Measuring extracellular volume fraction by MRI: First verification of values given by clinical sequences

Abstract

Purpose: To verify MR measurements of myocardial extracellular volume fraction (ECV) based on clinically applicable T1-mapping sequences against ECV measurements by radioisotope tracer in pigs and to relate the results to those obtained in volunteers.

Methods: Between May 2016 and March 2017, 8 volunteers (25 ± 4 years, 3 female) and 8 pigs (4 female) underwent ECV assessment with SASHA, MOLLI5(3b)3, MOLLI5(3s)3, and MOLLI5s(3s)3s. Myocardial ECV was measured independently in pigs using a radioisotope tracer method.

Results: In pigs, ECV in normal myocardium was not different between radioisotope (average ± standard deviation; 19 ± 2%) and SASHA (21 ± 2%; P = 0.086). ECV was higher by MOLLI5(3b)3 (26 ± 2%), MOLLI5(3s)3 (25 ± 2%), and MOLLI5s(3s)3s (25 ± 2%) compared with SASHA or radioisotope (P ≤ 0.001 for all). ECV in volunteers was higher by MOLLI5(3b)3 (26 ± 3%) and MOLLI5(3s)3 (26 ± 3%) than by SASHA (22 ± 3%; P = 0.022 and P = 0.033). No difference was found between MOLLI5s(3s)3s (25 ± 3%) and SASHA (P = 0.225). Native T1 of blood and myocardium as well as postcontrast T1 of myocardium was consistently lower using MOLLI compared with SASHA. ECV increased over time as measured by MOLLI5(3b)3 and MOLLI5(3s)3 for pigs (0.08% and 0.07%/min; P = 0.004 and P = 0.013) and by MOLLI5s(3s)3s for volunteers (0.07%/min; P = 0.032) but did not increase as measured by SASHA.

Conclusion: Clinically available MOLLI and SASHA techniques can be used to accurately estimate ECV in normal myocardium where MOLLI-sequences show minor overestimation driven by underestimation of postcontrast T1 when compared with SASHA. The timing of imaging after contrast administration affected the measurement of ECV using some variants of the MOLLI sequence.

Keywords: ECV; MOLLI; SASHA; T1-mapping; extracellular volume.

Figure 1

ECV in pigs by CMR and isotope and in volunteers by CMR. Note the higher ECV as measured by MOLLI versus both isotope and SASHA in pigs and versus SASHA in volunteers. The error bars show mean ± standard deviation. One symbol: P < 0.05. Two symbols: P < 0.01. Three symbols: P < 0.001

Figure 2

Modified Bland‐Altman plots of the difference in ECV by radioisotope and different T1 mapping techniques on the y‐axis and ECV by radioisotope on the x‐axis using data from 8 pigs. Note that SASHA shows a lesser bias than MOLLI compared with the reference method

Figure 3

Example images of the first TI in the series for pre‐ and postcontrast T1‐mapping sequences. The postcontrast images are taken from 0 to 10 min after contrast administration. Regions of interest were added for blood (yellow) and myocardial septum (blue) and then copied to all TIs for all acquisition time points in that subject. Red lines show endocardium and green lines show epicardium

Figure 4

Typical case in a porcine experiment and in a volunteer of postcontrast T1 values for myocardium to the left and blood to the right. Note that the MOLLI sequences show lower T1 in myocardium compared with SASHA and that there is no difference in T1 between the sequences for blood. For data on all the experiments and volunteers, see Supporting Information Figures S7 and S8

Figure 5

ECV over time after Gd calculated using 4 different T1‐mapping sequences. Red shows ECV of healthy volunteers (n = 8), green shows ECV of pigs (n = 8), and the black box shows ECV as measured by radioisotope in pigs. The first ECV‐measurements are shown at the actual mean time of acquisition while, for visualization purposes, the following data points are shown at intervals of 10 min. Error bars show mean ± standard deviation. Note that all MOLLI sequences (panels A‐C) show consistently higher ECV compared with the radioisotope reference method while SASHA agrees more closely with radioisotope. For information on individual data of ECV by sequence for each experiment and volunteer, please see Supporting Information Figure S6