Pub Date : 2025-02-01DOI: 10.1016/j.artmed.2024.103041
Yumeng Yang , Hongfei Lin , Zhihao Yang , Yijia Zhang , Di Zhao , Ling Luo
Medical coding involves assigning codes to clinical free-text documents, specifically medical records that average over 3,000 markers, in order to track patient diagnoses and treatments. This is typically accomplished through manual assignments by healthcare professionals. To improve efficiency and accuracy while reducing the workload on these professionals, researchers have employed a multi-label classification approach. Since the long-tail phenomenon impacts tens of thousands of ICD codes, whereby only a few codes (representative of common diseases) are frequently assigned, while the majority of codes (representative of rare diseases) are infrequently assigned, this paper presents an LCDL model that addresses the challenge at hand by examining the LongFormer pre-trained language model and the disease label co-occurrence map. To enhance the performance of automated medical coding in the biomedical domain, hierarchies with medical knowledge, synonyms and abbreviations are introduced, improving the medical knowledge representation. Test evaluations are extensively conducted on the benchmark dataset MIMIC-III, and obtained the competitive performance compared to the previous state-of-the-art methods.
{"title":"LCDL: Classification of ICD codes based on disease label co-occurrence dependency and LongFormer with medical knowledge","authors":"Yumeng Yang , Hongfei Lin , Zhihao Yang , Yijia Zhang , Di Zhao , Ling Luo","doi":"10.1016/j.artmed.2024.103041","DOIUrl":"10.1016/j.artmed.2024.103041","url":null,"abstract":"<div><div>Medical coding involves assigning codes to clinical free-text documents, specifically medical records that average over 3,000 markers, in order to track patient diagnoses and treatments. This is typically accomplished through manual assignments by healthcare professionals. To improve efficiency and accuracy while reducing the workload on these professionals, researchers have employed a multi-label classification approach. Since the long-tail phenomenon impacts tens of thousands of ICD codes, whereby only a few codes (representative of common diseases) are frequently assigned, while the majority of codes (representative of rare diseases) are infrequently assigned, this paper presents an LCDL model that addresses the challenge at hand by examining the LongFormer pre-trained language model and the disease label co-occurrence map. To enhance the performance of automated medical coding in the biomedical domain, hierarchies with medical knowledge, synonyms and abbreviations are introduced, improving the medical knowledge representation. Test evaluations are extensively conducted on the benchmark dataset MIMIC-III, and obtained the competitive performance compared to the previous state-of-the-art methods.</div></div>","PeriodicalId":55458,"journal":{"name":"Artificial Intelligence in Medicine","volume":"160 ","pages":"Article 103041"},"PeriodicalIF":6.1,"publicationDate":"2025-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142820289","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-02-01DOI: 10.1016/j.artmed.2024.103064
Zeki Kuş , Musa Aydin , Berna Kiraz , Alper Kiraz
Deep neural networks have significantly advanced medical image classification across various modalities and tasks. However, manually designing these networks is often time-consuming and suboptimal. Neural Architecture Search (NAS) automates this process, potentially finding more efficient and effective models. This study provides a comprehensive comparative analysis of our two NAS methods, PBC-NAS and BioNAS, across multiple biomedical image classification tasks using the MedMNIST dataset. Our experiments evaluate these methods based on classification performance (Accuracy (ACC) and Area Under the Curve (AUC)) and computational complexity (Floating Point Operation Counts). Results demonstrate that BioNAS models slightly outperform PBC-NAS models in accuracy, with BioNAS-2 achieving the highest average accuracy of 0.848. However, PBC-NAS models exhibit superior computational efficiency, with PBC-NAS-2 achieving the lowest average FLOPs of 0.82 GB. Both methods outperform state-of-the-art architectures like ResNet-18 and ResNet-50 and AutoML frameworks such as auto-sklearn, AutoKeras, and Google AutoML. Additionally, PBC-NAS and BioNAS outperform other NAS studies in average ACC results (except MSTF-NAS), and show highly competitive results in average AUC. We conduct extensive ablation studies to investigate the impact of architectural parameters, the effectiveness of fine-tuning, search space efficiency, and the discriminative performance of generated architectures. These studies reveal that larger filter sizes and specific numbers of stacks or modules enhance performance. Fine-tuning existing architectures can achieve nearly optimal results without separating NAS for each dataset. Furthermore, we analyze search space efficiency, uncovering patterns in frequently selected operations and architectural choices. This study highlights the strengths and efficiencies of PBC-NAS and BioNAS, providing valuable insights and guidance for future research and practical applications in biomedical image classification.
{"title":"Neural Architecture Search for biomedical image classification: A comparative study across data modalities","authors":"Zeki Kuş , Musa Aydin , Berna Kiraz , Alper Kiraz","doi":"10.1016/j.artmed.2024.103064","DOIUrl":"10.1016/j.artmed.2024.103064","url":null,"abstract":"<div><div>Deep neural networks have significantly advanced medical image classification across various modalities and tasks. However, manually designing these networks is often time-consuming and suboptimal. Neural Architecture Search (NAS) automates this process, potentially finding more efficient and effective models. This study provides a comprehensive comparative analysis of our two NAS methods, PBC-NAS and BioNAS, across multiple biomedical image classification tasks using the MedMNIST dataset. Our experiments evaluate these methods based on classification performance (Accuracy (ACC) and Area Under the Curve (AUC)) and computational complexity (Floating Point Operation Counts). Results demonstrate that BioNAS models slightly outperform PBC-NAS models in accuracy, with BioNAS-2 achieving the highest average accuracy of 0.848. However, PBC-NAS models exhibit superior computational efficiency, with PBC-NAS-2 achieving the lowest average FLOPs of 0.82 GB. Both methods outperform state-of-the-art architectures like ResNet-18 and ResNet-50 and AutoML frameworks such as auto-sklearn, AutoKeras, and Google AutoML. Additionally, PBC-NAS and BioNAS outperform other NAS studies in average ACC results (except MSTF-NAS), and show highly competitive results in average AUC. We conduct extensive ablation studies to investigate the impact of architectural parameters, the effectiveness of fine-tuning, search space efficiency, and the discriminative performance of generated architectures. These studies reveal that larger filter sizes and specific numbers of stacks or modules enhance performance. Fine-tuning existing architectures can achieve nearly optimal results without separating NAS for each dataset. Furthermore, we analyze search space efficiency, uncovering patterns in frequently selected operations and architectural choices. This study highlights the strengths and efficiencies of PBC-NAS and BioNAS, providing valuable insights and guidance for future research and practical applications in biomedical image classification.</div></div>","PeriodicalId":55458,"journal":{"name":"Artificial Intelligence in Medicine","volume":"160 ","pages":"Article 103064"},"PeriodicalIF":6.1,"publicationDate":"2025-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142959368","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-02-01DOI: 10.1016/j.artmed.2024.103050
Oleksandr Kovalyk-Borodyak, Juan Morales-Sánchez, Rafael Verdú-Monedero, José-Luis Sancho-Gómez
In this work, we present a multi-modal machine learning method to automate early glaucoma diagnosis. The proposed methodology introduces two novel aspects for automated diagnosis not previously explored in the literature: simultaneous use of ocular fundus images from both eyes and integration with the patient’s additional clinical data. We begin by establishing a baseline, termed monocular mode, which adheres to the traditional approach of considering the data from each eye as a separate instance. We then explore the binocular mode, investigating how combining information from both eyes of the same patient can enhance glaucoma diagnosis accuracy. This exploration employs the PAPILA dataset, comprising information from both eyes, clinical data, ocular fundus images, and expert segmentation of these images. Additionally, we compare two image-derived data modalities: direct ocular fundus images and morphological data from manual expert segmentation. Our method integrates Gradient-Boosted Decision Trees (GBDT) and Convolutional Neural Networks (CNN), specifically focusing on the MobileNet, VGG16, ResNet-50, and Inception models. SHAP values are used to interpret GBDT models, while the Deep Explainer method is applied in conjunction with SHAP to analyze the outputs of convolutional-based models. Our findings show the viability of considering both eyes, which improves the model performance. The binocular approach, incorporating information from morphological and clinical data yielded an AUC of 0.796 ( at a 95% confidence interval), while the CNN, using the same approach (both eyes), achieved an AUC of 0.764 ( at a 95% confidence interval).
{"title":"Glaucoma detection: Binocular approach and clinical data in machine learning","authors":"Oleksandr Kovalyk-Borodyak, Juan Morales-Sánchez, Rafael Verdú-Monedero, José-Luis Sancho-Gómez","doi":"10.1016/j.artmed.2024.103050","DOIUrl":"10.1016/j.artmed.2024.103050","url":null,"abstract":"<div><div>In this work, we present a multi-modal machine learning method to automate early glaucoma diagnosis. The proposed methodology introduces two novel aspects for automated diagnosis not previously explored in the literature: simultaneous use of ocular fundus images from both eyes and integration with the patient’s additional clinical data. We begin by establishing a baseline, termed <em>monocular mode</em>, which adheres to the traditional approach of considering the data from each eye as a separate instance. We then explore the <em>binocular mode</em>, investigating how combining information from both eyes of the same patient can enhance glaucoma diagnosis accuracy. This exploration employs the PAPILA dataset, comprising information from both eyes, clinical data, ocular fundus images, and expert segmentation of these images. Additionally, we compare two image-derived data modalities: direct ocular fundus images and morphological data from manual expert segmentation. Our method integrates Gradient-Boosted Decision Trees (GBDT) and Convolutional Neural Networks (CNN), specifically focusing on the MobileNet, VGG16, ResNet-50, and Inception models. SHAP values are used to interpret GBDT models, while the Deep Explainer method is applied in conjunction with SHAP to analyze the outputs of convolutional-based models. Our findings show the viability of considering both eyes, which improves the model performance. The binocular approach, incorporating information from morphological and clinical data yielded an AUC of 0.796 (<span><math><mrow><mo>±</mo><mn>0</mn><mo>.</mo><mn>003</mn></mrow></math></span> at a 95% confidence interval), while the CNN, using the same approach (both eyes), achieved an AUC of 0.764 (<span><math><mrow><mo>±</mo><mn>0</mn><mo>.</mo><mn>005</mn></mrow></math></span> at a 95% confidence interval).</div></div>","PeriodicalId":55458,"journal":{"name":"Artificial Intelligence in Medicine","volume":"160 ","pages":"Article 103050"},"PeriodicalIF":6.1,"publicationDate":"2025-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142866491","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-02-01DOI: 10.1016/j.artmed.2024.103030
A. Mencattini , T. Tocci , M. Nuccetelli , M. Pieri , S. Bernardini , E. Martinelli
The Anti-Nuclear Antibodies (ANA) test using Human Epithelial type 2 (HEp-2) cells in the Indirect Immuno-Fluorescence (IIF) assay protocol is considered the gold standard for detecting Connective Tissue Diseases. Computer-assisted systems for HEp-2 image analysis represent a growing field that harnesses the potential offered by novel machine learning techniques to address the classification of HEp-2 images and ANA patterns.
In this study, we introduce an innovative platform based on transfer learning with pre-trained deep learning models. This platform combines the power of unsupervised deep description of HEp-2 images, a novel feature selection approach designed for unbalanced datasets, and independent testing using two distinct datasets from different hospitals to tackle cross-hardware compatibility issues. To enhance the trustworthiness of our method, we also present a modified version of gradient-weighted class activation mapping for regional explainability and introduce a new sample quality index based on the Jensen-Shannon divergence to enhance method reliability and quantify sample heterogeneity.
The results we provide demonstrate exceptionally high performance in intensity and ANA pattern recognition when compared to state-of-the-art approaches. Our method's ability to eliminate the need for cell segmentation in favor of statistical analysis of the sample makes it applicable, robust, and versatile. Our future work will focus on addressing the challenge of mitotic spindle recognition by expanding our proposed approach to cover mixed patterns.
{"title":"Automatic classification of HEp-2 specimens by explainable deep learning and Jensen-Shannon reliability index","authors":"A. Mencattini , T. Tocci , M. Nuccetelli , M. Pieri , S. Bernardini , E. Martinelli","doi":"10.1016/j.artmed.2024.103030","DOIUrl":"10.1016/j.artmed.2024.103030","url":null,"abstract":"<div><div>The Anti-Nuclear Antibodies (ANA) test using Human Epithelial type 2 (HEp-2) cells in the Indirect Immuno-Fluorescence (IIF) assay protocol is considered the gold standard for detecting Connective Tissue Diseases. Computer-assisted systems for HEp-2 image analysis represent a growing field that harnesses the potential offered by novel machine learning techniques to address the classification of HEp-2 images and ANA patterns.</div><div>In this study, we introduce an innovative platform based on transfer learning with pre-trained deep learning models. This platform combines the power of unsupervised deep description of HEp-2 images, a novel feature selection approach designed for unbalanced datasets, and independent testing using two distinct datasets from different hospitals to tackle cross-hardware compatibility issues. To enhance the trustworthiness of our method, we also present a modified version of gradient-weighted class activation mapping for regional explainability and introduce a new sample quality index based on the Jensen-Shannon divergence to enhance method reliability and quantify sample heterogeneity.</div><div>The results we provide demonstrate exceptionally high performance in intensity and ANA pattern recognition when compared to state-of-the-art approaches. Our method's ability to eliminate the need for cell segmentation in favor of statistical analysis of the sample makes it applicable, robust, and versatile. Our future work will focus on addressing the challenge of mitotic spindle recognition by expanding our proposed approach to cover mixed patterns.</div></div>","PeriodicalId":55458,"journal":{"name":"Artificial Intelligence in Medicine","volume":"160 ","pages":"Article 103030"},"PeriodicalIF":6.1,"publicationDate":"2025-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142787959","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-02-01DOI: 10.1016/j.artmed.2024.103056
Zhiyue Zhang , Yao Zhao , Yanxun Xu
In applications such as biomedical studies, epidemiology, and social sciences, recurrent events often co-occur with longitudinal measurements and a terminal event, such as death. Therefore, jointly modeling longitudinal measurements, recurrent events, and survival data while accounting for their dependencies is critical. While joint models for the three components exist in statistical literature, many of these approaches are limited by heavy parametric assumptions and scalability issues. Recently, incorporating deep learning techniques into joint modeling has shown promising results. However, current methods only address joint modeling of longitudinal measurements at regularly-spaced observation times and survival events, neglecting recurrent events. In this paper, we develop TransformerLSR, a flexible transformer-based deep modeling and inference framework to jointly model all three components simultaneously. TransformerLSR integrates deep temporal point processes into the joint modeling framework, treating recurrent and terminal events as two competing processes dependent on past longitudinal measurements and recurrent event times. Additionally, TransformerLSR introduces a novel trajectory representation and model architecture to potentially incorporate a priori knowledge of known latent structures among concurrent longitudinal variables. We demonstrate the effectiveness and necessity of TransformerLSR through simulation studies and analyzing a real-world medical dataset on patients after kidney transplantation.
{"title":"TransformerLSR: Attentive joint model of longitudinal data, survival, and recurrent events with concurrent latent structure","authors":"Zhiyue Zhang , Yao Zhao , Yanxun Xu","doi":"10.1016/j.artmed.2024.103056","DOIUrl":"10.1016/j.artmed.2024.103056","url":null,"abstract":"<div><div>In applications such as biomedical studies, epidemiology, and social sciences, recurrent events often co-occur with longitudinal measurements and a terminal event, such as death. Therefore, jointly modeling longitudinal measurements, recurrent events, and survival data while accounting for their dependencies is critical. While joint models for the three components exist in statistical literature, many of these approaches are limited by heavy parametric assumptions and scalability issues. Recently, incorporating deep learning techniques into joint modeling has shown promising results. However, current methods only address joint modeling of longitudinal measurements at regularly-spaced observation times and survival events, neglecting recurrent events. In this paper, we develop TransformerLSR, a flexible transformer-based deep modeling and inference framework to jointly model all three components simultaneously. TransformerLSR integrates deep temporal point processes into the joint modeling framework, treating recurrent and terminal events as two competing processes dependent on past longitudinal measurements and recurrent event times. Additionally, TransformerLSR introduces a novel trajectory representation and model architecture to potentially incorporate <em>a priori</em> knowledge of known latent structures among concurrent longitudinal variables. We demonstrate the effectiveness and necessity of TransformerLSR through simulation studies and analyzing a real-world medical dataset on patients after kidney transplantation.</div></div>","PeriodicalId":55458,"journal":{"name":"Artificial Intelligence in Medicine","volume":"160 ","pages":"Article 103056"},"PeriodicalIF":6.1,"publicationDate":"2025-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142873489","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Current clinical decision support systems (DSS) are trained and validated on observational data from the clinic in which the DSS is going to be applied. This is problematic for treatments that have already been validated in a randomized clinical trial (RCT), but have not yet been introduced in any clinic. In this work, we report on a method for training and validating the DSS core before introduction to a clinic, using the RCT data themselves. The key challenges we address are of missingness, foremost: missing rationale when assigning a treatment to a patient (the assignment is at random), and missing verification evidence, since the effectiveness of a treatment for a patient can only be verified (ground truth) if the treatment was indeed assigned to the patient — but then the assignment was at random.
Materials:
We use the data of a multi-armed clinical trial that investigated the effectiveness of single treatments and combination treatments for 240+ tinnitus patients recruited and treated in 5 clinical centres.
Methods:
To deal with the ‘missing rationale for treatment assignment’ challenge, we re-model the target variable that measures the outcome of interest, in order to suppress the effect of the individual treatment, which was at random, and control on the effect of treatment in general. To deal with missing features for many patients, we use a learning core that is robust to missing features. Further, we build ensembles that parsimoniously exploit the small patient numbers we have for learning. To deal with the ‘missing verification evidence’ challenge, we introduce counterfactual treatment verification, a verification scheme that juxtaposes the effectiveness of the recommendations of our approach to the effectiveness of the RCT assignments in the cases of agreement/disagreement between the two.
Results and limitations:
We demonstrate that our approach leverages the RCT data for learning and verification, by showing that the DSS suggests treatments that improve the outcome. The results are limited through the small number of patients per treatment; while our ensemble is designed to mitigate this effect, the predictive performance of the methods is affected by the smallness of the data.
Outlook:
We provide a basis for the establishment of decision supporting routines on treatments that have been tested in RCTs but have not yet been deployed clinically. Practitioners can use our approach to train and validate a DSS on new treatments by simply using the RCT data available to them. More work is needed to strengthen the robustness of the predictors. Since there are no further data available to this purpose, but those already used, the potential of synthetic data generation seems an appropriate alternative.
{"title":"Training and validating a treatment recommender with partial verification evidence","authors":"Vishnu Unnikrishnan , Clara Puga , Miro Schleicher , Uli Niemann , Berthold Langguth , Stefan Schoisswohl , Birgit Mazurek , Rilana Cima , Jose Antonio Lopez-Escamez , Dimitris Kikidis , Eleftheria Vellidou , Rüdiger Pryss , Winfried Schlee , Myra Spiliopoulou","doi":"10.1016/j.artmed.2024.103062","DOIUrl":"10.1016/j.artmed.2024.103062","url":null,"abstract":"<div><h3>Background:</h3><div>Current clinical decision support systems (DSS) are trained and validated on observational data from the clinic in which the DSS is going to be applied. This is problematic for treatments that have already been validated in a randomized clinical trial (RCT), but have not yet been introduced in any clinic. In this work, we report on a method for training and validating the DSS core before introduction to a clinic, using the RCT data themselves. The key challenges we address are of missingness, foremost: missing rationale when assigning a treatment to a patient (the assignment is at random), and missing verification evidence, since the effectiveness of a treatment for a patient can only be verified (ground truth) if the treatment was indeed assigned to the patient — but then the assignment was at random.</div></div><div><h3>Materials:</h3><div>We use the data of a multi-armed clinical trial that investigated the effectiveness of single treatments and combination treatments for 240+ tinnitus patients recruited and treated in 5 clinical centres.</div></div><div><h3>Methods:</h3><div>To deal with the ‘missing rationale for treatment assignment’ challenge, we re-model the target variable that measures the outcome of interest, in order to suppress the effect of the individual treatment, which was at random, and control on the effect of treatment in general. To deal with missing features for many patients, we use a learning core that is robust to missing features. Further, we build ensembles that parsimoniously exploit the small patient numbers we have for learning. To deal with the ‘missing verification evidence’ challenge, we introduce <em>counterfactual treatment verification</em>, a verification scheme that juxtaposes the effectiveness of the recommendations of our approach to the effectiveness of the RCT assignments in the cases of agreement/disagreement between the two.</div></div><div><h3>Results and limitations:</h3><div>We demonstrate that our approach leverages the RCT data for learning and verification, by showing that the DSS suggests treatments that improve the outcome. The results are limited through the small number of patients per treatment; while our ensemble is designed to mitigate this effect, the predictive performance of the methods is affected by the smallness of the data.</div></div><div><h3>Outlook:</h3><div>We provide a basis for the establishment of decision supporting routines on treatments that have been tested in RCTs but have not yet been deployed clinically. Practitioners can use our approach to train and validate a DSS on new treatments by simply using the RCT data available to them. More work is needed to strengthen the robustness of the predictors. Since there are no further data available to this purpose, but those already used, the potential of synthetic data generation seems an appropriate alternative.</div></div>","PeriodicalId":55458,"journal":{"name":"Artificial Intelligence in Medicine","volume":"160 ","pages":"Article 103062"},"PeriodicalIF":6.1,"publicationDate":"2025-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142959321","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-02-01DOI: 10.1016/j.artmed.2024.103065
Yiqiu Qi , Guangyuan Li , Jinzhu Yang , Honghe Li , Qi Yu , Mingjun Qu , Hongxia Ning , Yonghuai Wang
Left ventricular systolic dysfunction (LVSD) and its severity are correlated with the prognosis of cardiovascular diseases. Early detection and monitoring of LVSD are of utmost importance. Left ventricular ejection fraction (LVEF) is an essential indicator for evaluating left ventricular function in clinical practice, the current echocardiography-based evaluation method is not avaliable in primary care and difficult to achieve real-time monitoring capabilities for cardiac dysfunction. We propose a two-branch deep learning model (ECGEFNet) for calculating LVEF using electrocardiogram (ECG), which holds the potential to serve as a primary medical screening tool and facilitate long-term dynamic monitoring of cardiac functional impairments. It integrates original numerical signal and waveform plots derived from the signals in an innovative manner, enabling joint calculation of LVEF by incorporating diverse information encompassing temporal, spatial and phase aspects. To address the inadequate information interaction between the two branches and the lack of efficiency in feature fusion, we propose the fusion attention mechanism (FAT) and the two-branch feature fusion module (BFF) to guide the learning, alignment and fusion of features from both branches. We assemble a large internal dataset and perform experimental validation on it. The accuracy of cardiac dysfunction screening is 92.3%, the mean absolute error (MAE) in LVEF calculation is 4.57%. The proposed model performs well and outperforms existing basic models, and is of great significance for real-time monitoring of the degree of cardiac dysfunction.
{"title":"ECGEFNet: A two-branch deep learning model for calculating left ventricular ejection fraction using electrocardiogram","authors":"Yiqiu Qi , Guangyuan Li , Jinzhu Yang , Honghe Li , Qi Yu , Mingjun Qu , Hongxia Ning , Yonghuai Wang","doi":"10.1016/j.artmed.2024.103065","DOIUrl":"10.1016/j.artmed.2024.103065","url":null,"abstract":"<div><div>Left ventricular systolic dysfunction (LVSD) and its severity are correlated with the prognosis of cardiovascular diseases. Early detection and monitoring of LVSD are of utmost importance. Left ventricular ejection fraction (LVEF) is an essential indicator for evaluating left ventricular function in clinical practice, the current echocardiography-based evaluation method is not avaliable in primary care and difficult to achieve real-time monitoring capabilities for cardiac dysfunction. We propose a two-branch deep learning model (ECGEFNet) for calculating LVEF using electrocardiogram (ECG), which holds the potential to serve as a primary medical screening tool and facilitate long-term dynamic monitoring of cardiac functional impairments. It integrates original numerical signal and waveform plots derived from the signals in an innovative manner, enabling joint calculation of LVEF by incorporating diverse information encompassing temporal, spatial and phase aspects. To address the inadequate information interaction between the two branches and the lack of efficiency in feature fusion, we propose the fusion attention mechanism (FAT) and the two-branch feature fusion module (BFF) to guide the learning, alignment and fusion of features from both branches. We assemble a large internal dataset and perform experimental validation on it. The accuracy of cardiac dysfunction screening is 92.3%, the mean absolute error (MAE) in LVEF calculation is 4.57%. The proposed model performs well and outperforms existing basic models, and is of great significance for real-time monitoring of the degree of cardiac dysfunction.</div></div>","PeriodicalId":55458,"journal":{"name":"Artificial Intelligence in Medicine","volume":"160 ","pages":"Article 103065"},"PeriodicalIF":6.1,"publicationDate":"2025-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142985312","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-01-31DOI: 10.1016/j.artmed.2025.103074
Mostafa Abdelrahim , Mohamed Khudri , Ahmed Elnakib , Mohamed Shehata , Kate Weafer , Ashraf Khalil , Gehad A. Saleh , Nihal M. Batouty , Mohammed Ghazal , Sohail Contractor , Gregory Barnes , Ayman El-Baz
Autism Spectrum Disorder (ASD) is a neurological condition, with recent statistics from the CDC indicating a rising prevalence of ASD diagnoses among infants and children. This trend emphasizes the critical importance of early detection, as timely diagnosis facilitates early intervention and enhances treatment outcomes. Consequently, there is an increasing urgency for research to develop innovative tools capable of accurately and objectively identifying ASD in its earliest stages. This paper offers a short overview of recent advancements in non-invasive technology for early ASD diagnosis, focusing on an imaging modality, structural MRI technique, which has shown promising results in early ASD diagnosis. This brief review aims to address several key questions: (i) Which imaging radiomics are associated with ASD? (ii) Is the parcellation step of the brain cortex necessary to improve the diagnostic accuracy of ASD? (iii) What databases are available to researchers interested in developing non-invasive technology for ASD? (iv) How can artificial intelligence tools contribute to improving the diagnostic accuracy of ASD? Finally, our review will highlight future trends in ASD diagnostic efforts.
{"title":"AI-based non-invasive imaging technologies for early autism spectrum disorder diagnosis: A short review and future directions","authors":"Mostafa Abdelrahim , Mohamed Khudri , Ahmed Elnakib , Mohamed Shehata , Kate Weafer , Ashraf Khalil , Gehad A. Saleh , Nihal M. Batouty , Mohammed Ghazal , Sohail Contractor , Gregory Barnes , Ayman El-Baz","doi":"10.1016/j.artmed.2025.103074","DOIUrl":"10.1016/j.artmed.2025.103074","url":null,"abstract":"<div><div>Autism Spectrum Disorder (ASD) is a neurological condition, with recent statistics from the CDC indicating a rising prevalence of ASD diagnoses among infants and children. This trend emphasizes the critical importance of early detection, as timely diagnosis facilitates early intervention and enhances treatment outcomes. Consequently, there is an increasing urgency for research to develop innovative tools capable of accurately and objectively identifying ASD in its earliest stages. This paper offers a short overview of recent advancements in non-invasive technology for early ASD diagnosis, focusing on an imaging modality, structural MRI technique, which has shown promising results in early ASD diagnosis. This brief review aims to address several key questions: (i) Which imaging radiomics are associated with ASD? (ii) Is the parcellation step of the brain cortex necessary to improve the diagnostic accuracy of ASD? (iii) What databases are available to researchers interested in developing non-invasive technology for ASD? (iv) How can artificial intelligence tools contribute to improving the diagnostic accuracy of ASD? Finally, our review will highlight future trends in ASD diagnostic efforts.</div></div>","PeriodicalId":55458,"journal":{"name":"Artificial Intelligence in Medicine","volume":"161 ","pages":"Article 103074"},"PeriodicalIF":6.1,"publicationDate":"2025-01-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143358294","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-01-22DOI: 10.1016/j.artmed.2025.103067
Jing-Jie Peng , Yi-Yue Zhang , Rui-Feng Li , Wen-Jun Zhu , Hong-Rui Liu , Hui-Yin Li , Bin Liu , Dong-Sheng Cao , Jun Peng , Xiu-Ju Luo
Multiple cell death mechanisms are triggered during ischemic stroke and they are interconnected in a complex network with extensive crosstalk, complicating the development of targeted therapies. We therefore propose a novel framework for identifying disease-specific drug-target interaction (DTI), named strokeDTI, to extract key nodes within an interconnected graph network of activated pathways via leveraging transcriptomic sequencing data. Our findings reveal that the drugs a model can predict are highly representative of the characteristics of the database the model is trained on. However, models with comparable performance yield diametrically opposite predictions in real testing scenarios. Our analysis reveals a correlation between the reported literature on drug-target pairs and their binding scores. Leveraging this correlation, we introduced an additional module to assess the predictive validity of our model for each unique target, thereby improving the reliability of the framework's predictions. Our framework identified Cerdulatinib as a potential anti-stroke drug via targeting multiple cell death pathways, particularly necroptosis and apoptosis. Experimental validation in in vitro and in vivo models demonstrated that Cerdulatinib significantly attenuated stroke-induced brain injury via inhibiting multiple cell death pathways, improving neurological function, and reducing infarct volume. This highlights strokeDTI's potential for disease-specific drug-target identification and Cerdulatinib's potential as a potent anti-stroke drug.
{"title":"Hybrid approach for drug-target interaction predictions in ischemic stroke models","authors":"Jing-Jie Peng , Yi-Yue Zhang , Rui-Feng Li , Wen-Jun Zhu , Hong-Rui Liu , Hui-Yin Li , Bin Liu , Dong-Sheng Cao , Jun Peng , Xiu-Ju Luo","doi":"10.1016/j.artmed.2025.103067","DOIUrl":"10.1016/j.artmed.2025.103067","url":null,"abstract":"<div><div>Multiple cell death mechanisms are triggered during ischemic stroke and they are interconnected in a complex network with extensive crosstalk, complicating the development of targeted therapies. We therefore propose a novel framework for identifying disease-specific drug-target interaction (DTI), named strokeDTI, to extract key nodes within an interconnected graph network of activated pathways via leveraging transcriptomic sequencing data. Our findings reveal that the drugs a model can predict are highly representative of the characteristics of the database the model is trained on. However, models with comparable performance yield diametrically opposite predictions in real testing scenarios. Our analysis reveals a correlation between the reported literature on drug-target pairs and their binding scores. Leveraging this correlation, we introduced an additional module to assess the predictive validity of our model for each unique target, thereby improving the reliability of the framework's predictions. Our framework identified Cerdulatinib as a potential anti-stroke drug via targeting multiple cell death pathways, particularly necroptosis and apoptosis. Experimental validation in in vitro and in vivo models demonstrated that Cerdulatinib significantly attenuated stroke-induced brain injury via inhibiting multiple cell death pathways, improving neurological function, and reducing infarct volume. This highlights strokeDTI's potential for disease-specific drug-target identification and Cerdulatinib's potential as a potent anti-stroke drug.</div></div>","PeriodicalId":55458,"journal":{"name":"Artificial Intelligence in Medicine","volume":"161 ","pages":"Article 103067"},"PeriodicalIF":6.1,"publicationDate":"2025-01-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143168061","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-01-18DOI: 10.1016/j.artmed.2025.103066
Marwa Saady , Mahmoud Eissa , Ahmed S. Yacoub , Ahmed B. Hamed , Hassan Mohamed El-Said Azzazy
Introduction
There is a growing interest in leveraging artificial intelligence (AI) technologies to enhance various aspects of clinical trials. The goal of this systematic review is to assess the impact of implementing AI approaches on different aspects of oncology clinical trials.
Methods
Pertinent keywords were used to find relevant articles published in PubMed, Scopus, and Google Scholar databases, which described the clinical application of AI approaches. A quality evaluation utilizing a customized checklist specifically adapted was conducted. This study is registered with PROSPERO (CRD42024537153).
Results
Out of the identified 2833 studies, 72 studies satisfied the inclusion criteria. Clinical Trial Enrollment & Eligibility were among the most commonly studied clinical trial aspects with 30 papers. The prediction of outcomes was covered in 25 studies of which 15 addressed the prediction of patients' survival and 10 addressed the prediction of drug outcomes. The trial design was studied in 10 articles. Three studies addressed each of the personalized treatments and decision-making, while one addressed data management. The results demonstrate using AI in cancer clinical trials has the potential to increase clinical trial enrollment, predict clinical outcomes, improve trial design, enhance personalized treatments, and increase concordance in decision-making. Additionally, automating some areas and tasks, clinical trials were made more efficient, and human error was minimized. Nevertheless, concerns and restrictions related to the application of AI in clinical studies are also noted.
Conclusion
AI tools have the potential to revolutionize the design, enrollment rate, and outcome prediction of oncology clinical trials.
{"title":"Implementation of artificial intelligence approaches in oncology clinical trials: A systematic review","authors":"Marwa Saady , Mahmoud Eissa , Ahmed S. Yacoub , Ahmed B. Hamed , Hassan Mohamed El-Said Azzazy","doi":"10.1016/j.artmed.2025.103066","DOIUrl":"10.1016/j.artmed.2025.103066","url":null,"abstract":"<div><h3>Introduction</h3><div>There is a growing interest in leveraging artificial intelligence (AI) technologies to enhance various aspects of clinical trials. The goal of this systematic review is to assess the impact of implementing AI approaches on different aspects of oncology clinical trials.</div></div><div><h3>Methods</h3><div>Pertinent keywords were used to find relevant articles published in PubMed, Scopus, and Google Scholar databases, which described the clinical application of AI approaches. A quality evaluation utilizing a customized checklist specifically adapted was conducted. This study is registered with PROSPERO (CRD42024537153).</div></div><div><h3>Results</h3><div>Out of the identified 2833 studies, 72 studies satisfied the inclusion criteria. Clinical Trial Enrollment & Eligibility were among the most commonly studied clinical trial aspects with 30 papers. The prediction of outcomes was covered in 25 studies of which 15 addressed the prediction of patients' survival and 10 addressed the prediction of drug outcomes. The trial design was studied in 10 articles. Three studies addressed each of the personalized treatments and decision-making, while one addressed data management. The results demonstrate using AI in cancer clinical trials has the potential to increase clinical trial enrollment, predict clinical outcomes, improve trial design, enhance personalized treatments, and increase concordance in decision-making. Additionally, automating some areas and tasks, clinical trials were made more efficient, and human error was minimized. Nevertheless, concerns and restrictions related to the application of AI in clinical studies are also noted.</div></div><div><h3>Conclusion</h3><div>AI tools have the potential to revolutionize the design, enrollment rate, and outcome prediction of oncology clinical trials.</div></div>","PeriodicalId":55458,"journal":{"name":"Artificial Intelligence in Medicine","volume":"161 ","pages":"Article 103066"},"PeriodicalIF":6.1,"publicationDate":"2025-01-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143016910","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
扫码关注我们
求助内容：
应助结果提醒方式：
京公网安备 11010802042870号