Fast myocardial T₁ mapping using shortened inversion recovery based schemes

Figure 1. Numerical simulations of T1 accuracy and precision of all T1 mapping schemes. T1 accuracy (a) and precision (b) are shown as a function of T1 using a typical myocardial T2 time of 45 msec for all T1 mapping schemes. Impact of T2 times on T1 accuracy (c) and precision (d) are shown for 2P‐2 and MOLLI.

Figure 2. T1 accuracy (a), spatial variability (b), and repeatability (c) of all T1 mapping schemes in phantom experiments. There were no statistically significant differences between all schemes in terms of accuracy (P = 1.00), spatial variability (P = 0.71) and repeatability (P = 0.75).

Figure 3. Example native T1 maps of a 29‐year‐old male healthy volunteer (HR 52 bpm) along the short axis using all T1 mapping schemes. Similar image quality and native T1 range were obtained across all slices using all schemes.

Figure 4. Native T1 times, spatial variability, and repeatability for myocardium (a–c, respectively) and blood (d–f, respectively) obtained using all T1 mapping schemes in 16 healthy volunteers. Average (bar plots) and SD (error bars) over all healthy volunteers are presented. There were no statistically significant differences between native myocardial and blood T1 times (P = 0.21 and P = 0.79, respectively) and repeatability (P = 0.87 and 0.41, respectively) obtained using all schemes. All shortened T1 mapping schemes led to increased myocardial and blood T1 spatial variability with respect to MOLLI (P < 0.0001).

Figure 5. Segment‐wise native myocardial T1 times, spatial variability, and repeatability using 2P‐2 and MOLLI in 16 healthy volunteers. Data are shown as average ± SD over all healthy volunteers. No statistically significant differences were found between segmental values of native T1 times, spatial variability, and repeatability obtained using both methods.

Figure 6. Example native myocardial T1 maps of a 38‐year‐old male patient (HR 84 bpm) admitted with syncope using all T1 mapping schemes. All schemes provided similar T1 map image quality and similar characteristics for native myocardial T1 times. Shortened schemes tended to have lower spatial homogeneity than MOLLI in the blood pool.

Figure 7. Example postcontrast myocardial T1 maps obtained in a 32‐year‐old female patient (HR 61 bpm) with severe left ventricular systolic dysfunction and pericardial effusion using all T1 mapping schemes. All schemes provided similar T1 map image quality as well as similar myocardial and blood T1 ranges across all slices.

Figure 8. Native/postcontrast myocardial/blood T1 times (a–d, respectively) in 24 patients using all T1 mapping schemes. Average (bar plots) and SD (error bars) over all patients are presented. All methods led to similar range of native/postcontrast myocardial/blood T1 times (P ≥ 0.19).

Figure 9. Correlation and agreement tests between T1 times obtained using 2P‐2 and MOLLI in all subjects including 16 healthy volunteers and 24 patients. Subfigure (a–d) illustrate Pearson correlation analysis and Bland–Altman plot of native myocardial T1 times, native blood T1 times, postcontrast myocardial T1 times, and postcontrast blood T1 times, respectively. Strong correlation and good agreement were found between T1 times obtained using 2P‐2 and MOLLI. In Pearson correlation analysis plots, confidence interval (solid lines) and identity line (y = x, dashed line) are also plotted besides the linear regression line (solid line). Correlation information including the Pearson correlation coefficient (r‐value), linear regression relationship (y as a function of x), and coefficient of determination (r2) is also displayed in Pearson correlation plots. In Bland–Altman plots, "average" stands for (T12P‐2 + T1MOLLI)/2 and "difference" stands for (T12P‐2‐T1MOLLI).