T1 and extracellular volume fraction mapping in cardiac magnetic resonance: estimation of accuracy and precision of a novel algorithm. (11th February 2019)

• Record Type: Journal Article
• Authors: Borrazzo, Cristian; Pacilio, Massimiliano; Galea, Nicola; Preziosi, Enrico; Carnì, Marco; Francone, Marco; Catalano, Carlo; Carbone, Iacopo
• Is Part Of: Physics in medicine & biology. Volume 64:Number 4(2019:Feb.)
• Publication Date: 2019-02-11
• Subjects: T1 mapping -- ECV mapping -- extracellular volume fraction -- cardiac magnetic resonance -- MOLLI sequence; Biophysics -- Periodicals; Medical physics -- Periodicals; 610.153
• DOI: 10.1088/1361-6560/aafcca
• Languages: English
• ISSNs: 0031-9155

Abstract

Abstract: Magnetic resonance imaging is effective for non-invasive detection of myocardial diseases by extracellular volume fraction (ECV) estimation. A new methodology for T1 and ECV mapping is tested in this work, comparing results with other well-consolidated methods. The associated level of uncertainty for data was also estimated, to assess the reliability of the technique. A phantom with known T1 values was used for reference, and 60 subjects (40 controls, 20 diseased patients) were examined, using the modified look-locker inversion-recovery (MOLLI) method. Obtained T1 data were studied in terms of accuracy (difference with reference T1), paired comparisons with other methods, and Gamma-tool analysis with tolerances criteria of 4.21 mm for distance-to-agreement, and between 2%–5% for T1 difference. Accuracy and precision of the T1 mapping was analysed by phantom measurements, and the uncertainty of the ECV was estimated by T1 error propagation. Differences (in paired comparisons) of T1 datasets were not significant neither for statistical tests, nor for Bland–Altman analysis. T1 accuracy was between −12% and −1% across methods, slightly better for the tested method (mean accuracy in the T1 range of interest better than 2%). The Gamma analysis confirm slightly better results for the tested method than other methodologies. The new method improves the computational efficiency by a factor of 25 (at least), revealing to be also more suitable for Big Data-related applications.