| Part I Quantitative myocardial T1 mapping:in vitro and in vivo study at 1.5 TeslaObjectives:To compare 11 heart beat (HB) and 17 HB modified lock-locker inversion recovery (MOLLI) pulse sequence at 1.5T and to assess the variability of myocardial T1 relaxation times in the normal human population estimated with recently proposed 11 HB Modified Look-Locker Inversion recovery T1 mapping technique.Materials and Methods:1 Eight phantoms containing agarose gel doped with diluted Gadolinium DTPA were scanned at 1.5T MR using both 17 (3-3-5)HB and 11 (5-3-3,4-3-2)HB MOLLI sequence with simulated heart rates from 40 to 110 beats per minute (BPM) in increments of 10 heartbeats. The reference T1 time of each phantom was determined by standard inversion-recovery spin echo sequence with the same FOV and matrix size with 6 different TIs with a TR of 10 sec, and a TE of 9.4 msec. Bland-Altman plots were used to describe the difference of T1 values between 17 HB and 11 HB protocols. The T1 values obtained by two protocols were statistically compared using a general linear mixed model. The MOLLI protocol was included in the model as a fixed effect, while the phantom was included as a random effect. Percentage error in T1 estimation at different simulated heart rates as compared to reference values was analyzed. The standard deviation (SD) at certain heart rate was also compared between the protocols.2 Thirty healthy human subjects without cardiovascular disease were scanned at 1.5T MR using 11 HB MOLLI (5-3-3) T1 mapping sequence both during diastole and systole, of which 23 volunteers were scanned using both 17 HB and 11 HB MOLLI sequence during diastole. The image quality, breath hold time and the T1 values were compared between the 11 HB and 17 HB protocol. The T1 values of all the myocardial segments obtained during diastole and systole were compared using student t test. The variation of septal and non-septal segment and different scanning slices was also analyzed.Results:1 The reference T1 values of eight phantoms were as follows:343±2.5 ms,412±3.2 ms,527±4.4ms,678±5.3ms,745±3.2ms,992±6.4ms,1060±7.2ms,1424±13 ms. Bland-Altman Plots showed good agreement as comparing 17 HB MOLLI and 11 HB (5-3-3) MOLLI T1 values at 70 bpm. As compared with 11 HB MOLLI using 8 images, the original MOLLI (17) protocol using 11 images has excellent precision albeit the accuracy degrades for long T1 values. When the simulated heart rates were more than 90 bpm, the accuracy increased with the matrix decreased from 256 to 192. Percentage error increased with the heart rate increased for all protocols. The maximum of percentage error for MOLLI (3-3-5) and MOLLI (5-3-3) was 13.3% and 7.9% respectively, while the post-contrast MOLLI (4-3-2) had the largest variation compared the others two sequences.2 Mean myocardial T1 values obtained with original 17 HB MOLLI did not differ with that of 11 HB MOLLI(964.9±36.7ms vs 967.6±30.9ms, t=-0.422, P>0.05) significantly, while the breath hold time of 17 HB MOLLI was significantly longer than that of 11 HB MOLLI. Bland-Altman plots showed good agreement between the T1 values obtained with the two protocols.Mean segmental T1 values between systole and diastole differed significantly (965.2±30.0ms vs 982.3±25.6ms, P<.05). The T1 values of septal segments (8,9) of left ventricle in diastole were longer significantly than that of non-septal(11,12) segments, while the T1 values of septal segments (8,9) of left ventricle in systole were longer significantly than that of non-septal(11)segments. With respect to the different scanning slices, the T1 values of middle slice were significantly longer than that of basal and apical slices, while the T1 values of four-chamber, three-chamber and two-chamber did not differ significantly.Conclusion:1 Both the 11HB MOLLI and 17 HB MOLLI have some dependency on heart rate and matrix.11 HB MOLLI (5-3-3) is a faster method for high-resolution myocardial T1 mapping at 1.5T with accuracy.2 The 11 HB MOLLI (4-3-2) can only be used for short T1 values due to its large percentage error for long T1 value.3 Due to variations in T1 time during the cardiac cycle and in different myocardial regions, as well as different scanning slices, T1 measurements should be obtained at the same cardiac phase and myocardial region in order to obtain consistent results.Part Ⅱ Myocardial T1 Mapping:Application to patients with chronic total coronary occlusionObjectives:The aim of this study was to assess and delineate myocardial scar in chronic total coronary occlusion patients using CMR late gadolinium enhancement and T1 mapping techniques.Materials and Methods:A total of 83 patients suspected of having at least one CTO lesion according to clinical symptoms or coronary CTA findings underwent CMR scanning and invasive coronary angiography. The CMR examinations were performed at 1.5T applying traditional sequences and a modified Look-Locker Inversion Recovery sequence with 11 heart beats before and 15 minutes after contrast at 3 short-axis slice positions. Late gadolinium enhancement using an inversion recovery gradient recalled echo sequence was deemed as standard of reference. The prevalence of LGE and the patterns of scar were analyzed in CTO patients confirmed by invasive coronary angiography, as well as the relationship between the incidences of LGE and the collateral vessels Rentrop Grades. T1 values of myocardial scar, the myocardium remote from scar and blood pool were measured on pre-contrast and post-contrast T1 maps. Sensitivities and specificities for detection of myocardial scar on pre-contrast and post-contrast T1 maps were calculated. Receiver operating characteristic (ROC) analysis was performed for post-contrast T1 values for discrimination of myocardial scar. The scar size (if detected) on pre-contrast and post-contrast T1 maps was compared with which measured on LGE imaging using Bland-Altman plots.Results:The MR images of 62 patients with CTO lesions confirmed by invasive coronary angiography were analyzed. Forty-five patients (72.6%,45/62) were found with varying patterns of abnormal enhancement on LGE imaging. In detail,7 subjects manifested patch style (15.5%,7/45),25 cases manifested linear enhancement (55.6%,25/45), and the remaining 13 patients showed transmural enhancement (28.9%,13/45). There were 4 patients without infarct-related artery to the myocardial scar (8.9%,4/45). Those who definitely had a history of myocardial infarction (MI) based on laboratory, echocardiography, and imaging findings had more frequency of transmural enhancement on LGE images (76.5%,13/17), while patch and linear style was prone to be detected in patients without MI history (84.2%,32/38). The incidences of LGE in patients with good coronary collateral circulation (n=41, Rentrop Ⅱ-Ⅲ) and poor collateral circulation (n=21, Rentrop 0-1) were 73.2% and 71.4%, respectively (p>0.05). The comparison of T1 values of myocardial scar and myocardium remote from scar revealed significant differences in pre-contrast and post-contrast scans (1138.5±52.4ms vs 987.3±79.2ms, P<0.05, and 275.6±24.6ms vs 379.7±45.6ms, P<0.01). Sensitivities and specificities for detection of myocardial scar were 40.3% and 100% in pre-contrast T1 mapping scans and 91.2% and 99.3% in post-contrast images, respectively. ROC analysis revealed that the cutoff values was 307.5 milliseconds or less with sensitivities and specificities being 93.3% and 93.3% and the area under the curve (AUC) 0.96 (95%CI:0.91-1.00) for detecting myocardial scar by post-contrast T1 values.Conclusion:The myocardial scar of CTO patients has certain patterns on CMR LGE images, and there is no significant correlation between the presence of myocardial scar and the grades of coronary collateral circulation. Post-contrast T1mapping allow for accurate detection of myocardial scar of CTO patients with higher sensitivity and specificity, while pre-contrast T1 values lack accuracy in delineation of myocardial scar. Part Ⅲ Extracellular volume of myocardial scar and remodelingby MR T1mapping in CTO patients:initial resultsObjectives:To evaluate the value of ECV measurement by T1 mapping in myocardial scar and remodeling remote from regions of scar in chronic total occlusion patients.Materials and Methods:Cardiac magnetic resonance T1 mapping was performed in 12 healthy volunteers and 62 patients with at least one CTO lesion in coronary vessels before and after injection of gadolinium contrast. ECVs of myocardial scar and normal appearing myocardium remote from regions of scar, as well as the myocardium of healthy volunteers, were calculated according to the pre-contrast and post-contrast T1 values and hematocrit. Two-tailed Studen t t test was used for statistical analysis of ECVs in different myocardial regions of CTO patients. The relationship between left ventricular ejection fraction and the ECV of myocardial scar and myocardium remote from scar was analyzed using Pearson’s correlation coefficient. Receiver operating characteristic (ROC) analysis was performed for ECV for discrimination of myocardial scar in CTO patients. The dynamic changes of ECVs in myocardial scar and non-scar region over time after contrast media injection were also analyzed.Results:Average ECV for remote normal myocardium and myocardial scar in CTO patients were 28.8±3.2%(24.0%-40.3%) and 51.6±7.0%(36.5%-63.6%) (P<0.05), while the Average ECV was 26.9±2.2%(23.8%-31.2%) for normal myocardium in healthy volunteers. Sensitivities and specificities were 91.1% and 100% for detecting myocardial scar in CTO patients for ECV, with cutoff values being greater than 41.3%. Left ventricular ejection fraction decreased as the ECV of’normal appearing’ myocardium remote from myocardial scar increased (r=-0.5878, P<0.05). There was no relationship between left ventricular ejection fraction and the ECV of myocardial scar(r=-0.08, P=0.59). The Kruskal-Wallis test of ECV of five time point after contrast injection showed that the ECV of myocardial scar and ’normal appearing’ myocardium remote from myocardial scar was not dependent on imaging time (P=0.8050 for ’normal appearing’ myocardium, P=0.2640 for myocardial scar).Conclusion:Extracellular volume fraction obtained from MR pre- and post-T1mapping can quantitatively characterize myocardial scar, and subtle diffuse myocardial abnormalities due to remodeling remote from myocardial scar. In a relatively broad span, the ECV of myocardial scar and ’normal appearing’ myocardium remote from myocardial scar is stable. |
PDF Full Text Request