A microfluidic biosensor for the diagnosis of chronic wasting disease.

Cervids are affected by a neurologic disease that is always fatal to individuals and has population effects. This disease is called chronic wasting disease (CWD) and is caused by a misfolded prion protein. The disease is transmitted via contact with contaminated body fluids and tissue or exposure to the environment, such as drinking water or food. Current CWD diagnosis depends on ELISA screening of cervid lymph nodes and subsequent immunohistochemistry (IHC) confirmation of ELISA-positive results. The disease has proven to be difficult to control in part because of sensitivity and specificity issues with the current test regimen. We have investigated an accurate, rapid, and low-cost microfluidic microelectromechanical system (MEMS) biosensing device for the detection of CWD pathologic prions in retropharyngeal lymph nodes (RLNs), which is the current standard type of CWD diagnostic sample. The device consists of three novel regions for concentrating, trapping, and detecting the prion. The detection region includes an array of electrodes coated with a monoclonal antibody against pathologic prions. The experimental conditions were optimized using an engineered prion control antigen. Testing could be completed in less than 1 hour with high sensitivity and selectivity. The biosensor detected the engineered prion antigen at a 1:24 dilution, while ELISA detected the same antigen at a 1:8 dilution. The relative limit of detection (rLOD) of the biosensor was a 1:1000 dilution of a known strong positive RLN sample, whereas ELISA showed a rLOD of 1:100 dilution. Thus, the biosensor was 10 times more sensitive than ELISA, which is the currently approved CWD diagnostic test. The biosensor's specificity and selectivity were confirmed using known negative RPLN samples, a negative control antibody (monoclonal antibody against bovine coronavirus BCV), and two negative control antigens (bluetongue virus and Epizootic hemorrhagic disease virus). The biosensor's ability to detect pathogenic prions was verified by testing proteinase-digested positive RLN samples.

A microfluidic biosensor architecture for the rapid detection of COVID-19.

The lack of enough diagnostic capacity to detect severe acute respiratory syndrome coronavirus 2 (SARS-COV-2) has been one of the major challenges in the control the 2019 COVID pandemic; this led to significant delay in prompt treatment of COVID-19 patients or accurately estimate disease situation. Current methods for the diagnosis of SARS-COV-2 infection on clinical specimens (e.g. nasal swabs) include polymerase chain reaction (PCR) based methods, such as real-time reverse transcription (rRT) PCR, real-time reverse transcription loop-mediated isothermal amplification (rRT-LAMP), and immunoassay based methods, such as rapid antigen test (RAT). These conventional PCR methods excel in sensitivity and specificity but require a laboratory setting and typically take up to 6 h to obtain the results whereas RAT has a low sensitivity (typically at least 3000 TCID50/ml) although with the results with 15 min. We have developed a robust micro-electro-mechanical system (MEMS) based impedance biosensor fit for rapid and accurate detection of SARS-COV-2 of clinical samples in the field with minimal training. The biosensor consisted of three regions that enabled concentrating, trapping, and sensing the virus present in low quantities with high selectivity and sensitivity in 40 min using an electrode coated with a specific SARS-COV-2 antibody cross-linker mixture. Changes in the impedance value due to the binding of the SARS-COV-2 antigen to the antibody will indicate positive or negative result. The testing results showed that the biosensor's limit of detection (LoD) for detection of inactivated SARS-COV-2 antigen in phosphate buffer saline (PBS) was as low as 50 TCID50/ml. The biosensor specificity was confirmed using the influenza virus while the selectivity was confirmed using influenza polyclonal sera. Overall, the results showed that the biosensor is able to detect SARS-COV-2 in clinical samples (swabs) in 40 min with a sensitivity of 26 TCID50/ml.

A microfluidic biosensor for rapid simultaneous detection of waterborne pathogens.

A microfluidic based biosensor was investigated for rapid and simultaneous detection of Salmonella, Legionella, and Escherichia coli O157:H7 in tap water and wastewater. The biosensor consisted of two sets of focusing electrodes connected in parallel and three sets of interdigitated electrodes (IDE) arrays. The electrodes enabled the biosensor to concentrate and detect bacteria at both low and high concentrations. The focusing region was designed with vertical metal sidewall pairs and multiple tilted thin-film finger pairs to generate positive dielectrophoresis (p-DEP) to force the bacteria moving toward the microchannel centerline. As a result, the bacterial pathogens were highly concentrated when they reached the detection electrode arrays. The detection IDE arrays were coated with three different antibodies against the target bacterial pathogens and a cross-linker to enhance the binding of antibodies to the detection electrode. As the binding of bacterial pathogen to its specific antibodies took place, the impedance value changed. The results demonstrated that the biosensors were capable of detecting Salmonella, Legionella, and E. coli 0157:H7 simultaneously with a detection limit of 3 bacterial cells/ml in 30 - 40 min.

Review of deep learning: concepts, CNN architectures, challenges, applications, future directions.

In the last few years, the deep learning (DL) computing paradigm has been deemed the Gold Standard in the machine learning (ML) community. Moreover, it has gradually become the most widely used computational approach in the field of ML, thus achieving outstanding results on several complex cognitive tasks, matching or even beating those provided by human performance. One of the benefits of DL is the ability to learn massive amounts of data. The DL field has grown fast in the last few years and it has been extensively used to successfully address a wide range of traditional applications. More importantly, DL has outperformed well-known ML techniques in many domains, e.g., cybersecurity, natural language processing, bioinformatics, robotics and control, and medical information processing, among many others. Despite it has been contributed several works reviewing the State-of-the-Art on DL, all of them only tackled one aspect of the DL, which leads to an overall lack of knowledge about it. Therefore, in this contribution, we propose using a more holistic approach in order to provide a more suitable starting point from which to develop a full understanding of DL. Specifically, this review attempts to provide a more comprehensive survey of the most important aspects of DL and including those enhancements recently added to the field. In particular, this paper outlines the importance of DL, presents the types of DL techniques and networks. It then presents convolutional neural networks (CNNs) which the most utilized DL network type and describes the development of CNNs architectures together with their main features, e.g., starting with the AlexNet network and closing with the High-Resolution network (HR.Net). Finally, we further present the challenges and suggested solutions to help researchers understand the existing research gaps. It is followed by a list of the major DL applications. Computational tools including FPGA, GPU, and CPU are summarized along with a description of their influence on DL. The paper ends with the evolution matrix, benchmark datasets, and summary and conclusion.

Novel Transfer Learning Approach for Medical Imaging with Limited Labeled Data.

Deep learning requires a large amount of data to perform well. However, the field of medical image analysis suffers from a lack of sufficient data for training deep learning models. Moreover, medical images require manual labeling, usually provided by human annotators coming from various backgrounds. More importantly, the annotation process is time-consuming, expensive, and prone to errors. Transfer learning was introduced to reduce the need for the annotation process by transferring the deep learning models with knowledge from a previous task and then by fine-tuning them on a relatively small dataset of the current task. Most of the methods of medical image classification employ transfer learning from pretrained models, e.g., ImageNet, which has been proven to be ineffective. This is due to the mismatch in learned features between the natural image, e.g., ImageNet, and medical images. Additionally, it results in the utilization of deeply elaborated models. In this paper, we propose a novel transfer learning approach to overcome the previous drawbacks by means of training the deep learning model on large unlabeled medical image datasets and by next transferring the knowledge to train the deep learning model on the small amount of labeled medical images. Additionally, we propose a new deep convolutional neural network (DCNN) model that combines recent advancements in the field. We conducted several experiments on two challenging medical imaging scenarios dealing with skin and breast cancer classification tasks. According to the reported results, it has been empirically proven that the proposed approach can significantly improve the performance of both classification scenarios. In terms of skin cancer, the proposed model achieved an F1-score value of 89.09% when trained from scratch and 98.53% with the proposed approach. Secondly, it achieved an accuracy value of 85.29% and 97.51%, respectively, when trained from scratch and using the proposed approach in the case of the breast cancer scenario. Finally, we concluded that our method can possibly be applied to many medical imaging problems in which a substantial amount of unlabeled image data is available and the labeled image data is limited. Moreover, it can be utilized to improve the performance of medical imaging tasks in the same domain. To do so, we used the pretrained skin cancer model to train on feet skin to classify them into two classes-either normal or abnormal (diabetic foot ulcer (DFU)). It achieved an F1-score value of 86.0% when trained from scratch, 96.25% using transfer learning, and 99.25% using double-transfer learning.