Diastolic dysfunction assessed by cardiac magnetic resonance imaging tissue tracking on normal-thickness wall segments in hypertrophic cardiomyopathy

We retrospectively enrolled 63 HCM patients who underwent contrast-enhanced CMR imaging in our institution from May 2018 to January 2020 and 30 healthy controls matched for age and sex (Fig. 1). The control group showed no abnormalities on contrast-enhanced CMR. The HCM inclusion criteria were: (1) maximum LV wall thickness ≥ 15 mm, or ≥ 13 mm with family history, and (2) age ≥ 18 years old. The exclusion criteria were: (1) the presence of other conditions that could lead to LV hypertrophy, (2) image quality too poor for segmental analysis, and (3) LVEF < 50%. The clinical history, serum markers, and echocardiographic data were collected for all subjects. This study was approved by the Institutional Ethics Committee, and written informed consent was waived due to the retrospective design.

All subjects underwent a CMR scan with an 18-channel phased-array body coil on a 3 T scanner (MAGNETOM Skyra, Siemens Healthcare, Erlangen, Germany). CMR protocols included cine in the short- and long-axis (2-, 3-, and 4-chamber), native and post-contrast T1 mapping, T2 mapping in three short axis slices (basal, middle, and apical), and late gadolinium enhancement (LGE) images at the exact location as the short-axis cine slices.

A standard breath-hold cine imaging with steady-state free precession (SSFP) sequence was acquired. The parameters were as follows: echo time (TE) = 1.4 ms; repetition time (TR) = 3.1 ms; flip angle (FA) = 55°; slice thickness = 8 mm; voxel size = 1.9 × 1.9 mm²; slice gap = 2 mm; bandwidth = 965 Hz/pixel; FOV = 360 × 360 mm²; and 25 calculated phases per heartbeat.

T1 mapping was acquired by using a single breath-hold Modified Look-Locker Inversion recovery sequence. In addition, 5b(3b)3b and 4b(1b)3b(1b)2b schemes were used for the native and post-contrast T1 imaging, respectively, 15–20 min after administration of a contrast agent dose (0.2 mL/kg of MultiHance, Bracco Diagnostics). The T1 parameters were as follows: TE = 1.2 ms; TR = 3.8 ms; FA = 35°; slice thickness = 5 mm; and voxel size = 1.4 × 1.4 mm². Furthermore, T2 mapping was generated using a T2-prepared bSSFP sequence with three T2 preparation times: 0, 24, and 55 ms. The T2 parameters were as follows: TE = 1.41 ms; TR = 3.3 ms; FA = 12°; slice thickness = 5 mm; voxel size = 1.9 × 1.9 mm²; and acquisition matrix = 206 × 256.

LGE imaging was performed 10–15 min after administration of the contrast agent using an inversion recovery prepared pulse and segmented FLASH sequence with a phase-sensitive reconstruction technique. Imaging parameters were as follows: TE = 1.2 ms; TR = 5.2 ms; FA = 55°; slice thickness = 8.0 mm; and voxel size = 1.5 × 1.5 mm².

All CMR images were analyzed with commercial software (CVI 42, version 5.11, Circle Cardiovascular Imaging, Calgary, Canada) according to the 16-segment American Heart Association (AHA) model [8].

Contours in the endo- and epi-myocardium at the LV end-diastolic phase without papillary muscle were automatically sketched and manually corrected on a short-axis cine by a radiologist with 3 years of experience using the short-3D module (Figure 2). In addition, the LV functional parameters were obtained automatically.

Then, the cine images of the short and long axes were loaded into the Tissue Tracking module to outline the endocardium and epicardium and generate segmental strain parameters, including the peak strain (PS), systolic peak strain rate (PSR) and diastolic PSR in the radial, circumferential, and longitudinal directions.

According to the maximum thickness of the myocardium, the total HCM segments were divided into group A (< 12mm) [15], group B (12–15 mm), and group C (> 15mm) [8, 14]. We further compared the strain within the subgroups of obstructive and non-obstructive HCM to exclude the influence of an obstructive left ventricular outflow tract (LVOT) on this strain parameter. Notably, a maximum LVOT gradient of ≥ 30 mmHg at rest or with provocation represented the presence of obstruction, based on the result of echocardiography [16].

Segmental native T1 and T2 values were obtained by manually delineating the endocardium and epicardium of the entire LV myocardial region (which included the LGE-positive regions) on the T1 and T2 mapping images [17]. Then, the extracellular volume (ECV) was calculated using the hematocrit (Hct) and the myocardial and blood intensity before and after enhancement: ECV = (1/T1_{myo post}− 1/T1_{myo pre}) × (1 − Hct)/(1/T1_{blood post}− 1/T1_{blood pre})¹⁶. T1_{myo post} and T1_{myo pre} represented the post- and pre-contrast myocardial T1 values, respectively, whereas T1_{blood post} and T1 _{blood pre} were the post- and pre-contrast blood T1 value [18]. The Hct used was from the routine blood test conducted on the day of CMR scanning. Positive LGE lesions were defined as a signal intensity greater than six standard deviations above the mean signal intensity of the remote reference myocardium [19, 20]. In addition, the LGE ratios were calculated according to the LGE volume and segmental LV myocardial volume.

We randomly selected 20 patients for inter- and intra-observer agreement analysis. The strain parameters were measured independently by two blinded observers (J.H.Q. and P.J.Z, with 3 and 5 years of CMR diagnostic experience, respectively), one of whom re-measured the strains after one month (J.H.Q).

All statistical analyses were performed using SPSS Statistics 22 (IBM, Chicago, USA). Categorical variables were presented as frequencies (percentages) and analyzed with Pearson’s chi-squared test. The normality distribution of the continuous data was determined by the Shapiro–Wilk test. Continuous variables were expressed as mean ± standard deviation or median (interquartile range). An independent t-test or Mann–Whitney U-test was performed to compare two groups of continuous variables. The Kruskal–Wallis rank test was used to analyze the differences among three or more groups. In addition, Bonferroni’s correction was conducted to perform post-hoc comparisons. The association between myocardial thickness and other quantitative myocardial parameters (e.g., strain indexes and tissue characteristics) was evaluated with the Spearman correlation, which had the following values: Spearman correlation value of 0.00–0.10 was considered a negligible correlation; 0.10–0.39, weak correlation; 0.40–0.69, moderate correlation; 0.70–0.89, strong correlation; and 0.90–1.00, very strong correlation [21]. Furthermore, the intraclass correlation coefficient (ICC) was applied to assess the inter- and intra-observer reproducibility variability, in which a value > 0.75 represented good agreement. Lastly, a p-value < 0.05 was considered statistically significant.

Compared with the controls, patients with HCM had a significantly increased myocardial mass index, cardiac output, and stroke volume (p < 0.001, p = 0.03, and p = 0.04, respectively) (Table 2). However, there were no statistically significant differences between the HCM patients and controls in LVEF, end-diastolic volume, or end-systolic volume (p = 0.07, p = 0.10, and p = 0.90, respectively).

There were 716 segments in group A, 111 in group B, and 181 in group C. As shown in Table 3, the diastolic PSR values in group A were all lower than in the control group in the radial, circumferential, and longitudinal directions (all p < 0.05). Notably, there was a significant difference in the native T1 (p < 0.001) between group A and the control group but no such difference in T2 (p = 0. 59) or ECV (p = 0. 99).

Further analysis of segmental strain in the normal-thickness wall segments are shown in Fig. 3 for patients with obstructive HCM, non-obstructive HCM, and controls. There were decreased diastolic PSRs in both subgroups compared with the controls (both p < 0.05). In addition, the demographic characteristics of the subgroups are shown in the supplementary materials (Additional file 1: Table S1).

As demonstrated in Table 4, the radial diastolic PSR was moderately associated with the end-diastolic wall thickness (EDTH) (r = 0.44, p < 0.001). The EDTH was weakly correlated with the extent of LGE (r = 0.35, p < 0.001) and native T1 (r = 0.34, p < 0.001).

As shown in Table 5, all inter-observer ICCs were above 0.848 and the intra-observer ICCs were higher than 0.862, indicating good observer agreement and reliability.