Effect of tibialis anterior muscle resistance training on ankle and foot dorsum extension function in hypertensive cerebral hemorrhage hemiplegia patients: A randomized controlled trial.

BACKGROUND: This study aimed to investigate the effect of tibialis anterior muscle resistance training on ankle and foot dorsum extension function in patients with hemiplegia caused by hypertensive cerebral hemorrhage. METHODS: Fifty cases of hypertensive cerebral hemorrhage in patients with hemiplegia were selected according to the random number table method. The patients were divided into the treatment group and control group. Each group included 25 cases, and the treatment group was given routine rehabilitation treatment and passive and active foot back stretch training (300 times/d). The control group received conventional rehabilitation treatment. The conventional rehabilitation treatment included stretching, muscle strengthening and other conventional rehabilitation treatment techniques. Surface electromyography was used to evaluate the muscle strength and tension of the triceps and tibialis anterior muscles of the affected side of the patients before and after treatment. The root mean square value of the surface electromyography (RMS) of the passive triceps extension before and after treatment was used to evaluate the muscle strength and tension of the affected side. The ratio of passive traction and relaxation of the triceps before and after treatment and the ratio of active contraction and relaxation of the tibialis anterior muscle before and after treatment were recorded. RESULTS: There was no significant difference in surface electromyography between the 2 groups before treatment (P > .05). After 2 months of treatment, the results of patients in both groups improved compared with those before treatment. The RMS of triceps in the treatment group was significantly lower than that in the control group, and the ratio of RMS of triceps in the treatment group was significantly lower than that in the control group. The RMS during active contraction and the RMS ratio between active contraction and relaxation of the tibialis anterior muscle in the treatment group were significantly higher than those in the control group (P < .05). CONCLUSIONS: Tibialis anterior muscle resistance training can effectively improve the strength of the tibialis anterior muscle in patients with hemiplegia caused by hypertensive cerebral hemorrhage, reduce tension in the triceps calf muscle, and improve ankle joint function and foot dorsum extension.

DeepRetina: Layer Segmentation of Retina in OCT Images Using Deep Learning.

Purpose: To automate the segmentation of retinal layers, we propose DeepRetina, a method based on deep neural networks. Methods: DeepRetina uses the improved Xception65 to extract and learn the characteristics of retinal layers. The Xception65-extracted feature maps are inputted to an atrous spatial pyramid pooling module to obtain multiscale feature information. This information is then recovered to capture clearer retinal layer boundaries in the encoder-decoder module, thus completing retinal layer auto-segmentation of the retinal optical coherence tomography (OCT) images. Results: We validated this method using a retinal OCT image database containing 280 volumes (40 B-scans per volume) to demonstrate its effectiveness. The results showed that the method exhibits excellent performance in terms of the mean intersection over union and sensitivity (Se), which are as high as 90.41 and 92.15%, respectively. The intersection over union and Se values of the nerve fiber layer, ganglion cell layer, inner plexiform layer, inner nuclear layer, outer plexiform layer, outer nuclear layer, outer limiting membrane, photoreceptor inner segment, photoreceptor outer segment, and pigment epithelium layer were found to be above 88%. Conclusions: DeepRetina can automate the segmentation of retinal layers and has great potential for the early diagnosis of fundus retinal diseases. In addition, our approach will provide a segmentation model framework for other types of tissues and cells in clinical practice. Translational Relevance: Automating the segmentation of retinal layers can help effectively diagnose and monitor clinical retinal diseases. In addition, it requires only a small amount of manual segmentation, significantly improving work efficiency.

FecalNet: Automated detection of visible components in human feces using deep learning.

PURPOSE: To automate the detection and identification of visible components in feces for early diagnosis of gastrointestinal diseases, we propose FecalNet, a method using multiple deep neural networks. METHODS: FecalNet uses the ResNet152 residual network to extract and learn the characteristics of visible components in fecal microscopic images, acquire feature maps in combination with the feature pyramid network, apply the full convolutional network to classify and locate the fecal components, and implement the improved focal loss function to reoptimize the classification results. This allowed the complete automation of the detection and identification of the visible components in feces. RESULTS: We validated this method using a fecal database of 1,122 patients. The results indicated a mean average precision (mAP) of 92.16% and an average recall (AR) of 93.56%. The average precision (AP) and AR of erythrocyte, leukocyte, intestinal mucosal epithelial cells, hookworm eggs, ascarid eggs, and whipworm eggs were 92.82% and 93.38%, 93.99% and 96.11%, 90.71% and 92.41%, 89.95% and 93.88%, 96.90% and 91.21%, and 88.61% and 94.37%, respectively. The average times required by the GPU and the CPU to analyze a fecal microscopic image are approximately 0.14 and 1.02 s, respectively. CONCLUSION: FecalNet can automate the detection and identification of visible components in feces. It also provides a detection and identification framework for detecting several other types of cells in clinical practice.

[Artificial intelligence empowers laboratory medicine in Industry 4.0].

Since 2017, China, the United States, and the European Union have successively issued national-level artificial intelligence (AI) strategic development plans, and the human history is about to witness the 4th industrial revolution with the theme of "intelligence". In the field of medical testing, the explosive growth of AI theories and technologies also provide a new direction for the development of medical testing theory, methods and applications. We review the evolution of AI and the recent progress in three major elements of AI, namely algorithms, data and computing power, and elaborate on the combined innovation of "AI + testing" in light of the key application dimensions of medical testing. The major applications include specimen collection robots, sample dilution robots and sample transfer robots involved in the processing of test specimens; test item mining such as tumor markers and pharmacogenomics; cytomorphology, laboratory medicine data processing, auxiliary diagnostic models, and internet-based medical tests. With the advent of the era of Industry 4.0, AI technology will promote the development of medical testing from automation to a highly intelligent stage.

Inspection of visible components in urine based on deep learning.

PURPOSE: Urinary particles are particularly important parameters in clinical urinalysis, especially for the diagnosis of nephropathy. Therefore, it is highly important to precisely detect urinary particles in the clinical setting. However, artificial microscopy is subjective and time consuming, and various previous detection algorithms lack the adequate accuracy. In this study, a method is proposed for the analysis of urinary particles based on deep learning. METHODS: We used seven cellular components (i.e., erythrocytes, leukocytes, epithelial, low-transitional epithelium, casts, crystal, and squamous epithelial cells) in the microscopic imaging of urine as the detection targets. After the extraction of features using Resnet50, feature maps of different sizes are obtained in the last few layers of the feature pyramid net (FPN). The feature maps are then input into the classification subnetwork and regression subnetwork for classification and localization respectively, and detection results are obtained. First, we introduce the basic model (RetinaNet) to detect the cellular components in urinary particles, and the features of the objects can then be extracted more effectively by replacing different basic networks. Lastly, the effects of different weight initialization methods and different anchor scales on the performance of the model are investigated. RESULTS: We obtained the optimal network structure based on the adjustment of the loss functional parameters, thereby achieving the best results in the test set of urinary particles. The experimental data yielded an accuracy of 88.65% with a processing time of only 0.2 s for each image on a GeForce GTX 1080 graphics processing unit (GPU). Our results demonstrate that this method cannot only achieve the speed of the first-stage target detector, but also the accuracy of the two-stage target algorithm in the analysis of urinary particles. CONCLUSION: This study developed new automated analysis urinary particles based on deep learning, and this method is expected to be used for the automated analysis and detection of urinary particles. Moreover, our approach will be useful for the detection of other cells in the clinical setting.

Novel Biochemical Insights in the Cerebrospinal Fluid of Patients with Neurosyphilis Based on a Metabonomics Study.

Neurosyphilis is a chronic central nervous system infectious disease caused by Treponema pallidum. Our aim was to study the metabolic profiling in the cerebrospinal fluid of neurosyphilis patients and identify specific potential biomarkers. Fifteen cerebrospinal fluid samples from neurosyphilis patients and 14 non-neurosyphilis samples were analyzed by liquid chromatography-mass spectrometer (LC-MS). The LC-MS data were preprocessed by supervised pattern recognition to obtain diagnostic models. Both orthogonal projections to a latent structures discriminant analysis (OPLS-DA) and a t test were used to obtain specific metabolites for neurosyphilis. LC-MS data showed that the metabolites in cerebrospinal fluid (CSF) from neurosyphilis are different from the non-neurosyphilis group. The OPLS-DA model parameters R2Y and Q2Y are both more than 0.7 and indicated a satisfactory diagnostic performance. Bilirubin, L-histidine, prostaglandin E2, alpha-kamlolenic acid, and butyryl-L-carnitine and palmitoyl-L-carnitine were identified as novel potential biomarkers for neurosyphilis. The metabolic study of CSF may provide a new way to explore the pathogenesis of neurosyphilis.

Automatic Tracking of Muscle Cross-Sectional Area Using Convolutional Neural Networks with Ultrasound.

OBJECTIVES: The purpose of this study was to develop an automatic tracking method for the muscle cross-sectional area (CSA) on ultrasound (US) images using a convolutional neural network (CNN). The performance of the proposed method was evaluated and compared with that of the state-of-the art muscle segmentation method. METHODS: A real-time US image sequence was obtained from the rectus femoris muscle during voluntary contraction. A CNN was built to segment the rectus femoris muscle and calculate the CSA in each US frame. This network consisted of 2 stages: feature extraction and score map reconstruction. The training of the network was divided into 3 steps with output score map resolutions of one-fourth, one-half, and all of the original image. We evaluated the segmentation performance of our method with 5-fold cross-validation. The mean precision, recall, and dice similarity score were calculated. RESULTS: The mean precision, recall, and Dice's coefficient (DSC) ± SD were 0.936 ± 0.029, 0.882 ± 0.045, and 0.907 ± 0.023, respectively. Compared with the state-of-the-art muscle segmentation method (constrained mutual-information-based free-form deformation), the proposed method using CNN showed high performance. CONCLUSIONS: The automated method proposed in this study provides an accurate and efficient approach to the estimation of the muscle CSA during muscle contraction.

Tumor Segmentation in Contrast-Enhanced Magnetic Resonance Imaging for Nasopharyngeal Carcinoma: Deep Learning with Convolutional Neural Network.

OBJECTIVES: To evaluate the application of a deep learning architecture, based on the convolutional neural network (CNN) technique, to perform automatic tumor segmentation of magnetic resonance imaging (MRI) for nasopharyngeal carcinoma (NPC). MATERIALS AND METHODS: In this prospective study, 87 MRI containing tumor regions were acquired from newly diagnosed NPC patients. These 87 MRI were augmented to >60,000 images. The proposed CNN network is composed of two phases: feature representation and scores map reconstruction. We designed a stepwise scheme to train our CNN network. To evaluate the performance of our method, we used case-by-case leave-one-out cross-validation (LOOCV). The ground truth of tumor contouring was acquired by the consensus of two experienced radiologists. RESULTS: The mean values of dice similarity coefficient, percent match, and their corresponding ratio with our method were 0.89±0.05, 0.90±0.04, and 0.84±0.06, respectively, all of which were better than reported values in the similar studies. CONCLUSIONS: We successfully established a segmentation method for NPC based on deep learning in contrast-enhanced magnetic resonance imaging. Further clinical trials with dedicated algorithms are warranted.

Fully Automated Delineation of Gross Tumor Volume for Head and Neck Cancer on PET-CT Using Deep Learning: A Dual-Center Study.

Purpose: In this study, we proposed an automated deep learning (DL) method for head and neck cancer (HNC) gross tumor volume (GTV) contouring on positron emission tomography-computed tomography (PET-CT) images. Materials and Methods: PET-CT images were collected from 22 newly diagnosed HNC patients, of whom 17 (Database 1) and 5 (Database 2) were from two centers, respectively. An oncologist and a radiologist decided the gold standard of GTV manually by consensus. We developed a deep convolutional neural network (DCNN) and trained the network based on the two-dimensional PET-CT images and the gold standard of GTV in the training dataset. We did two experiments: Experiment 1, with Database 1 only, and Experiment 2, with both Databases 1 and 2. In both Experiment 1 and Experiment 2, we evaluated the proposed method using a leave-one-out cross-validation strategy. We compared the median results in Experiment 2 (GTVa) with the performance of other methods in the literature and with the gold standard (GTVm). Results: A tumor segmentation task for a patient on coregistered PET-CT images took less than one minute. The dice similarity coefficient (DSC) of the proposed method in Experiment 1 and Experiment 2 was 0.481∼0.872 and 0.482∼0.868, respectively. The DSC of GTVa was better than that in previous studies. A high correlation was found between GTVa and GTVm (R = 0.99, P < 0.001). The median volume difference (%) between GTVm and GTVa was 10.9%. The median values of DSC, sensitivity, and precision of GTVa were 0.785, 0.764, and 0.789, respectively. Conclusion: A fully automatic GTV contouring method for HNC based on DCNN and PET-CT from dual centers has been successfully proposed with high accuracy and efficiency. Our proposed method is of help to the clinicians in HNC management.

Metabonomics screening of serum identifies pyroglutamate as a diagnostic biomarker for nonalcoholic steatohepatitis.

OBJECTIVE: A key step in managing non-alcoholic fatty liver disease (NAFLD) is to differentiate nonalcoholic steatohepatitis (NASH) from simple steatosis (SS). METHOD: Serum samples were collected from three groups: NASH patients (N=21), SS patients (N=38) and healthy controls (N=31). High performance liquid chromatography-mass spectrometry (HPLC-MS) was used to analyse the metabolic profile of the serum samples. The acquired data were processed by multivariate principal component analysis (PCA) and orthogonal partial least-squares-discriminant analysis (OPLS-DA) to identify novel metabolites. The potential biomarkers were quantitatively determined and their diagnostic power was further validated. RESULTS: A total of 56 metabolites were capable of distinguishing NASH from SS samples based on the OPLS-DA model. Pyroglutamate was found to be the most promising factor in distinguishing the NASH from SS groups. With an optimal cut-off value of 4.82mmol/L, the sensitivity and specificity of the diagnosis of NASH were 72% and 85%, respectively. The area under the receiver operating characteristic (AUROC) of the pyroglutamate levels of NASH versus SS patients was more than those of tumor necrosis factor-α, adiponectin and interleukin-8. CONCLUSION: These data suggest that pyroglutamate may be a new and useful biomarker for the diagnosis of NASH.

Flexible and Highly Photosensitive Electrolyte-Gated Organic Transistors with Ionogel/Silver Nanowire Membranes.

Flexible and low-voltage photosensors with high near-infrared (NIR) sensitivity are critical for realization of interacting humans with robots and environments by thermal imaging or night vision techniques. In this work, we for the first time develop an easy and cost-effective process to fabricate flexible and ultrathin electrolyte-gated organic phototransistors (EGOPTs) with high transparent nanocomposite membranes of high-conductivity silver nanowire (AgNW) networks and large-capacitance iontronic films. A high responsivity of 1.5 × 103 A·W1-, high sensitivity of 7.5 × 105, and 3 dB bandwidth of ∼100 Hz can be achieved at very low operational voltages. Experimental studies in temporal photoresponse characteristics reveal the device has a shorter photoresponse time at lower light intensity since strong interactions between photoexcited hole carriers and anions induce extra long-lived trap states. The devices, benefiting from fast and air-stable operations, provide the possibility of the organic photosensors for constructing cost-effective and smart optoelectronic systems in the future.

A Cross-Modality Learning Approach for Vessel Segmentation in Retinal Images.

This paper presents a new supervised method for vessel segmentation in retinal images. This method remolds the task of segmentation as a problem of cross-modality data transformation from retinal image to vessel map. A wide and deep neural network with strong induction ability is proposed to model the transformation, and an efficient training strategy is presented. Instead of a single label of the center pixel, the network can output the label map of all pixels for a given image patch. Our approach outperforms reported state-of-the-art methods in terms of sensitivity, specificity and accuracy. The result of cross-training evaluation indicates its robustness to the training set. The approach needs no artificially designed feature and no preprocessing step, reducing the impact of subjective factors. The proposed method has the potential for application in image diagnosis of ophthalmologic diseases, and it may provide a new, general, high-performance computing framework for image segmentation.

Use of optical flow to estimate continuous changes in muscle thickness from ultrasound image sequences.

Muscle thickness is one of the most widely used parameters for quantifying muscle function. Ultrasonography is frequently used to estimate changes in muscle thickness in both static and dynamic contractions. Conventionally, most such measurements are conducted by manual analysis of ultrasound images. This manual approach is time consuming, subjective, susceptible to error and not suitable for measuring dynamic change. In this study, we developed an automated tracking method based on an optical flow algorithm using an affine motion model. The goal of the study was to evaluate the performance of the proposed method by comparing it with the manual approach and by determining its repeatability. Real-time B-mode ultrasound was used to examine the rectus femoris during voluntary contraction. The coefficient of multiple correlation (CMC) was used to quantify the level of agreement between the two methods and the repeatability of the proposed method. The two methods were also compared by linear regression and Bland-Altman analysis. The findings indicated that the results obtained using the proposed method were in good agreement with those obtained using the manual approach (CMC = 0.97 ± 0.02, difference = -0.06 ± 0.22 mm) and were highly repeatable (CMC = 0.91 ± 0.07). In conclusion, the automated method proposed here provides an accurate, highly repeatable and efficient approach to the estimation of muscle thickness during muscle contraction.

Determining the concentration of procalcitonin using a magnetic particles-based chemiluminescence assay for the clinical diagnosis of sepsis.

Our objective is to develop an assay based on magnetic particles (MPs) to determine the concentration of procalcitonin (PCT) using a chemiluminescence immunoassay (CLIA). Fluorescein isothiocyanate (FITC) and N-(aminobutyl)-N-(ethylisoluminol) (ABEI) were used to label two different anti-procalcitonin (PCT) monoclonal antibodies. The labeled antibodies, the PCT antigen, and the anti-FITC antibody-coated MPs formed a double-sandwiched immunocomplex. The measured relative light units (RLUs) of ABEI in the substrate solution were directly proportional to the amount of PCT present in the samples. The proposed method was linear to 600 ng/mL with a detection limit of 0.03 ng/mL. The coefficient of variation (CV) was <5% and <6% for the intra- and inter-assay precision, respectively. The average recoveries were between 95 and 107%. The linearity-dilution effect gave a linear correlation coefficient of 0.9912. This proposed assay provided an alternative method to quantitatively measure PCT in serum for the diagnosis of sepsis.

Continuous detection of muscle aspect ratio using keypoint tracking in ultrasonography.

Muscle aspect ratio of cross-sectional area is one of the most widely used parameters for quantifying muscle function in both diagnosis and rehabilitation assessment. Ultrasound imaging has been frequently used to noninvasively study the characteristics of human muscles as a reliable method. However, the aspect ratio measurement is traditionally conducted by the manual digitization of reference points; thus, it is subjective, time-consuming, and prone to errors. In this paper, a novel method is proposed to continuously detect the muscle aspect ratio. Two keypoint pairs are manually digitized on the lateral and longitudinal borders at the first frame, and automatically tracked by an optical flow technique at the subsequent frames. The muscle aspect ratio is thereby obtained based on the estimated muscle width and thickness. Six ultrasound sequences from different subjects are used to evaluate this method, and the overall coefficient of multiple correlation of the results between manual and proposed methods is 0.97 ± 0.02. The linear regression shows that a good linear correlation between the results of the two methods is obtained (R(2) = 0.974), with difference -0.01 ± 0.16. The method proposed here provides an accurate, high repeatable, and efficient approach for estimating muscle aspect ratio during human motion, thus justifying its application in biological sciences.

Development of a rapid and high-performance chemiluminescence immunoassay based on magnetic particles for protein S100B in human serum.

Protein S100B is a clinically useful non-invasive biomarker for brain cell damage. A rapid chemiluminescence immunoassay (CLIA) for S100B in human serum has been developed. Fluorescein isothiocyanate (FITC) and N-(aminobutyl)-N-(ethylisoluminol) (ABEI) are used to label two different monoclonal antibodies of anti-S100B. Protein S100B in serum combines with labeled antibodies and can form a sandwiched immunoreaction. A simplified separation procedure based on the use of magnetic particles (MPs) that were coated with anti-FITC antibody is performed to remove the unwanted materials. After adding the substrate solution, the relative light unit (RLU) of ABEI is measured and is found to be directly proportional to the concentration of S100B in serum. The relevant variables involved in the CLIA signals are optimized and the parameters of the proposed method are evaluated. The results demonstrate that the method is linear to 25 ng/mL S100B with a detection limit of 0.02 ng/mL. The coefficient of variation (CV) is < 5% and < 6% for intra- and interassay precision, respectively. The average recoveries are between 97 and 107%. The linearity-dilution effect produces a linear correlation coefficient of 0.9988. Compared with the commercial kit, the proposed method shows a correlation of 0.9897. The proposed method displays acceptable performance for quantification of S100B and is appropriate for use in clinical diagnosis.

[Design of acoustic radiation force module for ultrasound elastography].

Developing an acoustic radiation force excitation module including 64 channels based in FPGA for ultrasound elastography. The circuit of the module was derived in bipolar, and the parameters such as excitation frequency, pulse repetition frequency, pulse number, element number and focus depth were adjustable. The acoustic field for special parameter was experimented with OptiSon laser acoustic field system with a result which reflects the width of focal spot is about 3 mm. The acoustic power was experimented with RFB2000 radiation force balance with a result which reflects acoustic power is increasing linearly with the number of pulses and the number of elements, and is increasing squarely with the peak-to-peak value of excitation voltage. The module is promising in factual application which can be triggered externally in synchronously, and can be combined with B-mode ultrasound system for ultrasound elastography.

[The development and transplantation of LCD module-based interface for medical diagnosis instrument].

Based on LCD Module and Visual C++ development environment, this paper proposes a new method which can quickly develop the human-machine interface .We define a LCD module programming interface by designing Serial Communication Class(SCS). On this basis,we achieve the transplantation on an Embedded ARM Platform to fulfil the requirements of Medical Diagnostic Instruments (MDI). Experimental results show that this method has advantages of short development cycle and high level transplantation which has broad application prospects in the field of Medical Diagnosis Instrument.