Dataset Information

Recurrent disease progression networks for modelling risk trajectory of heart failure.

ABSTRACT:

Motivation

Recurrent neural networks (RNN) are powerful frameworks to model medical time series records. Recent studies showed improved accuracy of predicting future medical events (e.g., readmission, mortality) by leveraging large amount of high-dimensional data. However, very few studies have explored the ability of RNN in predicting long-term trajectories of recurrent events, which is more informative than predicting one single event in directing medical intervention.

Methods

In this study, we focus on heart failure (HF) which is the leading cause of death among cardiovascular diseases. We present a novel RNN framework named Deep Heart-failure Trajectory Model (DHTM) for modelling the long-term trajectories of recurrent HF. DHTM auto-regressively predicts the future HF onsets of each patient and uses the predicted HF as input to predict the HF event at the next time point. Furthermore, we propose an augmented DHTM named DHTM+C (where "C" stands for co-morbidities), which jointly predicts both the HF and a set of acute co-morbidities diagnoses. To efficiently train the DHTM+C model, we devised a novel RNN architecture to model disease progression implicated in the co-morbidities.

Results

Our deep learning models confers higher prediction accuracy for both the next-step HF prediction and the HF trajectory prediction compared to the baseline non-neural network models and the baseline RNN model. Compared to DHTM, DHTM+C is able to output higher probability of HF for high-risk patients, even in cases where it is only given less than 2 years of data to predict over 5 years of trajectory. We illustrated multiple non-trivial real patient examples of complex HF trajectories, indicating a promising path for creating highly accurate and scalable longitudinal deep learning models for modeling the chronic disease.

SUBMITTER: Lu XH

PROVIDER: S-EPMC7787457 | biostudies-literature | 2021

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Recurrent disease progression networks for modelling risk trajectory of heart failure.

Lu Xing Han XH Liu Aihua A Fuh Shih-Chieh SC Lian Yi Y Guo Liming L Yang Yi Y Marelli Ariane A Li Yue Y

PloS one 20210106 1

<h4>Motivation</h4>Recurrent neural networks (RNN) are powerful frameworks to model medical time series records. Recent studies showed improved accuracy of predicting future medical events (e.g., readmission, mortality) by leveraging large amount of high-dimensional data. However, very few studies have explored the ability of RNN in predicting long-term trajectories of recurrent events, which is more informative than predicting one single event in directing medical intervention.<h4>Methods</h4>I ...[more]

PMID: 33406155

Similar Datasets

Project description:Early identification of individuals at risk of developing Alzheimer's disease (AD) dementia is important for developing disease-modifying therapies. In this study, given multimodal AD markers and clinical diagnosis of an individual from one or more timepoints, we seek to predict the clinical diagnosis, cognition and ventricular volume of the individual for every month (indefinitely) into the future. We proposed and applied a minimal recurrent neural network (minimalRNN) model to data from The Alzheimer's Disease Prediction Of Longitudinal Evolution (TADPOLE) challenge, comprising longitudinal data of 1677 participants (Marinescu et al., 2018) from the Alzheimer's Disease Neuroimaging Initiative (ADNI). We compared the performance of the minimalRNN model and four baseline algorithms up to 6 years into the future. Most previous work on predicting AD progression ignore the issue of missing data, which is a prevalent issue in longitudinal data. Here, we explored three different strategies to handle missing data. Two of the strategies treated the missing data as a "preprocessing" issue, by imputing the missing data using the previous timepoint ("forward filling") or linear interpolation ("linear filling). The third strategy utilized the minimalRNN model itself to fill in the missing data both during training and testing ("model filling"). Our analyses suggest that the minimalRNN with "model filling" compared favorably with baseline algorithms, including support vector machine/regression, linear state space (LSS) model, and long short-term memory (LSTM) model. Importantly, although the training procedure utilized longitudinal data, we found that the trained minimalRNN model exhibited similar performance, when using only 1 input timepoint or 4 input timepoints, suggesting that our approach might work well with just cross-sectional data. An earlier version of our approach was ranked 5th (out of 53 entries) in the TADPOLE challenge in 2019. The current approach is ranked 2nd out of 63 entries as of June 3rd, 2020.

Project description:AimsHeart failure with mildly reduced ejection fraction (HFmrEF) is associated with a favourable prognosis compared with heart failure (HF) with reduced ejection fraction (EF). We assessed whether left ventricular ejection fraction (LVEF) trajectory can be used to identify groups of patients with HFmrEF who have different clinical outcomes in a large retrospective study of patients with serial imaging.Methods and resultsPatients with HF and ≥2 echocardiograms performed ≥6 months apart were included if the LVEF measured 40-49% on the second study. Patients were classified as HFmrEF-Increasing if LVEF had increased ≥10% (n = 450), HFmrEF-Decreasing if LVEF had decreased ≥10% (n = 512), or HFmrEF-Stable if they did not meet other criteria (n = 389). The primary outcome was all-cause mortality or cardiovascular hospitalization after the second echocardiogram. Associations with time to first event were assessed with multivariable Cox analyses adjusted for age, co-morbidities, and medications. In total, 1351 patients with HFmrEF (median age 74, 64.2% male) were included with 28.8% exhibiting stable LVEF. During median follow-up of 15.3 months, the composite outcome occurred in 811 patients. During follow-up, patients with HFmrEF-Increasing were less likely to experience the primary outcome [adjusted hazard ratio (HR) 0.72, 95% confidence interval (CI) 0.60-0.88, P < 0.001] compared with HFmrEF-Stable. Patients with HFmrEF-Decreasing were more likely to experience the composite outcome in unadjusted analyses (unadjusted HR 1.19, 95% CI 1.01-1.40, P = 0.040) but not adjusted analyses (adjusted HR 1.16, 95% CI 0.98-1.37, P = 0.092). Associations with death or HF hospitalizations were similar (HFmrEF-Increasing: adjusted HR 0.72, 95% CI 0.59-0.88, P = 0.005; HFmrEF-Decreasing: adjusted HR 1.20, 95% CI 1.01-1.44, P = 0.044). Patients with HFmrEF-Decreasing had a similar risk of the composite outcome as patients with HF with reduced EF (adjusted HR 1.03, 95% CI 0.89-1.20, P = 0.670). Patients with HFmrEF-Increasing were less likely to experience the composite outcome compared with patients with HF with preserved EF (adjusted HR 0.73, 95% CI 0.62-0.87, P < 0.001).ConclusionsAmongst patients with HFmrEF, those exhibiting positive LVEF trajectory were less likely to experience adverse outcomes after correcting for important confounders including medical therapy. Categorizing HFmrEF patients based on LVEF trajectory provides meaningful clinical information and may assist clinicians with management decisions.

Dataset Information

Recurrent disease progression networks for modelling risk trajectory of heart failure.

Motivation

Methods

Results

Publications

Recurrent disease progression networks for modelling risk trajectory of heart failure.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets