Arrhythmic mitral valve prolapse phenotype: An unsupervised machine learning analysis using a multicenter cardiac MRI registry

Ralph Kwame Akyea; Stefano Figliozzi; Pedro M. Lopes; Klemens B. Bauer; Sara Moura-Ferreira; Lara Tondi; Saima Mushtaq; Stefano Censi; Anna Giulia Pavon; Ilaria Bassi; Laura Galian-Gay; Arco J. Teske; Federico Biondi; Domenico Filomena; Vasileios Stylianidis; Camilla Torlasco; Denisa Muraru; Pierre Monney; Giuseppina Quattrocchi; Viviana Maestrini; Luciano Agati; Lorenzo Monti; Patrizia Pedrotti; Bert Vandenberk; Angelo Squeri; Massimo Lombardi; Antonio M. Ferreira; Juerg Schwitter; Giovanni Donato Aquaro; Gianluca Pontone; Amedeo Chiribiri; José F.Rodríguez Palomares; Ali Yilmaz; Daniele Andreini; Anca Rezeda Florian; Marco Francone; Tim Leiner; João Abecasis; Luigi Paolo Badano; Jan Bogaert; Georgios Georgiopoulos; Pier Giorgio Masci

doi:10.1148/ryct.230247

Arrhythmic mitral valve prolapse phenotype: An unsupervised machine learning analysis using a multicenter cardiac MRI registry

Ralph Kwame Akyea, Stefano Figliozzi, Pedro M. Lopes, Klemens B. Bauer, Sara Moura-Ferreira, Lara Tondi, Saima Mushtaq, Stefano Censi, Anna Giulia Pavon, Ilaria Bassi, Laura Galian-Gay, Arco J. Teske, Federico Biondi, Domenico Filomena, Vasileios Stylianidis, Camilla Torlasco, Denisa Muraru, Pierre Monney, Giuseppina Quattrocchi, Viviana MaestriniLuciano Agati, Lorenzo Monti, Patrizia Pedrotti, Bert Vandenberk, Angelo Squeri, Massimo Lombardi, Antonio M. Ferreira, Juerg Schwitter, Giovanni Donato Aquaro, Gianluca Pontone, Amedeo Chiribiri, José F.Rodríguez Palomares, Ali Yilmaz, Daniele Andreini, Anca Rezeda Florian, Marco Francone, Tim Leiner, João Abecasis, Luigi Paolo Badano, Jan Bogaert, Georgios Georgiopoulos, Pier Giorgio Masci^*

^*Corresponding author for this work

Research output: Contribution to journal › Article › peer-review

2 Citations (Scopus)

61 Downloads (Pure)

Abstract

Purpose: To use unsupervised machine learning to identify phenotypic clusters with increased risk of arrhythmic mitral valve prolapse (MVP). Materials and Methods: This retrospective study included patients with MVP without hemodynamically significant mitral regurgitation or left ventricular (LV) dysfunction undergoing late gadolinium enhancement (LGE) cardiac MRI between October 2007 and June 2020 in 15 European tertiary centers. The study end point was a composite of sustained ventricular tachycardia, (aborted) sudden cardiac death, or unexplained syncope. Unsupervised data-driven hierarchical k-mean algorithm was utilized to identify phenotypic clusters. The association between clusters and the study end point was assessed by Cox proportional hazards model. Results: A total of 474 patients (mean age, 47 years ± 16 [SD]; 244 female, 230 male) with two phenotypic clusters were identified. Patients in cluster 2 (199 of 474, 42%) had more severe mitral valve degeneration (ie, bileaflet MVP and leaflet displacement), left and right heart chamber remodeling, and myocardial fibrosis as assessed with LGE cardiac MRI than those in cluster 1. Demographic and clinical features (ie, symptoms, arrhythmias at Holter monitoring) had negligible contribution in differentiating the two clusters. Compared with cluster 1, the risk of developing the study end point over a median follow-up of 39 months was significantly higher in cluster 2 patients (hazard ratio: 3.79 [95% CI: 1.19, 12.12], P = .02) after adjustment for LGE extent. Conclusion: Among patients with MVP without significant mitral regurgitation or LV dysfunction, unsupervised machine learning enabled the identification of two phenotypic clusters with distinct arrhythmic outcomes based primarily on cardiac MRI features. These results encourage the use of in-depth imaging-based phenotyping for implementing arrhythmic risk prediction in MVP.

Original language	English
Article number	e230247
Journal	Radiology: Cardiothoracic Imaging
Volume	6
Issue number	3
Early online date	20 Jun 2024
DOIs	https://doi.org/10.1148/ryct.230247
Publication status	Published - Jun 2024

Keywords

Cardiac
Cardiac MRI
Cluster Analysis
Mitral Valve Prolapse
MR Imaging
Sudden Cardiac Death
Unsupervised Machine Learning
Ventricular Arrhythmia

UN SDGs

This output contributes to the following UN Sustainable Development Goals (SDGs)

Access to Document

10.1148/ryct.230247

Arrhythmic Mitral Valve Prolapse_Akyea_publishedonline20June2024_Green_AAMAccepted author manuscript, 1.13 MBLicence: CC BY

Cite this

Akyea, R. K., Figliozzi, S., Lopes, P. M., Bauer, K. B., Moura-Ferreira, S., Tondi, L., Mushtaq, S., Censi, S., Pavon, A. G., Bassi, I., Galian-Gay, L., Teske, A. J., Biondi, F., Filomena, D., Stylianidis, V., Torlasco, C., Muraru, D., Monney, P., Quattrocchi, G., ... Masci, P. G. (2024). Arrhythmic mitral valve prolapse phenotype: An unsupervised machine learning analysis using a multicenter cardiac MRI registry. Radiology: Cardiothoracic Imaging, 6(3), Article e230247. https://doi.org/10.1148/ryct.230247

@article{d46adaaf7a7048edbe2d3be1d3ed2791,

title = "Arrhythmic mitral valve prolapse phenotype: An unsupervised machine learning analysis using a multicenter cardiac MRI registry",

abstract = "Purpose: To use unsupervised machine learning to identify phenotypic clusters with increased risk of arrhythmic mitral valve prolapse (MVP). Materials and Methods: This retrospective study included patients with MVP without hemodynamically significant mitral regurgitation or left ventricular (LV) dysfunction undergoing late gadolinium enhancement (LGE) cardiac MRI between October 2007 and June 2020 in 15 European tertiary centers. The study end point was a composite of sustained ventricular tachycardia, (aborted) sudden cardiac death, or unexplained syncope. Unsupervised data-driven hierarchical k-mean algorithm was utilized to identify phenotypic clusters. The association between clusters and the study end point was assessed by Cox proportional hazards model. Results: A total of 474 patients (mean age, 47 years ± 16 [SD]; 244 female, 230 male) with two phenotypic clusters were identified. Patients in cluster 2 (199 of 474, 42%) had more severe mitral valve degeneration (ie, bileaflet MVP and leaflet displacement), left and right heart chamber remodeling, and myocardial fibrosis as assessed with LGE cardiac MRI than those in cluster 1. Demographic and clinical features (ie, symptoms, arrhythmias at Holter monitoring) had negligible contribution in differentiating the two clusters. Compared with cluster 1, the risk of developing the study end point over a median follow-up of 39 months was significantly higher in cluster 2 patients (hazard ratio: 3.79 [95% CI: 1.19, 12.12], P = .02) after adjustment for LGE extent. Conclusion: Among patients with MVP without significant mitral regurgitation or LV dysfunction, unsupervised machine learning enabled the identification of two phenotypic clusters with distinct arrhythmic outcomes based primarily on cardiac MRI features. These results encourage the use of in-depth imaging-based phenotyping for implementing arrhythmic risk prediction in MVP.",

keywords = "Cardiac, Cardiac MRI, Cluster Analysis, Mitral Valve Prolapse, MR Imaging, Sudden Cardiac Death, Unsupervised Machine Learning, Ventricular Arrhythmia",

author = "Akyea, {Ralph Kwame} and Stefano Figliozzi and Lopes, {Pedro M.} and Bauer, {Klemens B.} and Sara Moura-Ferreira and Lara Tondi and Saima Mushtaq and Stefano Censi and Pavon, {Anna Giulia} and Ilaria Bassi and Laura Galian-Gay and Teske, {Arco J.} and Federico Biondi and Domenico Filomena and Vasileios Stylianidis and Camilla Torlasco and Denisa Muraru and Pierre Monney and Giuseppina Quattrocchi and Viviana Maestrini and Luciano Agati and Lorenzo Monti and Patrizia Pedrotti and Bert Vandenberk and Angelo Squeri and Massimo Lombardi and Ferreira, {Antonio M.} and Juerg Schwitter and Aquaro, {Giovanni Donato} and Gianluca Pontone and Amedeo Chiribiri and Palomares, {Jos{\'e} F.Rodr{\'i}guez} and Ali Yilmaz and Daniele Andreini and Florian, {Anca Rezeda} and Marco Francone and Tim Leiner and Jo{\~a}o Abecasis and Badano, {Luigi Paolo} and Jan Bogaert and Georgios Georgiopoulos and Masci, {Pier Giorgio}",

note = "Publisher Copyright: {\textcopyright} RSNA, 2024",

year = "2024",

month = jun,

doi = "10.1148/ryct.230247",

language = "English",

volume = "6",

journal = "Radiology: Cardiothoracic Imaging",

issn = "2638-6135",

publisher = "Radiological Society of North America Inc.",

number = "3",

}

Akyea, RK, Figliozzi, S, Lopes, PM, Bauer, KB, Moura-Ferreira, S, Tondi, L, Mushtaq, S, Censi, S, Pavon, AG, Bassi, I, Galian-Gay, L, Teske, AJ, Biondi, F, Filomena, D, Stylianidis, V, Torlasco, C, Muraru, D, Monney, P, Quattrocchi, G, Maestrini, V, Agati, L, Monti, L, Pedrotti, P, Vandenberk, B, Squeri, A, Lombardi, M, Ferreira, AM, Schwitter, J, Aquaro, GD, Pontone, G, Chiribiri, A, Palomares, JFR, Yilmaz, A, Andreini, D, Florian, AR, Francone, M, Leiner, T, Abecasis, J, Badano, LP, Bogaert, J, Georgiopoulos, G & Masci, PG 2024, 'Arrhythmic mitral valve prolapse phenotype: An unsupervised machine learning analysis using a multicenter cardiac MRI registry', Radiology: Cardiothoracic Imaging, vol. 6, no. 3, e230247. https://doi.org/10.1148/ryct.230247

TY - JOUR

T1 - Arrhythmic mitral valve prolapse phenotype

T2 - An unsupervised machine learning analysis using a multicenter cardiac MRI registry

AU - Akyea, Ralph Kwame

AU - Figliozzi, Stefano

AU - Lopes, Pedro M.

AU - Bauer, Klemens B.

AU - Moura-Ferreira, Sara

AU - Tondi, Lara

AU - Mushtaq, Saima

AU - Censi, Stefano

AU - Pavon, Anna Giulia

AU - Bassi, Ilaria

AU - Galian-Gay, Laura

AU - Teske, Arco J.

AU - Biondi, Federico

AU - Filomena, Domenico

AU - Stylianidis, Vasileios

AU - Torlasco, Camilla

AU - Muraru, Denisa

AU - Monney, Pierre

AU - Quattrocchi, Giuseppina

AU - Maestrini, Viviana

AU - Agati, Luciano

AU - Monti, Lorenzo

AU - Pedrotti, Patrizia

AU - Vandenberk, Bert

AU - Squeri, Angelo

AU - Lombardi, Massimo

AU - Ferreira, Antonio M.

AU - Schwitter, Juerg

AU - Aquaro, Giovanni Donato

AU - Pontone, Gianluca

AU - Chiribiri, Amedeo

AU - Palomares, José F.Rodríguez

AU - Yilmaz, Ali

AU - Andreini, Daniele

AU - Florian, Anca Rezeda

AU - Francone, Marco

AU - Leiner, Tim

AU - Abecasis, João

AU - Badano, Luigi Paolo

AU - Bogaert, Jan

AU - Georgiopoulos, Georgios

AU - Masci, Pier Giorgio

PY - 2024/6

Y1 - 2024/6

N2 - Purpose: To use unsupervised machine learning to identify phenotypic clusters with increased risk of arrhythmic mitral valve prolapse (MVP). Materials and Methods: This retrospective study included patients with MVP without hemodynamically significant mitral regurgitation or left ventricular (LV) dysfunction undergoing late gadolinium enhancement (LGE) cardiac MRI between October 2007 and June 2020 in 15 European tertiary centers. The study end point was a composite of sustained ventricular tachycardia, (aborted) sudden cardiac death, or unexplained syncope. Unsupervised data-driven hierarchical k-mean algorithm was utilized to identify phenotypic clusters. The association between clusters and the study end point was assessed by Cox proportional hazards model. Results: A total of 474 patients (mean age, 47 years ± 16 [SD]; 244 female, 230 male) with two phenotypic clusters were identified. Patients in cluster 2 (199 of 474, 42%) had more severe mitral valve degeneration (ie, bileaflet MVP and leaflet displacement), left and right heart chamber remodeling, and myocardial fibrosis as assessed with LGE cardiac MRI than those in cluster 1. Demographic and clinical features (ie, symptoms, arrhythmias at Holter monitoring) had negligible contribution in differentiating the two clusters. Compared with cluster 1, the risk of developing the study end point over a median follow-up of 39 months was significantly higher in cluster 2 patients (hazard ratio: 3.79 [95% CI: 1.19, 12.12], P = .02) after adjustment for LGE extent. Conclusion: Among patients with MVP without significant mitral regurgitation or LV dysfunction, unsupervised machine learning enabled the identification of two phenotypic clusters with distinct arrhythmic outcomes based primarily on cardiac MRI features. These results encourage the use of in-depth imaging-based phenotyping for implementing arrhythmic risk prediction in MVP.

AB - Purpose: To use unsupervised machine learning to identify phenotypic clusters with increased risk of arrhythmic mitral valve prolapse (MVP). Materials and Methods: This retrospective study included patients with MVP without hemodynamically significant mitral regurgitation or left ventricular (LV) dysfunction undergoing late gadolinium enhancement (LGE) cardiac MRI between October 2007 and June 2020 in 15 European tertiary centers. The study end point was a composite of sustained ventricular tachycardia, (aborted) sudden cardiac death, or unexplained syncope. Unsupervised data-driven hierarchical k-mean algorithm was utilized to identify phenotypic clusters. The association between clusters and the study end point was assessed by Cox proportional hazards model. Results: A total of 474 patients (mean age, 47 years ± 16 [SD]; 244 female, 230 male) with two phenotypic clusters were identified. Patients in cluster 2 (199 of 474, 42%) had more severe mitral valve degeneration (ie, bileaflet MVP and leaflet displacement), left and right heart chamber remodeling, and myocardial fibrosis as assessed with LGE cardiac MRI than those in cluster 1. Demographic and clinical features (ie, symptoms, arrhythmias at Holter monitoring) had negligible contribution in differentiating the two clusters. Compared with cluster 1, the risk of developing the study end point over a median follow-up of 39 months was significantly higher in cluster 2 patients (hazard ratio: 3.79 [95% CI: 1.19, 12.12], P = .02) after adjustment for LGE extent. Conclusion: Among patients with MVP without significant mitral regurgitation or LV dysfunction, unsupervised machine learning enabled the identification of two phenotypic clusters with distinct arrhythmic outcomes based primarily on cardiac MRI features. These results encourage the use of in-depth imaging-based phenotyping for implementing arrhythmic risk prediction in MVP.

KW - Cardiac

KW - Cardiac MRI

KW - Cluster Analysis

KW - Mitral Valve Prolapse

KW - MR Imaging

KW - Sudden Cardiac Death

KW - Unsupervised Machine Learning

KW - Ventricular Arrhythmia

UR - http://www.scopus.com/inward/record.url?scp=85203260294&partnerID=8YFLogxK

U2 - 10.1148/ryct.230247

DO - 10.1148/ryct.230247

M3 - Article

C2 - 38900026

SN - 2638-6135

VL - 6

JO - Radiology: Cardiothoracic Imaging

JF - Radiology: Cardiothoracic Imaging

IS - 3

M1 - e230247

ER -

Arrhythmic mitral valve prolapse phenotype: An unsupervised machine learning analysis using a multicenter cardiac MRI registry

Abstract

Keywords

UN SDGs

Access to Document

Other files and links

Fingerprint

Cite this