Harnessing the power of artificial intelligence in predicting all-cause mortality in transcatheter aortic valve replacement: a systematic review and meta-analysis

Abstract

Objectives: In recent years, the use of artificial intelligence (AI) models to generate individualised risk assessments and predict patient outcomes post-Transcatheter Aortic Valve Implantation (TAVI) has been a topic of increasing relevance in literature. This study aims to evaluate the predictive accuracy of AI algorithms in forecasting post-TAVI mortality as compared to traditional risk scores.

Methods: Following the Preferred Reporting Items for Systematic Reviews and Meta-analyses for Systematic Reviews (PRISMA) standard, a systematic review was carried out. We searched four databases in total-PubMed, Medline, Embase, and Cochrane-from 19 June 2023-24 June, 2023.

Results: From 2,239 identified records, 1,504 duplicates were removed, 735 manuscripts were screened, and 10 studies were included in our review. Our pooled analysis of 5 studies and 9,398 patients revealed a significantly higher mean area under curve (AUC) associated with AI mortality predictions than traditional score predictions (MD: -0.16, CI: -0.22 to -0.10, p < 0.00001). Subgroup analyses of 30-day mortality (MD: -0.08, CI: -0.13 to -0.03, p = 0.001) and 1-year mortality (MD: -0.18, CI: -0.27 to -0.10, p < 0.0001) also showed significantly higher mean AUC with AI predictions than traditional score predictions. Pooled mean AUC of all 10 studies and 22,933 patients was 0.79 [0.73, 0.85].

Conclusion: AI models have a higher predictive accuracy as compared to traditional risk scores in predicting post-TAVI mortality. Overall, this review demonstrates the potential of AI in achieving personalised risk assessment in TAVI patients.

Registration and protocol: This systematic review and meta-analysis was registered under the International Prospective Register of Systematic Reviews (PROSPERO), under the registration name "All-Cause Mortality in Transcatheter Aortic Valve Replacement Assessed by Artificial Intelligence" and registration number CRD42023437705. A review protocol was not prepared. There were no amendments to the information provided at registration.

Systematic review registration: https://www.crd.york.ac.uk/, PROSPERO (CRD42023437705).

Keywords: aortic valve replacement; artificial intelligence; machine learning; mortality; systematic review; transcatheter; transcatheter aortic valve prosthesis.

Figure 1

PRISMA flow chart showing systematic search. 2,234 articles were discovered on initial search, 738 remained after duplicates were removed. After our exclusion criteria was applied to record screening based on title, abstract and full-text assessment, 10 articles remained for inclusion in our analysis.

Figure 2

Diagram illustrating the different AI algorithms: figure summarising four of the different AI algorithms used in the included studies. (A) The random forest model first divides the dataset into subsets using bootstrap sampling, then generates different decision trees from the same subset. The trees are then averaged to generate the final decision tree. (B) Gradient Boosting uses a stage-wise progression to combine weak learning models sequentially to produce an ensemble of different models with minimal prediction errors. In each iteration, the weight of wrong predictions is increased in order to improve the learning model in the successive iteration. (C) Multilayer Perceptron, a type of artificial neural network, inputs data into at least 3 layers of nodes: an input layer, a hidden layer and an output layer. Backpropogation of data is used for learning to minimise prediction errors. (D) Logistic Regression fits a logistic function onto a dataset. Maximum likelihood estimation is used to produce a curve with maximum likelihood.

Figure 3

Forest plot comparing AI and traditional scores: forest plot comparing mean AUC values of AI mortality predictions to that of traditional scores. Subgroup analysis was performed, dividing 5 studies into 3.1.1 30-day mortality, 3.1.2 1-year mortality and 3.1.3 5-year mortality, with AI showing an overall better performance than traditional scores.

Figure 4

Pooled mean AUC of included studies: forest plot of pooled mean AUC values for AI-predicted post-TAVI mortality, with intra-hospital, 30-day, 1-year and 5-year mortality subgroups.

Figure 5

Pooled mean AUC of traditional risk scores: forest plot of pooled mean AUC values for traditional risk score-predicted post-TAVI mortality, with intra-hospital, 30-day, 1-year and 5-year mortality subgroups.