Predicting post-operative right ventricular failure using video-based deep learning

Abstract

Despite progressive improvements over the decades, the rich temporally resolved data in an echocardiogram remain underutilized. Human assessments reduce the complex patterns of cardiac wall motion, to a small list of measurements of heart function. All modern echocardiography artificial intelligence (AI) systems are similarly limited by design – automating measurements of the same reductionist metrics rather than utilizing the embedded wealth of data. This underutilization is most evident where clinical decision making is guided by subjective assessments of disease acuity. Predicting the likelihood of developing post-operative right ventricular failure (RV failure) in the setting of mechanical circulatory support is one such example. Here we describe a video AI system trained to predict post-operative RV failure using the full spatiotemporal density of information in pre-operative echocardiography. We achieve an AUC of 0.729, and show that this ML system significantly outperforms a team of human experts at the same task on independent evaluation.

Artificial Intelligence in Echocardiography: The Time is Now

“…Accurate and reproducible quantification of RV function can be difficult due to its irregular crescent shape, poor echocardiographic visualisation of the RV and inconsistencies in the analysis of RV parameters [41]. AI has shown promise in rapid and accurate assessment of RV function [42,43]. [36] Evaluated diagnostic accuracy of AI assisted clinical decision system for diagnosing diastolic heart failure.…”

“…AI has also been used to develop predictive tools in assessing RV failure post-implantation of a left ventricular assist device (LVAD). A third of all LVAD implantations are complicated by RV failure post-operatively, this is in part due to increased RV preload from the device and excessive leftward shift of the interventricular septum, reducing its contribution to RV contraction [43,44]. RV failure post LVAD implantation is difficult to predict and is currently based on pre-existing echocardiographic assessment, biomarkers and clinical judgement [43].…”

“…A third of all LVAD implantations are complicated by RV failure post-operatively, this is in part due to increased RV preload from the device and excessive leftward shift of the interventricular septum, reducing its contribution to RV contraction [43,44]. RV failure post LVAD implantation is difficult to predict and is currently based on pre-existing echocardiographic assessment, biomarkers and clinical judgement [43]. Shad et al [43] used video-based DL to predict the likelihood of developing RV failure post device insertion using only 2D echocardiographic data, significantly outperforming a team of human experts at the same task (p = 0.016).…”

“…RV failure post LVAD implantation is difficult to predict and is currently based on pre-existing echocardiographic assessment, biomarkers and clinical judgement [43]. Shad et al [43] used video-based DL to predict the likelihood of developing RV failure post device insertion using only 2D echocardiographic data, significantly outperforming a team of human experts at the same task (p = 0.016). Although these clinical decisions are never performed with only echocardiographic data, the algorithm clearly demonstrates its worth as an adjunct tool in predicting RV dysfunction, allowing appropriate measures to be taken pre-emptively.…”

Machine learning-based 3D modeling and volumetry of human posterior vitreous cavity of optical coherence tomographic images

“…This difficulty has been addressed by implementing machine learning (ML) techniques which are able to perform complex tasks of analyzing medical images accurately by employing multilayered artificial neural networks trained on a large data set of labeled images 16 , 17 . ML combined with OCT has been used in ophthalmology in various ways such as noise reduction of the images and automatic diagnosis of eye diseases 18 – 23 .…”