Computed tomography vs cardiovascular magnetic resonance imaging derived extracellular volume fraction in patients with stable new-onset chest pain

Background

Computed tomography (CT) is increasingly recognised as a diagnostic modality across a range of cardiovascular conditions and is now first-line for the investigation of stable new-onset chest pain. Determination of the myocardial extracellular volume fraction (ECV) has been shown to correlate well with the identification and prognostication of disease. Cardiovascular magnetic resonance (CMR) imaging remains the gold standard for the measurement of myocardial ECVCMR using T1-mapping, but there is increasing evidence for the use of ECV by cardiac CT (ECVCT).

Purpose

To assess the performance of ECVCT against the reference standard of ECVCMR.

Methods

Patients with a history of chest pain and no previously documented coronary disease referred for invasive angiography were recruited as part of the EVINCI Heart-QIT study. A cohort of these patients (n=33) underwent CMR at 1.5T (Siemens Aera, Siemens Healthcare, Erlangen/Germany) with T1 mapping of a mid-ventricular short axis slice (by MOdified Look-Locker Inversion recovery [MOLLI]) before and 15 minutes after a bolus of gadolinium contrast (0.1 mmol/kg gadoterate meglumine), followed by whole-heart ECVCT quantification (Somatom Force, Siemens Healthcare, Erlangen/Germany) using a 5-min post-iodine-contrast acquisition protocol. To account for data clustering on a patient level and volumetric discrepancy on a modality level, comparisons were made using mid-ventricular pooled ECVCT and ECVCMR. Bland-Altman analysis was used to determine the limits of agreement and identify systematic differences between both measures.

Results

A total of 33 patients (70% male, mean age 56.8±12.6yr) underwent the combined CMR and CT. ECVCMR and ECVCT were then analysed retrospectively (Figure 1). The average pooled ECV for the 6 mid-ventricular segments for CMR and CT were (27.6±2.4 and 26.8±2.2 respectively). Bland-Altman analysis demonstrated a marginally higher CMR-ECV (0.8±2.1) vs CT-ECV, which is in keeping with the longer delay-time encountered in CMR protocols (Figure 2).

Conclusions

ECVCT obtained from 5-minute post-contrast CT protocols show good agreement with ECVCMR in a stable chest pain patient cohort.

Funding Acknowledgement

Type of funding sources: Private grant(s) and/or Sponsorship. Main funding source(s): Siemens Helthineers Educational Grant Figure 1. CMR (L) and CT (R) ECV mapsFigure 2. Bland-Altman plot