Parametric Imaging and Fingerprinting

1356288 - Parametric Mapping Standardization in the DCM Precision CMR Study: A Real-World Multi-Site Experience

Saturday, January 28, 2023

9:55 AM – 10:10 AM

Location: Silver Pearl 1

Presenting Author(s)

SS

Suzanne Smart, BSc, CCRP

Research Associate
The Ohio State University, United States

Primary Author(s)

SS

Suzanne Smart, BSc, CCRP

Research Associate
The Ohio State University, United States

Co-Author(s)

Orlando P. Simonetti, PhD

Professor, Medicine and Radiology
The Ohio State University
Columbus, Ohio, United States
Ray Hershberger, MD

Professor
The Ohio State University, United States
YL

Yingmin Liu, PhD

Research Associate
The Ohio State University
Columbus, Ohio, United States
NJ

Ning Jin, PhD

Senior Key Expert
Siemens Medical Solutions USA, Inc., Ohio, United States
Juliet Varghese, PhD

Research Assistant Professor
The Ohio State University
Columbus, Ohio, United States
Karolina M. Zareba, MD

Associate Professor
The Ohio State University
Columbus, Ohio, United States

Disclosure(s):

Orlando P. Simonetti, PhD: No relevant disclosure to display

Ray Hershberger, MD: No financial relationships to disclose

Karolina M. Zareba, MD: No financial relationships to disclose

Background: Dilated cardiomyopathy (DCM) often has a genetic basis and detecting early evidence of DCM in at-risk family members holds promise to prevent advanced disease. Cardiovascular magnetic resonance (CMR) including functional and tissue characterization has the potential to detect early disease (1). The DCM Precision CMR study was designed to determine if CMR may detect a pre-DCM phenotype in individuals at increased genetic risk. As part of this study, T1 and T2 mapping is measured across multiple sites, necessitating standardization across CMR scanners.

Methods: To ensure consistent CMR image and data quality across sites, imaging was restricted to 1.5T systems from a single MRI vendor (Siemens Healthcare, Erlangen, Germany). Each site was provided a study protocol customized specifically for their scanner model and software version (Table 1). We utilized two T1MES (T1 Mapping ECV Standardization Program) phantoms (Resonance Health, Burswood, Australia) to test mapping consistency across 6 sites (2,3). The two phantoms were initially scanned at study’s CMR Core Lab, and subsequently shipped to each site during site initiation. Sites were provided specific scan parameters and phantom scan instructions to reduce variability. The scan protocol includes a high-resolution 3D sequence (to confirm phantom integrity), followed by 10 repetitions each of T2, pre-contrast T1, and post-contrast T1 mapping sequences matched across all centers, and identical to those used to scan study participants. All maps were analyzed using validated software (cvi42, Circle Cardiovascular Imaging, Calgary). A circular region of interest (area of 3 cm²) was placed in the center of each of the nine phantom tubes (six for T2) and the mean value was recorded. Each tube’s 10 repeat values were then averaged and compared to the values obtained at the Core Lab on the same phantom. If any tubes exhibited >5% difference from the Core Lab reference, the images and scan parameters were reviewed and adjusted if necessary, and the phantom rescanned.

Results: The T1MES phantoms were scanned at 6 sites across the US during the site initiation process. Both T1MES phantoms had comparable T1 and T2 values (≤5% difference) when scanned at the Core Lab. Using identical pulse sequences and scan parameters, all 6 sites demonstrated consistent mapping values within ≤5% of the Core Lab standard (Figure 1). To test T1MES phantom stability over time, the phantom was rescanned at Core Lab twelve months later and demonstrated consistent values (Table 2). This will be performed at all sites on a yearly basis.

Conclusion: CMR based tissue characterization across multiple sites is feasible when pulse sequence design and scan parameters are carefully matched. Utilizing the T1MES phantom for standardization, initial data confirmed consistency across participating sites for T1 and T2 mapping within the proposed 5% deviation. Additionally, T1MES phantom stability over time was also demonstrated.