced cleaning protocols, such UV disinfection, to their standard cleaning protocols, which when properly implemented have been shown to decrease the spread of HAIs. A significant challenged faced by healthcare providers implementing enhanced cleaning protocols is verifying their efficacy of these interventions. Sampling using conventional microorganism counting methods, such as culturing and colony- counting methods, polymerase chain reaction methods, and immunoassay approaches, can be used, but these methods are labor-intensive, sample only limited areas, and do not provide real-time results. Spectroscopic and spectral imaging techniques have become popular and attractive due to minimal sample preparation and rapid data acquisition. Fluorescence spectroscopy uses an ultraviolet light source to excite electrons in molecules in microorganisms and measures the visible light emitted. Different microorganisms will produce different fluorescence signatures based on their constituent molecules, such as proteins, vitamins, and coenzymes. Fluorescence spectroscopy tools based on conventional spectrometers are currently used to quantify the bioburden in pharmaceutical and food production applications. However, these systems based on conventional spectrometers have a limited sampling area. Applying fluorescence spectroscopy to large areas requires a hyperspectral imager that measures both spatial and spectral information for each pixel in the image. In laboratory settings, this hyperspectral imager is typically a scanning image spectrometer, which is bulky, expensive, requires the sample and sensor to be still, and can take minutes to capture a single image. Nanohmics has developed a solid-state chip-scale hyperspectral imager that provides real-time full-frame data collection and spatially registered spectral data. Nanohmics proposes to develop a handheld, extensive-area, real-time fluorescence imaging detector (HEART-FID) to enable mapping of the bioburden in healthcare settings. The key components of the HEART-FID system are a custom fluorescence imager with excitation sources controlled by an embedded image acquisition and processing that uses spectral fingerprints and machine learning to differentiate between bacteria, fungi, and other organic materials. In the Phase I program, the team will demonstrate that the system can measure different concentrations of target microorganisms on relevant surfaces typically found in hospitals and compare these results to established methods of microbial monitoring. The goal of the Phase II program will be the design, optimization, and performance demonstration of a HEART-FID system that can incorporated into established disinfection routines. The prototype will be advanced to TRL 5-6 over the course of the Phase II program with the ability image a 100cm x 100cm in less than 10 seconds and distinguish between bacteria, fungi, and nonharmful organic materials. The system will provide healthcare providers with new data that will allow them to easily evaluate their cleaning procedures and quickly track and prevent potential outbreaks.