Skip to Main content Skip to Navigation
Poster communications

Proteomic signature for early diagnosis of left ventricular remodeling after myocardial infarction

Wilfried Heyse 1, 2 Vincent Vandewalle 3, 2 Philippe Amouyel 1 Guillemette Marot 3, 2 Christophe Bauters 1 Florence Pinet 1
2 MODAL - MOdel for Data Analysis and Learning
LPP - Laboratoire Paul Painlevé - UMR 8524, Université de Lille, Sciences et Technologies, Inria Lille - Nord Europe, METRICS - Evaluation des technologies de santé et des pratiques médicales - ULR 2694, Polytech Lille - École polytechnique universitaire de Lille
Abstract : Heart failure (HF) remains a main cause of mortality worldwide. The most common cause of HF is coronary artery disease and particularly myocardial infarction (MI). Left ventricular remodelling (LVR) is a progressive dilatation of the left ventricle that occurs in response to MI and is difficult to predict in clinical practice based on infarct size, infarct location or LV ejection fraction. Several studies have identified LVR as a powerful indicator of a high risk of HF or death after MI. The aim is to identify plasmatic proteins that could predict the occurrence and severity of LVR in order to prevent HF. The REVE and REVE-2 studies have included respectively, 215 and 246 patients with a first anterior MI. The patients have been followed-up with serial echocardiography during one year to quantify LVR. Plasma samples have been collected during hospitalization for both studies, and at 1, 3 and 12 months after MI for REVE-2, allowing to measure 5284 proteins thanks to a high throughput proteomic approach (SOMASCAN). Due to the high dimension of data (more variables than individuals), we used statistical methods performing variable selection to build a proteomic signature of LVR. We showed that REVE and REVE-2 studies share common statistical characteristics (distributions, correlations) allowing us to perform analysis on REVE-2 and use REVE for validation. We confirmed the difficulty to predict LVR using only clinical data, with prediction models explaining at the best 11% of LVR (R²=0,106). Using the proteomic data we explained 30% of LVR (R²=0,297) using a 22-proteins based score built with nested models. We confirmed that LVR is complex to predict even with a huge number of potential biomarkers to explore. Still, we enhanced the predictability of LVR and we intend to find a protein profile of LVR during the year after MI by studying data collected during this period.
Complete list of metadata
Contributor : Wilfried Heyse Connect in order to contact the contributor
Submitted on : Friday, January 29, 2021 - 1:27:05 PM
Last modification on : Monday, February 1, 2021 - 12:34:13 PM
Long-term archiving on: : Friday, April 30, 2021 - 6:18:04 PM


Files produced by the author(s)


  • HAL Id : hal-03124837, version 1



Wilfried Heyse, Vincent Vandewalle, Philippe Amouyel, Guillemette Marot, Christophe Bauters, et al.. Proteomic signature for early diagnosis of left ventricular remodeling after myocardial infarction. Printemps de la Cardiologie 2020, Oct 2020, Grenoble / Virtual, France. ⟨hal-03124837⟩



Les métriques sont temporairement indisponibles