Emmanuel Messas
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Emmanuel Messas has a total of 281 co-authors. Here are the top co-authors by number of shared publications:
- Tristan Mirault
Physics for Medicine Paris, INSERM U1273, ESPCI Paris, CNRS FRE 2031, PSL Research University, Paris, France
12 shared publications - Guillaume Goudot
Hopital Europeen Georges Pompidou Chirurgie cardio vasculaire
9 shared publications - Mickael Tanter
Physics for Medicine, ESPCI, INSERM, CNRS, PSL University
6 shared publications Guillaume Goudot
Cardiovascular Department, Hôpital Européen Georges-Pompidou, Université Paris-Cité, Paris, France
5 shared publicationsMathieu Pernot
Physics for Medicine Paris, Inserm/ESPCI Paris-PSL/CRNS, Paris, France
5 shared publications- Mathieu Pernot
Physics for Medicine, ESPCI, INSERM, CNRS, PSL University
5 shared publications
Latest Publications
- Open Access
- 11 August 2026
- Published by Wolters Kluwer Health in Circulation
- Vol. 154 (6) , 587-590
Prospective, Single-Arm Pivotal Study for the Treatment of Subjects With Severe Symptomatic Calcific Aortic Valve Stenosis Using the Valvosoft Noninvasive Ultrasound Therapy
- 11 August 2026
- Published by Wolters Kluwer Health in Circulation
- Vol. 154 (6) , 587-590
Open AccessProspective, Single-Arm Pivotal Study for the Treatment of Subjects With Severe Symptomatic Calcific Aortic Valve Stenosis Using the Valvosoft Noninvasive Ultrasound Therapy
- Open Access
- 21 May 2026
- Published by IOS Press
- Vol. 336, 363-367
Machine Learning for Cardiovascular Prevention Prescriptions: Real-World vs. Synthetic Data
Cardiovascular diseases remain the leading cause of death worldwide, primarily driven by atherosclerosis, which is targeted by lipid-lowering agents such as statins and berberine. This study investigates the use of machine learning (ML) to predict expert-level prescriptions of these therapies using both real-world clinical data and synthetic datasets generated with two approaches: Avatar-based (SD_avatar) and Statistical Data Vault (SD_sdv). After automatic and manual feature selection, multiple ML and neural network models were trained and evaluated through cross-validation using the F1-score as the main performance metric. The best-performing configurations were subsequently tested on a held-out dataset. Real-world data (RD) achieved the highest predictive performance (F1 = 0.67), comparable to both the hybrid dataset (RD + SD_avatar, F1 = 0.67) and the SD-avatar dataset alone (F1 = 0.67), while SDV-based data performed significantly lower (F1 = 0.39). These findings confirm the feasibility of modeling prescription behavior using ML, underline the potential of avatar-generated synthetic data, and highlight the remaining gap between structured data and the nuanced reasoning underlying expert clinical prescriptions.
- 21 May 2026
- Published by IOS Press
- Vol. 336, 363-367
Open AccessMachine Learning for Cardiovascular Prevention Prescriptions: Real-World vs. Synthetic Data
Cardiovascular diseases remain the leading cause of death worldwide, primarily driven by atherosclerosis, which is targeted by lipid-lowering agents such as statins and berberine. This study investigates the use of machine learning (ML) to predict expert-level prescriptions of these therapies using both real-world clinical data and synthetic datasets generated with two approaches: Avatar-based (SD_avatar) and Statistical Data Vault (SD_sdv). After automatic and manual feature selection, multiple ML and neural network models were trained and evaluated through cross-validation using the F1-score as the main performance metric. The best-performing configurations were subsequently tested on a held-out dataset. Real-world data (RD) achieved the highest predictive performance (F1 = 0.67), comparable to both the hybrid dataset (RD + SD_avatar, F1 = 0.67) and the SD-avatar dataset alone (F1 = 0.67), while SDV-based data performed significantly lower (F1 = 0.39). These findings confirm the feasibility of modeling prescription behavior using ML, underline the potential of avatar-generated synthetic data, and highlight the remaining gap between structured data and the nuanced reasoning underlying expert clinical prescriptions.
- 14 January 2026
- Published by Oxford University Press (OUP) in European Journal of Preventive Cardiology
Cardiovascular risk among giant cell arteritis patients
Objectives: Giant cell arteritis (GCA) is a chronic large-vessel vasculitis affecting older adults, associated with significant cardiovascular (CV) complications. Understanding CV risk drivers among GCA, including classical CV risk factors and inflammation, is essential for improved patient management. We (i) assessed the association between GCA and risks of CV events, including aortic events, major adverse cardiovascular events (MACE), peripheral artery disease (PAD) events, and visceral artery events and (ii) assessed the impact of traditional CV risk factors and GCA disease activity on these outcomes. Methods and results: Using the French National Health Data System (SNDS), we included 23 193 patients aged ≥50 years diagnosed with incident GCA from 2012 to 2022. Patients were matched with a general population cohort (92 772 individuals) and a hospitalized cohort (92 772 individuals) using propensity score matching based on CV comorbidities. Primary outcomes were first incidence of aortic events, MACE, PAD events, and visceral artery events. Patients with GCA faced an increased risk of all CV outcomes compared with the general population cohort: aortic events (HR: 5.33; 95%CI, 4.50–6.32), MACE (HR: 2.15; 95% CI, 2.04–2.26), PAD events (HR: 2.72; 95%CI, 2.47–2.99), and visceral artery events (HR: 3.04; 95%CI, 2.33–3.97). When compared with the hospitalized cohort, GCA patients had increased risk of all outcomes except for MACE risk, which did not significantly differ (HR: 1.01; 95%CI, 0.96–1.06). GCA disease activity was associated with increased MACE (HR: 1.98; 95%CI, 1.63–2.42). Conclusion: GCA significantly increases risks of vascular complications, highlighting the importance of CV risk management and inflammation control strategies.
- 14 January 2026
- Published by Oxford University Press (OUP) in European Journal of Preventive Cardiology
Cardiovascular risk among giant cell arteritis patients
Objectives: Giant cell arteritis (GCA) is a chronic large-vessel vasculitis affecting older adults, associated with significant cardiovascular (CV) complications. Understanding CV risk drivers among GCA, including classical CV risk factors and inflammation, is essential for improved patient management. We (i) assessed the association between GCA and risks of CV events, including aortic events, major adverse cardiovascular events (MACE), peripheral artery disease (PAD) events, and visceral artery events and (ii) assessed the impact of traditional CV risk factors and GCA disease activity on these outcomes. Methods and results: Using the French National Health Data System (SNDS), we included 23 193 patients aged ≥50 years diagnosed with incident GCA from 2012 to 2022. Patients were matched with a general population cohort (92 772 individuals) and a hospitalized cohort (92 772 individuals) using propensity score matching based on CV comorbidities. Primary outcomes were first incidence of aortic events, MACE, PAD events, and visceral artery events. Patients with GCA faced an increased risk of all CV outcomes compared with the general population cohort: aortic events (HR: 5.33; 95%CI, 4.50–6.32), MACE (HR: 2.15; 95% CI, 2.04–2.26), PAD events (HR: 2.72; 95%CI, 2.47–2.99), and visceral artery events (HR: 3.04; 95%CI, 2.33–3.97). When compared with the hospitalized cohort, GCA patients had increased risk of all outcomes except for MACE risk, which did not significantly differ (HR: 1.01; 95%CI, 0.96–1.06). GCA disease activity was associated with increased MACE (HR: 1.98; 95%CI, 1.63–2.42). Conclusion: GCA significantly increases risks of vascular complications, highlighting the importance of CV risk management and inflammation control strategies.
- Open Access
Is Multimodal Better? A Systematic Review of Multimodal versus Unimodal Machine Learning in Clinical Decision-Making
Machine learning has demonstrated success in clinical decision-making, yet the added value of multimodal approaches over unimodal models remains unclear. This systematic review evaluates studies comparing multimodal and unimodal ML algorithms for diagnosis, prognosis, or prescription. A comprehensive search of MEDLINE up to January 2025 identified 97 studies across 12 medical specialties, with oncology being the most represented. The most common data fusion involved tabular data and images (67%). A risk of bias assessment using PROBAST revealed that 57% of studies had a low risk of bias, while 41% had a high risk. Multimodality outperformed unimodality in 91% cases. No correlation between dataset sample size and added performance has been observed. However, considerable methodological heterogeneity and potential publication bias warrant caution in interpretation. Further research is needed to refine evaluation metrics and hybrid model architectures based on specific clinical tasks. MeSH Terms: Humans [B01.050.150.900.649.313.988.400.112.400.400], Machine Learning [L01.224.050.375.530], Clinical Decision-Making [E01.055], Systematic Review [V03.850].
Open AccessIs Multimodal Better? A Systematic Review of Multimodal versus Unimodal Machine Learning in Clinical Decision-Making
Machine learning has demonstrated success in clinical decision-making, yet the added value of multimodal approaches over unimodal models remains unclear. This systematic review evaluates studies comparing multimodal and unimodal ML algorithms for diagnosis, prognosis, or prescription. A comprehensive search of MEDLINE up to January 2025 identified 97 studies across 12 medical specialties, with oncology being the most represented. The most common data fusion involved tabular data and images (67%). A risk of bias assessment using PROBAST revealed that 57% of studies had a low risk of bias, while 41% had a high risk. Multimodality outperformed unimodality in 91% cases. No correlation between dataset sample size and added performance has been observed. However, considerable methodological heterogeneity and potential publication bias warrant caution in interpretation. Further research is needed to refine evaluation metrics and hybrid model architectures based on specific clinical tasks. MeSH Terms: Humans [B01.050.150.900.649.313.988.400.112.400.400], Machine Learning [L01.224.050.375.530], Clinical Decision-Making [E01.055], Systematic Review [V03.850].
- 15 August 2024
- Published by Elsevier in The Lancet
- Vol. 404 (10453) , 654
Limitations of non-invasive ultrasound therapy for valve stenosis – Authors' reply
- 15 August 2024
- Published by Elsevier in The Lancet
- Vol. 404 (10453) , 654
Limitations of non-invasive ultrasound therapy for valve stenosis – Authors' reply
- Open Access
- 7 August 2024
- Published by MDPI AG in Journal of Clinical Medicine
- Vol. 13 (16) , 4607
Non-Invasive Ultrasound Therapy for Severe Aortic Stenosis: Early Effects on the Valve, Ventricle, and Cardiac Biomarkers (A Case Series)
Background: Transcatheter aortic valve replacement (TAVR) was developed for inoperable patients with severe aortic stenosis. However, despite TAVR advancements, some patients remain untreated due to complex comorbidities, necessitating less-invasive approaches. Non-invasive ultrasound therapy (NIUT), a new treatment modality, has the potential to address this treatment gap, delivering short ultrasound pulses that create cavitation bubble clouds, aimed at softening embedded calcification in stiffened valve tissue. Methods: In the prospective Valvosoft® Serbian first-in-human study, we assessed the safety and efficacy of NIUT and its impact on aortic valve hemodynamics, on the left ventricle, and on systemic inflammation in patients with severe symptomatic aortic stenosis not eligible for TAVR or surgery. Results: Ten patients were included. Significant improvements were observed in hemodynamic parameters from baseline to one month, including a 39% increase in the aortic valve area (from 0.5 cm2 to 0.7 cm2, p = 0.001) and a 23% decrease in the mean transvalvular gradient (from 54 mmHg to 38 mmHg, p = 0.01). Additionally, left ventricular global longitudinal strain significantly rose, while global wasted work significantly declined at one month. A dose–response relationship was observed between treatment parameters (peak acoustic power, intensity spatial-peak pulse-average, and mean acoustic energy) and hemodynamic outcomes. NIUT was safely applied, with no clinically relevant changes in high-sensitivity troponin T or C-reactive protein and with a numerical, but not statistically significant, reduction in brain natriuretic peptide (from 471 pg/mL at baseline to 251 pg/mL at one month). Conclusions: This first-in-human study demonstrates that NIUT is safe and confers statistically significant hemodynamic benefits both on the valve and ventricle.
- 7 August 2024
- Published by MDPI AG in Journal of Clinical Medicine
- Vol. 13 (16) , 4607
Open AccessNon-Invasive Ultrasound Therapy for Severe Aortic Stenosis: Early Effects on the Valve, Ventricle, and Cardiac Biomarkers (A Case Series)
Background: Transcatheter aortic valve replacement (TAVR) was developed for inoperable patients with severe aortic stenosis. However, despite TAVR advancements, some patients remain untreated due to complex comorbidities, necessitating less-invasive approaches. Non-invasive ultrasound therapy (NIUT), a new treatment modality, has the potential to address this treatment gap, delivering short ultrasound pulses that create cavitation bubble clouds, aimed at softening embedded calcification in stiffened valve tissue. Methods: In the prospective Valvosoft® Serbian first-in-human study, we assessed the safety and efficacy of NIUT and its impact on aortic valve hemodynamics, on the left ventricle, and on systemic inflammation in patients with severe symptomatic aortic stenosis not eligible for TAVR or surgery. Results: Ten patients were included. Significant improvements were observed in hemodynamic parameters from baseline to one month, including a 39% increase in the aortic valve area (from 0.5 cm2 to 0.7 cm2, p = 0.001) and a 23% decrease in the mean transvalvular gradient (from 54 mmHg to 38 mmHg, p = 0.01). Additionally, left ventricular global longitudinal strain significantly rose, while global wasted work significantly declined at one month. A dose–response relationship was observed between treatment parameters (peak acoustic power, intensity spatial-peak pulse-average, and mean acoustic energy) and hemodynamic outcomes. NIUT was safely applied, with no clinically relevant changes in high-sensitivity troponin T or C-reactive protein and with a numerical, but not statistically significant, reduction in brain natriuretic peptide (from 471 pg/mL at baseline to 251 pg/mL at one month). Conclusions: This first-in-human study demonstrates that NIUT is safe and confers statistically significant hemodynamic benefits both on the valve and ventricle.
- 31 May 2024
- Published by Elsevier in Archives of Cardiovascular Diseases
- Vol. 117 (5) , 301-303
Non-invasive ultrasound therapy for aortic valve stenosis: State of the art
- 31 May 2024
- Published by Elsevier in Archives of Cardiovascular Diseases
- Vol. 117 (5) , 301-303
Non-invasive ultrasound therapy for aortic valve stenosis: State of the art
- Open Access
- 15 March 2024
- Published by Elsevier in Journal of Cardiovascular Magnetic Resonance
- Vol. 26 (1) , 101030
Four-dimensional flow cardiovascular magnetic resonance aortic cross-sectional pressure changes and their associations with flow patterns in health and ascending thoracic aortic aneurysm
- 15 March 2024
- Published by Elsevier in Journal of Cardiovascular Magnetic Resonance
- Vol. 26 (1) , 101030
Open AccessFour-dimensional flow cardiovascular magnetic resonance aortic cross-sectional pressure changes and their associations with flow patterns in health and ascending thoracic aortic aneurysm
- 14 November 2023
- Published by Elsevier in The Lancet
- Vol. 402 (10419) , 2317-2325
Treatment of severe symptomatic aortic valve stenosis using non-invasive ultrasound therapy: a cohort study
- 14 November 2023
- Published by Elsevier in The Lancet
- Vol. 402 (10419) , 2317-2325
Treatment of severe symptomatic aortic valve stenosis using non-invasive ultrasound therapy: a cohort study
- Open Access
- 6 October 2023
- Published by American Association for the Advancement of Science (AAAS) in Science Advances
- Vol. 9 (40) , eadi4252
The fundamental mechanisms of the Korotkoff sounds generation
Blood pressure measurement is the most widely performed clinical exam to predict mortality risk. The gold standard for its noninvasive assessment is the auscultatory method, which relies on listening to the so-called “Korotkoff sounds” in a stethoscope placed at the outlet of a pneumatic arm cuff. However, more than a century after their discovery, the origin of these sounds is still debated, which implies a number of clinical limitations. We imaged the Korotkoff sound generation in vivo at thousands of images per second using ultrafast ultrasound. We showed with both experience and theory that Korotkoff sounds are paradoxically not sound waves emerging from the brachial artery but rather shear vibrations conveyed in surrounding tissues by the nonlinear pulse wave propagation. When these shear vibrations reached the stethoscope, they were synchronous, correlated, and comparable in intensity with the Korotkoff sounds. Understanding this mechanism could ultimately improve blood pressure measurement and provide additional understanding of arterial mechanical properties.
- 6 October 2023
- Published by American Association for the Advancement of Science (AAAS) in Science Advances
- Vol. 9 (40) , eadi4252
Open AccessThe fundamental mechanisms of the Korotkoff sounds generation
Blood pressure measurement is the most widely performed clinical exam to predict mortality risk. The gold standard for its noninvasive assessment is the auscultatory method, which relies on listening to the so-called “Korotkoff sounds” in a stethoscope placed at the outlet of a pneumatic arm cuff. However, more than a century after their discovery, the origin of these sounds is still debated, which implies a number of clinical limitations. We imaged the Korotkoff sound generation in vivo at thousands of images per second using ultrafast ultrasound. We showed with both experience and theory that Korotkoff sounds are paradoxically not sound waves emerging from the brachial artery but rather shear vibrations conveyed in surrounding tissues by the nonlinear pulse wave propagation. When these shear vibrations reached the stethoscope, they were synchronous, correlated, and comparable in intensity with the Korotkoff sounds. Understanding this mechanism could ultimately improve blood pressure measurement and provide additional understanding of arterial mechanical properties.