Articles retenus sur la thématique « ÉQUIPES FRANCOPHONES A L’HONNEUR »
(version limitée)
Édition MAI 2025
Retour en haut de page
Retour au choix des thèmes
Machine learning to detect recent recreational drug use in intensive cardiac care units - 05/2025
Source : Arch Cardiovasc Dis
Auteurs : Nathan El Bèze, Kenza Hamzi, Patrick Henry, Antonin Trimaille, Amine El Ouahidi, Cyril Zakine, Olivier Nallet, Clément Delmas, Victor Aboyans, Marc Goralski, Franck Albert, Eric Bonnefoy-Cudraz, Thomas Bochaton, Guillaume Schurtz, Pascal Lim, Antoine Lequipar, Trecy Gonçalves, Emmanuel Gall, Thibaut Pommier, Léo Lemarchand, Christophe Meune, Sonia Azzakani, Claire Bouleti, Jonas Amar, Jean-Guillaume Dillinger, P Gabriel Steg, Eric Vicaut, Solenn Toupin, Théo Pezel; ADDICT-ICCU investigators
Abstract :
Background: Although recreational drug use is a strong risk factor for acute cardiovascular events, systematic testing is currently not performed in patients admitted to intensive cardiac care units, with a risk of underdetection. To address this issue, machine learning methods could assist in the detection of recreational drug use.
Aims: To investigate the accuracy of a machine learning model using clinical, biological and echocardiographic data for detecting recreational drug use in patients admitted to intensive cardiac care units.
Methods: From 07 to 22 April 2021, systematic screening for all traditional recreational drugs (cannabis, opioids, cocaine, amphetamines, 3,4-methylenedioxymethamphetamine) was performed by urinary testing in all consecutive patients admitted to intensive cardiac care units in 39 French centres. The primary outcome was recreational drug detection by urinary testing. The framework involved automated variable selection by eXtreme Gradient Boosting (XGBoost) and model building with multiple algorithms, using 31 centres as the derivation cohort and eight other centres as the validation cohort.
Results: Among the 1499 patients undergoing urinary testing for drugs (mean age 63±15 years; 70% male), 161 (11%) tested positive (cannabis: 9.1%; opioids: 2.1%; cocaine: 1.7%; amphetamines: 0.7%; 3,4-methylenedioxymethamphetamine: 0.6%). Of these, only 57% had reported drug use. Using nine variables, the best machine learning model (random forest) showed good performance in the derivation cohort (area under the receiver operating characteristic curve=0.82) and in the validation cohort (area under the receiver operating characteristic curve=0.76).
Conclusions: In a large intensive cardiac care unit cohort, a comprehensive machine learning model exhibited good performance in detecting recreational drug use, and provided valuable insights into the relationships between clinical variables and drug use through explainable machine learning techniques.
Commentaire de la SFSN : /
Édition JUILLET 2025
Retour en haut de page
Retour au choix des thèmes
Utility of an Echocardiographic Machine Learning Model to Predict Outcomes in Intensive Cardiac Care Unit Patients - 04/2025
Source : J Am Soc Echocardiogr
Auteurs : Samy Aghezzaf, Augustin Coisne, Kenza Hamzi, Solenn Toupin, Claire Bouleti, Charles Fauvel, Jean-Baptiste Brette, David Montaigne, Reza Rossanaly Vasram, Antonin Trimaille, Gilles Lemesle, Guillaume Schurtz, Edouard Gerbaud, Clément Delmas, Marc Bedossa, Jean-Claude Dib, Vincent Roule, Etienne Puymirat, Martine Gilard, Marouane Boukhris, Nicolas Mansencal, Nabil Bouali, Stephane Andrieu, Trecy Gonçalves, Jean-Guillaume Dillinger, Patrick Henry, Theo Pezel
Abstract :
Introduction: The risk stratification at admission to the intensive cardiac care unit (ICCU) is crucial and remains challenging.
Objectives: We aimed to investigate the accuracy of a machine learning (ML)-model based on initial transthoracic echocardiography (TTE) to predict in-hospital major adverse events (MAEs) in a broad spectrum of patients admitted to ICCU.
Methods: All consecutive patients hospitalized in ICCUs with a complete TTE performed within the first 24 hours of admission were included in this prospective multicenter study (39 centers). Sixteen TTE parameters were evaluated. The ML model involved automated feature selection by random survival forest and model building with an extreme gradient boosting (XGBoost) algorithm. The primary outcome was in-hospital MAEs defined as all-cause death, resuscitated cardiac arrest, or cardiogenic shock.
Results: Of 1,499 consecutive patients (63 ± 15 years, 70% male), MAEs occurred in 67 patients (4.5%). The 5 TTE parameters selected in the model were left ventricular outflow tract velocity-time integral, E/e’ ratio, systolic pulmonary artery pressure, tricuspid annular plane systolic excursion, and left ventricular ejection fraction. Using the XGBoost, the ML model exhibited a higher area under the receiver operating curve compared with any existing scores (ML model, 0.83 vs logistic regression, 0.76, ACUTE-HF score:,0.66; thrombolysis in myocardial infarction score, 0.60; Global Registry of Acute Coronary Events score, 0.58, all P < .001). The ML model had an incremental prognostic value for predicting MAE over a traditional model including clinical and biological data (C index 0.80 vs 0.73, P = .012; chi-square 59.7 vs 32.4; P < .001).
Conclusion: The ML model based on initial TTE exhibited a higher prognostic value to predict in-hospital MAEs compared with existing scores or clinical and biological data in the ICCU.
Commentaire de la SFSN : /
Édition SEPTEMBRE 2025
Retour en haut de page
Retour au choix des thèmes
Consensus document from the Group Exercise Rehabilitation Sports-Prevention (GERS-P) of the French Society of Cardiology on cardiac rehabilitation outside cardiac rehabilitation centres, including light private rehabilitation structures and cardiac telerehabilitation - 06/2025
Source : Arch Cardiovasc Dis
Auteurs :Dany Marcadet, Bruno Pavy, Sonia Corone, Antoine Faure, Marie Michelle Six, Ines Cazaubiel, Jean-Michel Guy, Frédéric Schnell; behalf of GERS-P
Abstract :
/
Commentaire de la SFSN : /
Implementation of a remote symptom monitoring pathway in oncology care: analysis of real-world experience across 33 cancer centres in France and Belgium - 07/2024
Source : Lancet Reg Health Eur
Auteurs : Maria Alice Franzoi, Arlindo R Ferreira, Antoine Lemaire, Joseph Rodriguez 5, Jessica Grosjean, Joana M Ribeiro, Laura Polastro, Thomas Grellety, Xavier Artignan, Katell Le Du, Martina Pagliuca, Élodie Nouhaud, Maximilien Autheman, Fabrice André, Ethan Basch, Otto Metzger, Charles Ferté, Mario Di Palma, Florian Scotté, Ines Vaz-Luis
Abstract :
Background: Remote patient monitoring (RPM) of symptoms using electronic patient reported outcomes (ePROs) has been shown to reduce symptom burden and hospitalizations, increase dose intensity and improve quality of life of patients during systemic therapy being recommended by international guidelines in routine oncology practice. However, implementation in routine care has been slow and faces several challenges. In this study we report on the real-world multi-center implementation of a RPM pathway encompassing weekly patient symptom ePRO reporting with electronic alert notifications triggered to providers for severe or worsening symptoms.
Methods: An RPM pathway was implemented in 33 European cancer centers in France and Belgium between November 2021 and August 2023. The implementation process followed a standardized phasic process of Exploration, Preparation, Implementation and Sustainment. Patient-level and system-level implementation metrics were collected and evaluated according to the Reach, Effectiveness, Adoption, Implementation and Maintenance (RE-AIM) framework.
Findings: Across the 33 cancer centers, the RPM pathway was implemented for 3015 patients cared for by 168 providers. The RPM pathway enabled effective and timely symptom management with 94.6% of all alerts (10,132/10,711) evolving to an improvement two weeks later, among which 88.4% (9468/10,711) showed ≥2 grades of improvement on the 5-point scale of the Patient-Reported Outcomes Common Terminology (PRO-CTCAE). The median time to alert management by the care team was 13 h 41 min (25th percentile: 1 h 42 min, 75th percentile: 1 day + 19 h 54 min), with 80% (36,269/45,334) of alerts managed by a nurse navigator telephone call. Patient adherence with weekly ePRO reporting was 82% (2472/3015). In an experience survey, 87% (32/38) of providers were satisfied with integrating the solution into their organization and 90% (276/307) of the patients felt that ePRO reporting positively impacted their care. As of March 2024, the pathway has been maintained in all participating centers, with activation of an additional 18 centers following data lock, and reimbursement for this RPM pathway approved in France in October 2023.
Interpretation: These findings demonstrate the feasibility of implementing and maintaining an RPM pathway during routine care across a diverse group of cancer centers in the European setting, with high levels of patient and provider engagement, and positive clinical impact.
Funding: Part of this work was funded Breast Cancer Research Foundation (Career Development Award to Maria Alice Franzoi) and Resilience (nurse navigation and technology support).
Commentaire de la SFSN : /
Retour en haut de page
Retour au choix des thèmes