Simulating time to event prediction with spatiotemporal echocardiography deep learning

TL;DR

This paper explores the application of deep learning models, specifically spatiotemporal CNNs, to predict survival times from echocardiography videos, demonstrating their accuracy using simulated survival data.

Contribution

It demonstrates the feasibility of using deep learning for time-to-event prediction in echocardiography, extending survival analysis methods to imaging data.

Findings

Spatiotemporal CNNs accurately predict survival times.

Deep learning models trained on simulated data perform well.

The approach integrates imaging and survival analysis effectively.

Abstract

Integrating methods for time-to-event prediction with diagnostic imaging modalities is of considerable interest, as accurate estimates of survival requires accounting for censoring of individuals within the observation period. New methods for time-to-event prediction have been developed by extending the cox-proportional hazards model with neural networks. In this paper, to explore the feasibility of these methods when applied to deep learning with echocardiography videos, we utilize the Stanford EchoNet-Dynamic dataset with over 10,000 echocardiograms, and generate simulated survival datasets based on the expert annotated ejection fraction readings. By training on just the simulated survival outcomes, we show that spatiotemporal convolutional neural networks yield accurate survival estimates.