Browsing by Subject "Diagnosis"

Showing increased application in biological and medical fields, acoustofluidics is a combined technology between acoustics and microfluidics. The core function of acoustofluidics is a label-free and contact-free manipulation of particles in the fluid, which can be applied as active separation, active mixing, and active concentration. Since in therapeutic and diagnostic applications, contamination in the samples can significantly interference analysis results and treatment outcome, proper per-screening of the sample can significantly decrease the target detection threshold and avoiding interferences come from noise and misreading. The acoustofluidic technology derive a particle manipulation based on physical properties of the particles and fluids, specifically, the size of the particle, densities for the particles and fluid, and the viscosity of the fluid, which generate a screening system that can separate particles with different sizes and densities. By utilizing this property, acoustofluidics has been applied on separating multiple biological particles and objects including circulating cancer cells, red blood cells, and multiple populations of vesicles. These reagent-free and contact-free separations have been demonstrated biocompatible for cells and vesicles and can conserve the cell viabilities and vesicle cargoes including DNA, miRNA, and proteins. However, current achievements on acoustofluidic manipulation focus on general analysis of the separated components, which are not disease specific biomarkers, and the body fluid using for separation are limited to blood and artificial isotonic solutions including phosphate-buffered saline. Although these works demonstrated acoustofluidic technology is eligible for separating bio-particles that have diagnosis and therapeutic functions, lack of real cases related applications and diseases specific investigations still make the technology’s application abilities being restricted to possibilities but not promised functions. To deeply investigate and demonstrate the acoustofluidic technology’s potential on diagnostic application, the technology was evaluated by using samples related with multiple specific diseases. Since the acoustofluidic technology has been demonstrated eligible for isolating exosomes, which are 50-200 nm vesicles secreted from cells, pathology related exosomes were selected for diagnostic application investigation. Exosomes’ vesicle structures make them ideal candidate for diagnosis, since vesicles formed by lipid bilayer membrane contain both proteins or nucleic acids as cargoes inside and transmembrane or membrane proteins and polysaccharides on the surface. Furthermore, the forming and secreting pathologies of exosomes are highly dependent on endocytosis and exocytosis pathologies, which are influenced by cellular metabolism. Exosomes’ cargoes have been found specifically correlated with secreting cells populations, indicates depending on types of cells, like tumor cells or stem cells, the secreted exosomes will contain different molecules that can be used as biomarkers for reversed identifying secreting cells. Except high values on biological and medical research and applications, exosomes’ small size makes the vesicles difficult for isolation and increase the cost on both equipment and time aspects. Since acoustofluidics provides an active approach for separating nanometer sized particles and the isolation is a continuous procedure, the simple and rapid exosome isolation the acoustofluidics can provide makes the technology high valuable. Considering these improvements, the acoustofluidics can provide on exosome related fields, demonstrating acoustofluidic devices separated exosomes containing disease biomarkers and could be used for diagnostic applications become a necessary step for validating the technology’s ability. In this dissertation, the first attempt for validating acoustofluidic exosome separation’s diagnostic potential was made for isolating salivary exosomes aimed at human papillomavirus (HPV) induced oropharyngeal cancer diagnosis. Different with previous research that worked on blood exosome separation, a unique property of this study is achieving exosome separation from saliva, which is a more unstable system on components and physical properties than blood. By isolating salivary exosome using the acoustofluidic technology and processing down-stream digital droplet polymerase chain reaction (PCR) analysis, HPV-16 virus, which has been found can induce oropharyngeal cancer, was found majorly distributed in isolated exosome fractions. Since saliva has complex components that cause inaccuracy analysis result, the application of acoustofluidic technology can increase the diagnostic sensitive and enable saliva based liquid biopsy for early screening of oropharyngeal cancer. In the next work, we further demonstrate the acoustofluidic technology’s advantage on rapid isolation of exosomes benefits the time sensitive diagnosis. The acoustofluidic devices were applied for isolating exosomes from mice models that were induced to traumatic brain injury (TBI), which can develop to chronic diseases or deteriorate in short term. Since these outcomes induced by improper or untimely treatments, fast screening of TBI becomes critical for achieving ideal therapeutic outcomes. By collecting plasma from mice and deriving exosome isolation through the acoustofluidics devices, isolated exosome samples with less contamination were found compared with original plasma. Protein analysis further indicates isolated exosomes keeps several exosome specific and neuron damage specific proteins, indicates the acoustofluidic technology is biocompatible and low harmful for exosome structures and components. High isolation purity achieved by the acoustofluidic technology also benefits downstream analysis by decreasing detection noise. In flow cytometer analysis, the acoustofluidic devices isolated exosomes demonstrated TBI disease biomarker increasing in 24 h after the mice were induced to TBI, while the plasma sample cannot demonstrate this tendency. The success of revealing early stage TBI biomarker changes indicates the acoustofluidic technology not only can benefit diagnosis, but also eligible for achieving diagnosis in a very early stage of the pathology. Since the acoustofluidic technology had demonstrated a promising performance on biocompatibility and rapid separation, other time-sensitive samples, including live virus was applied for evaluating the device’s performance. To achieve better control and eliminate irrelevant variable, we use cultured reverse transcription virus that is used for mammal cells transfection as target for isolation. The acoustofluidic technology showed reliable isolation of the murine leukemia virus and majority of the virus particles were separated out from the original sample. Virus viability was further validated robust based on the transfection experiments that using acoustofluidic separated virus and original virus samples demonstrated similar level transfection rates. This work indicates except vesicles like exosomes, the acoustofluidic technology is also eligible for isolating virus and keeping its viability, which significantly expands the application of the technology. Next, to expend the acoustofluidic technology’s functions, we utilized the concentration and manipulation ability of the device for deriving high efficiency membrane degradation. By generating strong microstreaming and microstreaming derived shear stress, the acoustofluidic devices can generate strong vertex flow fields in channel that can capture and lyse mammal cells. Since the acoustofluidic cell lysis is totally a physical process without participation of any chemical reagent and demonstrates a high lysis efficiency, this acoustofluidics application has potential for achieving high efficiency cell analysis. Since the acoustofluidic technology has demonstrated potential for concentration and lysis effect by generating high flow rate microstreaming vertex, we further investigated whether similar effect can derive exosome concentration and lysis. By generating acoustofluidic vertex in droplet containing exosome, nanoparticles, and small molecule drugs, exosome concentration and lysis effects were utilized for high efficiency drug loading and carrier encapsulation. Derived by the acoustofluidic concentration effect, the porous nanoparticles and drug molecules are concentrated in small area of the fluid system and this active concentration increasing induces a high drug loading rate. Simultaneously, the acoustofluidic vertex disrupts exosome membrane and concentrates exosomes with the nanoparticles, which induces exosome encapsulation. These exosome encapsulated drug-loaded nanoparticles demonstrate high intake rate of cells and derive more efficient drug delivery rate. Since the drug loading and exosome encapsulation are physical processes, the acoustofluidic technology derived particle manipulation has potential for deriving loading and encapsulation for large varieties of drugs, particles, and vesicles, which significantly expand the technology’s application.

The semiconductor industry continues to relentlessly advance silicon technology scaling into the deep-submicron (DSM) era. High integration levels and structured design methods enable complex systems that can be manufactured in high volume. However, due to increasing integration densities and high operating speeds, subtle manifestation of defects leads to functional failures at the board level. Functional fault diagnosis is, therefore, necessary for board-level product qualification. However, ambiguous diagnosis results can lead to long debug times and wrong repair actions, which significantly increase repair cost and adversely impact yield.

A state-of-the-art diagnosis system involves several key components: (1) design of functional test programs, (2) collection of functional-failure syndromes, (3) building of the diagnosis engine, (4) isolation of root causes, and (5) evaluation of the diagnosis engine. Advances in each of these components can pave the way for a more effective diagnosis system, thus improving diagnosis accuracy and reducing diagnosis time. Machine-learning techniques offer an unprecedented opportunity to develop an automated and adaptive diagnosis system to increase diagnosis accuracy and speed. This dissertation targets all the above components of an advanced diagnosis system by leveraging various machine-learning techniques.

This thesis first describes a diagnosis system based on support-vector machines (SVMs), multi-kernel SVMs (MK-SVMs) and incremental learning. The MK-SVM method leverages a linear combination of single kernels to achieve accurate root-cause isolation. The MK-SVMs thus generated also can be updated based on incremental learning. Furthermore, a data-fusion technique, namely majority-weighted voting, is used to leverage multiple learning techniques for diagnosis.

The diagnosis time is considerable for complex boards due to the large number of syndromes that must be used to ensure diagnostic accuracy. Syndrome collection and analysis are major bottlenecks in state-of-the-art diagnosis procedures. Therefore, this thesis describes an adaptive diagnosis method based on decision trees (DT). The number of syndromes required for diagnosis can be significantly reduced compared to the number of syndromes used for system training. Furthermore, an incremental version of DTs is used to facilitate online learning, so as to bridge the knowledge obtained at test-design stage with the knowledge gained during volume production.

This dissertation also includes an evaluation and enhancement framework based on information theory for guiding diagnosis systems using syndrome and root-cause analysis. Syndrome analysis based on subset selection provides a representative set of syndromes. Root-cause analysis measures the discriminative ability of differentiating a given root cause from others. The metrics obtained from the proposed framework can provide guidelines for test redesign to enhance diagnosis. In addition, traditional diagnosis systems fail to provide appropriate repair suggestions when the diagnostic logs are fragmented and some syndromes are not available. The feature of handling missing syndromes based on imputation methods has therefore been added to the diagnosis system.

Finally, to tackle the bottleneck of data acquisition during the initial product ramp-up phase, a knowledge-discovery method and a knowledge-transfer method are proposed for enriching the training data set, thus facilitating board-level functional fault diagnosis. In summary, this dissertation targets the realization of an automated diagnosis system with the features of high accuracy, low diagnosis time, self-evaluation, self-learning, and ability of selective learning from other diagnosis systems. Machine learning and information-theoretic techniques have been adopted to enable the above-listed features. The proposed diagnosis system is expected to contribute to quality assurance, accelerated product release, and manufacturing-cost reduction in the semiconductor industry.