Twenty-eight patients with HOCM who underwent TA-BSM at Tongji Hospital from March 2022 to April 2023 were prospectively enrolled in this study. All patients met the clinical diagnosis of HCM according to published guidelines (6): ventricular wall thickness ≥15 mm in one or more segments or maximum wall thickness ≥13 mm in relatives in the absence of another disease that could account for the hypertrophy. HOCM was defined as a left ventricular outflow tract gradient (LVOTG) ≥ 30 mmHg at rest or provoked LVOTG ≥ 50 mmHg. The indications for TA-BSM were mainly (1) severe symptoms, with syncope or near-syncope despite the effectiveness of medical therapy; (2) rest or provoked LVOTG ≥ 50 mmHg; (3) age ≥ 12 years. The major exclusions were: (1) failure to undergo complete MRI examinations (e.g., a previous pacemaker or metal stent implantation, postoperative adverse events, and missed appointments); (2) interval of preoperative CMR before TA-BSM greater than 6 months and interval of postoperative CMR after TA-BSM less than 3 months; (3) concomitant myocardial infarction or coronary artery stenosis ≥ 50% at coronary angiography or computed tomographic angiography (CTA); (4) abnormal subendocardial perfusion ≥ 1 segment corresponding to a coronary vascular distribution; (5) comorbid with other diseases (congenital heart disease, valvular disease, and hypertension); (6) previous cardiac surgery, including alcohols septal ablation, percutaneous radiofrequency ablation, and surgical myectomy; (7) uninterpretable images for perfusion analysis. In addition, 21 healthy individuals matched for age and sex were included as the control group.

All patients underwent a standard cardiac MRI on a 3 T MRI system (MAGNETOM Skyra, Siemens Healthcare, Erlangen, Germany). The CMR protocol included short-axis cine, native T1 mapping, T2 mapping, first-pass perfusion, and late gadolinium-enhanced (LGE) imaging. Cine images were acquired with ECG gating and breath-holding using a segmented, balanced, steady-state free-precession sequence. 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, and flip angle = 46°. T1 mapping images were acquired using the modified look-locker recovery sequence with 5b(3b)3b (b for heartbeat) scheme at the left ventricular basal, mid, and apical levels. The acquisition parameters were: echo time = 1.2 ms; repetition time = 2.8 ms, inversion time = 197 ms, increase step = 80 ms, flip angle = 35°, field of view = 360 × 324 mm², matrix = 257 × 232, and slice thickness = 5 mm. The parameters for T2 mapping are as follows: echo time = 1.35 ms; flip angle = 12°, field of view = 360 × 323 mm², matrix = 189 × 170 mm, and slice thickness = 5 mm, T2 preparation duration = 0, 30, 55 ms. The CMR perfusion image used a gradient echo sequence with 0.3 mmol/kg of gadolinium-based contrast agent (Adobenate Dimeglumine, Berlin, Germany). In addition, the scan parameters were: matrix size = 211 × 181, field of view = 380 × 326 mm², slice thickness = 8 mm, repetition time = 2.0 ms, echo time = 0.97 ms, flip angle = 18°, and nominal inversion time = 130 ms. LGE used a phase-sensitive, inverse recovery sequence 10–15 min after contrast injection. The scan parameters were: matrix size = 192 × 173, field of view = 350 × 263 mm², slice thickness = 6.0 mm, repetition time = 3.2 ms, echo time = 1.27 ms, flip angle = 10°, and nominal inversion time = 255 ms.

CMR analysis was performed using CVI 42 software (version 5.14.0, Circle Cardiovascular Imaging Inc., Canada). The left atrial (LA) anteroposterior and LA left–right diameters were obtained on three- and four- 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 LA ejection fraction (LAEF) was calculated as follows: LAEF = [(LAV_max−LAV_min) / LAV_max] × 100%. Left ventricular (LV) structure and functional parameters were measured on short-axis cines, including the maximum wall thickness, LV ejection fraction (LVEF), LV end-diastolic volume (LVEDV), LV end-systolic volume (LVESV), stroke volume (SV), cardiac output (CO) and LV mass (LVM). Some parameters were normalized to body surface area: LAV_max index, LAV_min index, LV end-diastolic volume index (LVEDVi), cardiac index (CI) and LV mass index (LVMi). In addition, the percent weight of the resected myocardium was calculated by dividing the surgically removed myocardial mass by the mass of the left ventricular myocardium at end-diastole.

The myocardium was divided into sixteen American Heart Association (AHA) segments for the perfusion assessment. Time-intensity curves were acquired for each segment through the entire perfusion process. The time-to-maximal signal intensity (time_max), maximal upload of myocardial intensity enhancement (Slope_max), and maximal signal intensity (Sl_max) were measured (Figure 1). Strengthening myocardium was identified when LGE was analyzed as the range of a signal threshold ≥ five standard deviations of reference myocardium. Lastly, the native myocardial T1 and T2 values were measured using three slices generating T1/T2 maps from the base to the apex.

Statistical data analysis was conducted using SPSS 25.0 software (IBM SPSS Inc., Chicago, USA). The measurement data conforming to the normal distribution were expressed as means ± standard deviation (x ± s). The data at baseline and follow-up were compared using the paired t-test. Data that did not conform to the normal distribution were indicated by medians and quartile M (Q1, Q3), and the data before and after surgery were compared using the paired Wilcoxon signed-rank test. The Pearson correlation coefficient or Spearman's rank correlation coefficient (r) was used to assess the association of continuous variables.

Changes in each variable were represented by △, equaling the preoperative value minus the postoperative value. Potential variables associated with Slope_max and Sl_max were analyzed by univariate analysis. The variables with p < 0.05 by univariate analysis were inputted into the multivariate model as covariates. We randomly selected 20 subjects to assess the intra- and inter-observer reproducibility of left ventricular function, perfusion, and histological parameters by Bland–Altman analysis. Lastly, a p-value < 0.05 was considered statistically significant.

Twenty-eight HOCM patients were enrolled, including fourteen males and seven females, aged 12∼72 years. Twenty-four (85.7%) patients had asymmetric septal hypertrophy, and four (14.3%) had apical hypertrophy. The maximal LV wall thickness was 22.8 ± 1.1 mm. In total, 448 segments were analyzed. Perfusion, native T1 mapping and T2 mapping images were interpretable in all segments. Notably, the mean weight of the HOCM resected myocardium was 6.7 ± 0.6 g. Clinically, the primary symptom experienced by the patients was dyspnea (71.4%). The baseline characteristics of HOCM patients are shown in Table 1.

Clinical, ECG, and laboratory data were collected one week before and three months after TA-BSM. All participants experienced relief of symptoms after TA-BSM, and LVOTG was significantly reduced (87.2 ± 5.8 mmHg vs. 13.2 ± 1.3 mmHg, p < 0.001) (Figure 2A). At follow-up, we found that the LVOTG increased in some patients three months after surgery relative to the immediate postoperative pressure gradient, within the normal range. In addition, the preoperative cardiac function of eighteen patients (64.3%) was classified as NYHA grade Ⅲ or Ⅳ, while the proportion after surgery was reduced to 10.7% (Figure 2B). Finally, the number of participants with grade 2 and below mitral regurgitation increased from eight (28.6%) to twenty-five (89.2%) (Figure 2C).

Comparisons of the preoperative and postoperative LA parameters are shown in Table. 2. Postoperative LA anteroposterior diameter, LA left-right diameter, LAV_max index, and LAV_min index were decreased, while LAEF was increased.

The pre- and postoperative parameters of left ventricular function in the control and study groups are listed in Table 3. The end-diastolic wall thickness of the left ventricle was significantly reduced after TA-BSM (22.8 mm vs. 17.7 mm, p < 0.001). In addition, LVEF, LVMi, SV, CI, native T1 value, T2 value and LGE volume were decreased (all p < 0.05), while LVMi, native T1 value, and T2 value were higher than the control group (all p < 0.001).

The Slope_max and Sl_max were elevated, indicating that the blood filling rate and myocardial blood flow in the coronary microcirculation were significantly improved (Figures 3A,B). Similarly, the decrease in time_max means that the blood fully entered the cardiomyocytes more rapidly (Figure 3C). A representative case of myocardial perfusion before and after TA-BSM is shown in Figures 4, 5.

Patients were divided into two subgroups according to whether LGE was ≥15%. The rate of slope_max change in the LGE ≥ 15% group was significantly greater than in the LGE < 15% group (p = 0.002). However, there were no statistically significant differences in other perfusion parameters between the two subgroups (Figure 6). Given the collinearity of myocardial perfusion parameters, we only performed regression analysis for slope_max because it directly reflected the rate of blood flow filling of the myocardium. The results of the linear regression analysis are shown in Table 4. Age, weight of the resected myocardium, maximum wall thickness, LGE, and ΔLVOTG were significantly associated with the rate of slope_max change in the univariate and multivariate linear regression analyses (all p < 0.05).

Inter- and intra-observer agreement for the perfusion parameters was high (r = 0.94 (0.89–0.97) and r = 0.95 (0.91–0.98), p < 0.05).