Valvular Heart Disease
Christian Nitsche, MD, PhD
Cardiology Registrar
Barts Heart Centre at St Bartholomew's Hospital, United Kingdom
Matthias Koschutnik, MD
Cardiology Registrar
Medical University of Vienna, Austria
Hunain Shiwani, MD
Clinical Research Fellow
University College London and Barts Heart Centre, United Kingdom
Andreas Kammerlander, MD, PhD
Cardiology Registrar
Medical University of Vienna, Austria
Carolina Dona, MD
Cardiology Registrar
Medical University of Vienna, Austria
George D. Thornton, MBBS
Clinical Research Fellow
University College London, United Kingdom
Jonathan B. Bennett, MBBS
Clinical Research Fellow
University College London, United Kingdom
Kelvin Chow, PhD
Staff Scientist
Siemens Healthineers
Chicago, Illinois, United States
Hui Xue, PhD
Director, Imaging AI Program
National Institutes of Health
Bethesda, Maryland, United States
Peter Kellman, PhD
Senior Scientist
National Institutes of Health, Maryland, United States
James C. Moon, MD
Clinical Director, Imaging
Barts Heart Centre and UCL
London, England, United Kingdom
Julia Mascherbauer, MD
Professor
Medical University of Vienna, Austria
Thomas A. Treibel, MD, PhD
Consultant Cardiologist
University College London, England, United Kingdom
Rhodri Davies, MD, PhD
Associate clinical professor
University College London
London, Wales, United Kingdom
In aortic stenosis (AS) longitudinal function by echocardiography is a more sensitive marker for left ventricular (LV) systolic dysfunction than ejection fraction (LVEF). Global longitudinal shortening (GLS) and mitral annular plane systolic excursion (MAPSE) by cardiac magnetic resonance (CMR) represent markers of LV longitudinal function, but measurement is subjective and not frequently performed because of the lack of automated analysis. Automated artificial intelligence (AI)-measured MAPSE and GLS have recently been shown to yield highly reproducible results, outperforming clinical experts.1 We hypothesized that AI-based MAPSE and GLS could convey important prognostic information in severe AS beyond LVEF and identify early markers of adverse remodeling.
Methods: Patients with severe AS scheduled for treatment discussion in an interdisciplinary heart team board were prospectively recruited at a single center and underwent CMR on a 1.5T scanner (MAGNETOM AvantoFit, Siemens). An AI-based algorithm 1 was used for automated detection of landmark points from long-axis cine images covering the entire cardiac cycle (inferoseptal and anterolateral mitral annular hinge points from 4-chamber (4ch) view, anterior and inferior points from 2-chamber (2ch) view, apex from both views) to compute MAPSE and GLS (Figure 1). All LV volume measurements were made by a clinically proven AI algorithm.2 All-cause mortality served as the primary endpoint.
Results:
In total, 379 AS patients (80±7years, 50% female) were included. Management according to heart team decision was: transcatheter aortic valve implantation (n=328), surgical aortic valve replacement (n=17), and conservative care (n=34). Mean MAPSE2ch was 9.5±4.2mm, MAPSE4ch 9.9±2.8mm, GLS2ch 11.7±3.9%, GLS4ch 12.9±4.2%, and LVEF 59.5±15.4%.
After 2.2±1.5 years, 32% (n=120/379) had died. All markers of longitudinal function were significant univariate predictors of outcome. MAPSE4ch showed the strongest association; receiver operating characteristic curve determined 8.1mm as the optimal prognostic cut-off value (log-rank: p< 0.00001; hazard ratio [HR] 2.3 (95% confidence interval 1.6-3.4); Figure 2A). Preserved LVEF (≥50%) with reduced MAPSE ( < 8.1mm) was observed in 37 patients. After multivariate Cox regression analysis (adjustment for demographics, comorbidities, valve replacement, NT-proBNP, LVEF), MAPSE4ch≤8.1mm remained significantly associated with mortality (adj HR 2.0 [1.3-2.9]). This association was consistent in patients with preserved LVEF (log-rank: p=0.004, adj HR 1.9 [1.1-3.3]; Figure 2B).
Conclusion: In severe AS, AI-measured MAPSE from CMR is readily available, detects early LV dysfunction before a reduction in LVEF, and provides incremental prognostic information. MAPSEAI may therefore have potential to serve as an important marker in the management process of severe AS.