Tachykardia komorowa
Rokowania, prognozy i postęp choroby

Tachykardia komorowa (VT) stanowi poważne zagrożenie życia, będąc przyczyną około 80% przypadków nagłej śmierci sercowej (SCD), która rocznie powoduje ponad 300 000 zgonów w USA. Wczesne przewidywanie epizodów VT jest kluczowe dla zmniejszenia śmiertelności, jednak tradycyjne metody stratyfikacji ryzyka oparte na statycznych pomiarach nie uwzględniają dynamicznych interakcji między substratem arytmogennym a czynnikami wyzwalającymi. Modele predykcyjne wykorzystujące 24-godzinne monitorowanie elektrycznej czynności serca, takie jak mobilna telemetria kardiologiczna (MCT), wykazują wysoką wartość predykcyjną – w najwyższym kwintylu ryzyka około 20% pacjentów doświadcza epizodu VT trwającego ≥10 pobudzeń w ciągu 29 dni, a w najniższym kwintylu wartość predykcyjna ujemna wynosi 98%.

Tachykardia komorowa – Prognoza (przewidywanie przebiegu)

Tachykardia komorowa (VT) jest potencjalnie zagrażającą życiu tachyarytmią, charakteryzującą się szybką pracą serca spowodowaną nieprawidłową aktywnością elektryczną komór serca. Jest to niebezpieczne zaburzenie rytmu, ponieważ może prowadzić do niedociśnienia, migotania komór, asystolii i nagłej śmierci sercowej. 1 Nagła śmierć sercowa (SCD) powoduje ponad 300 000 zgonów rocznie w Stanach Zjednoczonych, a spontaniczne tachyarytmie komorowe (VTA) stanowią około 80% wszystkich przypadków nagłej śmierci sercowej. 2

Modele przewidywania tachykardii komorowej

Wczesne przewidywanie epizodów tachykardii komorowej jest kluczowe dla zmniejszenia śmiertelności poprzez umożliwienie wdrożenia odpowiedniej profilaktyki. Obecne metody stratyfikacji ryzyka arytmii komorowych opierają się głównie na statycznych pomiarach, które nie uwzględniają w wystarczającym stopniu dynamicznych interakcji między substratem arytmogennym a czynnikami wyzwalającymi występujących w czasie. 3

Badania naukowe wskazują, że modele predykcyjne wykorzystujące dane z 24-godzinnego monitorowania czynności elektrycznej serca mogą skutecznie przewidywać ryzyko wystąpienia epizodów VT. W modelu opartym na 24-godzinnych danych z mobilnej telemetrii kardiologicznej (MCT), w najwyższym kwintylu przewidywanego ryzyka, u około 1 na 5 pacjentów wystąpił epizod VT trwający ≥10 pobudzeń w ciągu kolejnych 29 dni monitorowania. W najniższym kwintylu wartość predykcyjna ujemna wynosiła 98%. 4 5

Zaawansowane metody predykcji

Współczesne badania koncentrują się na wykorzystaniu zaawansowanych technik uczenia maszynowego (ML) i sztucznych sieci neuronowych (ANN) do poprawy dokładności przewidywania epizodów tachykardii komorowej:

  • Dynamiczne modele ML i sieci neuronowe skutecznie wykorzystują rutynowo zbierane długoterminowe zapisy EKG do spersonalizowanych i aktualizowanych przewidywań złośliwych arytmii komorowych 6
  • Średnia zmienna w czasie powierzchnia pod krzywą ROC dla modelu dynamicznego wynosi 0,738±0,07, w porównaniu do 0,639±0,03 dla modelu statycznego (opartego wyłącznie na danych wyjściowych) 7
  • Duże ilości spersonalizowanych danych elektrofizjologicznych zbieranych w czasie są skutecznie wykorzystywane przez modele nadzorowane i nienadzorowane do ułatwienia dynamicznych przewidywań złośliwych arytmii komorowych 8

Wczesne przewidywanie VT

Szczególnie obiecujący jest model predykcyjny umożliwiający przewidywanie epizodu tachykardii komorowej na godzinę przed jego wystąpieniem. Model ten wykorzystuje sztuczną sieć neuronową (ANN) generowaną przy użyciu 14 parametrów uzyskanych z analizy zmienności rytmu serca (HRV) i zmienności częstości oddechów (RRV). 9

Opracowany model przewidywania VT wykazał wysoką skuteczność osiągając czułość 0,88, swoistość 0,82 i pole pod krzywą ROC (AUC) wynoszące 0,93. 10 Wyniki badań pokazują, że wykorzystanie zarówno sygnałów EKG, jak i oddechowych zwiększa skuteczność wykrywania VT na godzinę przed jej wystąpieniem. 11

Znaczenie kliniczne zaawansowanych modeli predykcyjnych

Możliwość przewidywania epizodu tachykardii komorowej z jednogodzinnym wyprzedzeniem ma istotne znaczenie kliniczne:

  • Zapewnia wystarczająco dużo czasu na podjęcie szybkich działań medycznych, które mogą zapobiec poważnym konsekwencjom VT 12
  • Taki model może być przydatny zarówno dla pacjentów przebywających w domu, jak i w warunkach szpitalnych 13
  • Pozwala na lepsze triażowanie pacjentów pod względem potrzeby przedłużonego monitorowania 14

Korzyści z modeli dynamicznych w porównaniu do statycznych

Lepsza wydajność predykcyjna modeli dynamicznych w porównaniu z modelami statycznymi, oraz fakt, że te przewidywania są głównie napędzane przez zmienne w czasie ukryte reprezentacje EKG, wskazuje na potencjalne korzyści z integracji zmienności czasowej zmiennych w modelach predykcyjnych. 15

Wykorzystując dynamiczne modele ML i autoenkodery wariacyjne, informacje prognostyczne z rutynowo zbieranych zapisów EKG mogą być wyodrębniane i wykorzystywane do dostarczania spersonalizowanych przewidywań złośliwych arytmii komorowych u pacjentów z implantowanym kardiowerterem-defibrylatorem (ICD). 16

Ograniczenia i perspektywy

Głównym ograniczeniem niektórych badań nad modelami predykcyjnymi dla tachykardii komorowej jest niewielka liczba danych (np. 104 nagrania w jednym z badań), co ogranicza moc statystyczną analizy. 17 Pomimo tych ograniczeń, wyniki są obiecujące i mogą przyczynić się do opracowania lepszych modeli predykcyjnych dla tachykardii komorowych. 18

Aby opracować lepszy model predykcyjny tachykardii komorowej, konieczne jest wykorzystanie wielu parametrów z różnych metod analizy zmienności rytmu serca (HRV) oraz stworzenie klasyfikatora, który może radzić sobie ze złożonymi wzorcami składającymi się z takich parametrów. 19

Kolejne rozdziały

Zapraszamy do dalszego czytania naszego leksykonu.

Wybierz kolejny rozdział z menu poniżej, aby otworzyć nową podstronę kompedium wiedzy i uzyskać szczegółowe informację o leku, substancji lub chorobie.

  1. 16.04.2026
  2. www.leksykon.com.pl

Materiały źródłowe

  • #1 Prediction of Ventricular Tachycardia One Hour before Occurrence Using Artificial Neural Networks | Scientific ReportsClose bannerClose banner
    https://www.nature.com/articles/srep32390
    Prediction of Ventricular Tachycardia One Hour before Occurrence Using Artificial Neural Networks […] Ventricular tachycardia (VT) is a potentially fatal tachyarrhythmia, which causes a rapid heartbeat as a result of improper electrical activity of the heart. This is a potentially life-threatening arrhythmia because it can cause low blood pressure and may lead to ventricular fibrillation, asystole, and sudden cardiac death. To prevent VT, we developed an early prediction model that can predict this event one hour before its onset using an artificial neural network (ANN) generated using 14 parameters obtained from heart rate variability (HRV) and respiratory rate variability (RRV) analysis. […] The developed VT prediction model proved its performance by achieving a sensitivity of 0.88, specificity of 0.82, and AUC of 0.93.
  • #2 Prediction of Ventricular Tachycardia One Hour before Occurrence Using Artificial Neural Networks | Scientific ReportsClose bannerClose banner
    https://www.nature.com/articles/srep32390
    Sudden cardiac death (SCD) causes more than 300,000 deaths annually in the United States. […] In addition, spontaneous ventricular tachyarrhythmia (VTA) is a main cause of SCD, contributing to about 80% of SCDs. […] Accordingly, early prediction of VT will help in reducing mortality from SCD by allowing for preventive care of VTA. […] To make a better prediction model of VT, it is essential to utilize multiple parameters from various methods of HRV analysis and to generate a classifier that can deal with complex patterns composed of such parameters. […] Our group previously reported an ANN-based prediction model for VTAs by utilizing parameters obtained from HRV analysis. […] However, predicting VT 10 seconds before it occurs is not sufficiently valuable in clinical practice. […] To predict VT events earlier than in our previous work, gathering physiological signals including ECG from patients for a longer time period prior to VT events is essential.
  • #3 Dynamic prediction of malignant ventricular arrhythmias using neural networks in patients with an implantable cardioverter-defibrillator – PMC Lock
    https://pmc.ncbi.nlm.nih.gov/articles/PMC10772563/
    Risk stratification for ventricular arrhythmias currently relies on static measurements that fail to adequately capture dynamic interactions between arrhythmic substrate and triggers over time. […] Dynamic ML models and neural networks effectively leverage routinely collected longitudinal ECG recordings for personalised and updated predictions of malignant ventricular arrhythmias, outperforming static models. […] The mean time-varying area under the ROC curve for the dynamic model was 0.738±0.07, compared to 0.639±0.03 for a static (i.e. baseline-only model). […] High volumes of personalised electrophysiological data collected over time are effectively leveraged by supervised and unsupervised models to facilitate dynamic predictions of malignant ventricular arrhythmias. […] The improved predictive performance of the dynamic model over static models, and these predictions being predominantly driven by time-varying latent ECG representations, points towards the potential benefits of integrating temporal variability of covariates within prediction models. […] Utilising dynamic ML models and variational autoencoders, prognostic information from routinely collected ECGs can be extracted and leveraged to provide personalised predictions of malignant ventricular arrhythmias in patients with an ICD.
  • #4 Ventricular tachycardia risk prediction with an abbreviated duration mobile cardiac telemetry – PMC Lock
    https://pmc.ncbi.nlm.nih.gov/articles/PMC10461200/
    Ventricular tachycardia (VT) occurs intermittently, unpredictably, and has potentially lethal consequences. […] Our aim was to derive a risk prediction model for VT episodes ≥10 beats detected on 30-day mobile cardiac telemetry based on the first 24 hours of the recording. […] Our model can predict risk of VT ≥10 beats in the near term using variables derived from 24-hour electrocardiography, and could be used to triage patients to extended monitoring. […] Using only data easily derived from 24 hours of MCT, we derived a risk prediction model for NSVT episodes ≥10 beats on MCT monitoring in the next 29 days. In the top quintile of the predicted risk score roughly 1 in 5 patients had an episode of VT ≥10 beats during the subsequent monitoring. In the bottom quintile, the negative predictive value was 98% for both models.
  • #5 Ventricular tachycardia risk prediction with an abbreviated duration mobile cardiac telemetry – PMC Lock
    https://pmc.ncbi.nlm.nih.gov/articles/PMC10461200/
    A risk score based on variables from 24 hours of MCT can predict a high risk of VT ≥10 beats within 30 days and can be used to triage patients according to their need for extended monitoring. In the top quintile of the risk score, a VT event with a duration of ≥10 beats was detected in 1 in 5 patients.
  • #6 Dynamic prediction of malignant ventricular arrhythmias using neural networks in patients with an implantable cardioverter-defibrillator – PMC Lock
    https://pmc.ncbi.nlm.nih.gov/articles/PMC10772563/
    Risk stratification for ventricular arrhythmias currently relies on static measurements that fail to adequately capture dynamic interactions between arrhythmic substrate and triggers over time. […] Dynamic ML models and neural networks effectively leverage routinely collected longitudinal ECG recordings for personalised and updated predictions of malignant ventricular arrhythmias, outperforming static models. […] The mean time-varying area under the ROC curve for the dynamic model was 0.738±0.07, compared to 0.639±0.03 for a static (i.e. baseline-only model). […] High volumes of personalised electrophysiological data collected over time are effectively leveraged by supervised and unsupervised models to facilitate dynamic predictions of malignant ventricular arrhythmias. […] The improved predictive performance of the dynamic model over static models, and these predictions being predominantly driven by time-varying latent ECG representations, points towards the potential benefits of integrating temporal variability of covariates within prediction models. […] Utilising dynamic ML models and variational autoencoders, prognostic information from routinely collected ECGs can be extracted and leveraged to provide personalised predictions of malignant ventricular arrhythmias in patients with an ICD.
  • #7 Dynamic prediction of malignant ventricular arrhythmias using neural networks in patients with an implantable cardioverter-defibrillator – PMC Lock
    https://pmc.ncbi.nlm.nih.gov/articles/PMC10772563/
    Risk stratification for ventricular arrhythmias currently relies on static measurements that fail to adequately capture dynamic interactions between arrhythmic substrate and triggers over time. […] Dynamic ML models and neural networks effectively leverage routinely collected longitudinal ECG recordings for personalised and updated predictions of malignant ventricular arrhythmias, outperforming static models. […] The mean time-varying area under the ROC curve for the dynamic model was 0.738±0.07, compared to 0.639±0.03 for a static (i.e. baseline-only model). […] High volumes of personalised electrophysiological data collected over time are effectively leveraged by supervised and unsupervised models to facilitate dynamic predictions of malignant ventricular arrhythmias. […] The improved predictive performance of the dynamic model over static models, and these predictions being predominantly driven by time-varying latent ECG representations, points towards the potential benefits of integrating temporal variability of covariates within prediction models. […] Utilising dynamic ML models and variational autoencoders, prognostic information from routinely collected ECGs can be extracted and leveraged to provide personalised predictions of malignant ventricular arrhythmias in patients with an ICD.
  • #8 Dynamic prediction of malignant ventricular arrhythmias using neural networks in patients with an implantable cardioverter-defibrillator – PMC Lock
    https://pmc.ncbi.nlm.nih.gov/articles/PMC10772563/
    Risk stratification for ventricular arrhythmias currently relies on static measurements that fail to adequately capture dynamic interactions between arrhythmic substrate and triggers over time. […] Dynamic ML models and neural networks effectively leverage routinely collected longitudinal ECG recordings for personalised and updated predictions of malignant ventricular arrhythmias, outperforming static models. […] The mean time-varying area under the ROC curve for the dynamic model was 0.738±0.07, compared to 0.639±0.03 for a static (i.e. baseline-only model). […] High volumes of personalised electrophysiological data collected over time are effectively leveraged by supervised and unsupervised models to facilitate dynamic predictions of malignant ventricular arrhythmias. […] The improved predictive performance of the dynamic model over static models, and these predictions being predominantly driven by time-varying latent ECG representations, points towards the potential benefits of integrating temporal variability of covariates within prediction models. […] Utilising dynamic ML models and variational autoencoders, prognostic information from routinely collected ECGs can be extracted and leveraged to provide personalised predictions of malignant ventricular arrhythmias in patients with an ICD.
  • #9 Prediction of Ventricular Tachycardia One Hour before Occurrence Using Artificial Neural Networks | Scientific ReportsClose bannerClose banner
    https://www.nature.com/articles/srep32390
    Prediction of Ventricular Tachycardia One Hour before Occurrence Using Artificial Neural Networks […] Ventricular tachycardia (VT) is a potentially fatal tachyarrhythmia, which causes a rapid heartbeat as a result of improper electrical activity of the heart. This is a potentially life-threatening arrhythmia because it can cause low blood pressure and may lead to ventricular fibrillation, asystole, and sudden cardiac death. To prevent VT, we developed an early prediction model that can predict this event one hour before its onset using an artificial neural network (ANN) generated using 14 parameters obtained from heart rate variability (HRV) and respiratory rate variability (RRV) analysis. […] The developed VT prediction model proved its performance by achieving a sensitivity of 0.88, specificity of 0.82, and AUC of 0.93.
  • #10 Prediction of Ventricular Tachycardia One Hour before Occurrence Using Artificial Neural Networks | Scientific ReportsClose bannerClose banner
    https://www.nature.com/articles/srep32390
    Prediction of Ventricular Tachycardia One Hour before Occurrence Using Artificial Neural Networks […] Ventricular tachycardia (VT) is a potentially fatal tachyarrhythmia, which causes a rapid heartbeat as a result of improper electrical activity of the heart. This is a potentially life-threatening arrhythmia because it can cause low blood pressure and may lead to ventricular fibrillation, asystole, and sudden cardiac death. To prevent VT, we developed an early prediction model that can predict this event one hour before its onset using an artificial neural network (ANN) generated using 14 parameters obtained from heart rate variability (HRV) and respiratory rate variability (RRV) analysis. […] The developed VT prediction model proved its performance by achieving a sensitivity of 0.88, specificity of 0.82, and AUC of 0.93.
  • #11 Prediction of Ventricular Tachycardia One Hour before Occurrence Using Artificial Neural Networks | Scientific ReportsClose bannerClose banner
    https://www.nature.com/articles/srep32390
    In our current study, we propose a model to predict VT events 1 hour in advance by utilizing ANN and parameters from HRV and RRV analysis. Such a model may help to reduce mortality from VT events. […] The results demonstrate that utilizing both ECG and respiration signals increases the performance of detecting VT one hour before its occurrence. […] Therefore, including RRV parameters may be important in predicting a VT event. […] In addition, the model we have implemented shows great value in predicting VT one hour before its occurrence. […] Therefore, prompt medical action is necessary. One hour may be enough time to visit the hospital if there is no one around to help; as such, this model may be suitable for the general population at home as well as those in hospital. […] The major limitation of this study is the small number of data (total: 104 recordings), which limits the statistical power of our analysis. […] We believe that this could help in developing better predictor of VTs.
  • #12 Prediction of Ventricular Tachycardia One Hour before Occurrence Using Artificial Neural Networks | Scientific ReportsClose bannerClose banner
    https://www.nature.com/articles/srep32390
    In our current study, we propose a model to predict VT events 1 hour in advance by utilizing ANN and parameters from HRV and RRV analysis. Such a model may help to reduce mortality from VT events. […] The results demonstrate that utilizing both ECG and respiration signals increases the performance of detecting VT one hour before its occurrence. […] Therefore, including RRV parameters may be important in predicting a VT event. […] In addition, the model we have implemented shows great value in predicting VT one hour before its occurrence. […] Therefore, prompt medical action is necessary. One hour may be enough time to visit the hospital if there is no one around to help; as such, this model may be suitable for the general population at home as well as those in hospital. […] The major limitation of this study is the small number of data (total: 104 recordings), which limits the statistical power of our analysis. […] We believe that this could help in developing better predictor of VTs.
  • #13 Prediction of Ventricular Tachycardia One Hour before Occurrence Using Artificial Neural Networks | Scientific ReportsClose bannerClose banner
    https://www.nature.com/articles/srep32390
    In our current study, we propose a model to predict VT events 1 hour in advance by utilizing ANN and parameters from HRV and RRV analysis. Such a model may help to reduce mortality from VT events. […] The results demonstrate that utilizing both ECG and respiration signals increases the performance of detecting VT one hour before its occurrence. […] Therefore, including RRV parameters may be important in predicting a VT event. […] In addition, the model we have implemented shows great value in predicting VT one hour before its occurrence. […] Therefore, prompt medical action is necessary. One hour may be enough time to visit the hospital if there is no one around to help; as such, this model may be suitable for the general population at home as well as those in hospital. […] The major limitation of this study is the small number of data (total: 104 recordings), which limits the statistical power of our analysis. […] We believe that this could help in developing better predictor of VTs.
  • #14 Ventricular tachycardia risk prediction with an abbreviated duration mobile cardiac telemetry – PMC Lock
    https://pmc.ncbi.nlm.nih.gov/articles/PMC10461200/
    Ventricular tachycardia (VT) occurs intermittently, unpredictably, and has potentially lethal consequences. […] Our aim was to derive a risk prediction model for VT episodes ≥10 beats detected on 30-day mobile cardiac telemetry based on the first 24 hours of the recording. […] Our model can predict risk of VT ≥10 beats in the near term using variables derived from 24-hour electrocardiography, and could be used to triage patients to extended monitoring. […] Using only data easily derived from 24 hours of MCT, we derived a risk prediction model for NSVT episodes ≥10 beats on MCT monitoring in the next 29 days. In the top quintile of the predicted risk score roughly 1 in 5 patients had an episode of VT ≥10 beats during the subsequent monitoring. In the bottom quintile, the negative predictive value was 98% for both models.
  • #15 Dynamic prediction of malignant ventricular arrhythmias using neural networks in patients with an implantable cardioverter-defibrillator – PMC Lock
    https://pmc.ncbi.nlm.nih.gov/articles/PMC10772563/
    Risk stratification for ventricular arrhythmias currently relies on static measurements that fail to adequately capture dynamic interactions between arrhythmic substrate and triggers over time. […] Dynamic ML models and neural networks effectively leverage routinely collected longitudinal ECG recordings for personalised and updated predictions of malignant ventricular arrhythmias, outperforming static models. […] The mean time-varying area under the ROC curve for the dynamic model was 0.738±0.07, compared to 0.639±0.03 for a static (i.e. baseline-only model). […] High volumes of personalised electrophysiological data collected over time are effectively leveraged by supervised and unsupervised models to facilitate dynamic predictions of malignant ventricular arrhythmias. […] The improved predictive performance of the dynamic model over static models, and these predictions being predominantly driven by time-varying latent ECG representations, points towards the potential benefits of integrating temporal variability of covariates within prediction models. […] Utilising dynamic ML models and variational autoencoders, prognostic information from routinely collected ECGs can be extracted and leveraged to provide personalised predictions of malignant ventricular arrhythmias in patients with an ICD.
  • #16 Dynamic prediction of malignant ventricular arrhythmias using neural networks in patients with an implantable cardioverter-defibrillator – PMC Lock
    https://pmc.ncbi.nlm.nih.gov/articles/PMC10772563/
    Risk stratification for ventricular arrhythmias currently relies on static measurements that fail to adequately capture dynamic interactions between arrhythmic substrate and triggers over time. […] Dynamic ML models and neural networks effectively leverage routinely collected longitudinal ECG recordings for personalised and updated predictions of malignant ventricular arrhythmias, outperforming static models. […] The mean time-varying area under the ROC curve for the dynamic model was 0.738±0.07, compared to 0.639±0.03 for a static (i.e. baseline-only model). […] High volumes of personalised electrophysiological data collected over time are effectively leveraged by supervised and unsupervised models to facilitate dynamic predictions of malignant ventricular arrhythmias. […] The improved predictive performance of the dynamic model over static models, and these predictions being predominantly driven by time-varying latent ECG representations, points towards the potential benefits of integrating temporal variability of covariates within prediction models. […] Utilising dynamic ML models and variational autoencoders, prognostic information from routinely collected ECGs can be extracted and leveraged to provide personalised predictions of malignant ventricular arrhythmias in patients with an ICD.
  • #17 Prediction of Ventricular Tachycardia One Hour before Occurrence Using Artificial Neural Networks | Scientific ReportsClose bannerClose banner
    https://www.nature.com/articles/srep32390
    In our current study, we propose a model to predict VT events 1 hour in advance by utilizing ANN and parameters from HRV and RRV analysis. Such a model may help to reduce mortality from VT events. […] The results demonstrate that utilizing both ECG and respiration signals increases the performance of detecting VT one hour before its occurrence. […] Therefore, including RRV parameters may be important in predicting a VT event. […] In addition, the model we have implemented shows great value in predicting VT one hour before its occurrence. […] Therefore, prompt medical action is necessary. One hour may be enough time to visit the hospital if there is no one around to help; as such, this model may be suitable for the general population at home as well as those in hospital. […] The major limitation of this study is the small number of data (total: 104 recordings), which limits the statistical power of our analysis. […] We believe that this could help in developing better predictor of VTs.
  • #18 Prediction of Ventricular Tachycardia One Hour before Occurrence Using Artificial Neural Networks | Scientific ReportsClose bannerClose banner
    https://www.nature.com/articles/srep32390
    In our current study, we propose a model to predict VT events 1 hour in advance by utilizing ANN and parameters from HRV and RRV analysis. Such a model may help to reduce mortality from VT events. […] The results demonstrate that utilizing both ECG and respiration signals increases the performance of detecting VT one hour before its occurrence. […] Therefore, including RRV parameters may be important in predicting a VT event. […] In addition, the model we have implemented shows great value in predicting VT one hour before its occurrence. […] Therefore, prompt medical action is necessary. One hour may be enough time to visit the hospital if there is no one around to help; as such, this model may be suitable for the general population at home as well as those in hospital. […] The major limitation of this study is the small number of data (total: 104 recordings), which limits the statistical power of our analysis. […] We believe that this could help in developing better predictor of VTs.
  • #19 Prediction of Ventricular Tachycardia One Hour before Occurrence Using Artificial Neural Networks | Scientific ReportsClose bannerClose banner
    https://www.nature.com/articles/srep32390
    Sudden cardiac death (SCD) causes more than 300,000 deaths annually in the United States. […] In addition, spontaneous ventricular tachyarrhythmia (VTA) is a main cause of SCD, contributing to about 80% of SCDs. […] Accordingly, early prediction of VT will help in reducing mortality from SCD by allowing for preventive care of VTA. […] To make a better prediction model of VT, it is essential to utilize multiple parameters from various methods of HRV analysis and to generate a classifier that can deal with complex patterns composed of such parameters. […] Our group previously reported an ANN-based prediction model for VTAs by utilizing parameters obtained from HRV analysis. […] However, predicting VT 10 seconds before it occurs is not sufficiently valuable in clinical practice. […] To predict VT events earlier than in our previous work, gathering physiological signals including ECG from patients for a longer time period prior to VT events is essential.