Chinese Journal of Biomedical Engineering

Home

Editorial Board

Download

Quick Search

Office Online

Current Issue

Download Articles

Download

Links

2020 Vol. 39, No. 5 Published: 2020-10-20

Reviews Communications Regular Papers

Regular Papers

	513	Multiparametric Magnetic Resonance Imaging Based Radiomics for Prediction of Histological Information of Breast Cancer
		Lou Xiaofang, Fan Ming, Xu Maosheng, Wang Shiwei, Li Lihua
		DOI: 10.3969/j.issn.0258-8021.2020.05.001
		To create a prediction model based on multiparametric magnetic resonance imaging (MRI) radiomics features extracted from dynamic enhanced magnetic resonance imaging (DCE-MRI),T2 weighted imaging (T2WI) and diffusion weighted imaging (DWI) to predict molecular subtypes,histological grade and Ki-67 expression of breast cancer. In this study,150 cases of breast invasive ductal carcinoma before surgery and chemotherapy were collected,and multiparametric images of DCE-MRI,T2WI and DWI were obtained. Breast tumor areas in the different parametric images were segmented and multiparametric imaging features were extracted. The best imaging feature subset was obtained using support vector machine recursive feature elimination (SVM-RFE) algorithm,and a prediction model based on SVM was created using the training set of each parameter imaging series. The performance of prediction model was tested in the test set. The prediction models for all parameter imaging series were fused using the probabilistic averaging method,the probabilistic voting method,and the probabilistic model optimization method. The prediction performance was evaluated by calculating the area under the ROC curve (AUC). The single-parametric imaging models discriminated among the Luminal A,Luminal B,HER2,and Basal-like subtypes with the best AUC values of 0.672 1,0.694 0,0.677 7,and 0.708 6,respectively,and the prediction performance of multiparametric imaging models was increased to AUC of 0.799 5,0.727 9,0.737 5 and 0.792 5,respectively. The single-parametric imaging models discriminated among histological grades with the best AUC values of 0.753 3,and the prediction performance of multiparametric imaging model was increased to AUC of 0.801 7. The single-parametric imaging models discriminated among Ki-67 expression with the best AUC values of 0.664 7,and the prediction performance of multiparametric imaging model was increased to AUC of 0.771 8. The prediction accuracy of multiparametric imaging models was increased significantly compared to single-parameter models (P<0.05). Our results showed that the combination of multiparametric imaging (DCE-MRI,T2WI,and DWI) radiomics could significantly improve the performance of single-parameter imaging model in predicting pathological information of breast cancer,which is of great significance for the diagnosis and selection of personalized treatment plan for breast cancer.
		2020 Vol. 39 (5): 513-523 [Abstract] ( 552 ) HTML (1 KB) PDF (1098 KB) ( 400 )

	524	Study of Temporal Pattern of Embryonic Angiogenesis Based on Interpretable Convolutional Neural Networks
		Lv Xueqian, Zhao Shenjia, Li Peilun, Fang Luping, Ning Gangmin, Pan Qing
		DOI: 10.3969/j.issn.0258-8021.2020.05.002
		Understanding temporal patterns of vascular network angiogenesis facilitates the study of biological development mechanisms and the pathophysiology of tumors. In this paper,an interpretable convolutional neural network (CNN) was proposed to study the temporal pattern of angiogenesis of the chicken vitelline vascular networks. We constructed a model for classifying the vascular images of chicken vitelline at 3 days post-fertilization (3dpf) and 4 days post-fertilization (4dpf) based on CNN,explained the classification results using gradient-weighted class activation mapping (Grad-CAM),and analyzed the angiogenic pattern from 3dpf to 4dpf based on the model. A total of 17 fertilized eggs were observed in the experiment. Results showed that the accuracy of the optimal model to classify 3dpf and 4dpf vascular images was 98.62%. Using the Grad-CAM technique,we found out that the main manifestation of the vascular network development from 3dpf to 4dpf was the growth of capillary networks. Between 3dpf and 4dpf,the process of angiogenesis was more intense during the first 12 hours and then tended to be stable. This study provided new approaches for the researches on angiogenesis,assisting the physiological studies including angiogenesis mechanisms,tumor growth and organ aging.
		2020 Vol. 39 (5): 524-531 [Abstract] ( 393 ) HTML (1 KB) PDF (11628 KB) ( 279 )

	532	Breast Cancer Image Classification Based on Fusion Multi-Network Deep Convolution Features and Sparse Double Relation Regularization Method
		Wang Yongjun, Huang Fanglin, Huang Shan, Jiang Feng, Lei Baiying, Wang Tianfu
		DOI: 10.3969/j.issn.0258-8021.2020.05.003
		Breast cancer is one of the leading causes of cancer death worldwide. Existing diagnostic methods are mainly dependent on the observation with histopathological images,which is laborious,time-consuming,and relies on the doctor's professional knowledge and experience,making the diagnosis efficiency unsatisfactory. In view of these problems,this paper aimed to improve the breast cancer diagnostic accuracy and reduce the workload of doctors by devising a deep learning framework based on histological image. Specifically,this paper developed a classification model based on multi-network feature fusion and sparse double-relation regularized learning. First,the breast cancer pathological images were preprocessed by sub-image clipping and color enhancement. Then,three deep convolutional neural networks (InceptionV3,ResNet-50,and VGG-16) typical of deep learning model were used to extract multi-network deep convolution features of breast cancer pathological images. Third,by using two relations (“sample-sample” relation and “feature-feature” relation) and l_F regularization,we proposed a supervised double relation regularization learning method to reduce feature dimension. Support vector machines was used to distinguish breast cancer pathological images into four categories:normal,benign,carcinoma in situ,and invasive carcinoma. In the experiment,by using 400 breast cancer pathological images in the ICIAR 2018 public data set to verify the proposed method,93% classification accuracy was obtained. Results showed that multi-network deep convolution fusion features could effectively capture rich image information,and sparse dual-relation regularization learning could effectively reduce feature honor and reduce noise interference,which will effectively improve the classification performance of the model.
		2020 Vol. 39 (5): 532-540 [Abstract] ( 414 ) HTML (1 KB) PDF (5627 KB) ( 453 )

	541	Prognostic Study of Tumor-Microenvironment Interaction Risk Model Based on Digital Histopathological Images in Oropharyngeal Cancer
		Zhang Dan Lu Cheng, Dai Cai, Wu Yuxin, Lu Qianshuai, Lei Xiujuan
		DOI: 10.3969/j.issn.0258-8021.2020.05.004
		Digital whole slide histopathology image provides a new opportunity for computerized quantitative analysis. More and more studies have shown that interactions between immune cells and cancerous cells in the tumor microenvironment is an important prognostic indicator. HPV⁺ oropharyngeal cancer is a common malignant tumor of head and neck. At present,there are not any ideal prognostic indicators for HPV⁺ oropharyngeal cancer. In this study,the computer image processing and pattern recognition technology were used to quantitatively extract the nuclear morphology from the whole slide image of digital histopathology,and the nuclear morphological features were used to measure the degree of interaction between the tumor microenvironment and the cancerous region and construct a recurrence risk model of oropharyngeal cancer. The histopathological sections and corresponding follow-up data of 234 patients with oropharyngeal cancer were collected retrospectively from the pathology files of the University of Washington Medical Center. We found out that the recurrent risk model of oropharyngeal cancer constructed by image quantitative analysis could distinguish relapsed and non-relapsed patients significantly,and the average AUC of 100-fold 5-fold cross-validated classification results reached 0.67±0.02;In univariate (HR (95%CI)=1.76 (0.99～3.13),P=0.0352) and multivariate analysis (HR (95%CI)=3.27(1.12～5.46),P=0.039),the analysis results showed that patients with oropharyngeal cancer with stronger interactions between the tumor microenvironment and the cancerous area had lower risk of recurrence and longer survival than those with low interaction. This finding revealed that the interaction between immune cells and cancerous cells in the tumor microenvironment could serve as an independent prognostic indicator to guide the treatment of oropharyngeal cancer.
		2020 Vol. 39 (5): 541-549 [Abstract] ( 391 ) HTML (1 KB) PDF (6806 KB) ( 351 )

	550	Three-Type Classification Method for Wearable ECG Signal Quality
		Wang Shuai, Zhao Zhongyao, Zhang Xiangyu, Zhao Lina, Li Jianqing, Liu Chengyu
		DOI: 10.3969/j.issn.0258-8021.2020.05.005
		This study proposed a three-type classification method for wearable ECG signal quality assessment. Rationale for three-type classification is from the clinical requirement for diagnosis,and the three types for signal quality are:1) clinically useful signals with good signal quality;2) clinically useful signals with poor signal quality;3) clinically useless signals. The method firstly extracted 12 signal quality features from time-domain,frequency-domain and nonlinear-domain,and constructed the feature matrix. Then a support vector machine (SVM) classifier based on a radial basis kernel function was trained on the collected wearable ECG signals with manual annotations. Results from 375 independent test ECG samples with clinical labels showed that the proposed method achieved an F measure of 0.909,0.827 and 0.973 for the three-type quality signals respectively,and the overall classification accuracy was 92.3%,which was 2.2% and 6.4% higher than the comparable methods,i.e.,the CNN-based model and the traditional SVM model. This study demonstrated that the new three-type classification model of signal quality had application potentials in wearable dynamic ECG signal quality assessment.
		2020 Vol. 39 (5): 550-556 [Abstract] ( 469 ) HTML (1 KB) PDF (1442 KB) ( 721 )

	557	Design and Research of Bionic Hand Exoskeleton Based on Flexible Hinge
		Meng Qiaoling, Shen Zhijia, Chen Zhongzhe, Nie Zhiyang
		DOI: 10.3969/j.issn.0258-8021.2020.05.006
		In order to help patients with hand dysfunction perform hand rehabilitation training and activities better in daily life,this paper proposed a bionic hand exoskeleton based on flexible hinge. Firstly,based on elastic mechanics,the mapping relation model of stiffness between the straight beam and semi-circular beam-type flexible hinge was established. Then the stiffness design equation of semi-circular beam-type flexible hinge was improved by the finite element model and the universality and effectiveness of the equation were proved by the finite element model of three-hinge series finger exoskeleton. Secondly,based on the improved equation,this paper designed a bionic hand exoskeleton based on flexible hinge,which could realize buckling motion through ropes and stretch motion through flexure hinge. And the kinematics model was established by homogeneous coordinate transformation to study the motion characteristics of rope drive. Finally,experiments of bending performance and grasping ability were carried out on the prototype. The simulation results of finger exoskeleton showed that the stiffness design equation of single hinge could be used to calculate the stiffness of each hinge in the complex mechanism,and the error value was no more than 3.5%. In this study,the weight of hand executive component of prototype was only 125 g,which was convenient for users to carry in daily life. The experimental results of bending performance showed that the improved stiffness design equation reduced the calculation error of 40% stiffness of semi-circular beam-type flexible hinge,and made the error within 5%,so that the geometric dimensions of the hinge satisfying the bending condition could be obtained quickly. The experimental results of grasping ability showed that two healthy subjects could measure the stable output of 8 N fingertip force from the single finger of the exoskeleton hand,which could meet the needs of daily life. Therefore,the hand exoskeleton could not only provide normal range of motion of the joints,but also assisted patients with hand dysfunction in hand rehabilitation training and daily activities due to its characteristics of light weight,portability and high grasping ability.
		2020 Vol. 39 (5): 557-565 [Abstract] ( 448 ) HTML (1 KB) PDF (3556 KB) ( 638 )

	566	Study on the Effect of Electrode Spacing and Electrode Diameter on Bipolar Radiofrequency Fat Melting at Constant Power
		Zang Lianru, Zhou Yu, Kang Jia, Lin Haixiao, Li Yuan, Xue Yinmin
		DOI: 10.3969/j.issn.0258-8021.2020.05.007
		So far there is no uniform specification for some design parameters of bipolar radiofrequency (RF) devices for noninvasive treatment of local fat accumulation. In this work,the finite element method and external experiment were used to analyze the influence of different parameters on the fat melting effect of tissues in bipolar RF,seeking for the fat melting configuration with a large range of fat melting effect without causing thermal damage to skin layer as far as possible. The finite element analysis of thermoelectric coupling of biological tissues was carried out by COMSOL Multiphysics. To verify the validity of the model,a self-developed single-channel bipolar RF output device was used to perform RF output of pig abdominal tissues. The results of finite element analysis showed that the final temperature of the skin layer was lower than the thermal damage threshold temperature and part of the fat layer was in the thermal damage area under the fat melting configuration with the power of 10 W,the diameter of the electrode sphere of 3,5,8 mm and electrode spacing of 2 and 3 cm after 30 min heating with bipolar RF;the domain point probe showed that the temperature curve in the thermal damage area of the fat layer met the requirement of the fat melting temperature. The temperature distribution in tissues under bipolar RF heating was significantly affected by different fat melting configurations. At a power of 10 W,when using a spherical electrode with a diameter of 8 mm and pressing the skin to a depth of 1 mm,under the condition of electrode spacing of 2 and 3 cm,the largest area of continuous thermal damage area and spot-shaped thermal damage area within the fat layer will be generated,and the area of thermal damage area is 2.84 and 2.55 cm²,respectively.External experiments with the same fat melting configuration showed that the final temperature of the thermocouple probe at the same position as the tissue model was 0.92±0.43℃ different from that of the corresponding domain point probe. The finite element analysis results were consistent with the experimental results. According to the results of simulation and experimental validation,the above configuration made the bipolar RF not cause thermal damage to the skin layer,at the same time while generate effective thermal damage area in the fat layer. The temperature in the thermal damage area could meet the requirement of the fat melting temperature. The reasonableconfiguration is a critical factor for the success of bipolar RF fat melting.
		2020 Vol. 39 (5): 566-576 [Abstract] ( 428 ) HTML (1 KB) PDF (7145 KB) ( 107 )

	577	Simulation Analysis of Microporous Characteristics of Spherical Cells Based on Electroporation and Pore Radii Change Equation
		Guo Fei, Zhang Lin, Liu Xin, Peng Hao
		DOI: 10.3969/j.issn.0258-8021.2020.05.008
		In order to explore the theoretical mechanism of electroporation (EP),a 2D-axisymmetric model of single cell EP was established in this paper. The asymptotic electroporation equation and the pore radii evolution equation were included in the model to express the pore density and radii of EP. The perforation area was calculated more accurate by the axisymmetry of the model. Thus the temporal and spatial distribution characteristics of the micropores can be obtained,and the influence of the field strength and pulse width on EP was discussed. The results showed that 7 862 micropores were produced under the microsecond pulse electric field (μsPEF) of 100 μs and 2 kV/cm,and the perforation area accounted for 6.3% of the cell surface;the temporal and spatial distribution of the parameters of EP were consistent with the results of literature,which verified the correctness of the constructed model;in the range of 1 to 5 kV/cm,the number of pores was directly proportional to the field strength,the pore radius at P₁ was in inverse proportion to the field strength,while the ratio of pore area to cell area increased from 1.3% to 12.9%;two groups of nsPEF and μsPEF with the same energy were selected for comparative study,it was found that at the end of the pulse,the number of pores generated by the former was 353.1 times of the latter one,and the pore radius at P₁ of the latter was 19.3 times of the former one,indicated that nsPEF was conducive to the growth of the number of pores,while the μsPEF was conducive to the expansion of the pore radii. The simulation results showed that microporous characteristics determined the occurrence and development of EP,and the accurate calculation of microporosity was the key to explain the effect of EP.
		2020 Vol. 39 (5): 577-586 [Abstract] ( 413 ) HTML (1 KB) PDF (4480 KB) ( 463 )

Reviews

	587	Advances in Real-Time Functional Magnetic Resonance Imaging Neurofeedback
		Yang Dongmei, Zhang Wenhai, Ding Qiang
		DOI: 10.3969/j.issn.0258-8021.2020.05.009
		Neurofeedback refers to a self-regulation technique that provides individuals with feedback about specific brain activity in connection with a related behavior. Real-time Functional Magnetic Resonance Imaging Neurofeedback (rtfMRI-nf) is a novel neurofeedback technique,requiring participants to adjust blood oxygen level-dependent signal index during training to regulate their brain activity. Recently,rtfMRI-nf has made significant progress in data acquisition and analysis. Two new rtfMRI-nf technologies have begun available,decoded neurofeedback and functional connectivity-based neurofeedback. This paper introduced the two new rtfMRI-nf technologies and summarizes their progress in key method areas such as implicit protocol,multivariate analysis,and connectivity analysis. At the same time,this paper reviewed the current status of the two rtfMRI-nf technologies in basic research fields such as perceptual learning and metacognition,as well as clinical research fields such as fear elimination,depression,autism spectrum disorder and nicotine addiction. Finally,we discussed the two potential problems of “one to many” relationship and dimension curse of these two rtfMRI-nf technologies,and propose solutions.
		2020 Vol. 39 (5): 587-594 [Abstract] ( 471 ) HTML (1 KB) PDF (2734 KB) ( 486 )

	595	Research Progress of Multimodal Functional Neural ImagingTechnology Based on EEG
		Zhou Yijie, Song Xizi, He Feng, Wan Baikun, Ming Dong
		DOI: 10.3969/j.issn.0258-8021.2020.05.010
		Functional neural imaging is able to non-invasively detect the changes in structure and function of living brain. The scalp electroencephalogram has the advantages of high time resolution,easy operation and direct reflection of the brain's neurophysiological activities. It has been integrated into the new multimodal functional neural imaging technology with both high temporal and spatial resolution. In this paper,electroencephalogram functional magnetic resonance imaging (EEG-fMRI),electroencephaloelectric near-infrared spectroscopy (EEG-NIRS) and brain electric transcranial focused ultrasound (EEG-tFUS) were taken as examples to reflect the current state of the development. The main contents included technical principles,technical characteristics and the latest progress in neural function research. In addition,the shortcomings and development trend of the three new models were discussed in detail to promote the research and application of multimodal functional neural imaging technology in brain science.
		2020 Vol. 39 (5): 595-602 [Abstract] ( 433 ) HTML (1 KB) PDF (2665 KB) ( 1436 )

	603	Progress in Computer-Aided Diagnosis of Parkinson′s Disease Based on Magnetic Resonance Imaging
		Yang Yifeng, Hu Ying, Nie Shengdong
		DOI: 10.3969/j.issn.0258-8021.2020.05.011
		In recent years,due to the clinical complexity of Parkinson’s disease (PD) and the high-dimensional nature of multi-modemagnetic resonance (MR) images,how to effectively use the specific image biomarkers and establish an efficient Computer-aided Diagnosis (CAD) model for disease diagnosis is a challenging problem in PD research. This paper reviewed the research progress,and summarized key techniques of CAD modeling based on traditional machine learning methods such as feature extraction,feature selection andthe classifier model. This paper also briefly introduced the recent research and application of deep learning in early PD classification diagnosis. It is pointed out that based on multi-modal images,CAD model constructed by machine learning or deep learning can recognize PD patients and normal people objectively and accurately,which has great value and application prospect to improve the accuracy of early PD diagnosis. Future researches should be carried out to explore the potential biomarkers of PD in multi-modality images,and to develop higher-order CAD models to assist the clinical intelligent diagnosis of early PD.
		2020 Vol. 39 (5): 603-610 [Abstract] ( 361 ) HTML (1 KB) PDF (788 KB) ( 599 )

	611	Progress in Tissue Engineering Scaffolds Based on Polycaprolactone Fibers
		Wu Xiaotong, He Er, Liu Laijun, Li Chaojing, Wang Lu, Wang Fujun
		DOI: 10.3969/j.issn.0258-8021.2020.05.012
		Polycaprolactone (PCL) is a kind of biocompatible,absorbable and easily modified polyester. Tissue engineering scaffolds based on PCL fibers are widely used because of their high specific surface area,good mechanical properties,and easy control of pore size,porosity and fiber orientation. In this paper,main application defects of PCL fiber scaffolds,such as poor cell affinity,slow degradation rate and low mechanical strength and the improvement methods were reviewed. At the same time,the latest development of tissue engineering scaffolds based on PCL fiber in the regeneration of skin,blood vessel,nerve,tendon,ligament and cartilage was summarized. It is shown that most of the current studies focused on improving cell-scaffold interaction and regulating the degradation behavior of scaffolds by introducing bioactive substances or drugs,or changing the physical structure of scaffolds by using different spinning processes and parameters to regulate the mechanical properties and cell-behavior induction of scaffolds. In addition,most of the studies are still staying in the laboratory stage. It is an important development direction in the future to promote tissue engineering scaffolds based on PCL fibers with the advantages of low cost and easy processing.
		2020 Vol. 39 (5): 611-620 [Abstract] ( 541 ) HTML (1 KB) PDF (838 KB) ( 415 )

Communications

	621	Classification of X-Ray Phase-Contrast CT Images in Liver Cancer Based on Machine Learning
		Wang Kun, Zhang Xueliang, Zhang Suixia, Ji Xuewen, Liu Huiqiang
		DOI: 10.3969/j.issn.0258-8021.2020.05.013

		2020 Vol. 39 (5): 621-625 [Abstract] ( 380 ) HTML (1 KB) PDF (2817 KB) ( 346 )

	626	Application in Respiratory Tract Nine Combined Tests of Fluorescence Microscopic Image Recognition Methods
		Wang Chao, Liu Cong, Hou Jianping, Zhao Wanli, Duan Yirui
		DOI: 10.3969/j.issn.0258-8021.2020.05.014

		2020 Vol. 39 (5): 626-630 [Abstract] ( 251 ) HTML (1 KB) PDF (10279 KB) ( 30 )

	631	Design and Implementation of 128-Channel Neural Signal Simulator Based on FPGA and Real Datasets
		Zheng Yanyu, Wang Minmin, Gao Xiang, Zhou Fan, Zhang Shaomin
		DOI: 10.3969/j.issn.0258-8021.2020.05.015

		2020 Vol. 39 (5): 631-635 [Abstract] ( 412 ) HTML (1 KB) PDF (2831 KB) ( 420 )

	636	The Influence of COMT Gene Polymorphism and other Factors on Aggressive Behavior in Schizophrenia
		Cui Lijun, Li Liang, Li Jianhua, Zhu Jielin, SongJuanfang, Sun Jushui
		DOI: 10.3969/j.issn.0258-8021.2020.05.016

		2020 Vol. 39 (5): 636-640 [Abstract] ( 249 ) HTML (1 KB) PDF (715 KB) ( 129 )

Copyright © Editorial Board of Chinese Journal of Biomedical Engineering
Supported by:Beijing Magtech