CMR-Analysis (including machine learning)
Andrew Loe
Summer Student
Libin Cardiovascular Institute of Alberta, University of Calgary
Okotoks, Alberta, Canada
Fereshteh Hasanzadeh, MSc
Master Student in Cardiovascular and Respiratory Science
Cumming School of Medicine, University of Calgary
Calgary, Alberta, Canada
Dina Labib, MD, PhD, FSCMR
PhD student
Libin Cardiovascular Institute of Alberta, University of Calgary
Calgary, Alberta, Canada
Steven Dykstra, MSc, BSc
PhD Student
Libin Cardiovascular Institute of Alberta, University of Calgary
Calgary, Alberta, Canada
Rylan Marianchuk
Intern
University of Calgary, Canada
Sandra Rivest, RN
Research nurse
Libin Cardiovascular Institute of Alberta, University of Calgary
Calgary, Alberta, Canada
Jacqueline Flewitt, MSc
Research Collaborations Coordinator
Libin Cardiovascular Institute of Alberta, University of Calgary
Calgary, Alberta, Canada
Andrew G. Howarth, MD, PhD
Clinical Co-director
Libin Cardiovascular Institute of Alberta, University of Calgary
Calgary, Alberta, Canada
Carmen P. Lydell, MD
Clinical Co-director
Libin Cardiovascular Institute of Alberta, University of Calgary
Calgary, Alberta, Canada
Bobby Heydari, MD
Associate Professor
University of Calgary
Calgary, Alberta, Canada
James White, MD
Professor of Cardiology
Stephenson Cardiac Imaging Centre
Calgary, Alberta, Canada
Easter Prosia, MD
Core Lab Technician
University of Calgary, Canada
Rosa Sandonato, RN
General Associate AHS Research/Clinical Admin
University of Calgary
Calgary, Alberta, Canada
Louis Kolman, MD
Clinical Co-director
Libin Cardiovascular Institute of Alberta, University of Calgary, Alberta, Canada
The burden of fibrosis by CMR has been associated with elevated risk of clinical outcomes in hypertrophic cardiomyopathy (HCM). However, obstacles to adoption have been acknowledged for signal threshold-based quantification of late gadolinium enhancement (LGE) in this population. HCM commonly displays patchy LGE with inconsistent reference tissue signal profiles, making threshold-based approaches subjectively influenced by user adjustment of this region, often aimed at matching visual perceptions of LGE. Whether visual perception alone, when guided by a standardized framework, can deliver risk stratification in HCM has not been previously explored. In this study we tested a graphical 72 sub-segmental fibrosis scoring interface within routine practice for standardized reporting of visually estimated LGE (veLGE) burden. We assessed capacity for this marker to predict major adverse cardiovascular outcomes (MACE).
Methods:
Patients with CMR-confirmed diagnosis of HCM were identified from the CIROC Registry. All completed a standardized CMR protocol inclusive of LGE imaging. CMR images were clinically interpreted using a standardized reporting interface (cardioDITM, Cohesic Inc., Calgary) with fibrosis scored by presence, pattern(s) and with LGE burden visually scored to estimate veLGE (Figure 1), this expressed as a percentage of 72 LV subsegments (4 subsegments assigned to each of 16 AHA segments, 1 to the true apex, and 2 to each RV insertion point). Patients were followed for the composite outcome of all-cause death, survived sudden cardiac arrest, ventricular tachycardia, heart failure hospitalization, and new-onset atrial fibrillation. Univariable and multi-variable analyses were performed to identify associations between visually estimated fibrosis burden and time to first MACE outcome.
Results: A total of 586 patients with HCM were studied. Baseline characteristics, health profiles and cardiac MRI characteristics are shown in Table 1. At a median follow-up of 1153 days (IQR 531-1752), 73 patients (13%) experienced the primary composite outcome. Patients with an event were older, had more comorbidities, and poorer renal function. Patients with events had higher indexed LV mass and LA volumes. Any LGE was scored in 77% of patients with mid-wall patchy the predominant pattern. Patients with events showed significantly higher veLGE (12.4% vs 8.6%; HR=1.05 per 1%, p< 0.001). Following adjustment for age, HTN, DM, GFR, indexed LV mass, and indexed LA volume, veLGE remained independently associated with the primary outcome (aHR 1.03 per 1%, p=0.03). Figure 2 shows Kaplan Meier event-free survival curves stratified by a veLGE threshold of ≥5%, the latter based on prior publications, demonstrating cumulative event rates of 17% versus 11% (HR 1.60, p=0.04) at the median follow-up of 1153 days.
Conclusion:
In patients with HCM, veLGE is a pragmatic and prognostically relevant marker of future cardiovascular outcomes that can be effectively incorporated into routine clinical reporting.