Early reverse remodeling of left heart morphology and function evaluated by cardiac magnetic resonance in hypertrophic obstructive cardiomyopathy after transapical beating-heart septal myectomy

We experimentally performed 10 cases of TA-BSM approved by the Tongji Hospital ethics committee and collected clinical data from the initial 10 HOCM patients as a pre-test result. The sample size was calculated according to the formula of the paired experimental design, taking into account the actual clinical surgical volume of this procedure (\(n=\frac{{\left( {Z}_{\frac{\alpha }{2}}+{Z}_{\beta }\right)}^{2}{* \sigma }^{2}}{{\delta }^{2}}\), α = 0.05, β = 0.1). Finally, the study prospectively enrolled 80 HOCM patients who underwent TA-BSM in Tongji Hospital from April 2022 to January 2023 (Fig. 1). All patients were diagnosed as having HCM based on the published guidelines (maximum wall thickness ≥ 15 mm or ≥ 13 mm in patients with a family history of HCM, in the absence of other cardiovascular diseases) [1]. The indications for TA-BSM were: (1) resting or provoked left ventricular outflow tract pressure gradient (LVOTG) ≥ 50 mmHg, and (2) severe symptoms or poor response to medical therapy. The major exclusion criteria were: (1) failure to undergo complete CMR examinations before and three months after TA-BSM (e.g., a previous pacemaker or metal stent implantation, postoperative adverse events, and missed appointments), (2) poor image quality due to arrhythmias, (3) other procedures performed simultaneously, such as transcatheter aortic valve implantation, (4) combined severe coronary artery disease [coronary artery stenosis ≥ 50%], (5) previous cardiac surgery, including alcohol septal ablation, percutaneous radiofrequency ablation, the Liwen procedure, surgical myectomy, and valve replacement. Patients with previous ventricular septal myectomy or valve repair or replacement were excluded. This prospective single-center study was approved by the Tongji Hospital ethics committee (approval numbers: 2022-S013, 2022-S013-1, 2022-S013-2, 2022-S013-3, and 2022-S013-4). Informed consent was obtained from all patients for this study.

Standard transthoracic echocardiography (TTE) was performed to evaluate LVOT stenosis within one week before TA-BSM and to assess the surgical outcome three months after the operation. Conventional cardiac parameters were measured by two experienced echocardiographers. The peak LVOT velocity was acquired by continuous-wave Doppler echocardiography in the apical long-axis or five-chamber view in the resting and provoked states, respectively. The LVOT peak gradient was calculated using the simplified Bernoulli equation. Assessments of LV diastolic function performed were the ratio between early mitral inflow velocity and early diastolic mitral annular velocity (E/e') and mitral inflow early-to-late filling velocity ratio (E/A). Mitral regurgitation was graded as 0 (none), 1 + (mild), 2 + (moderate), 3 + (moderate to severe), 4 + (severe) [10]. Systolic anterior motion (SAM) was graded as 0 (none), 1 (leaflet-septal distance > 10 mm), 2 (leaflet–septal distance ≤ 10 mm but no leaflet-septal contact), 3 (leaflet-septal contact < 30% of the systolic duration), 4 (leaflet-septal contact ≥ 30% of the systolic duration), according to the recommendations of the American Society of Echocardiography [11].

All patients underwent a standard CMR examination on a 3T system (MAGNETOM Skyra, Siemens Healthcare, Erlangen, Germany). Cine images (four-chamber, three-chamber, and short-axial views) were acquired with ECG-gated and breath-holding using a segmented, balanced, steady-state free-precession sequence. Short-axial cine covered the LV from the apex to the mitral valval ring. The typical parameters were: section thickness = 8 mm, section gap = 2 mm, echo time = 1.39 ms, repetition time = 3.2 ms, field of view = 360 × 360 mm², matrix size = 189 × 154, flip angle = 46°, temporal resolution 32 ms, and calculated phases 25.

All CMR function analyses were performed using commercial cardiac software (CVI 42, version 5.14.0, Circle Cardiovascular Imaging Inc., Calgary, Alberta, Canada). Epicardial and endocardial LV contours from the base to the apex were automatically delineated on short-axial cines. The maximum wall thickness of surgical and non-surgical segments was measured manually according to the sixteen American Heart Association (AHA) segments. Left ventricular end-diastolic diameter (LVEDD) was measured from anteroseptal wall to inferolateral wall in the middle short-axial slice at the end-diastolic phase. Left ventricular end-diastolic volume (LVEDV), left ventricular end-systolic volume (LVESV), stroke volume (SV), cardiac output (CO), left ventricular ejection fraction (LVEF), and left ventricular mass (LVM) were obtained through end-diastolic and end-diastolic delineation. In addition, LVEDV, LVESV, CO, and LVM were indexed to the body surface area. The left ventricular remodeling index (LVRI) was calculated as the ratio between LVM and LVEDV (LVRI = LVM/LVEDV) [12, 13] which reflects the interplay between morphology and function during cardiac remodeling.

The left atrial (LA) anteroposterior and LA left–right diameters were obtained on four- and three-chamber cine images, respectively. The maximum LA volume (LAV_max) and minimum LA volume (LAV_min) were acquired with a combination of two- and four-chamber cine images at the end of ventricular systole and diastole. The left atrial ejection fraction (LAEF) was calculated as follows:

Statistical analysis of preoperative and postoperative CMR data was performed using SPSS statistical software (version 25.0, IBM SPSS Inc., Chicago, IL, USA). Continuous and normally distributed data were expressed as means ± standard deviations (SD), while medians [interquartile range (Q1, Q3)] were used for variables with non-normal distributions. Categorical variables were presented as percentage frequencies and compared using the chi-squared test. Parameters at baseline and follow-up CMR were compared using the paired t-test if the differences conformed to a normal distribution. Otherwise, non-normally distributed variables were compared using paired Wilcoxon signed-ranks test. In addition, comparisons of normally distributed continuous variables between HOCM patients and the general population were performed using the independent sample t-test. Mann–Whitney U test were used for comparation of non-normally distributed continuous variables. Statistical significance was indicated by p < 0.05.

The change in each variable was denoted by △, equal to the preoperative value minus the postoperative value. The Pearson correlation coefficient (r) and Spearman correlation coefficients (r_s) were used to investigate the relationships between the parameters. Predictors of the left ventricular mass index (LVMI) reduction and LVRI were calculated using a stepwise multiple linear regression model entered as covariate factors. Last, variables with p < 0.05 were entered into the multivariable analysis.

Twenty patients were random selected from HOCM patients and controls respectively for the intra- and interobserver reproducibility assessment of CMR parameters by the intraclass correlation coefficient (ICC). All measurements were performed by two observers who were unaware of the values obtained during the selection process. The intra-observer measurements were repeated after 2 weeks by observers unawarded of the previous data.

Forty-one HOCM patients were enrolled in this study, with a mean age of 45.9 ± 2.4 years. A total of 65.9% were male. The baseline characteristics of the 41 HOCM patients are listed in Table 1. There were 18 patients with a sigmoid septum and 23 with a reverse septal curvature. The primary symptom was dyspnea (80.5%), and the incidence of syncope was 22.0%. Preoperative New York Heart Association (NYHA) functional class 3 and above was present in 58.6% of patients.

All patients underwent TA-BSM performed by the same cardiac surgeon with 25 years of experience. Minimally invasive myocardial resection without extracorporeal circulation was performed to remove abnormal hypertrophic myocardium. The mean myocardial resection mass was 5.6 (3.3, 10.6) g. Almost all patients experienced an improvement in symptoms after TA-BSM, and intraoperative catheter manometry showed a decrease in LVOTG to normal levels (< 30 mmHg). A total of 92.7% of postoperative patients were classified as NYHA I or II compared to 41.5% before myectomy (Fig. 2b).

The echocardiographic results are summarized in Table 2. The TTE showed that the LVOTG of all HOCM patients was significantly reduced after TA-BSM (89 ± 4.9 mmHg vs. 16 ± 1.4 mmHg, p < 0.001) (Fig. 2a). Mitral regurgitation occurred in all patients, of which moderate to severe accounted for 78.0% preoperatively and decreased to 7.3% postoperatively. (Fig. 2c). Almost all patients presented with mitral systolic anterior motion (SAM). SAM grades 3–4 occurred in 80.7% of patients but decreased to 2.4% postoperatively.

As shown in Table 3, LA anteroposterior diameters, LA left–right diameters, LAV_max, and LAV_min were significantly decreased postoperatively (all p < 0.001). LAEF increased postoperatively and was slightly lower than that of the healthy controls [pre-TA-BSM, 48 ± 1.4% vs. post-TA-BSM, 54 ± 1.5% (p < 0.001), Controls, 63 ± 1.2% (both p < 0.001 for pre-TA-BSM and post-TA-BSM)].

The LV morphological and functional parameters of the study groups are displayed in Table 4. SV, CI, LVMI, LVRI, LVEF, maximum wall thickness, maximum wall thickness in non-surgical segments, and LVMI decreased after TA-BSM (all p < 0.05). However, SV, LVMI, LVRI, maximum wall thickness, and non-surgical segment maximum wall thickness were still higher than in the normal controls (all p < 0.05). A representative case of cine images before and after TA-BSM is shown in Fig. 3. After the procedure, the thickness of the myocardial segment without surgical resection decreased significantly [pre-TA-BSM, 22 ± 1.0 mm vs. post-TA-BSM, 16 ± 0.7 mm (p < 0.001), Controls, 9.1 ± 0.1 mm (both p < 0.001)], and the high-speed blood flow signals due to LVOT obstruction disappeared.

There was no significant preoperative difference between the two subgroups in the function parameters. Patients with a reverse septal curvature had a significantly higher LVMI and maximum wall thickness than patients with a sigmoid septum preoperatively (121 ± 6.5 g/m² vs. 71 (68, 107) g/m², 26 ± 1.1 mm vs. 19 ± 1.0 mm, both p < 0.001), but their LVEDD was smaller (45 ± 1.1 mm vs. 51 (48, 57) mm, p = 0.001). In addition, SV, CI, LVEF, LVMI, LVRI, maximum wall thickness, and maximum wall thickness in the non-surgical segments were significantly reduced postoperatively in both subgroups (all p < 0.05). However, the left ventricular end-diastolic volume index (LVEDVI) and LVEDD decreased in patients with a sigmoid septum postoperatively (pre-TA-BSM, 83 ml/m² vs. post-TA-BSM, 73 ml/m², pre-TA-BSM, 51 (48, 57) mm vs. post-TA-BSM, 47 (43, 50) mm, both p < 0.001) but increased in patients with reverse septal curvature (pre-TA-BSM, 80 ml/m² vs. post-TA-BSM, 87 ml/m², pre-TA-BSM, 45 ± 1.1 mm vs. post-TA-BSM, 51 ± 0.9, both p < 0.001) (Fig. 4 and Table 5).

The correlations between LVRI and conventional parameters are shown in Fig. 5. Among the morphological and functional parameters, LVMI and maximum wall thickness were significantly correlated with LVRI preoperatively (r_s = 0.734 and r = 0.679, both p < 0.001). After TA-BSM, both remained significantly associated with LVRI. The correlation between LVMI and LVRI was enhanced (r = 0.853, p < 0.001), but the correlation between maximal wall thickness and LVRI did not change significantly (r = 0.696, p < 0.001).

Clinical variables, including patient demographics, echocardiographic parameters, and CMR measurement parameters, were analyzed to identify the predictors of reverse remodeling. In the univariate analysis, the weight of the resected myocardium, maximum wall thickness, △SAM, △LVOTG, baseline LVEDVI, and baseline LVMI appeared to predict a reduction in LVMI (all p < 0.05). However, in the multivariate analysis, only △LVOTG (adjusted β = 0.323, p = 0.018) and baseline LVMI (adjusted β = 0.436, p < 0.040) were independent predictors of LV mass regression (Table 6). In addition, the weight of the resected myocardium (adjusted β = 0.476, p = 0.005) and △mitral regurgitation degree (adjusted β = − 0.245, p = 0.040) were associated with △LVRI.

The LA and LV parameters in the two subgroups before and after TA-BSM showed excellent reproducibility (intra-observer ICC: 0.90–0.99; inter-observer ICC: 0.92–0.99).