Potential of non-contrast stress T1 mapping for the assessment of myocardial injury in hypertrophic cardiomyopathy

We retrospectively enrolled 31 patients with HCM and 14 control subjects who underwent CMR imaging to evaluate suspected CAD between October 2018 and June 2022. In the HCM group, the electrocardiogram of all patients revealed ST-T changes including ST depression or negative T waves. Five patients presented with atypical chest discomfort and the remaining were asymptomatic. None of patients had a history of non-sustained ventricular tachycardia or syncope. HCM was diagnosed according to the JCS/JHFS guideline and was defined as left ventricular wall thickness ≥ 15 mm (in the absence of a family history of HCM), or 13–14 mm (with a family history of HCM) [15]. Patients with other cardiac diseases including storage, infiltrative, or systemic diseases were excluded [15]. Patients with stress perfusion defects underwent MR coronary angiography (MRCA) or computed tomography coronary angiography (CTCA). None of the patients had significant coronary artery stenosis. In the control subjects, 2 subjects presented with dyspnea on exertion and 12 presented with non-exertional chest discomfort. Subjects who exhibited no abnormalities on CMR imaging and had no history of cardiovascular diseases were included in the control group. Patients with known ischemic heart disease, more than moderate valvular heart disease, and contraindications for CMR were excluded from this study. This retrospective study of clinically acquired data was approved by the Institutional Review Board of Tokyo Medical University, and the need for written informed consent was waived.

CMR imaging was performed using a Magnetom Skyra 3T system (Siemens Healthineers, Erlangen, Germany) with a 60-channel body coil. This included cine images, T1 mapping at rest and during ATP stress, late gadolinium enhancement (LGE), and phase-contrast (PC) cine imaging of coronary sinus flow at rest and during ATP stress to assess the coronary flow reserve (CFR). Patients were instructed not to consume caffeine for 12 h before ATP stress CMR. PC cine imaging was performed on 25 subjects (17 patients with HCM and 8 control subjects). Short-axis images covering the left ventricle (LV) from base to apex and 2-chamber, 3-chamber, and 4-chamber long-axis cine images were acquired with steady-state free precession to evaluate cardiac function and myocardial mass (repetition time (TR): 28.2 ms, echo time (TE): 1.6 ms, flip angle: 60°, field of view (FOV): 360 × 270 mm, acquisition matrix: 224 × 224, slice thickness: 6 mm, number of cardiac cycle phases: 30). T1 mapping at rest was performed on short-axis slices at the base, mid-ventricle, and apex (Fig. 1). A modified Look-Locker Inversion recovery (MOLLI) sequence with a 5(3)3 scheme was used (TR: 349 ms, TE: 1.1 ms, flip angle: 35°, FOV: 360 × 306 mm, acquisition matrix: 256 × 169, slice thickness: 8 mm, number of cardiac cycle phases: 1). Stress CMR was performed with a 3-min continuous intravenous injection of ATP (140 μg/kg/min). After a 3-min infusion, stress T1 mapping was performed on the same short-axis slices as ones of T1 mapping at rest using MOLLI sequence with a 5(3)3 scheme (Fig. 1). The MOLLI sequence was modified to reduce heart rate sensitivity [16, 17], although the reduction in heart rate sensitivity was insufficient compared to the shortened MOLLI sequence and the 5s(3s)3s-MOLLI sequence used in previous studies [12, 14, 18, 19]. Rest and stress PC cine imaging were performed immediately after rest and stress T1 mapping. The imaging plane of the PC cine images was positioned perpendicular to the coronary sinus 2 cm from the ostium [8]. PC cine images were acquired during shallow breath-holding (TR: 37.8 ms, TE: 2.6 ms, flip angle: 20°, FOV: 340 × 231 mm, acquisition matrix: 192 × 173, slice thickness: 6.0 mm, number of cardiac cycle phases: 40). Stress perfusion imaging was performed immediately after the stress PC cine imaging. Three short-axis slices identical to T1 mapping were obtained with breath-hold by intravenous administration of 0.1 mmol/kg of gadobutrol (Gadovist, Bayer Healthcare, Leverkusen, Germany) at a rate of 3 ml/s, followed by a 20 ml saline flush (TR: 410 ms, TE: 1.1 ms, flip angle: 35°, FOV: 360 × 306 mm, acquisition matrix: 256 × 169, slice thickness: 8 mm, number of cardiac cycle phases: 1), and ATP infusion was terminated. Ten minutes later, rest perfusion imaging was performed using the same protocol as that for stress perfusion. LGE images were obtained using an inversion-recovery gradient echo sequence 10 min after rest perfusion (TR: 700 ms, TE: 1.9 ms, flip angle: 12°, FOV: 360 × 270 mm, acquisition matrix: 192 × 157, slice thickness: 6 mm, number of cardiac cycle phases: 1). The inversion time was adjusted to null normal myocardium for each patient.

CMR images were analyzed in a blinded fashion by two observers using commercially available CMR postprocessing software (Ziostation 2, Ziosoft, Tokyo, Japan). Left ventricular volume, ejection fraction, and mass were calculated by manual tracing of endocardial and epicardial contours of the end-diastolic and end-systolic short-axis cine images [20]. Perfusion imaging and T1 mapping were performed using a 16-segment model. The presence of a perfusion defect was defined by hypoenhancement in each segment that persisted for at least 3 consecutive frames after maximal myocardial enhancement [9, 21]. A total of 626 segments for T1 reactivity and 689 segments for native T1 values from 45 patients were available for analysis. The segments in the HCM group were classified as hypertrophied and non-hypertrophied myocardial segments. Hypertrophy was defined as left ventricular wall thickness of ≥ 13 mm. Native T1 values were measured by placing a region of interest in each segment, and T1 reactivity was calculated as follows: (T1stress − T1rest)/T1rest × 100 (%). Each region of interest was carefully drawn avoiding the adjacent blood pool and epicardium. The numbers of segments for T1 reactivity and native T1 were different because there were more segments after stress that could not be measured as T1 values, relative to the segments at rest, due to insufficient image quality. Segments with or without LGE in the HCM group were visually assessed. Blood flow in the coronary sinus was measured by tracing the contour of the coronary sinus on the magnitude images in each frame. Coronary sinus flow (CSF) was corrected using the rate pressure product as follows: rate pressure product = systolic blood pressure × heart rate, corrected CSF = CSF/rate pressure product × 7500, and CFR was calculated as corrected CSF during stress divided by corrected CSF at rest [7, 20]. Intraobserver and interobserver agreements of native T1 measurements from a random sample of 10 subjects were assessed by two independent observers. Interstudy reproducibility was assessed in three subjects.

All data are expressed as a mean ± standard error of the mean. A chi-square analysis was performed to evaluate the association between the collected clinical information. The Mann–Whitney U test was used to compare age, body mass index (BMI), and CMR data between the HCM and control groups. This test was also used to compare native T1 values at rest (rest T1 values) and T1 reactivity between LGE-positive and LGE-negative groups and segments. The difference in T1 reactivity between the three groups was calculated using a 1-way analysis of variance with Bonferroni’s post hoc test. The Kruskal–Wallis test was used to compare resting T1 values between the three and four groups and T1 reactivity among the four groups, followed by Dunn’s post hoc test with Bonferroni correction. Univariable and multivariable linear regression analyses were used to evaluate the relationship between T1 reactivity and the clinical and imaging parameters. All variables associated with T1 reactivity (p < 0.2) were included in multivariable linear regression analysis [22]. Area under the curve (AUC) was calculated to assess the comparison between T1 reactivity and myocardial perfusion defect. Intraobserver and interobserver agreements of native T1 measurements were assessed with intraclass correlation coefficients. The Shapiro–Wilk test was used to assess normality for each variable. All analyses were performed using the SPSS software (SPSS 26.0, Chicago, IL, USA). Statistical significance was set at p-value < 0.05.

The clinical characteristics of the patients are summarized in Table 1. There were no significant differences in age, sex, BMI, smoking frequency, hypertension, diabetes mellitus, or hyperlipidemia between the HCM and control groups. The HCM group included six patients with apical hypertrophy and two patients with left ventricular outflow obstruction; aneurysm formation was not seen in this group.

The CMR data are listed in Table 2. There were no significant differences in stress and resting heart rate, left ventricular ejection fraction (LVEF), and LV volumes between the HCM and control groups. The LV mass index (LVMI) was significantly greater in the HCM group compared to the control group (60.5 ± 3.4 g/m² vs. 39.6 ± 1.6 g/m², p < 0.001). There were no statistically significant differences in systolic and diastolic blood pressures between the HCM and control groups (systolic pressure: 137 ± 3 mmHg vs. 144 ± 11 mmHg, p = 0.923, diastolic pressure: 85 ± 2 mmHg vs. 83 ± 7 mmHg, p = 0.468).

Resting T1 values in the HCM group were significantly higher compared to those in the control group (1244 ± 9 ms vs. 1197 ± 10 ms, p = 0.004, Fig. 1A). Native T1 values during stress (stress T1 values) significantly increased in both the HCM and control groups compared to those at rest (HCM: 1285 ± 10 ms, p < 0.001 vs. rest, control: 1248 ± 10 ms, p = 0.002 vs. rest, Fig. 2A). T1 reactivity in the HCM group was significantly lower compared to the control group (4.2 ± 0.3% vs. 5.6 ± 0.5%, p = 0.044, Fig. 2B). Nineteen percent of segments in the HCM group showed myocardial perfusion defect, while there were no segments positive for perfusion defect in the control group. T1 reactivity showed an AUC of 0.71 (95% confidence interval (CI), 0.66–0.77) for detecting myocardial perfusion defects; the cut-off value of T1 reactivity was 4.3% (Fig. 3). Using this cut-off value, sensitivity and specificity to differentiate myocardial injury was 56% and 77%, respectively. The interclass correlation coefficient for intraobserver agreements of native T1 (rest T1 and stress T1) measurements were 0.990 (95% CI 0.963–0.997) and 0.993 (95% CI 0.974–0.998). The interclass correlation coefficient for interobserver agreements of rest T1 and stress T1 measurements were 0.984 (95% CI 0.936–0.996) and 0.996 (95% CI 0.985–0.999). Interstudy reproducibility was assessed in three subjects. Mean interval between the first and second scans was 16.3 ± 5.2 months. There was no change in clinical findings between the first and second scans. T1 reactivity of the first scan was not significantly different compared to that of the second scan (4.3 ± 1.0% vs. 4.2 ± 0.8%, p = 0.474) and there was no significant bias on the Bland–Altman analysis (0.14 ± 0.27%).

In HCM patients, resting T1 values and T1 reactivity showed no significant differences between LGE-positive and LGE-negative groups (resting T1: 1248 ± 11 ms vs. 1232 ± 11 ms, p = 0.464, T1 reactivity: 4.0 ± 0.4% vs. 5.0 ± 0.7%, p = 0.203). With regards segmental analysis, segments with LGE showed significantly higher rest T1 values and lower T1 reactivity than those without LGE (resting T1: 1300 ± 9 ms vs. 1235 ± 3 ms, p < 0.001, T1 reactivity: 3.4 ± 0.4% vs. 4.9 ± 0.2%, p < 0.001).

CFR was significantly lower in the HCM group compared to the control group (2.0 ± 0.3 vs. 4.2 ± 0.8, p = 0.003). Univariate analysis showed that resting T1 values, LVMI, and CFR were correlated with T1 reactivity. However, there was no significant correlation between T1 reactivity and age, BMI, or LVEF. Multiple linear regression analysis demonstrated that only CFR was independently correlated with T1 reactivity (β = 0.449, p = 0.032; Table 3), but no independent correlation was found between T1 reactivity, and rest T1 values and LVMI.

Resting T1 values in the hypertrophied myocardium in the HCM group (143 segments: 1280 ± 5.6 ms) were significantly higher compared to those in the non-hypertrophied myocardium in the HCM (335 segments: 1229 ± 3.4 ms, p < 0.001, Fig. 4A) and control (211 segments: 1198 ± 3.8 ms, p < 0.001, Fig. 4A) groups. Furthermore, resting T1 values in the non-hypertrophied myocardium in the HCM group was significantly higher compared to those in the control group (p < 0.001, Fig. 4A). T1 reactivity in the hypertrophied myocardium in the HCM group (139 segments: 3.1 ± 0.2%) was significantly lower than that in the non-hypertrophied myocardium in the HCM (301 segments: 4.8 ± 0.2%, p < 0.001, Fig. 4B) and control (169 segments: 5.4 ± 0.3%, p < 0.001, Fig. 4B) groups. However, there was no statistically significant difference in T1 reactivity between the non-hypertrophied myocardium in the HCM and control groups (p = 0.290, Fig. 4B).

In the HCM group, the hypertrophied and non-hypertrophied myocardium with LGE had significantly higher resting T1 values compared to those without LGE (p < 0.001). Resting T1 values in the hypertrophied myocardium without LGE were significantly higher compared to those in the non-hypertrophied myocardium without LGE (p = 0.014). There was no statistically significant difference in the rest T1 values between the hypertrophied myocardium with LGE and the non-hypertrophied myocardium with LGE (Fig. 4C). The non-hypertrophied myocardium without LGE revealed significantly higher T1 reactivity than the hypertrophied myocardium with and without LGE (p < 0.001) and the non-hypertrophied myocardium with LGE (p = 0.024). There were no statistically significant differences among the three groups (Fig. 4D).

This study has some limitations. First, the sample size was small. Therefore, a larger prospective study is required to confirm these findings. Second, MRCA or CTCA was performed to evaluate CAD in subjects with positive stress perfusion. Patients with significant CAD were excluded, but invasive coronary angiography was not performed. Therefore, the exclusion of significant CAD may have been incomplete in this study. Third, the perfusion defect was assessed visually and the diagnostic accuracy for detecting myocardial perfusion defect by T1 reactivity was moderate. The measurement of myocardial blood flow or myocardial perfusion reserve is warranted in future investigations to examine the comparison between T1 reactivity and myocardial hypoperfusion. Finally, 13.1% of myocardial segments were excluded because of insufficient image quality during ATP stress. The MOLLI sequence used in this study was modified to reduce heart rate sensitivity [16, 17], but the reduction in heart rate sensitivity was insufficient compared to the shortened MOLLI sequence and the 5s(3s)3s-MOLLI sequence used in previous studies [12, 14, 18, 19]. However, the proportion of myocardial segments with insufficient image quality in this study was comparable to that reported in previous studies. Therefore, further investigation is needed to improve diagnostic performance.