<i>Deep Survival Machines</i>: Fully Parametric Survival Regression and Representation Learning for Censored Data With Competing Risks

Nagpal, Chirag; Li, Xinyu; Dubrawski, Artur

doi:10.1109/jbhi.2021.3052441

Cited by 93 publications

(72 citation statements)

References 33 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Neural networks represent stateoftheart survival analysis. [16][17][18][19] If applied to realworld medical data, the model's increased complexity could facilitate the inte gration of polygenic information for primary prevention of cardiovascular disease by inherently accounting for the interaction of the polygenic information and the clinical parameters.…”

Section: Introductionmentioning

confidence: 99%

“…This study presents the development and validation of a novel neural networkbased cardiovascular disease risk model, NeuralCVD, based on Deep Survival Machines, 19 for primary prevention based on a set of established cardiovascular disease risk factors. Comparing our model against existing risk scores and a Cox proportional hazards model 20 trained on the same data over the entire study population, we first demonstrated its discriminative capabilities.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Neural network-based integration of polygenic and clinical information: development and validation of a prediction model for 10-year risk of major adverse cardiac events in the UK Biobank cohort

Steinfeldt

Buergel

Loock

et al. 2022

The Lancet Digital Health

View full text Add to dashboard Cite

Background In primary cardiovascular disease prevention, early identification of high-risk individuals is crucial. Genetic information allows for the stratification of genetic predispositions and lifetime risk of cardiovascular disease. However, towards clinical application, the added value over clinical predictors later in life is crucial. Currently, this genotype-phenotype relationship and implications for overall cardiovascular risk are unclear. MethodsIn this study, we developed and validated a neural network-based risk model (NeuralCVD) integrating polygenic and clinical predictors in 395 713 cardiovascular disease-free participants from the UK Biobank cohort. The primary outcome was the first record of a major adverse cardiac event (MACE) within 10 years. We compared the NeuralCVD model with both established clinical scores (SCORE, ASCVD, and QRISK3 recalibrated to the UK Biobank cohort) and a linear Cox-Model, assessing risk discrimination, net reclassification, and calibration over 22 spatially distinct recruitment centres. Findings The NeuralCVD score was well calibrated and improved on the best clinical baseline, QRISK3 (∆Concordance index [C-index] 0•01, 95% CI 0•009-0•011; net reclassification improvement (NRI) 0•0488, 95% CI 0•0442-0•0534) and a Cox model (∆C-index 0•003, 95% CI 0•002-0•004; NRI 0•0469, 95% CI 0•0429-0•0511) in risk discrimination and net reclassification. After adding polygenic scores we found further improvements on population level (∆C-index 0•006, 95% CI 0•005-0•007; NRI 0•0116, 95% CI 0•0066-0•0159). Additionally, we identified an interaction of genetic information with the pre-existing clinical phenotype, not captured by conventional models. Additional high polygenic risk increased overall risk most in individuals with low to intermediate clinical risk, and age younger than 50 years. Interpretation Our results demonstrated that the NeuralCVD score can estimate cardiovascular risk trajectories for primary prevention. NeuralCVD learns the transition of predictive information from genotype to phenotype and identifies individuals with high genetic predisposition before developing a severe clinical phenotype. This finding could improve the reprioritisation of otherwise low-risk individuals with a high genetic cardiovascular predisposition for preventive interventions.

show abstract

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Neural network-based integration of polygenic and clinical information: development and validation of a prediction model for 10-year risk of major adverse cardiac events in the UK Biobank cohort

Steinfeldt

Buergel

Loock

et al. 2022

The Lancet Digital Health

View full text Add to dashboard Cite

show abstract

“…This is in contrast to the classical statistical literature, in which a wide variety of methods are available [20][21][22][23][24] , and in which it is widely agreed that a properly conducted competing-risks analysis is often necessary to avoid biased estimation results and/or predictions 36 . Although several adaptations to DNN architectures have been proposed recently 9,11,25 , these adaptions rely on a huge number of parameters, making network training and regularization a challenging task. In this work, we showed that an imputation strategy based on subdistribution weights could convert the competing risks survival data into a dataset that has only one event in the presence of censoring.…”

Section: Discussionmentioning

confidence: 99%

“…Similarly, Gupta et al 11 use separate subnetworks per event. In another work, Nagpal et al 25 proposed a Deep Survival Machine (DSM), to learn a mixture of primitive distributions in order to estimate the conditional survival function S(t|x) = P(T > t). Again, in this model an additional set of parameters are added in order to describe the event distribution for each competing risk.…”

Section: Introductionmentioning

confidence: 99%

Subdistribution-Based Imputation for Deep Survival Analysis with Competing Events

Zadeh

Behning

Schmid

2021

Preprint

View full text Add to dashboard Cite

With the popularity of deep neural networks (DNNs) in recent years, many researchers have proposed DNNs for the analysis of survival data (time-to-event data). These networks learn the distribution of survival times directly from the predictor variables without making strong assumptions on the underlying stochastic process. In survival analysis, it is common to observe several types of events, also called competing events. The occurrences of these competing events are usually not independent of one another and have to be incorporated in the modeling process in addition to censoring. In classical survival analysis, a popular method to incorporate competing events is the subdistribution hazard model, which is usually fitted using weighted Cox regression. In the DNN framework, only few architectures have been proposed to model the distribution of time to a specific event in a competing events situation. These architectures are characterized by a separate subnetwork/pathway per event, leading to large networks with huge amounts of parameters that may become difficult to train. In this work, we propose a novel imputation strategy for data preprocessing that incorporates the subdistribution weights derived from the classical model. With this, it is no longer necessary to add multiple subnetworks to the DNN to handle competing events. Our experiments on synthetic and real-world datasets show that DNNs with multiple subnetworks per event can simply be replaced by a DNN designed for a single-event analysis without loss in accuracy.

show abstract

“…To consider applicable approaches to various data forms, we exclude networks that are difficult to adjust to new forms of data (e.g., focusing on specific modality [33] or specific architecture [34], and relying on time-varying information [35]) because there are increasingly diverse types of input for survival prediction. Because medical data are difficult to model, we also avoid approaches that attempt to model the distribution of data or distribution of outcome or outcome censoring (e.g., sampling from generative modeling [36], Perturbation [37], assuming distributions of outcome [38], and training with artificial target responses [39]) because data modeling and augmentation are difficult to validate and introduce risks of learning from invalid data. Deepsurv and our baselines under our investigations fit all these criteria.…”

Section: Introductionmentioning

confidence: 99%

Regularizing the Deepsurv Network Using Projection Loss for Medical Risk Assessment

et al. 2022

View full text Add to dashboard Cite

State-of-the-art deep survival prediction approaches expand network parameters to accommodate performance over a fine discretization of output time. For medical applications where data are limited, the regression-based Deepsurv approach is more advantageous because its continuous output design limits unnecessary network parameters. Despite the practical advantage, the typical network lacks control over the feature distribution causing the network to be more prone to noisy information and occasional poor prediction performance. We propose a novel projection loss as a regularizing objective to improve the timeto-event Deepsurv model. The loss formulation maximizes the lower bound of the multiple-correlation coefficient between the network's features and the desired hazard value. Reducing the loss also theoretically lowers the upper bound on the likelihood of discordant pair and improves C-index performance. We observe superior performances and robustness of regularized Deepsurv over many state-of-the-art approaches in our experiments with five public medical datasets and two cross-cohort validation tasks.

show abstract

Deep Survival Machines: Fully Parametric Survival Regression and Representation Learning for Censored Data With Competing Risks

Cited by 93 publications

References 33 publications

Neural network-based integration of polygenic and clinical information: development and validation of a prediction model for 10-year risk of major adverse cardiac events in the UK Biobank cohort

Neural network-based integration of polygenic and clinical information: development and validation of a prediction model for 10-year risk of major adverse cardiac events in the UK Biobank cohort

Subdistribution-Based Imputation for Deep Survival Analysis with Competing Events

Regularizing the Deepsurv Network Using Projection Loss for Medical Risk Assessment

Contact Info

Product

Resources

About