OBJECTIVES: To design a deep learning model that balances model complexity and performance to enable its integration into wearable ECG monitoring devices for automated diagnosis of atrial fibrillation. METHODS: This study was performed based on data from 84 patients with atrial fibrillation, 25 patients with atrial fibrillation, and 18 subjects without obvious arrhythmia collected from the publicly available datasets LTAFDB, AFDB, and NSRDB, respectively. A lightweight attention network based on depthwise separable convolution and fusion of channel-spatial information, namely DSC-AttNet, was proposed. Depthwise separable convolution was introduced to replace standard convolution and reduce model parameters and computational complexity to realize high efficiency and light weight of the model. The multilayer hybrid attention mechanism was embedded to compute the attentional weights of the channels and spatial information at different scales to improve the feature expression ability of the model. Ten-fold cross-validation was performed on LTAFDB, and external independent testing was conducted on AFDB and NSRDB datasets. RESULTS: DSC-AttNet achieved a ten-fold average accuracy of 97.33% and a precision of 97.30% on the test set, both of which outperformed the other 4 comparison models as well as the 3 classical models. The accuracy of the model on the external test set reached 92.78%, better than those of the 3 classical models. The number of parameters of DSC-AttNet was 1.01M, and the computational volume was 27.19G, both smaller than the 3 classical models. CONCLUSIONS This proposed method has a smaller complexity, achieves better classification performance, and has a better generalization ability for atrial fibrillation classification.
Atrial Fibrillation/diagnosis* ; Humans ; Electrocardiography ; Deep Learning ; Wearable Electronic Devices ; Neural Networks, Computer

OBJECTIVES: We propose a myocardial infarction (MI) detection and localization model for improving the diagnostic accuracy for MI to provide assistance to clinical decision-making. METHODS: The proposed model was constructed based on multi-scale field residual blocks fusion modified channel attention (MSF-RB-MCA). The model utilizes lead II electrocardiogram (ECG) signals to detect and localize MI, and extracts different levels of feature information through the multi-scale field residual block. A modified channel attention for automatic adjustment of the feature weights was introduced to enhance the model's ability to focus on the MI region, thereby improving the accuracy of MI detection and localization. RESULTS: A 5-fold cross-validation test of the model was performed using the publicly available Physikalisch-Technische Bundesanstalt (PTB) dataset. For MI detection, the model achieved an accuracy of 99.96% on the test set with a specificity of 99.84% and a sensitivity of 99.99%. For MI localization, the accuracy, specificity and sensitivity were 99.81%, 99.98% and 99.65%, respectively. The performances of the model for MI detection and localization were superior to those of other comparison models. CONCLUSIONS The proposed MSF-RB-MCA model shows excellent performance in AI detection and localization based on lead II ECG signals, demonstrating its great potential for application in wearable devices.
Myocardial Infarction/diagnosis* ; Humans ; Electrocardiography/methods* ; Signal Processing, Computer-Assisted ; Algorithms ; Sensitivity and Specificity

To address the limitations of the existing automatic electrocardiogram(ECG)diagnosis models in learning long-term dependencies,an automatic 12-lead long-term ECG signal diagnosis model which combines bidirectional selective state space model(bidirectional mamba,BiMamba)with residual multi-scale receptive field block(RMSF)is proposed:(1)designing a multi-scale receptive field module with residual connections to realize more extensive feature extraction and fusion;(2)introducing BiMamba block to enhance the model's temporal modeling capability by employing both forward and backward temporal processing;(3)using the classifier to process features from BiMamba for accomplishing multi-label ECG classification.Five major diagnostic categories from the PTB-XL dataset are extracted and subjected to 5-fold cross-validation experiments.The experimental results from the comparative study show that BiMamba-RMSF achieves an average accuracy of 89.42%,an average AUC of 0.9356,and an average F1 score of 72.85%,outperforming the other 4 automatic ECG diagnosis models.Additionally,ablation study further validates the effectiveness of BiMamba block.It is demonstrated that the proposed model has a high precision in the multi-label classification for 12-lead long-term ECG signals.

To address the limitations of the existing automatic electrocardiogram(ECG)diagnosis models in learning long-term dependencies,an automatic 12-lead long-term ECG signal diagnosis model which combines bidirectional selective state space model(bidirectional mamba,BiMamba)with residual multi-scale receptive field block(RMSF)is proposed:(1)designing a multi-scale receptive field module with residual connections to realize more extensive feature extraction and fusion;(2)introducing BiMamba block to enhance the model's temporal modeling capability by employing both forward and backward temporal processing;(3)using the classifier to process features from BiMamba for accomplishing multi-label ECG classification.Five major diagnostic categories from the PTB-XL dataset are extracted and subjected to 5-fold cross-validation experiments.The experimental results from the comparative study show that BiMamba-RMSF achieves an average accuracy of 89.42%,an average AUC of 0.9356,and an average F1 score of 72.85%,outperforming the other 4 automatic ECG diagnosis models.Additionally,ablation study further validates the effectiveness of BiMamba block.It is demonstrated that the proposed model has a high precision in the multi-label classification for 12-lead long-term ECG signals.

The superiority of magnetic resonance(MR)images in soft tissue imaging makes them indispensable for medical diagnosis and radiotherapy,but factors such as acquisition cost and contraindications limit their widespread application.In contrast,computed tomography(CT)scanning has the advantages of fast imaging speed and low cost.Herein,this review summarizes the research progress of generative deep learning models in the field of medical CT to MR image synthesis,and especially analyzes the technical characteristics,performance advantages,and challenges of various MR image synthesis methods from clinical scenarios such as spinal lesions,acute ischemic stroke,and tumor segmentation.Furthermore,the application value and future research prospects of medical image synthesis are discussed.

The superiority of magnetic resonance(MR)images in soft tissue imaging makes them indispensable for medical diagnosis and radiotherapy,but factors such as acquisition cost and contraindications limit their widespread application.In contrast,computed tomography(CT)scanning has the advantages of fast imaging speed and low cost.Herein,this review summarizes the research progress of generative deep learning models in the field of medical CT to MR image synthesis,and especially analyzes the technical characteristics,performance advantages,and challenges of various MR image synthesis methods from clinical scenarios such as spinal lesions,acute ischemic stroke,and tumor segmentation.Furthermore,the application value and future research prospects of medical image synthesis are discussed.

Objective To develop a method for accurate identification of multiscale carotid plaques in ultrasound images.Methods We proposed a two-stage carotid plaque detection method based on deep convolutional neural network(SM-YOLO).A series of algorithms such as median filtering,histogram equalization,and Gamma transformation were used to preprocess the dataset to improve image quality.In the first stage of the model construction,a candidate plaque set was built based on the YOLOX_l target detection network,using multiscale image training and multiscale image prediction strategies to accommodate carotid artery plaques of different shapes and sizes.In the second stage,the Histogram of Oriented Gradient(HOG)features and Local Binary Pattern(LBP)features were extracted and fused,and a Support Vector Machine(SVM)classifier was used to screen the candidate plaque set to obtain the final detection results.This model was compared quantitatively and visually with several target detection models(YOLOX_l,SSD,EfficientDet,YOLOV5_l,Faster R-CNN).Results SM-YOLO achieved a recall of 89.44%,an accuracy of 90.96%,a F1-Score of 90.19%,and an AP of 92.70%on the test set,outperforming other models in all performance indicators and visual effects.The constructed model had a much shorter detection time than the Faster R-CNN model(only one third of that of the latter),thus meeting the requirements of real-time detection.Conclusion The proposed carotid artery plaque detection method has good performance for accurate identification of carotid plaques in ultrasound images.

Objective To establish a deep learning model for testing the feasibility of combining magnetic resonance imaging(MRI)deep learning features with clinical features for preoperative prediction of cytokeratin 19(CK19)status of hepatocellular carcinoma(HCC).Methods A retrospective experiment was conducted based on the data of 116 HCC patients with confirmed CK19 status.A single sequence multi-scale feature fusion deep learning model(MSFF-IResnet)and a multi-scale and multi-modality feature fusion model(MMFF-IResnet)were established based on the hepatobiliary phase(HBP),diffusion weighted imaging(DWI)sequences of enhanced MRI images,and the clinical features significantly correlated with CK19 status.The classification performance of the models were evaluated to assess the effectiveness of the deep learning models for predicting CK19 status of HCC before surgery.Results Multivariate analysis showed that an increased neutrophil-to-lymphocyte ratio(P=0.029)and incomplete tumor capsule(P=0.028)were independent predictors of CK19 expression in HCC.The deep learning models improved by multi-scale feature fusion and multi-modality feature fusion methods achieved better classification results than the traditional machine learning models and baseline models,and the final MMFF-IResnet model showed the best classification performance with an AUC of 84.2%,an accuracy of 80.6%,a sensitivity of 80.1%and a specificity of 81.2%.Conclusion The multi-scale and multi-modality feature fusion model based on MRI and clinical parameters is capable of predicting CK19 status of HCC,demonstrating the feasibility of combining deep learning methods with MRI and clinical features for preoperative prediction of CK19 status.

Objective To develop a method for accurate identification of multiscale carotid plaques in ultrasound images.Methods We proposed a two-stage carotid plaque detection method based on deep convolutional neural network(SM-YOLO).A series of algorithms such as median filtering,histogram equalization,and Gamma transformation were used to preprocess the dataset to improve image quality.In the first stage of the model construction,a candidate plaque set was built based on the YOLOX_l target detection network,using multiscale image training and multiscale image prediction strategies to accommodate carotid artery plaques of different shapes and sizes.In the second stage,the Histogram of Oriented Gradient(HOG)features and Local Binary Pattern(LBP)features were extracted and fused,and a Support Vector Machine(SVM)classifier was used to screen the candidate plaque set to obtain the final detection results.This model was compared quantitatively and visually with several target detection models(YOLOX_l,SSD,EfficientDet,YOLOV5_l,Faster R-CNN).Results SM-YOLO achieved a recall of 89.44%,an accuracy of 90.96%,a F1-Score of 90.19%,and an AP of 92.70%on the test set,outperforming other models in all performance indicators and visual effects.The constructed model had a much shorter detection time than the Faster R-CNN model(only one third of that of the latter),thus meeting the requirements of real-time detection.Conclusion The proposed carotid artery plaque detection method has good performance for accurate identification of carotid plaques in ultrasound images.

Objective To establish a deep learning model for testing the feasibility of combining magnetic resonance imaging(MRI)deep learning features with clinical features for preoperative prediction of cytokeratin 19(CK19)status of hepatocellular carcinoma(HCC).Methods A retrospective experiment was conducted based on the data of 116 HCC patients with confirmed CK19 status.A single sequence multi-scale feature fusion deep learning model(MSFF-IResnet)and a multi-scale and multi-modality feature fusion model(MMFF-IResnet)were established based on the hepatobiliary phase(HBP),diffusion weighted imaging(DWI)sequences of enhanced MRI images,and the clinical features significantly correlated with CK19 status.The classification performance of the models were evaluated to assess the effectiveness of the deep learning models for predicting CK19 status of HCC before surgery.Results Multivariate analysis showed that an increased neutrophil-to-lymphocyte ratio(P=0.029)and incomplete tumor capsule(P=0.028)were independent predictors of CK19 expression in HCC.The deep learning models improved by multi-scale feature fusion and multi-modality feature fusion methods achieved better classification results than the traditional machine learning models and baseline models,and the final MMFF-IResnet model showed the best classification performance with an AUC of 84.2%,an accuracy of 80.6%,a sensitivity of 80.1%and a specificity of 81.2%.Conclusion The multi-scale and multi-modality feature fusion model based on MRI and clinical parameters is capable of predicting CK19 status of HCC,demonstrating the feasibility of combining deep learning methods with MRI and clinical features for preoperative prediction of CK19 status.