SCMR Registry

There are two ways to participate in the SCMR Registry: as a data contributor, and as a researcher.
More information on the processes required for both can be found in these slides:

How to participate in the SCMR Registry | PDF

If you wish to join and contribute data to the Registry, the first step is for your institution to sign the Participation Agreement:

SCMR Registry Participation Agreement | PDF

Participating Centers

• Ascension St. Vincent Southside Hospital, Jacksonville, Florida, USA
• Carolinas Medical Center, Charlotte, North Carolina, USA
• Cedars Sinai Medical Center, Los Angeles, California, USA
• Cleveland Clinic, Cleveland, Ohio, USA
• Connecticut Children’s Medical Center, Connecticut, USA
• Duke University, Durham, North Carolina, USA
• Georgetown University, Washington, D.C., USA
• Houston Methodist Research Institute, Houston, Texas, USA
• University of Illinois Chicago, Chicago, Illinois, USA
• Indiana University, Bloomington, Indiana, USA
• Kings College, London, England, UK
• University of Minnesota, Minneapolis, Minnesota, USA
• New York-Presbyterian Brooklyn Methodist Hospital, Brooklyn, New York, USA
• Ohio State University, Columbus, Ohio, USA
• Piedmont Heart Institute, Atlanta, Georgia, USA
• Salinas Valley Memorial Hospital, Salinas, California, USA
• Seton Heart Institute, Austin, Texas, USA
• St. Vincent Heart Center of Indiana, Indianapolis, Indiana, USA
• University Hospitals of Cleveland, Cleveland, Ohio, USA
• Vanderbilt University, Nashville, Tennessee, USA
• Virginia Commonwealth University, Richmond, Virginia, USA
• Weill-Cornell Medical Center, New York, New York, USA

COVID-19 Participating Centers

• Elisabeth-Krankenhaus, Essen, Germany
• German Heart Institute Berlin and Charite University Medicine Berlin, Berlin, Germany
• Leeds Institute of Cardiovascular and Metabolic Medicine, Advanced Imaging Centre
• University of Medical Sciences in Poznan, Poznan, Poland
• University of Oxford, Oxford, United Kingdom
• University of Pennsylvania, Philadelphia, PA, USA

Project: Pulmonary Hypertension Outcomes and Risk Assessment (PHORA) imaging model
Principal Investigator(s): Raymond Benza, MD, Charles Fauvel, MD
Lead Institution(s): Ohio State University

We aim to build an imaging model based on combination of several CMR variables to improve pre-existing clinical model for pulmonary arterial hypertension (PAH) patients (PHORA 2.0). Accurate risk stratification is essential before making individualized treatment decisions for pulmonary arterial hypertension. Several PAH risk stratification tools are available nowadays but remain imperfect since they have neglecting modern tools such as cardiac magnetic resonance. Further, we are now facing more complex patients that the “classical PAH phenotype description from which traditional risk variables were derived. Our ongoing work demonstrates the advantages of machine learning methodology in overcoming several of these limitations. In addition, right heart dysfunction is closely linked to PAH patient’s prognosis but it is still disregarded in actual risk assessment tools. Public health impact. This study represents the largest and most powerful imaging risk assessment model among PAH patients. It will allow physician to make appropriate diagnostic work up, appropriate risk stratification and then tailor individualized therapeutic decisions.

Project: Right Ventricular Thrombus on Cardiac Magnetic Resonance Imaging: Prevalence, Markers and Long Term Outcomes in Patients with Right Ventricular Systolic Dysfunction
Principal Investigator(s): Chetan Shenoy, MD
Lead Institution(s): University of Minnesota

Right ventricular thrombus has been described in case reports but there have been no systematic studies yet because it is difficult to reliably detect using echocardiography. Our results would provide the prevalence, which would tell us if it is a common condition. They will identify the markers, which would tell us which patients would benefit from CMR in place or, or in addition to echocardiography. Finally, they will identify the prognostic implications, which would inform us if it is a clinically important condition.

Project: Sex-based Differences in Left Ventricular Remodeling in Patients with Chronic Aortic Regurgitation
Principal Investigator(s): Deborah Kwon, MD
Lead Institution(s): The Cleveland Clinic

CMR arguably provides the most comprehensive phenotypic evaluation of patients with chronic aortic regurgitation, providing gold standard measurements of LV volume, quantitative measurements of AR volume and regurgitant fraction, and precise evaluation of the thoracic aorta for the assessment of concomitant aortopathy.

However, there are no definitive clinical/societal guidelines for defining CMR thresholds to categorize severity of aortic regurgitation, or threshold of LV dilation measured by CMR included in the criteria for surgical indications for aortic valve intervention. Routine CMR assessment in the setting of moderate or severe aortic regurgitation CMR may be useful in assessing LV volumes and function as well as quantifying AR, particularly when it is not clear if patients fulfill adequate indications for surgical intervention. Currently, clinical guidelines do not include CMR measurements of left ventricular dilation as criteria for identifying patients who should be referred for surgical intervention.

Project: Prevalence and Prognostic Significance of Small Myocardial Infarcts Detected by CMR in 18,000 Patients with Normal Contractile Function: A Multicenter Study with Ten Years of Follow-up
Principal Investigator(s): Han W. Kim, MD
Lead Institution(s): Duke University

This research project includes patients who underwent magnetic resonance imaging (MRI) of the heart (also known as cardiovascular magnetic resonance or cardiac MRI) at several hospitals within the United States. The aim is to understand how often heart damage or scarring occurs in patients who appear to have normally functioning hearts. In addition to determining how commonly scarring is present, the study will examine if the presence of scarring impacts patient longevity in the long term. If a long term impact can be identified, researchers could then evaluate strategies to reduce the likelihood of developing scarring or employ treatments for people who are at risk for developing scarring.

Project: Cardiac Magnetic Resonance to Assess Impact of Aortic Regurgitation and Myocardial Scar on Clinical Outcomes The AR SCAR Multicenter Study
Principal Investigator(s): Dipan Shah, MD; Maan Malahfji, MD
Lead Institution(s): Houston Methodist Institute

The impact of myocardial scarring on survival in chronic AR has not been well established, and the effect of AVR on this association has not been previously studied. We hypothesize that myocardial scarring detected by CMR is independently associated with mortality in patients with chronic AR. We propose a multicenter study to evaluate the association of myocardial scar with ventricular remodeling and outcomes in chronic aortic regurgitation. Specific objectives will be: 1) To study the prevalence, burden, and predictors of myocardial scar by CMR in a large population of patients with chronic AR. 2) Investigate the association of myocardial scar with total mortality, cardiovascular death, and development of symptoms or other indications for AVR. 3) Explore the impact of management strategy (AVR vs. medical management) on outcomes in patients with AR and myocardial scar. 4) Examine the association of myocardial scarring with AR related ventricular remodeling.

Project: Tetralogy of Fallot Biventricular Shape Atlas
Principal Investigator(s): Michael Jay Campbell, MD; Alistair Young, MD
Lead Institution(s): Duke University; King’s College London

Tetralogy of Fallot is a common type of congenital heart disease, which is in turn the most common type of birth defect. Although surgical corrections have improved, there is a critical need for enhanced evaluation in these patients. An important clinical decision is whether to replace the pulmonary valve, which is often damaged by the surgical correction process. If left too late, there is a chronic regurgitation of blood back into the right ventricle, leading to clinical decline and heart failure. If performed too early, additional interventions may be required later, and complications arising from multiple surgeries over time can lead to adverse outcomes. This proposal will develop new methods for evaluating heart function in patients with tetralogy of Fallot, which can be used to better optimize the timing of interventions.

Project: Using Neural Networking to Develop a Novel Cardiac Magnetic Resonance-based Model to Predict Cardiac Amyloid Subtype
Principal Investigator(s): Jeremy A. Slivnick, MD; Karolina M. Zareba, MD
Lead Institution(s): Ohio State University

Cardiac Amyloidosis (CA) most commonly occurs due to deposition of either abnormally folded immunoglobulin (AL-CA) or transthyretin (TTR-CA) protein in the myocardium. Potentially life-saving treatment are now available for CA but differ greatly based on which of the two subtypes is present. Therefore, the accurate differentiation of CA subtype is crucial to streamline effective care in this population. Currently, determining CA subtype requires either nuclear scans or invasive tissue biopsy. However, additional testing may be costly and potentially be associated with complications. While CMR is commonly utilized in the evaluation of CA due to its diagnostic and prognostic utility, its role in determining CA subtype is currently unclear. The purpose of our study is to utilize machine learning to develop a Cardiac Magnetic Resonance (CMR) based model to differentiate CA subtype. We hope to leverage the power of the Society of CMR (SCMR) registry to identify a large enough cohort of patients to develop a neural network for this purpose. This project is anticipated to take 9 months. If successful, a CMR-based model could potentially improve cost-effectiveness and reduce complications in the diagnostic evaluation of patients with CA.

Project: Platform-independent Automated Analysis of Stress CMR Perfusion Datasets Using Deep Learning Without the Need for Motion Correction
Principal Investigator(s): Behzad Sharif, PhD
Lead Institution(s): Indiana University, School of Medicine, Indianapolis

We have recently developed a deep-learning approach for rapid and objective analysis of perfusion datasets that does not require motion correction and is highly robust to poor image quality and variation in sequence parameters. We propose to apply this approach for automatic analysis of a large subset of Stress Perfusion CMR studies, i.e., first-pass perfusion image series available in the SCMR Registry from multiple sites and multiple vendors. The goal of the project is to evaluate the ability of the developed A.I. models to objectively assess the presence and prognostic significance of stress-induced perfusion defects. This project has the potential to create and disseminate an open-source vendor-agnostic A.I. tool that is robust to variations in image quality and automatically detects/grades stress-induced perfusion defects. All results and trained A.I. models will be published as open source for the wider SCMR community to use including for future multi-vendor/multi-site clinical trials.

Project: AI Algorithm Development for Quantitative Cardiac Structure and Function Analysis in the SCMR Registry
Principal Investigator(s): Qian Tao, Ph.D.
Lead Institution(s): Delft University of Technology

The purpose of this research project is to develop and validate a highly reliable artificial intelligence (AI) method that can automatically extract quantitative structural and functional parameters from short-axis cine MRI in the SCMR Registry. By integrating this AI tool into the SCMR Registry, we aim to provide a standardized, reliable, and objective solution to extract the quantitative parameters from the high-volume, high-complexity imaging data. The solution is not only meant for the existing data in the SCMR Registry, but also for upcoming data in a scalable manner. Ultimately, we aim to make the quantitative CMR parameters accessible to the SCMR community to facilitate a wide range of cardiovascular disease research.

Technical research goals include: (1) to address the domain shift problem in cardiac MRI and make the algorithm vendor- and center-independent, and (2) to investigate uncertainty metrics in addition to the quantification as a necessary means of quality control in rare cases of AI failure.

Project: Functional Tricuspid Regurgitation Quantified by Cardiovascular Magnetic Resonance
Principal Investigator(s): Dipan J. Shah; Andrada C. Guta
Lead Institution(s): Houston Methodist DeBakey Heart and Vascular Center

This study is sought to evaluate the impact of FTR on mortality using CMR in a large multi-center cohort. FTR is highly prevalent and is associated with increased morbidity and mortality, independently of the etiology. More than 90% of the TR cases are secondary right heart remodeling. Surgical interventions for the tricuspid valve are rather rare and carry a high risk. Several percutaneous options are under investigation; however, patient selection criteria are vague and accurate quantification of TR is essential. Additionally, different FTR etiologies associate with different outcomes and require distinctive management. When evaluating FTR, several challenges emerge. Current thresholds for severe TR are extrapolated from the mitral valve and quantification of TR severity using echocardiography is objectively challenging due to the dynamicity of TR throughout systole and the respiratory cycle with unpredictable regurgitant orifice geometry. CMR can accurately quantify the regurgitant volume and regurgitant fraction without any anatomical and functional assumptions. CMR is the gold standard when valuating right heart size and function showing different remodeling patterns that associate with FTR. This project will extend over 5 years. Results will help identifying a more accurate grading system for FTR severity and better patient selection for tricuspid valve intervention.

1. Review both the SCMR Registry Data Access Policy  and the SCMR Registry Data Access Policy – Highlights.
2. Submit a Search Request Form.
3. Search requests will be reviewed according to the Search Request Review Process.
4. Once your research project is approved, submit a  Data Access Application.
5. Requests will be scored based-on set criteria described in the  Data Access Review Process and  Score Sheet.