A deep learning-based method for cervical transformation zone classification in colposcopy images.

BACKGROUND: Colposcopy is one of the common methods of cervical cancer screening. The type of cervical transformation zone is considered one of the important factors for grading colposcopic findings and choosing treatment. OBJECTIVE: This study aims to develop a deep learning-based method for automatic classification of cervical transformation zone from colposcopy images. METHODS: We proposed a multiscale feature fusion classification network to classify cervical transformation zone, which can extract features from images and fuse them at multiple scales. Cervical regions were first detected from original colposcopy images and then fed into our multiscale feature fusion classification network. RESULTS: The results on the test dataset showed that, compared with the state-of-the-art image classification models, the proposed classification network had the highest classification accuracy, reaching 88.49%, and the sensitivity to type 1, type 2 and type 3 were 90.12%, 85.95% and 89.45%, respectively, higher than the comparison methods. CONCLUSIONS: The proposed method can automatically classify cervical transformation zone in colposcopy images, and can be used as an auxiliary tool in cervical cancer screening.

A lightweight hierarchical convolution network for brain tumor segmentation.

BACKGROUND: Brain tumor segmentation plays a significant role in clinical treatment and surgical planning. Recently, several deep convolutional networks have been proposed for brain tumor segmentation and have achieved impressive performance. However, most state-of-the-art models use 3D convolution networks, which require high computational costs. This makes it difficult to apply these models to medical equipment in the future. Additionally, due to the large diversity of the brain tumor and uncertain boundaries between sub-regions, some models cannot well-segment multiple tumors in the brain at the same time. RESULTS: In this paper, we proposed a lightweight hierarchical convolution network, called LHC-Net. Our network uses a multi-scale strategy which the common 3D convolution is replaced by the hierarchical convolution with residual-like connections. It improves the ability of multi-scale feature extraction and greatly reduces parameters and computation resources. On the BraTS2020 dataset, LHC-Net achieves the Dice scores of 76.38%, 90.01% and 83.32% for ET, WT and TC, respectively, which is better than that of 3D U-Net with 73.50%, 89.42% and 81.92%. Especially on the multi-tumor set, our model shows significant performance improvement. In addition, LHC-Net has 1.65M parameters and 35.58G FLOPs, which is two times fewer parameters and three times less computation compared with 3D U-Net. CONCLUSION: Our proposed method achieves automatic segmentation of tumor sub-regions from four-modal brain MRI images. LHC-Net achieves competitive segmentation performance with fewer parameters and less computation than the state-of-the-art models. It means that our model can be applied under limited medical computing resources. By using the multi-scale strategy on channels, LHC-Net can well-segment multiple tumors in the patient's brain. It has great potential for application to other multi-scale segmentation tasks.

Segmentation of the cervical lesion region in colposcopic images based on deep learning.

Background: Colposcopy is an important method in the diagnosis of cervical lesions. However, experienced colposcopists are lacking at present, and the training cycle is long. Therefore, the artificial intelligence-based colposcopy-assisted examination has great prospects. In this paper, a cervical lesion segmentation model (CLS-Model) was proposed for cervical lesion region segmentation from colposcopic post-acetic-acid images and accurate segmentation results could provide a good foundation for further research on the classification of the lesion and the selection of biopsy site. Methods: First, the improved Faster Region-convolutional neural network (R-CNN) was used to obtain the cervical region without interference from other tissues or instruments. Afterward, a deep convolutional neural network (CLS-Net) was proposed, which used EfficientNet-B3 to extract the features of the cervical region and used the redesigned atrous spatial pyramid pooling (ASPP) module according to the size of the lesion region and the feature map after subsampling to capture multiscale features. We also used cross-layer feature fusion to achieve fine segmentation of the lesion region. Finally, the segmentation result was mapped to the original image. Results: Experiments showed that on 5455 LSIL+ (including cervical intraepithelial neoplasia and cervical cancer) colposcopic post-acetic-acid images, the accuracy, specificity, sensitivity, and dice coefficient of the proposed model were 93.04%, 96.00%, 74.78%, and 73.71%, respectively, which were all higher than those of the mainstream segmentation model. Conclusion: The CLS-Model proposed in this paper has good performance in the segmentation of cervical lesions in colposcopic post-acetic-acid images and can better assist colposcopists in improving the diagnostic level.

Treatment of PD-1 Inhibitor-Associated Toxic Epidermal Necrolysis: A Case Report and Brief Review.

Background: Sintilimab is a fully human monoclonal antibody targeting PD-1, which has been considered well tolerated among patients and widely applied in malignancies. Case Presentation: We present a case report of a patient with gallbladder carcinoma treated with sintilimab who developed toxic epidermal necrolysis (TEN). A 72-year-old female presented with fever and maculopapular rash after receiving one dose of sintilimab for metastatic gallbladder carcinoma. Widespread maculopapular rashes with progressive skin detachment occurred within one week. Early skin biopsy of the patient showed apoptotic keratinocytes along with interface dermatitis. She was initially treated with escalating methylprednisolone (from 0.8 to 1.6 mg/kg/d) and subsequently in the combination of intravenous immunoglobulin. Her skin lesions significantly improved, and satisfying re-epithelialization was achieved after 43 days of hospitalization. Conclusion: Because of the high mortality of grade four immune related adverse event (irAE) on skin, we recommend early monitoring and recognition of symptoms. During management, high-dose glucocorticoids with combined intravenous immune globulin and supportive care may be helpful.

Design of lung nodules segmentation and recognition algorithm based on deep learning.

BACKGROUND: Accurate segmentation and recognition algorithm of lung nodules has great important value of reference for early diagnosis of lung cancer. An algorithm is proposed for 3D CT sequence images in this paper based on 3D Res U-Net segmentation network and 3D ResNet50 classification network. The common convolutional layers in encoding and decoding paths of U-Net are replaced by residual units while the loss function is changed to Dice loss after using cross entropy loss to accelerate network convergence. Since the lung nodules are small and rich in 3D information, the ResNet50 is improved by replacing the 2D convolutional layers with 3D convolutional layers and reducing the sizes of some convolution kernels, 3D ResNet50 network is obtained for the diagnosis of benign and malignant lung nodules. RESULTS: 3D Res U-Net was trained and tested on 1044 CT subcases in the LIDC-IDRI database. The segmentation result shows that the Dice coefficient of 3D Res U-Net is above 0.8 for the segmentation of lung nodules larger than 10 mm in diameter. 3D ResNet50 was trained and tested on 2960 lung nodules in the LIDC-IDRI database. The classification result shows that the diagnostic accuracy of 3D ResNet50 is 87.3% and AUC is 0.907. CONCLUSION: The 3D Res U-Net module improves segmentation performance significantly with the comparison of 3D U-Net model based on residual learning mechanism. 3D Res U-Net can identify small nodules more effectively and improve its segmentation accuracy for large nodules. Compared with the original network, the classification performance of 3D ResNet50 is significantly improved, especially for small benign nodules.

Cascading detection model for prediction of apnea-hypopnea events based on nasal flow and arterial blood oxygen saturation.

PURPOSE: Sleep apnea and hypopnea syndrome (SAHS) seriously affects sleep quality. In recent years, much research has focused on the detection of SAHS using various physiological signals and algorithms. The purpose of this study is to find an efficient model for detection of apnea-hypopnea events based on nasal flow and SpO2 signals. METHODS: A 60-s detector and a 10-s detector were cascaded for precise detection of apnea-hypopnea (AH) events. Random forests were adopted for classification of data segments based on morphological features extracted from nasal flow and arterial blood oxygen saturation (SpO2). Then the segments' classification results were fed into an event detector to locate the start and end time of every AH event and predict the AH index (AHI). RESULTS: A retrospective study of 24 subjects' polysomnography recordings was conducted. According to segment analysis, the cascading detection model reached an accuracy of 88.3%. While Pearson's correlation coefficient between estimated AHI and reference AHI was 0.99, in the diagnosis of SAHS severity, the proposed method exhibited a performance with Cohen's kappa coefficient of 0.76. CONCLUSIONS: The cascading detection model is able to detect AH events and provide an estimate of AHI. The results indicate that it has the potential to be a useful tool for SAHS diagnosis.

Robust closed-loop control of spike-and-wave discharges in a thalamocortical computational model of absence epilepsy.

In this paper, we investigate the abatement of spike-and-wave discharges in a thalamocortical model using a closed-loop brain stimulation method. We first explore the complex states and various transitions in the thalamocortical computational model of absence epilepsy by using bifurcation analysis. We demonstrate that the Hopf and double cycle bifurcations are the key dynamical mechanisms of the experimental observed bidirectional communications during absence seizures through top-down cortical excitation and thalamic feedforward inhibition. Then, we formulate the abatement of epileptic seizures to a closed-loop tracking control problem. Finally, we propose a neural network based sliding mode feedback control system to drive the dynamics of pathological cortical area to track the desired normal background activities. The control system is robust to uncertainties and disturbances, and its stability is guaranteed by Lyapunov stability theorem. Our results suggest that the seizure abatement can be modeled as a tracking control problem and solved by a robust closed-loop control method, which provides a promising brain stimulation strategy.

The segmentation of bones in pelvic CT images based on extraction of key frames.

BACKGROUND: Bone segmentation is important in computed tomography (CT) imaging of the pelvis, which assists physicians in the early diagnosis of pelvic injury, in planning operations, and in evaluating the effects of surgical treatment. This study developed a new algorithm for the accurate, fast, and efficient segmentation of the pelvis. METHODS: The proposed method consists of two main parts: the extraction of key frames and the segmentation of pelvic CT images. Key frames were extracted based on pixel difference, mutual information and normalized correlation coefficient. In the pelvis segmentation phase, skeleton extraction from CT images and a marker-based watershed algorithm were combined to segment the pelvis. To meet the requirements of clinical application, physician's judgment is needed. Therefore the proposed methodology is semi-automated. RESULTS: In this paper, 5 sets of CT data were used to test the overlapping area, and 15 CT images were used to determine the average deviation distance. The average overlapping area of the 5 sets was greater than 94%, and the minimum average deviation distance was approximately 0.58 pixels. In addition, the key frame extraction efficiency and the running time of the proposed method were evaluated on 20 sets of CT data. For each set, approximately 13% of the images were selected as key frames, and the average processing time was approximately 2 min (the time for manual marking was not included). CONCLUSIONS: The proposed method is able to achieve accurate, fast, and efficient segmentation of pelvic CT image sequences. Segmentation results not only provide an important reference for early diagnosis and decisions regarding surgical procedures, they also offer more accurate data for medical image registration, recognition and 3D reconstruction.

Principal dynamic mode analysis of neural mass model for the identification of epileptic states.

The detection of epileptic seizures in Electroencephalography (EEG) signals is significant for the diagnosis and treatment of epilepsy. In this paper, in order to obtain characteristics of various epileptiform EEGs that may differentiate different states of epilepsy, the concept of Principal Dynamic Modes (PDMs) was incorporated to an autoregressive model framework. First, the neural mass model was used to simulate the required intracerebral EEG signals of various epileptiform activities. Then, the PDMs estimated from the nonlinear autoregressive Volterra models, as well as the corresponding Associated Nonlinear Functions (ANFs), were used for the modeling of epileptic EEGs. The efficient PDM modeling approach provided physiological interpretation of the system. Results revealed that the ANFs of the 1st and 2nd PDMs for the auto-regressive input exhibited evident differences among different states of epilepsy, where the ANFs of the sustained spikes' activity encountered at seizure onset or during a seizure were the most differentiable from that of the normal state. Therefore, the ANFs may be characteristics for the classification of normal and seizure states in the clinical detection of seizures and thus provide assistance for the diagnosis of epilepsy.

Suppression of seizures based on the multi-coupled neural mass model.

Epilepsy is one of the most common serious neurological disorders, which affects approximately 1% of population in the world. In order to effectively control the seizures, we propose a novel control methodology, which combines the feedback linearization control (FLC) with the underlying mechanism of epilepsy, to achieve the suppression of seizures. The three coupled neural mass model is constructed to study the property of the electroencephalographs (EEGs). Meanwhile, with the model we research on the propagation of epileptiform waves and the synchronization of populations, which are taken as the foundation of our control method. Results show that the proposed approach not only yields excellent performances in clamping the pathological spiking patterns to the reference signals derived under the normal state but also achieves the normalization of the pathological parameter, where the parameters are estimated from EEGs with Unscented Kalman Filter. The specific contribution of this paper is to treat the epilepsy from its pathogenesis with the FLC, which provides critical theoretical basis for the clinical treatment of neurological disorders.

Characterization of complexity in the electroencephalograph activity of Alzheimer's disease based on fuzzy entropy.

In this paper, experimental neurophysiologic recording and statistical analysis are combined to investigate the nonlinear characteristic and the cognitive function of the brain. Fuzzy approximate entropy and fuzzy sample entropy are applied to characterize the model-based simulated series and electroencephalograph (EEG) series of Alzheimer's disease (AD). The effectiveness and advantages of these two kinds of fuzzy entropy are first verified through the simulated EEG series generated by the alpha rhythm model, including stronger relative consistency and robustness. Furthermore, in order to detect the abnormality of irregularity and chaotic behavior in the AD brain, the complexity features based on these two fuzzy entropies are extracted in the delta, theta, alpha, and beta bands. It is demonstrated that, due to the introduction of fuzzy set theory, the fuzzy entropies could better distinguish EEG signals of AD from that of the normal than the approximate entropy and sample entropy. Moreover, the entropy values of AD are significantly decreased in the alpha band, particularly in the temporal brain region, such as electrode T3 and T4. In addition, fuzzy sample entropy could achieve higher group differences in different brain regions and higher average classification accuracy of 88.1% by support vector machine classifier. The obtained results prove that fuzzy sample entropy may be a powerful tool to characterize the complexity abnormalities of AD, which could be helpful in further understanding of the disease.

[Study on detection methods of interstitial fluid glucose concentration based on infrared attenuated total reflection].

Measuring the glucose concentrations in the interstitial fluid is currently the main method to achieve the continuous blood glucose monitoring. The MIR-ATR(Mid-infrared, Attenuated Total Reflection)Spectroscopy has prominent advantage on the analysis of small biological molecule for composition information like the glucose, but it is still an unresolved problem that how to detect the subcutaneous glucose concentration by using the MIR-ATR Spectroscopy. In the present paper, we carry out the experiment based on MIR-ATR for the detection of subcutaneous glucose information on both the natural state and the penetration state based on the theoryanalysis of MIR penetration depth. Firstly, collect spectral data of the subcutaneous glucose concentration of human finger on the natural state were collected as the light shined the skin directly, and it was discussed whether the MIR can penetrate the skin to get the information of subcutaneous glucose. On this basis, collect spectral data of the subcutaneous glucose concentration of human finger at the penetration state were collected when the Interstitial fluid is permeated to the surface layer by using low-frequency ultrasound and vacuum, then it analyzed that whether it can detect the glucose-specific information or not. As the two-dimensional correlation spectroscopy has high resolution and good versatility, it is widely used to analyze the inter-molecular reaction and judge the absorption peaks information in many fields including the MIR spectroscopy field, so we choose the Two-dimensional correlation spectroscopy to analyze the information of subcutaneous glucose concentration at the natural state and the penetration state. The experiment result shows that the MIR-ATR spectroscopy can't be applied in the detection of subcutaneous glucose concentrationdirectly, and it is a promising direction to make the Interstitial fluid permeated to the surface layer by the physical methods or chemical methods.

Expression of cytochrome P450 2A5 in a C57BL/6J mouse model of nonalcoholic fatty liver disease.

Cytochrome P450 2A5 (CYP2A5) expression is induced during the hepatotoxicity caused by various hepatotoxins and hepatitis. However, CYP2A5 expression during nonalcoholic fatty liver disease (NAFLD) is still unknown. In this study, serum biochemical parameters and liver histopathological analyses found that NAFLD had developed in C57BL/6J mice via administration of a high-fat diet continuously for 8 weeks. Subsequently, CYP2A5 expression was probed in the mice diagnosed with NAFLD that were treated with or without pyrazole, the inducer of chemical liver injury. It is shown that hepatic CYP2A5 mRNA, protein expression and coumarin 7-hydroxylase activity are enhanced with high-fat feed, and that pyrazole is able to further increase CYP2A5 expression and activity in mice with NAFLD. These results revealed that CYP2A5 is elevated during NAFLD and suggested that pyrazole and NAFLD act synergistically to induce the expression of CYP2A5 via an unclear mechanism.

[Identification of adulterated milk based on two-dimensional correlation near-infrared spectra parameterization and BP neural network].

Discriminant models of adulterated milk and pure milk were established using BP neural network combined with two-dimensional (2D) correlation near-infrared spectra parameterization. Forty pure milk samples, 40 adulterated milk samples with urea (1-20 g x L(-1)) and 40 adulterated milk samples with melamine (0.01-3 g x L(-1)) were prepared respectively. Based on the characteristics of 2D correlation near-infrared spectra of pure milk and adulterated milk, 5 apparent statistic parameters were calculated based on the parameterization theory. Using 5 characteristic parameters, discriminant models of urea adulterated milk, melamine adulterated milk and two types of adulterated milk were built by BP neural network The prediction rate of unknown samples were 95%, 100% and 96.7%, respectively. The results show that this method can extract effectively feature information of adulterant, reduce the input dimensions of BP neural network, and better realize qualitative analysis of adulterant in milk.

[Study on specificity of noninvasive glucose measurements based on two-dimensional correlation mid-infrared spectroscopy].

Glucose specificity is the premise of spectroscopic measurements of blood glucose concentration, and it is also paramount for feasibility study of a spectral measurement method. Two-dimensional correlation spectroscopy technology is widely used in many fields such as inter-/intra-molecular reaction, material phase transition and information extraction because of its high resolution and effective Noda's rule. In the present paper, the glucose specificity for noninvasive glucose measurements by mid-infrared spectra based on the 2D correlation spectroscopy was investigated. First, the feasibility of this method was validated by a series of in vitro experiments of glucose. Then the in vivo experiments of four volunteers were conducted and the characteristic information of glucose by mid-infrared spectra collected from human fingers was confirmed by 2D correlation spectroscopy analysis.

[Study of chance correlation in blood glucose sensing].

In the noninvasive blood glucose sensing by the near-infrared spectroscopy, chemometrics is applied to achieve the quantitative analysis of unknown samples. In modeling and validation process, however, there usually introduces a certain degree of chance correlation, thus affecting the stability of the model. In the present paper, normally distributed random numbers were used to simulate spectral data and reference concentration. In this way, it can investigate the probability level of chance correlation from the number of selected modeling wavelengths and different probable cross validation methods. Chance correlation exists in the process of modeling. In this paper, there has also given the best level of modeling wavelengths and the optimal cross validation method to reduce the chance correlation. In addition, the in vitro experiment of glucose aqueous solution at different temperature is conducted. In this experiment, the relationship between the temperature and the glucose concentration was obtained, according to which the temperature effect in practice was reduced.

Liver fatty acid composition in mice with or without nonalcoholic fatty liver disease.

BACKGROUND: Nonalcoholic fatty liver disease (NAFLD) is one of the most frequent causes of abnormal liver function. Because fatty acids can damage biological membranes, fatty acid accumulation in the liver may be partially responsible for the functional and morphological changes that are observed in nonalcoholic liver disease. The aim of this study was to use gas chromatography-mass spectrometry to evaluate the fatty acid composition of an experimental mouse model of NAFLD induced by high-fat feed and CCl4 and to assess the association between liver fatty acid accumulation and NAFLD. C57BL/6J mice were given high-fat feed for six consecutive weeks to develop experimental NAFLD. Meanwhile, these mice were given subcutaneous injections of a 40% CCl4-vegetable oil mixture twice per week. RESULTS: A pathological examination found that NAFLD had developed in the C57BL/6J mice. High-fat feed and CCl4 led to significant increases in C14:0, C16:0, C18:0 and C20:3 (P < 0.01), and decreases in C15:0, C18:1, C18:2 and C18:3 (P < 0.01) in the mouse liver. The treatment also led to an increase in SFA and decreases in other fatty acids (UFA, PUFA and MUFA). An increase in the ratio of product/precursor n-6 (C20:4/C18:2) and n-3 ([C20:5+C22:6]/C18:3) and a decrease in the ratio of n-6/n-3 (C20:4/[C20:5+C22:6]) were also observed. CONCLUSION: These data are consistent with the hypothesis that fatty acids are deranged in mice with non-alcoholic fatty liver injury induced by high-fat feed and CCl4, which may be involved in its pathogenesis and/or progression via an unclear mechanism.

[Super-resolution image reconstruction algorithm based on projection onto convex sets and wavelet fusion].

In this paper a new super-resolution image reconstruction algorithm was proposed. With the improvement of the classical projection onto convex sets (POCS) algorithm, as ground work, and with the combined use of POCS and wavelet fusion, a high resolution CT image was restored by using a group of low resolution CT images. The experimental results showed: the proposed algorithm improves the ability of fusing different information, the detail of the image is more prominent, and the image quality is better.

[ECG pattern classification by feature searching algorithm based on maximal divergence].

This paper presents a method of using feature searching algorithm based on maximal divergence value to get the optimized feature combinations at different dimensions from feature space. Feature space is obtained through wavelet transform on ECG beat. Then the feature vector is determined by analyzing the changes of divergence value of those optimized feature combinations along with the dimensions. BP artificial neural network is trained by the feature vector and four types of ECG beats(normal beat, left bundle branch block beat, right bundle branch block beat and paced beat) obtained from MIT-BIH database are classified with a success of 93.9%.

[Measurement of body surface Laplacian ECG and its signal processing].

Surface Laplacian of the body surface potential (Laplacian ECG--LECG) is a new approach to resolve spatially distributed bioelectrical source. In this paper, we discussed an LECG sensor which integrated triple concentric ring electrodes and signal adjustor on a printed board. The LECG is measured directly by this sensor. The frequency, amplitude and phase of the power line interference were detected by a nonlinear adaptive filter so that interference was eliminated. The wavelet shrinking technique was used to eliminate the rest of random noise. And we got the high quality LECG signal. It laid the foundation for heart disease diagnosis.