Brain-Tumor Interface-Based MRI Radiomics Models to Determine EGFR Mutation, Response to EGFR-TKI and T790M Resistance Mutation in Non-Small Cell Lung Carcinoma Brain Metastasis.

BACKGROUND: Preoperative assessment of epidermal growth factor receptor (EGFR) status, response to EGFR-tyrosine kinase inhibitors (TKI) and development of T790M mutation in non-small cell lung carcinoma (NSCLC) patients with brain metastases (BM) is important for clinical decision-making, while previous studies were only based on the whole BM. PURPOSE: To investigate values of brain-to-tumor interface (BTI) for determining the EGFR mutation, response to EGFR-TKI and T790M mutation. STUDY TYPE: Retrospective. POPULATION: Two hundred thirty patients from Hospital 1 (primary cohort) and 80 patients from Hospital 2 (external validation cohort) with BM and histological diagnosis of primary NSCLC, and with known EGFR status (biopsy) and T790M mutation status (gene sequencing). FIELD STRENGTH/SEQUENCE: Contrast-enhanced T1-weighted (T1CE) and T2-weighted (T2W) fast spin echo sequences at 3.0T MRI. ASSESSMENT: Treatment response to EGFR-TKI therapy was determined by the Response Evaluation Criteria in Solid Tumors. Radiomics features were extracted from the 4 mm thickness BTI and selected by least shrinkage and selection operator regression. The selected BTI features and volume of peritumoral edema (VPE) were combined to construct models using logistic regression. STATISTICAL TESTS: The performance of each radiomics model was evaluated using the area under the receiver operating characteristic (ROC) curve (AUC). RESULTS: A total of 7, 3, and 3 features were strongly associated with the EGFR mutation status, response to EGFR-TKI and T790M mutation status, respectively. The developed models combining BTI features and VPE can improve the performance than those based on BTI features alone, generating AUCs of 0.814, 0.730, and 0.774 for determining the EGFR mutation, response to EGFR-TKI and T790M mutation, respectively, in the external validation cohort. DATA CONCLUSION: BTI features and VPE were associated with the EGFR mutation status, response to EGFR-TKI and T790M mutation status in NSCLC patients with BM. EVIDENCE LEVEL: 3 Technical Efficacy: Stage 2.

[Classification Model of Corneal Opacity Based on Digital Image Features].

OBJECTIVE: According to the digital image features of corneal opacity, a multi classification model of support vector machine (SVM) was established to explore the objective quantification method of corneal opacity. METHODS: The cornea digital images of dead pigs were collected, part of the color features and texture features were extracted according to the previous experience, and the SVM multi classification model was established. The test results of the model were evaluated by precision, sensitivity and F1 scores. The optimal feature subset was found by SVM-RFE combined with cross validation to optimize the model. RESULTS: In the classification of corneal opacity, the highest F1 score was 0.974 4, and the number of features in the optimal feature subset was 126. CONCLUSIONS: The SVM multi classification model can classify the degree of corneal opacity.

The self-organization model reveals systematic characteristics of aging.

BACKGROUND: Aging is a fundamental biological process, where key bio-markers interact with each other and synergistically regulate the aging process. Thus aging dysfunction will induce many disorders. Finding aging markers and re-constructing networks based on multi-omics data (i.e. methylation, transcriptional and so on) are informative to study the aging process. However, optimizing the model to predict aging have not been performed systemically, although it is critical to identify potential molecular mechanism of aging related diseases. METHODS: This paper aims to model the aging self-organization system using a series of supervised learning methods, and study complex molecular mechanisms of aging at system level: i.e. optimizing the aging network; summarizing interactions between aging markers; accumulating patterns of aging markers within module; finding order-parameters in the aging self-organization system. RESULTS: In this work, the normal aging process is modeled based on multi-omics profiles across different tissues. In addition, the computational pipeline aims to model aging self-organizing systems and study the relationship between aging and related diseases (i.e. cancers), thus provide useful indicators of aging related diseases and could help to improve prediction abilities of diagnostics. CONCLUSIONS: The aging process could be studied thoroughly by modelling the self-organization system, where key functions and the crosstalk between aging and cancers were identified.

[Retinal Vessel Segmentation Based on Multiscale Matched Filtering].

Retinal vascular function is complex, morphological structure varies from person to person, and is susceptible to vascular diseases and systemic vascular diseases. Its accurate segmentation is of great significance for disease diagnosis and identification. In this paper, a multi-scale matching filtering algorithm is proposed for the uneven size of retinal blood vessels. On the basis of the traditional singlescale Gaussian matching filter, multiscale Gaussian matched filters with two sizes are used to enhance grayscale images. Enhancement is performed, and the superimposed image is binarized using a twodimensional maximum entropy threshold segmentation algorithm. The algorithm is tested in the DRIVE database with sensitivity, specificity and accuracy of 0.803, 0.959, 0.981, respectively. Comparing with the traditional algorithm, the algorithm has high sensitivity, fast running speed and rich details of segmentation results.

[Automatic Identification and Classification Diagnosis of Atrial Ventricular Hypertrophy Electrocardiogram Based on Convolutional Neural Network].

OBJECTIVE: Identifying Atrial Ventricular Hypertrophy Electrocardiogram (AVH ECG)and diagnosing the classification of theirs automatically. METHODS: The ECG data used in this experiment was collected from the First Affiliated Hospital of China Medical University. CNN are combined with conventional methods and a 10 layers of one dimensional CNN are created in this experiment to extract the features of ECG signals automatically and achieve the function of classifying. ROC, sensitivity and F1-score are used here to evaluate the effects of the model. RESULTS: In the experiment of identifying AVH ECG, the AUC of test dataset is 0.991, while in the experiment of classifying AVH ECG, the maximal F1-score can reach 0.992. CONCLUSIONS: The CNN model created in this experiment can achieve the auxiliary diagnosis of AVH ECG.

Associations of psychological distress with positive psychological variables and activities of daily living among stroke patients: a cross-sectional study.

BACKGROUND: Depression and anxiety result in psychological distress, which can further affect mental status and quality of life in stroke patients. Exploring the associations between positive psychological variables and symptoms of psychological distress following stroke is of great significance for further psychological interventions. METHODS: A total of 710 stroke patients from the five largest cities in Liaoning Province in China were enrolled into the present study in July 2014. All patients independently completed the questionnaires with respect to psychological distress and positive psychological variables. Depressive and anxiety symptoms were evaluated using Center for Epidemiologic Studies Depression Scale (CES-D) and Self-Rating Anxiety Scale, respectively. Positive psychological variables were evaluated using Perceived Social Support Scale, Adult Hope Scale (AHS), General Perceived Self-Efficacy Scale and Resilience Scale-14 (RS-14). Activities of Daily Living (ADL) was measured using Barthel Index. Factors associated with psychological variables and depressive and anxiety symptoms were identified using t-test, ANOVA, correlation and hierarchical linear regression analysis. RESULTS: Depressive and anxiety symptoms were present in 600 of 710 (84.51%) and 537 of 710 (75.63%) stroke patients enrolled, respectively. Social support (ß = - 0.111, p < 0.001) and hope (ß = - 0.120, p < 0.001) were negatively associated with both depressive and anxiety symptoms. Resilience (ß = - 0.179, p < 0.001) was negatively associated with depressive symptoms. Self-efficacy (ß = - 0.135, p < 0.001) was negatively associated with anxiety symptoms. Hierarchical regression analyses indicated that ADL accounted for 10.0 and 6.0% of the variance of depressive and anxiety symptoms, respectively. Social support, resilience, self-efficacy and hope as a whole accounted for 7.5 and 5.3% of the variance of depressive and anxiety symptoms. CONCLUSIONS: The high frequency of depressive and anxiety symptoms among Chinese stroke survivors should receive attentions from all stakeholders. Findings suggested that intervention strategies on ADL, social support, hope, resilience and self-efficacy could be developed to improve psychosocial outcomes for stroke survivors.

[On the Status of the Domestic Mobile Medical Software Supervision].

With the widespread use of smart phones, mobile medical class of applications use more widely. The regulation for medical applications is in the offing in domestic. How to effectively regulate such software to control its risks for patients is needed to solve. In this paper, the status of such simple software monitoring in domestic and abroad is analyzed, some problems of domestic mobile medical applications are summarized and some recommendations are proposed for the relevant departments.

[Metrology research on biomedical engineering publications from China in recent years].

The present paper is to evaluate the scientific research level and development trends of biomedical engineering in China using metrology analysis on Chinese biomedical engineering scientific literatures. Pubmed is used to search the biomedical engineering publications in recent 5 years which are indexed by Science Citation Index, and the number and cited times of these publications and the impact factor of the journals are analyzed. The results show that comparing with the world, although the number of the publication in China has increased in recent 5 years, there is still much room for improvement. Among Chinese mainland, Hongkong and Taiwan, Chinese mainland maintains the obvious advantage in this subject, but Hongkong has the highest average cited number. Shanghai and Beijing have better research ability than other areas in Chinese mainland.

ERTNet: an interpretable transformer-based framework for EEG emotion recognition.

Background: Emotion recognition using EEG signals enables clinicians to assess patients' emotional states with precision and immediacy. However, the complexity of EEG signal data poses challenges for traditional recognition methods. Deep learning techniques effectively capture the nuanced emotional cues within these signals by leveraging extensive data. Nonetheless, most deep learning techniques lack interpretability while maintaining accuracy. Methods: We developed an interpretable end-to-end EEG emotion recognition framework rooted in the hybrid CNN and transformer architecture. Specifically, temporal convolution isolates salient information from EEG signals while filtering out potential high-frequency noise. Spatial convolution discerns the topological connections between channels. Subsequently, the transformer module processes the feature maps to integrate high-level spatiotemporal features, enabling the identification of the prevailing emotional state. Results: Experiments' results demonstrated that our model excels in diverse emotion classification, achieving an accuracy of 74.23% ± 2.59% on the dimensional model (DEAP) and 67.17% ± 1.70% on the discrete model (SEED-V). These results surpass the performances of both CNN and LSTM-based counterparts. Through interpretive analysis, we ascertained that the beta and gamma bands in the EEG signals exert the most significant impact on emotion recognition performance. Notably, our model can independently tailor a Gaussian-like convolution kernel, effectively filtering high-frequency noise from the input EEG data. Discussion: Given its robust performance and interpretative capabilities, our proposed framework is a promising tool for EEG-driven emotion brain-computer interface.

Tumor-liver interface in MRI of liver metastasis enables prediction of EGFR mutation in patients with lung cancer: A proof-of-concept study.

BACKGROUND: Preoperative prediction of the epidermal growth factor receptor (EGFR) status in non-small-cell lung cancer (NSCLC) patients with liver metastasis (LM) may have potential clinical values for assisting in treatment decision-making. PURPOSE: To explore the value of tumor-liver interface (TLI)-based magnetic resonance imaging (MRI) radiomics for detecting the EGFR mutation in NSCLC patients with LM. METHODS: This retrospective study included 123 and 44 patients from hospital 1 (between Feb. 2018 and Dec. 2021) and hospital 2 (between Nov. 2015 and Aug. 2022), respectively. The patients received contrast-enhanced T1-weighted (CET1) and T2-weighted (T2W) liver MRI scans before treatment. Radiomics features were extracted from MRI images of TLI and the whole tumor region, separately. The least absolute shrinkage and selection operator (LASSO) regression was used to screen the features and establish radiomics signatures (RSs) based on TLI (RS-TLI) and the whole tumor (RS-W). The RSs were evaluated by the receiver operating characteristic (ROC) curve analysis. RESULTS: A total of 5 and 6 features were identified highly correlated with the EGFR mutation status from TLI and the whole tumor, respectively. The RS-TLI showed better prediction performance than RS-W in the training (AUCs, RS-TLI vs. RS-W, 0.842 vs. 0.797), internal validation (AUCs, RS-TLI vs. RS-W, 0.771 vs. 0.676) and external validation (AUCs, RS-TLI vs. RS-W, 0.733 vs. 0.679) cohort. CONCLUSION: Our study demonstrated that TLI-based radiomics can improve prediction performance of the EGFR mutation in lung cancer patients with LM. The established multi-parametric MRI radiomics models may be used as new markers that can potentially assist in personalized treatment planning.

Identification of crucial inflammaging related risk factors in multiple sclerosis.

Background: Multiple sclerosis (MS) is an immune-mediated disease characterized by inflammatory demyelinating lesions in the central nervous system. Studies have shown that the inflammation is vital to both the onset and progression of MS, where aging plays a key role in it. However, the potential mechanisms on how aging-related inflammation (inflammaging) promotes MS have not been fully understood. Therefore, there is an urgent need to integrate the underlying mechanisms between inflammaging and MS, where meaningful prediction models are needed. Methods: First, both aging and disease models were developed using machine learning methods, respectively. Then, an integrated inflammaging model was used to identify relative risk factors, by identifying essential "aging-inflammation-disease" triples. Finally, a series of bioinformatics analyses (including network analysis, enrichment analysis, sensitivity analysis, and pan-cancer analysis) were further used to explore the potential mechanisms between inflammaging and MS. Results: A series of risk factors were identified, such as the protein homeostasis, cellular homeostasis, neurodevelopment and energy metabolism. The inflammaging indices were further validated in different cancer types. Therefore, various risk factors were integrated, and even both the theories of inflammaging and immunosenescence were further confirmed. Conclusion: In conclusion, our study systematically investigated the potential relationships between inflammaging and MS through a series of computational approaches, and could present a novel thought for other aging-related diseases.

Brain metastasis magnetic resonance imaging-based deep learning for predicting epidermal growth factor receptor (EGFR) mutation and subtypes in metastatic non-small cell lung cancer.

Background: The preoperative identification of epidermal growth factor receptor (EGFR) mutations and subtypes based on magnetic resonance imaging (MRI) of brain metastases (BM) is necessary to facilitate individualized therapy. This study aimed to develop a deep learning model to preoperatively detect EGFR mutations and identify the location of EGFR mutations in patients with non-small cell lung cancer (NSCLC) and BM. Methods: We included 160 and 72 patients who underwent contrast-enhanced T1-weighted (T1w-CE) and T2-weighted (T2W) MRI at Liaoning Cancer Hospital and Institute (center 1) and Shengjing Hospital of China Medical University (center 2) to form a training cohort and an external validation cohort, respectively. A multiscale feature fusion network (MSF-Net) was developed by adaptively integrating features based on different stages of residual network (ResNet) 50 and by introducing channel and spatial attention modules. The external validation set from center 2 was used to assess the performance of MSF-Net and to compare it with that of handcrafted radiomics features. Receiver operating characteristic (ROC) curves, accuracy, precision, recall, and F1-score were used to evaluate the effectiveness of the models. Gradient-weighted class activation mapping (Grad-CAM) was used to demonstrate the attention of the MSF-Net model. Results: The developed MSF-Net generated a better diagnostic performance than did the handcrafted radiomics in terms of the microaveraged area under the curve (AUC) (MSF-Net: 0.91; radiomics: 0.80) and macroaveraged AUC (MSF-Net: 0.90; radiomics: 0.81) for predicting EGFR mutations and subtypes. Conclusions: This study provides an end-to-end and noninvasive imaging tool for the preoperative prediction of EGFR mutation status and subtypes based on BM, which may be helpful for facilitating individualized clinical treatment plans.

Habitat-Based Radiomics for Predicting EGFR Mutations in Exon 19 and 21 From Brain Metastasis.

RATIONALE AND OBJECTIVES: To explore and externally validate habitat-based radiomics for preoperative prediction of epidermal growth factor receptor (EGFR) mutations in exon 19 and 21 from MRI imaging of non-small cell lung cancer (NSCLC)-originated brain metastasis (BM). METHODS: A total of 170, 62 and 61 patients from center 1, center 2 and center 3, respectively were included. All patients underwent contrast-enhanced T1-weighted (T1CE) and T2-weighted (T2W) MRI scans. Radiomics features were extracted from the tumor active (TA) and peritumoral edema (PE) regions in each MRI slice. The most important features were selected by the least absolute shrinkage and selection operator regression to develop radiomics signatures based on TA (RS-TA), PE (RS-PE) and their combination (RS-Com). Receiver operating characteristic (ROC) curve analysis was performed to access performance of radiomics models for both internal and external validation cohorts. RESULTS: 10, four and six most predictive features were identified to be strongly associated with the EGFR mutation status, exon 19 and exon 21, respectively. The RSs derived from the PE region outperformed those from the TA region for predicting the EGFR mutation, exon 19 and exon 21. The RS-Coms generated the highest performance in the primary training (AUCs, RS-EGFR-Com vs. RS-exon 19-Com vs. RS-exon 21-Com, 0.955 vs. 0.946 vs. 0.928), internal validation (AUCs, RS-EGFR-Com vs. RS-exon 19-Com vs. RS-exon 21-Com, 0.879 vs. 0.819 vs. 0.882), external validation 1 (AUCs, RS-EGFR-Com vs. RS-exon 19-Com vs. RS-exon 21-Com, 0.830 vs. 0.825 vs. 0.822), and external validation 2 (AUCs, RS-EGFR-Com vs. RS-exon 19-Com vs. RS-exon 21-Com, 0.812 vs. 0.818 vs. 0.800) cohort. CONCLUSION: The developed habitat-based radiomics model can be used to accurately predict the EGFR mutation subtypes, which may potentially guide personalized treatments for NSCLC patients with BM.

Association of XPC polymorphisms with susceptibility and clinical outcome to chemotherapy in breast cancer patients.

The aim of the current study was to evaluate the relation between xeroderma pigmentosum complementation group C (XPC) polymorphisms and susceptibility to breast cancer (BC), the development and progression of disease, and response to different individualized drug treatments. We investigated two polymorphisms in XPC Ala499Val and Lys939Gln using PCR-RFLP assays including 618 cases and 622 controls. The frequency of the TT genotype of Ala499Val (adjusted odds ratio = 1.575; 95% confidence interval, 1.104-2.245; P = 0.012) and the AC genotype of Lys939Gln (adjusted odds ratio = 1.330; 95% confidence interval, 1.045-1.694; P = 0.020) were found to significantly increase the risk of developing BC. The CT+TT genotypes of Ala499Val were associated with estrogen receptor positive, and Her-2 and p53 negative status, and the AC+CC genotypes of Lys939Gln were associated with BRCA1 negative status. Moreover, a significantly increased chance of treatment with neoadjuvant anthracycline-based chemotherapy response was found in women carrying TT genotype of Ala499Val, or CC and AC genotypes of Lys939Gln. In addition, a significantly longer overall survival after chemotherapy was observed in patients who had XPC Lys939Gln AC+CC genotypes with estrogen receptor positive (log-rank test, P = 0.086) and p53 negative (log-rank test, P = 0.020). The current data suggested that XPC Ala499Val and Lys939Gln polymorphisms may contribute to the identification of patients with increased risk for BC. Moreover, the polymorphisms were associated with the prognosis of BC patients.

The Inflamm-Aging Model Identifies Key Risk Factors in Atherosclerosis.

Background: Atherosclerosis, one of the main threats to human life and health, is driven by abnormal inflammation (i.e., chronic inflammation or oxidative stress) during accelerated aging. Many studies have shown that inflamm-aging exerts a significant impact on the occurrence of atherosclerosis, particularly by inducing an immune homeostasis imbalance. However, the potential mechanism by which inflamm-aging induces atherosclerosis needs to be studied more thoroughly, and there is currently a lack of powerful prediction models. Methods: First, an improved inflamm-aging prediction model was constructed by integrating aging, inflammation, and disease markers with the help of machine learning methods; then, inflamm-aging scores were calculated. In addition, the causal relationship between aging and disease was identified using Mendelian randomization. A series of risk factors were also identified by causal analysis, sensitivity analysis, and network analysis. Results: Our results revealed an accelerated inflamm-aging pattern in atherosclerosis and suggested a causal relationship between inflamm-aging and atherosclerosis. Mechanisms involving inflammation, nutritional balance, vascular homeostasis, and oxidative stress were found to be driving factors of atherosclerosis in the context of inflamm-aging. Conclusion: In summary, we developed a model integrating crucial risk factors in inflamm-aging and atherosclerosis. Our computation pipeline could be used to explore potential mechanisms of related diseases.

Abnormal ECG detection based on an adversarial autoencoder.

Automatic detection and alarm of abnormal electrocardiogram (ECG) events play an important role in an ECG monitor system; however, popular classification models based on supervised learning fail to detect abnormal ECG effectively. Thus, we propose an ECG anomaly detection framework (ECG-AAE) based on an adversarial autoencoder and temporal convolutional network (TCN) which consists of three modules (autoencoder, discriminator, and outlier detector). The ECG-AAE framework is trained only with normal ECG data. Normal ECG signals could be mapped into latent feature space and then reconstructed as the original ECG signal back in our model, while abnormal ECG signals could not. Here, the TCN is employed to extract features of normal ECG data. Then, our model is evaluated on an MIT-BIH arrhythmia dataset and CMUH dataset, with an accuracy, precision, recall, F1-score, and AUC of 0.9673, 0.9854, 0.9486, 0.9666, and 0.9672 and of 0.9358, 0.9816, 0.8882, 0.9325, and 0.9358, respectively. The result indicates that the ECG-AAE can detect abnormal ECG efficiently, with its performance better than other popular outlier detection methods.

[Application of the mixed programming with Labview and Matlab in biomedical signal analysis].

This paper introduces the method of mixed programming with Labview and Matlab, and applies this method in a pulse wave pre-processing and feature detecting system. The method has been proved suitable, efficient and accurate, which has provided a new kind of approach for biomedical signal analysis.

Comparative Analysis of Multiple Neurodegenerative Diseases Based on Advanced Epigenetic Aging Brain.

Background: Neurodegenerative Diseases (NDs) are age-dependent and include Alzheimer's disease (AD), Parkinson's disease (PD), progressive supranuclear palsy (PSP), frontotemporal dementia (FTD), and so on. There have been numerous studies showing that accelerated aging is closely related (even the driver of) ND, thus promoting imbalances in cellular homeostasis. However, the mechanisms of how different ND types are related/triggered by advanced aging are still unclear. Therefore, there is an urgent need to explore the potential markers/mechanisms of different ND types based on aging acceleration at a system level. Methods: AD, PD, PSP, FTD, and aging markers were identified by supervised machine learning methods. The aging acceleration differential networks were constructed based on the aging score. Both the enrichment analysis and sensitivity analysis were carried out to investigate both common and specific mechanisms among different ND types in the context of aging acceleration. Results: The extracellular fluid, cellular metabolisms, and inflammatory response were identified as the common driving factors of cellular homeostasis imbalances during the accelerated aging process. In addition, Ca ion imbalance, abnormal protein depositions, DNA damage, and cytoplasmic DNA in macrophages were also revealed to be special mechanisms that further promote AD, PD, PSP, and FTD, respectively. Conclusion: The accelerated epigenetic aging mechanisms of different ND types were integrated and compared through our computational pipeline.

Automatic consecutive context perceived transformer GAN for serial sectioning image blind inpainting.

BACKGROUND AND OBJECTIVE: Serial sectioning is the routine method in histology study. In order to restore the defective images in section stack, and overcome the limitations of manual annotation of broken areas, we developed a fully automatic approach for locating and restoring the defective image stack. METHODS: We proposed a novel end-to-end framework named automatic consecutive context perceived transformer GAN (ACCP-GAN) for fully automatic serial sectioning image blind inpainting. The first stage network (auto-detection module) was designed to detect the broken areas and repair them roughly, then guided the second stage network (refined inpainting module) to generate these expected patches precisely; therefore, the segmentation part was integrated into restoring part. The transformer module (SPTransformer), based on self-attention mechanism, was introduced to make the refined inpainting module focus on the features from neighboring images to help in correcting inpainting results. Moreover, gated convolution was largely used to extract features from normal parts in the defective image. The framework was trained and validated on the N7 dataset (803 images), and the generalization ability of the model was tested on the E17 (701 images) and N5 (413 images) datasets, all of these images were collected for previous kidney study. RESULTS: N7 dataset was divided into training, validation, and test sets with a ratio of 6:2:2. Our model performed well in broken areas segmentation with the accuracy = 0.9995. The final restoration got the best performance with FSIM = 0.9478, MS-SSIM = 0.9592, PSNR = 29.7903, VIF = 0.8543, and FID = 47.2252 compared to the popular inpainting methods. The model was further tested on E15 and N5 datasets, and the generalization ability was satisfying. CONCLUSIONS: Our method could detect and restore the defective serial sectioning image stack automatically, even the broken patches were large on an individual image. The newly designed SPTransformer performed well in feature extraction. This method reduced the workload of manual annotation and improved the analysis or application of large scale sectioning image stack in histology research.

MRI Based Radiomics Approach With Deep Learning for Prediction of Vessel Invasion in Early-Stage Cervical Cancer.

This article aims to build deep learning-based radiomic methods in differentiating vessel invasion from non-vessel invasion in cervical cancer with multi-parametric MRI data. A set of 1,070 dynamic T1 contrast-enhanced (DCE-T1) and 986 T2 weighted imaging (T2WI) MRI images from 167 early-stage cervical cancer patients (January 2014 - August 2018) were used to train and validate deep learning models. Predictive performances were evaluated using receiver operating characteristic (ROC) curve and confusion matrix analysis, with the DCE-T1 showing more discriminative results than T2WI MRI. By adopting an attention ensemble learning strategy that integrates both MRI sequences, the highest average area was obtained under the ROC curve (AUC) of 0.911 (Sensitivity = 0.881 and Specificity = 0.752). The superior performances in this article, when compared to existing radiomic methods, indicate that a wealth of deep learning-based radiomics could be developed to aid radiologists in preoperatively predicting vessel invasion in cervical cancer patients.