Objective The trial was designed to evaluate the safety and performance of the ev3 Protégé TM stent in the treatment of de novo or re-stenotic common and/or internal carotid artery stenoses with adjunctive u...Objective The trial was designed to evaluate the safety and performance of the ev3 Protégé TM stent in the treatment of de novo or re-stenotic common and/or internal carotid artery stenoses with adjunctive use of a CE-marked filter embolic protection device.Methods This study was a prospective multi-center, single-arm trial. Between June and October 2003, 77 patients were enrolled in 8 investigational centers throughout Europe. The primary endpoint was the incidence of Major Neurological Events (MANE) through one month. Other endpoints were the ability to properly place the stent, and primary patency and MANE after six months. Eligible for the study were patients with a de novo or restenotic target lesion located in the common and/or internal carotid artery (>70% stenosis for asymptomatic and >50% stenosis for symptomatic patients). The ev3 Spider (Embolic Protection Filter was used in 75 of 77 cases. Results In 76 out of the 77 patients (99%), the stent could be successfully implanted with a residual stenosis ≤30% as criterion. Of the 74 patients that had a carotid ultrasound at one month follow-up, none had a re-stenosis of the target lesion. There were three MANEs during or immediately after the procedure (3.9%), two were major and one was a minor stroke. There were eight severe complications (9.1%); six of these happened during or immediately after the procedure and were related to the procedure, none was related to the device. They are resolved without sequelae. No deaths have occurred.Conclusions The Protégé stent is safe and performs well in the treatment of carotid artery stenosis. The technical success rate for placement of the Protégé stent as assessed by the residual stenosis post implant was very high and all stents were successfully deployed. The incidence of MANE was comparable with that in other recent carotid stent studies and still lower than standard CEA.展开更多
Background Medical informatics accumulated vast amounts of data for clinical diagnosis and treatment.However,limited access to follow-up data and the difficulty in integrating data across diverse platforms continue to...Background Medical informatics accumulated vast amounts of data for clinical diagnosis and treatment.However,limited access to follow-up data and the difficulty in integrating data across diverse platforms continue to pose significant barriers to clinical research progress.In response,our research team has embarked on the development of a specialized clinical research database for cardiology,thereby establishing a comprehensive digital platform that facilitates both clinical decision-making and research endeavors.Methods The database incorporated actual clinical data from patients who received treatment at the Cardiovascular Medicine Department of Chinese PLA General Hospital from 2012 to 2021.It included comprehensive data on patients'basic information,medical history,non-invasive imaging studies,laboratory test results,as well as peri-procedural information related to interventional surgeries,extracted from the Hospital Information System.Additionally,an innovative artificial intelligence(AI)-powered interactive follow-up system had been developed,ensuring that nearly all myocardial infarction patients received at least one post-discharge follow-up,thereby achieving comprehensive data management throughout the entire care continuum for highrisk patients.Results This database integrates extensive cross-sectional and longitudinal patient data,with a focus on higher-risk acute coronary syndrome patients.It achieves the integration of structured and unstructured clinical data,while innovatively incorporating AI and automatic speech recognition technologies to enhance data integration and workflow efficiency.It creates a comprehensive patient view,thereby improving diagnostic and follow-up quality,and provides high-quality data to support clinical research.Despite limitations in unstructured data standardization and biological sample integrity,the database's development is accompanied by ongoing optimization efforts.Conclusion The cardiovascular specialty clinical database is a comprehensive digital archive integrating clinical treatment and research,which facilitates the digital and intelligent transformation of clinical diagnosis and treatment processes.It supports clinical decision-making and offers data support and potential research directions for the specialized management of cardiovascular diseases.展开更多
Background The efficacy of percutaneous transluminal angioplasty and stenting(PTAS)relative to medical management in treating symptomatic intracranial arterial stenosis(ICAS)varies based on the qualifying artery.This ...Background The efficacy of percutaneous transluminal angioplasty and stenting(PTAS)relative to medical management in treating symptomatic intracranial arterial stenosis(ICAS)varies based on the qualifying artery.This study aims to evaluate PTAS compared with medical therapy alone in cases of ICAS involving the internal carotid artery(ICA),middle cerebral artery(MCA),vertebral artery(VA)and basilar artery(BA).Methods This study involves a thorough pooled analysis of individual patient data from two randomised controlled trials,evaluating the efficacy of PTAS in comparison to medical management for symptomatic ICAS with different qualifying arteries.The primary outcome was stroke or death within 30 days postenrolment,or stroke in the region of the qualifying artery beyond 30 days through 1 year.A methodology based on intention-to-treat was employed,and HR accompanied by 95%CIs were used to convey risk estimates.Results The data of 809 individuals were collected from Stenting vs Aggressive Medical Management for Preventing Recurrent Stroke in Intracranial Stenosis trial and China Angioplasty and Stenting for Symptomatic Intracranial Severe Stenosis trial.Four hundred were designated for PTAS,while 409 were assigned to medical therapy alone.For the primary outcome,patients with symptomatic BA stenosis had a significantly higher risk of receiving PTAS compared with medical therapy(17.17%vs 7.77%;9.40;HR,2.38(1.03 to 5.52);p=0.04).However,PTAS had no significant difference in patients with symptomatic ICA(26.67%vs 16.67%;HR,1.68(0.78 to 3.62);p=0.19),MCA(8.28%vs 9.79%;HR,0.85(0.42 to 1.74);p=0.66)and VA stenosis(9.52%vs 10.71%;HR,0.91(0.32 to 2.62);p=0.86)compared with medical therapy.Conclusions PTAS significantly increases the risk of both short-term and long-term stroke in patients with symptomatic BA stenosis.Without significant technological advancements to mitigate these risks,PTAS offers limited benefits.For symptomatic ICA,MCA and VA stenosis,PTAS provided no significant advantage.展开更多
BACKGROUND Lutetium has been shown to be an important potential innovation in pre-treated metastatic castration-resistant prostate cancer.Two clinical trials have evaluated lutetium thus far(therap and vision with 99 ...BACKGROUND Lutetium has been shown to be an important potential innovation in pre-treated metastatic castration-resistant prostate cancer.Two clinical trials have evaluated lutetium thus far(therap and vision with 99 and 385 patients,respectively),but their results are discordant.AIM To synthetize the available evidence on the effectiveness of lutetium in pre-treated metastatic castration-resistant prostate cancer;and to test the application of a new artificial intelligence technique that synthetizes effectiveness based on reconstructed patient-level data.METHODS We employed a new artificial intelligence method(shiny method)to pool the survival data of these two trials and evaluate to what extent the lutetium cohorts differed from one another.The shiny technique employs an original reconstruction of individual patient data from the Kaplan-Meier curves.The progression-free survival graphs of the two lutetium cohorts were analyzed and compared.RESULTS The hazard ratio estimated was in favor of the vision trial;the difference was statistically significant(P<0.001).These results indicate that further studies on lutetium are needed because the survival data of the two trials published thus far are conflicting.CONCLUSION Our study confirms the feasibility of reconstructing patient-level data from survival graphs in order to generate a survival statistics.展开更多
Purpose: Computed tomography is a leading imaging technique for head & neck and brain and most of these imaging protocols iodine-based contrast media are utilised. The chief aim of this research is to utilize the ...Purpose: Computed tomography is a leading imaging technique for head & neck and brain and most of these imaging protocols iodine-based contrast media are utilised. The chief aim of this research is to utilize the effects of the contrast media “CM” used in computed tomography “CT” which is used to enhance subject contrast on the delivered CT via its inclusion into the CT dose index “CTDI”, and to introduce a simple method to determine this effect via the available CT numbers at the imaged targets. Method: The CT dose increase is estimated theoretically and measured experimentally and then related to the average CT number in the volume of CM uptake. A factor dependent on CM concentration and beam energy is added to the CTDI equation to represent the increased dose burden. A simple holed Perspex phantom was built to measure the variation of imaged CT number. CT Gafchromic type film was alternately imaged in a reservoir of CM and water. The relative difference in the dose burden as obtained by scanning the two films represents the dose difference and hence the CM dependent increase. Results: Measured dose effects due to the inclusion of the CM varied depending on the concentration. The increase in dose is estimated to be about 17% for 20% contrast media in the target while that for 10% by volume is around 6.6%. These are estimated from the CT numbers. Patients’ data also shows influence of the CM on the CTDI values. Conclusion: The dosimetric effects of the contrast media are included into the CTDI and can be estimated by using the CT numbers obtained.展开更多
Background:Electronic Health Record(EHR)systems are used as an efficient and effective technique for sharing patient’s health records among different hospitals and various other key stakeholders of the healthcare ind...Background:Electronic Health Record(EHR)systems are used as an efficient and effective technique for sharing patient’s health records among different hospitals and various other key stakeholders of the healthcare industry to achieve better diagnosis and treatment of patients globally.However,the existing EHR systems mostly lack in providing appropriate security,entrusted access control and handling privacy and secrecy issues and challenges in current hospital infrastructures.Objective:To solve this delicate problem,we propose a Blockchain-enabled Hyperledger Fabric Architecture for different EHR systems.Methodology:In our EHR blockchain system,Peer nodes from various organizations(stakeholders)create a ledger network,where channels are created to enable secure and private communication between different stakeholders on the ledger network.Individual patients and other stakeholders are identified and registered on the network by unique digital certificates issued by membership service provider(MSP)component of the fabric architecture.Results:We created and implemented different Chaincodes to handle the business logic for executing separate EHR transactions on the network.The proposed fabric architecture provides a secure,transparent and immutable mechanism to store,share and exchange EHRs in a peer-to-peer network of different healthcare stakeholders.It ensures interoperability,scalability and availability in adapting the existing EHRs for strengthening and providing an effective and secure method to integrate and manage patient records among medical institutions in the healthcare ecosystem.展开更多
Publication biases and collection limitations are the main disadvantages of a traditional meta-analysis based on aggregate patient data（APD）from published articles.Individual patient data（IPD）meta-analysis,as the ...Publication biases and collection limitations are the main disadvantages of a traditional meta-analysis based on aggregate patient data（APD）from published articles.Individual patient data（IPD）meta-analysis,as the gold standard of systematic review,is a possible alternative in this context.However,the publications relative to IPD meta-analyses are still rare compared with the traditional ones,especially in the research oriented to Chinese medicine（CM）.In this article,the strengths and detailed functioning of IPD meta-analysis are described.Furthermore,the need for IPD meta-analysis to assess the treatments based on CM was also discussed.Compared with the traditional APD meta-analysis,the IPD meta-analysis might give a more accurate and unbiased assessment and is worth to be recommended to CM researchers.展开更多
Purpose-Patient treatment trajectory data are used to predict the outcome of the treatment to particular disease that has been carried out in the research.In order to determine the evolving disease on the patient and ...Purpose-Patient treatment trajectory data are used to predict the outcome of the treatment to particular disease that has been carried out in the research.In order to determine the evolving disease on the patient and changes in the health due to treatment has not considered existing methodologies.Hence deep learning models to trajectory data mining can be employed to identify disease prediction with high accuracy and less computation cost.Design/methodology/approach-Multifocus deep neural network classifiers has been utilized to detect the novel disease class and comorbidity class to the changes in the genome pattern of the patient trajectory data can be identified on the layers of the architecture.Classifier is employed to learn extracted feature set with activation and weight function and then merged on many aspects to classify the undetermined sequence of diseases as a new variant.The performance of disease progression learning progress utilizes the precision of the constituent classifiers,which usually has larger generalization benefits than those optimized classifiers.Findings-Deep learning architecture uses weight function,bias function on input layers and max pooling.Outcome of the input layer has applied to hidden layer to generate the multifocus characteristics of the disease,and multifocus characterized disease is processed in activation function using ReLu function along hyper parameter tuning which produces the effective outcome in the output layer of a fully connected network.Experimental results have proved using cross validation that proposed model outperforms methodologies in terms of computation time and accuracy.Originality/value-Proposed evolving classifier represented as a robust architecture on using objective function to map the data sequence into a class distribution of the evolving disease class to the patient trajectory.Then,the generative output layer of the proposed model produces the progression outcome of the disease of the particular patient trajectory.The model tries to produce the accurate prognosis outcomes by employing data conditional probability function.The originality of the work defines 70%and comparisons of the previous methods the method of values are accurate and increased analysis of the predictions.展开更多
Background:The electrocardiogram(ECG)is a valuable,noninvasive tool for monitoring heart-related conditions,providing critical insights.However,the interpretation of ECG data alongside patient information demands subs...Background:The electrocardiogram(ECG)is a valuable,noninvasive tool for monitoring heart-related conditions,providing critical insights.However,the interpretation of ECG data alongside patient information demands substantial medical expertise and resources.While deep learning methods help streamline this process,they often fall short in integrating patient data with ECG readings and do not provide the nuanced clinical suggestions and insights necessary for accurate diagnosis.Methods:Although recent advancements in multi-modal large language modeling have propelled their application scope beyond the natural language processing domain,their applicability to ECG processing remains largely unexplored,partly due to the lack of text–ECG data.To this end,we develop ECG-Language Model(ECG-LM),the first multi-modal large language model able to process natural language and understand ECG signals.The model employs a specialized ECG encoder that transforms raw ECG signals into a high-dimensional feature space,which is then aligned with the textual feature space derived from the large language model.To address the scarcity of text–ECG data,we generated text–ECG pairs by leveraging detailed ECG pattern descriptions from medical guidelines,creating a robust dataset for pre-training ECG-LM.Additionally,we fine-tune ECG-LM with public clinical conversation datasets and build an additional supervised fine-tuning dataset based on real clinical data from the hospital,aiming to provide a more comprehensive and customized user experience.Results:ECG-LM outperforms existing few-shot and zero-shot solutions in cardiovascular disease detection across all 3 tasks(diagnostic,rhythm,and form)while also demonstrating strong potential in ECG-related question answering.Conclusions:The results across various tasks demonstrate that ECG-LM effectively captures the intricate features of ECGs,showcasing its versatility in applications such as disease prediction and advanced question answering.展开更多
文摘Objective The trial was designed to evaluate the safety and performance of the ev3 Protégé TM stent in the treatment of de novo or re-stenotic common and/or internal carotid artery stenoses with adjunctive use of a CE-marked filter embolic protection device.Methods This study was a prospective multi-center, single-arm trial. Between June and October 2003, 77 patients were enrolled in 8 investigational centers throughout Europe. The primary endpoint was the incidence of Major Neurological Events (MANE) through one month. Other endpoints were the ability to properly place the stent, and primary patency and MANE after six months. Eligible for the study were patients with a de novo or restenotic target lesion located in the common and/or internal carotid artery (>70% stenosis for asymptomatic and >50% stenosis for symptomatic patients). The ev3 Spider (Embolic Protection Filter was used in 75 of 77 cases. Results In 76 out of the 77 patients (99%), the stent could be successfully implanted with a residual stenosis ≤30% as criterion. Of the 74 patients that had a carotid ultrasound at one month follow-up, none had a re-stenosis of the target lesion. There were three MANEs during or immediately after the procedure (3.9%), two were major and one was a minor stroke. There were eight severe complications (9.1%); six of these happened during or immediately after the procedure and were related to the procedure, none was related to the device. They are resolved without sequelae. No deaths have occurred.Conclusions The Protégé stent is safe and performs well in the treatment of carotid artery stenosis. The technical success rate for placement of the Protégé stent as assessed by the residual stenosis post implant was very high and all stents were successfully deployed. The incidence of MANE was comparable with that in other recent carotid stent studies and still lower than standard CEA.
基金Noncommunicable Chronic Diseases-National Science and Technology Major Project(2023ZD0503906)。
文摘Background Medical informatics accumulated vast amounts of data for clinical diagnosis and treatment.However,limited access to follow-up data and the difficulty in integrating data across diverse platforms continue to pose significant barriers to clinical research progress.In response,our research team has embarked on the development of a specialized clinical research database for cardiology,thereby establishing a comprehensive digital platform that facilitates both clinical decision-making and research endeavors.Methods The database incorporated actual clinical data from patients who received treatment at the Cardiovascular Medicine Department of Chinese PLA General Hospital from 2012 to 2021.It included comprehensive data on patients'basic information,medical history,non-invasive imaging studies,laboratory test results,as well as peri-procedural information related to interventional surgeries,extracted from the Hospital Information System.Additionally,an innovative artificial intelligence(AI)-powered interactive follow-up system had been developed,ensuring that nearly all myocardial infarction patients received at least one post-discharge follow-up,thereby achieving comprehensive data management throughout the entire care continuum for highrisk patients.Results This database integrates extensive cross-sectional and longitudinal patient data,with a focus on higher-risk acute coronary syndrome patients.It achieves the integration of structured and unstructured clinical data,while innovatively incorporating AI and automatic speech recognition technologies to enhance data integration and workflow efficiency.It creates a comprehensive patient view,thereby improving diagnostic and follow-up quality,and provides high-quality data to support clinical research.Despite limitations in unstructured data standardization and biological sample integrity,the database's development is accompanied by ongoing optimization efforts.Conclusion The cardiovascular specialty clinical database is a comprehensive digital archive integrating clinical treatment and research,which facilitates the digital and intelligent transformation of clinical diagnosis and treatment processes.It supports clinical decision-making and offers data support and potential research directions for the specialized management of cardiovascular diseases.
基金Beijing Hospitals Authority’s Ascent Plan(DFL20220702)National Natural Science Foundation of China(82101398).
文摘Background The efficacy of percutaneous transluminal angioplasty and stenting(PTAS)relative to medical management in treating symptomatic intracranial arterial stenosis(ICAS)varies based on the qualifying artery.This study aims to evaluate PTAS compared with medical therapy alone in cases of ICAS involving the internal carotid artery(ICA),middle cerebral artery(MCA),vertebral artery(VA)and basilar artery(BA).Methods This study involves a thorough pooled analysis of individual patient data from two randomised controlled trials,evaluating the efficacy of PTAS in comparison to medical management for symptomatic ICAS with different qualifying arteries.The primary outcome was stroke or death within 30 days postenrolment,or stroke in the region of the qualifying artery beyond 30 days through 1 year.A methodology based on intention-to-treat was employed,and HR accompanied by 95%CIs were used to convey risk estimates.Results The data of 809 individuals were collected from Stenting vs Aggressive Medical Management for Preventing Recurrent Stroke in Intracranial Stenosis trial and China Angioplasty and Stenting for Symptomatic Intracranial Severe Stenosis trial.Four hundred were designated for PTAS,while 409 were assigned to medical therapy alone.For the primary outcome,patients with symptomatic BA stenosis had a significantly higher risk of receiving PTAS compared with medical therapy(17.17%vs 7.77%;9.40;HR,2.38(1.03 to 5.52);p=0.04).However,PTAS had no significant difference in patients with symptomatic ICA(26.67%vs 16.67%;HR,1.68(0.78 to 3.62);p=0.19),MCA(8.28%vs 9.79%;HR,0.85(0.42 to 1.74);p=0.66)and VA stenosis(9.52%vs 10.71%;HR,0.91(0.32 to 2.62);p=0.86)compared with medical therapy.Conclusions PTAS significantly increases the risk of both short-term and long-term stroke in patients with symptomatic BA stenosis.Without significant technological advancements to mitigate these risks,PTAS offers limited benefits.For symptomatic ICA,MCA and VA stenosis,PTAS provided no significant advantage.
文摘BACKGROUND Lutetium has been shown to be an important potential innovation in pre-treated metastatic castration-resistant prostate cancer.Two clinical trials have evaluated lutetium thus far(therap and vision with 99 and 385 patients,respectively),but their results are discordant.AIM To synthetize the available evidence on the effectiveness of lutetium in pre-treated metastatic castration-resistant prostate cancer;and to test the application of a new artificial intelligence technique that synthetizes effectiveness based on reconstructed patient-level data.METHODS We employed a new artificial intelligence method(shiny method)to pool the survival data of these two trials and evaluate to what extent the lutetium cohorts differed from one another.The shiny technique employs an original reconstruction of individual patient data from the Kaplan-Meier curves.The progression-free survival graphs of the two lutetium cohorts were analyzed and compared.RESULTS The hazard ratio estimated was in favor of the vision trial;the difference was statistically significant(P<0.001).These results indicate that further studies on lutetium are needed because the survival data of the two trials published thus far are conflicting.CONCLUSION Our study confirms the feasibility of reconstructing patient-level data from survival graphs in order to generate a survival statistics.
文摘Purpose: Computed tomography is a leading imaging technique for head & neck and brain and most of these imaging protocols iodine-based contrast media are utilised. The chief aim of this research is to utilize the effects of the contrast media “CM” used in computed tomography “CT” which is used to enhance subject contrast on the delivered CT via its inclusion into the CT dose index “CTDI”, and to introduce a simple method to determine this effect via the available CT numbers at the imaged targets. Method: The CT dose increase is estimated theoretically and measured experimentally and then related to the average CT number in the volume of CM uptake. A factor dependent on CM concentration and beam energy is added to the CTDI equation to represent the increased dose burden. A simple holed Perspex phantom was built to measure the variation of imaged CT number. CT Gafchromic type film was alternately imaged in a reservoir of CM and water. The relative difference in the dose burden as obtained by scanning the two films represents the dose difference and hence the CM dependent increase. Results: Measured dose effects due to the inclusion of the CM varied depending on the concentration. The increase in dose is estimated to be about 17% for 20% contrast media in the target while that for 10% by volume is around 6.6%. These are estimated from the CT numbers. Patients’ data also shows influence of the CM on the CTDI values. Conclusion: The dosimetric effects of the contrast media are included into the CTDI and can be estimated by using the CT numbers obtained.
基金funded by the Deanship of Scientific Research at Princess Nourah bint Abdulrahman University through the Fast-track Research Funding Program.
文摘Background:Electronic Health Record(EHR)systems are used as an efficient and effective technique for sharing patient’s health records among different hospitals and various other key stakeholders of the healthcare industry to achieve better diagnosis and treatment of patients globally.However,the existing EHR systems mostly lack in providing appropriate security,entrusted access control and handling privacy and secrecy issues and challenges in current hospital infrastructures.Objective:To solve this delicate problem,we propose a Blockchain-enabled Hyperledger Fabric Architecture for different EHR systems.Methodology:In our EHR blockchain system,Peer nodes from various organizations(stakeholders)create a ledger network,where channels are created to enable secure and private communication between different stakeholders on the ledger network.Individual patients and other stakeholders are identified and registered on the network by unique digital certificates issued by membership service provider(MSP)component of the fabric architecture.Results:We created and implemented different Chaincodes to handle the business logic for executing separate EHR transactions on the network.The proposed fabric architecture provides a secure,transparent and immutable mechanism to store,share and exchange EHRs in a peer-to-peer network of different healthcare stakeholders.It ensures interoperability,scalability and availability in adapting the existing EHRs for strengthening and providing an effective and secure method to integrate and manage patient records among medical institutions in the healthcare ecosystem.
基金Supported by the Fundamental Research Funds for the CentralPublic Welfare Research Institutes of China(No.ZZ070818 andZ0259)National Natural Science Foundation of China(No.81072920 and 81303149)
文摘Publication biases and collection limitations are the main disadvantages of a traditional meta-analysis based on aggregate patient data（APD）from published articles.Individual patient data（IPD）meta-analysis,as the gold standard of systematic review,is a possible alternative in this context.However,the publications relative to IPD meta-analyses are still rare compared with the traditional ones,especially in the research oriented to Chinese medicine（CM）.In this article,the strengths and detailed functioning of IPD meta-analysis are described.Furthermore,the need for IPD meta-analysis to assess the treatments based on CM was also discussed.Compared with the traditional APD meta-analysis,the IPD meta-analysis might give a more accurate and unbiased assessment and is worth to be recommended to CM researchers.
文摘Purpose-Patient treatment trajectory data are used to predict the outcome of the treatment to particular disease that has been carried out in the research.In order to determine the evolving disease on the patient and changes in the health due to treatment has not considered existing methodologies.Hence deep learning models to trajectory data mining can be employed to identify disease prediction with high accuracy and less computation cost.Design/methodology/approach-Multifocus deep neural network classifiers has been utilized to detect the novel disease class and comorbidity class to the changes in the genome pattern of the patient trajectory data can be identified on the layers of the architecture.Classifier is employed to learn extracted feature set with activation and weight function and then merged on many aspects to classify the undetermined sequence of diseases as a new variant.The performance of disease progression learning progress utilizes the precision of the constituent classifiers,which usually has larger generalization benefits than those optimized classifiers.Findings-Deep learning architecture uses weight function,bias function on input layers and max pooling.Outcome of the input layer has applied to hidden layer to generate the multifocus characteristics of the disease,and multifocus characterized disease is processed in activation function using ReLu function along hyper parameter tuning which produces the effective outcome in the output layer of a fully connected network.Experimental results have proved using cross validation that proposed model outperforms methodologies in terms of computation time and accuracy.Originality/value-Proposed evolving classifier represented as a robust architecture on using objective function to map the data sequence into a class distribution of the evolving disease class to the patient trajectory.Then,the generative output layer of the proposed model produces the progression outcome of the disease of the particular patient trajectory.The model tries to produce the accurate prognosis outcomes by employing data conditional probability function.The originality of the work defines 70%and comparisons of the previous methods the method of values are accurate and increased analysis of the predictions.
基金sponsored by Tsinghua-Toyota Joint Research Institute Inter-disciplinary Program.
文摘Background:The electrocardiogram(ECG)is a valuable,noninvasive tool for monitoring heart-related conditions,providing critical insights.However,the interpretation of ECG data alongside patient information demands substantial medical expertise and resources.While deep learning methods help streamline this process,they often fall short in integrating patient data with ECG readings and do not provide the nuanced clinical suggestions and insights necessary for accurate diagnosis.Methods:Although recent advancements in multi-modal large language modeling have propelled their application scope beyond the natural language processing domain,their applicability to ECG processing remains largely unexplored,partly due to the lack of text–ECG data.To this end,we develop ECG-Language Model(ECG-LM),the first multi-modal large language model able to process natural language and understand ECG signals.The model employs a specialized ECG encoder that transforms raw ECG signals into a high-dimensional feature space,which is then aligned with the textual feature space derived from the large language model.To address the scarcity of text–ECG data,we generated text–ECG pairs by leveraging detailed ECG pattern descriptions from medical guidelines,creating a robust dataset for pre-training ECG-LM.Additionally,we fine-tune ECG-LM with public clinical conversation datasets and build an additional supervised fine-tuning dataset based on real clinical data from the hospital,aiming to provide a more comprehensive and customized user experience.Results:ECG-LM outperforms existing few-shot and zero-shot solutions in cardiovascular disease detection across all 3 tasks(diagnostic,rhythm,and form)while also demonstrating strong potential in ECG-related question answering.Conclusions:The results across various tasks demonstrate that ECG-LM effectively captures the intricate features of ECGs,showcasing its versatility in applications such as disease prediction and advanced question answering.
