Noncontrast cardiac computed tomography‐derived mitral annular calcification scores in mitral valve disease

Abstract Background and Aims Mitral annular calcification (MAC) by computed tomography (CT) is reported as an independent predictor of poor outcomes. However, it currently remains unclear if quantitative MAC parameters provide more value for mitral valve disease (MVD) management, therefore, we examined the prognostic value of MAC scores using noncontrast cardiac‐CT in MVD patients. Methods Between January 2020 and December 2021, we prospectively enrolled 300 consecutive patients with MVD (MAC‐present = 80 and MAC‐absent = 220) undergoing preoperative cardiac‐CT and mitral valve (MV) surgery. Noncontrast cardiac‐CT images were used to qualitatively detect MAC (present or absent) and evaluate MAC scores. For analyses, we also collected baseline clinical data, intraoperative conversion (from MV repair to MV replacement), and follow‐up arrhythmia data. Results Compared with the MAC‐absent group, MAC‐present patients were older (62 ± 7 vs. 58 ± 9 years, p < .001), mostly women (55% vs. 39.5%, p = .017), and also had aortic valve calcification (57.5% vs. 23.2%, p < .001), mitral stenosis (82.5% vs. 61.8%, p < .001), atrial fibrillation (30% vs. 11.8%, p < .001), and larger left atrial end‐diastolic dimension (LADD, 49 [44–56] versus 46 [41–50], p = .001]. Furthermore, MAC‐present patients underwent more MV replacements (61.8% vs. 82.5%, p = .001) and experienced a higher intraoperative conversion prevalence (11.8% vs. 61.3%, p < .001). Multiple logistic regression analyses showed that the female gender (odds ratio [OR]/95% confidence interval [CI]/p = 2.001/1.042–3.841/0.037) and MAC scores (OR/95% CI/p = 10.153/4.434–23.253/p < .001) were independent predictors of intraoperative conversion. During a follow‐up of 263 ± 134 days, MAC‐present patients had more arrhythmias (42.5% vs. 9.5%, p < .001). Also, MAC‐scores (hazard ratio [HR]/95% CI/p = 6.841/3.322–14.089/p < .001) and LADD (HR/95% CI/p = 1.039/1.018–1.060/p < .001) were independently associated with arrhythmias by Cox regression analyses. Conclusions Noncontrast cardiac CT‐derived MAC‐scores showed a high risk for intraoperative conversion and follow‐up arrhythmias in MVD‐patients.


| INTRODUCTION
Characterized as a progressive and chronic degenerative process in the fibrous annulus of the mitral valve (MV), mitral annular calcification (MAC) is often an incidental, asymptomatic, and underreported finding. 1,22][3][4] The condition is also associated with elevated left ventricular afterload, including hypertrophic cardiomyopathy with obstruction, hypertension, and valvular aortic stenosis (AS).MAC may lead to mitral stenosis (MS) and/or mitral regurgitation (MR), with concomitant severe AS requiring ameliorative doublevalve intervention. 5,6C is associated with elevated perioperative complications and all-cause mortality risks 1,2,5 ; it reportedly causes a sixfold increase in operative mortality in patients undergoing isolated MV surgery, 7 while early mortality rates, upon surgical MV replacement in MAC, reportedly as high as 28%. 1 Generally, replacing or repairing MV in severely affected patients with MAC is technically difficult, even when concomitant aortic valve replacement risks are removed.8,9 Multidetector computed tomography (MDCT) generates detailed MAC assessment before surgery and may alter therapeutic strategies.12][13] The approach, characterized by high X-ray calcium attenuation, excellent spatial resolution, and three-dimensional postprocessing analysis, assesses MAC characteristics, total calcium distribution, and coronary artery and aortic valve calcification (AVC).10,14 MAC is reportedly an independent predictor of poor outcomes, 1,2,5 with prognostic value for atrial fibrillation (AF) ablation 15 and transcatheter aortic valve implantation.3 Noncontrast cardiac CT is a semi-automated quantification method for calculating calcium burden (MAC scores), such as coronary calcification scores which are used to quantitatively evaluate calcification, and provide more accurate risk assessments and disease prognosis predictions for multicenter clinical research.However, the value of MAC scores from noncontrast cardiac CT for mitral valve diseases (MVDs) have been rarely reported.The literature is limited in determining if MAC is associated with surgical method choice and predicting postoperative arrhythmia in MVD. Terefore, new investigations must ascertain if MAC provides useful clinical information enabling early intervention and improving treatment strategies.In our retrospective study, we examined MAC incidence, explored the clinical value of MAC scores in selecting surgical methods, and identified its potential predictive power in patients with MVD.

| PATIENTS AND METHODS
After approval from our ethics committee, informed patient consent was waived due to the retrospective nature of our investigation.

| Patient selection
In our single-center hospital (in the Department of Cardiovascular Surgery), between January 2020 and December 2021, we evaluated 300 consecutively admitted patients with MVD.We adhered to the 2017 European Society of Cardiology/European Association for Cardio-Thoracic Surgery guidelines outlining MVD patient management. 16Upon admission and before surgery, patients underwent echocardiography and electrocardiography (ECG)/dynamic ECG. 17 Before surgery, patients also underwent cardiac computed tomography angiography to evaluate coronary artery and intracardiac disease.
To be included in the study, patients with MVD were ≥18 years or older, had noncontrast cardiac CT detection, and had undergone MV surgery.Patients were excluded if they had previous valvular surgery or ablation for AF, poor image quality, and had not undergone previous surgery.Patient basic characteristics, intraoperative conversion, follow-up data, and arrhythmia information were retrospectively collected from the 300 patients.

| Imaging
Imaging was performed using a 640-slice MDCT scanner (Aquilion | 1311 reconstruction phase = 75%.ECG editing technology was used to reconstruct images in severe arrhythmia.

| MAC
MAC was examined using noncontrast cardiac-gated CT.MAC scores and distribution were gathered using the Agatston method 18,19 (Supporting Information: Figure S1).Noncontrast images (0.5-mm slices and 0.5-mm increments) were assessed using semiautomatic software (VScore, Vitrea, Vital Images).We selected the diastolic phase of the cardiac cycle, with maximal MV plane, using 4-and 2-chamber views.
We recorded calcific deposit status in MVs or annulus segments.
Manual editing was performed to eliminate aortic or coronary calcium.
To define calcium areas, Agatston scores using a CT attenuation threshold = 130 Hounsfield units were used, 12 and the maximum CT attenuation in lesions was used to generate weighting scores.Weight = 1 indicated an attenuation of 130-199; 2 = 200-299; 3 = 300-399; and 4 ≥ 400. 13The weighting factor was multiplied by lesion area, with the total of lesions values used to determine total Agatston scores 13,20 (Figure 1).CT scans were separately and independently analyzed by two experienced and blinded cardioradiologists (≥5 and ≥3 years' experience, respectively).Discordance was settled by discussion and consensus.MAC scores were also recorded by cardioradiologists to identify intraclass correlation coefficients (ICCs) and evaluate the precision and accuracy of the MAC score method.

| Patient follow-up
Patients were followed up to December 2022.Primary study outcomes were arrhythmia (including atrioventricular block, AF, and bundle branch block) recurrence during routine follow-up (>3 months postsurgery), which required direct or drug current cardioversion.In the first 3 months postsurgery ("blank period"), arrhythmias were not recorded as adverse events.The period between the surgery date and arrhythmia recurrence was recorded as the time to event (arrhythmia recurrence).
Follow-up at outpatient visits or rehospitalization included echocardiography and ECG evaluations after surgery.

| Statistical analyses
Data were analyzed in SPSS v. 20.0 (IBM Corporation).Continuous data were represented as the mean ± standard deviation or median (quartiles), and analyzed using independent sample t-or Mann-Whitney U tests.Normal distributions across continuous variables were examined using Kolmogorov-Smirnov tests.Categorical data (numbers and percentages) were analyzed using Fisher's exact or Pearson's χ 2 tests (Tables 1 and 2).Inter-and intraobserver agreement data for subjectively assessing MAC and AVC occurrence were evaluated using cross-tabulation and kappa (κ) calculations.To determine significant independent predictors, multivariate logistic regression analyses were performed (Table 3).Also, to determine collinear covariates, multicollinearity analyses were performed.We used Cox regression for follow-up arrhythmia analyses after surgery, and parameters with significant effects in univariate Cox regression analysis underwent multivariate Cox regression (Table 4).We used receiver operating characteristic (ROC) analyses to examine predictive potential factors in a multivariate-adjusted logistic regression model.A p < .05value indicated statistical significance.

| Study population
Of the 300 patients (mean age = 59 ± 9 years and 43.7% were females), 80/300 (26.7%) were assigned to the MAC-present group 1 showed the case of a mitral valve disease with MAC.MAC score was evaluated by CT.MAC volume was measured at 1372 mm 3 , MAC score was measured at 1767.CT, computed tomography; MAC, mitral valve calcification.From surgery data, MV replacement (χ 2 = 11.409,p = .001),aortic valve replacement (χ 2 = 3.952, p = .047),and intraoperative conversion from MV repair to replacement ratios were higher in the MACpresent group when compared with the MAC-absent group (χ 2 = 76.455,p < .001).No significant differences (p > .05) in followup duration were observed between groups.During follow-up (263 ± 134 days), patients with MAC had a higher arrhythmia prevalence (χ 2 = 42.554,p < .001),while follow-up echocardiographic findings were significantly improved when compared with those at admission (p < .05).Additionally, a significant difference was observed in follow-up LADD between groups after follow-up (Z = −5.051,p < .001;Table 1).

| MV calcification assessment
We identified 80 patients (

| DISCUSSION
When compared with the MAC-absent group, patients with MAC were older and mainly female, and had AF, AVC, AS, MS, and larger LADD.Patients with MAC had a higher prevalence of intraoperative conversion from MV repair to replacement.MAC scores and female gender were independent predictors of intraoperative conversion.
During follow-up, MAC patients had an increased arrhythmia incidence.MAC scores and LADD were independent arrhythmia predictors.Thus, MAC was an important imaging index in MVD prognosis outcomes and treatment.MAC scores, based on quantitative nonenhanced cardiac CT evaluations, were important in predicting intraoperative conversion and postoperative arrhythmia events in patients with MVD.
Recent studies reported that MAC is an active and controlled molecular event associated with microscopic and macroscopic injury, lipid deposition, hemodynamic stress, chronic kidney disease, dysregulated bone and mineral metabolism regulators, and local inflammation. 21,22Baseline MAC burden was also related to disease | 1315 activity and disease progression rates. 21MAC appears to induce anatomical changes which culminate in either MS or combined MS and MR, while MS in severe MAC settings is caused by encroaching orifice areas, and rheumatic MS arises due to an absence of leaflet commissural union. 23MR is generated by an altered annulus during systole or leaflet coaptation distortion, which cause left atrium volume and pressure overload, leading to enlargement. 24,25Pawade et al. 13 reported that AVC should be measured using noncontrast CT and the Agatston approach.In the valve, the majority of data are related to Agatston scores and not calcium volume measurements.
Density weighting is likely advantageous, the denser the calcium deposits, the more likely they will cause hemodynamic obstruction and valve-leaflet stiffening.We recorded excellent inter-and intraobserver agreements between operators who measured MAC from cardiac CT images, consistent with previous studies. 2,26We showed that noncontrast cardiac-gated CT is a good semiquantitative method assessing MAC severity.MAC is common in cardiovascular imaging and postmortem and surgical samples, with an estimated 8%-42% prevalence. 11We also showed that MAC prevalence in patients with MVD was 25.9%, consistent with previous results. 11tients with MAC were advanced in age and more likely to be female, with hypertension and valvular heart disease.8][29] These observations suggested overlapping but distinct mechanisms underlying these pathologies.
Interestingly, MAC scores and female gender were independent risk markers for intraoperative conversion; indeed, the literature indicated that MAC was more prevalent in females.
T A B L E 1 Clinical characteristics of all patients.
31,30While surgical treatment in patients with MAC is technically complex, there is a need for annular reconstruction and adequate debridement before MV replacement or repair.31In such es, MV repair may not be undertaken due to difficulties suturing calcified sites and severe calcification, thereby requiring prosthetic valve replacement.PatientsT A B L E 3 T A B L E 4Abbreviations: AF, atrial fibrillation; AS, aortic stenosis; CI, confidence interval; HR, hazard ratio; LADD, left atrial end-diastolic dimension; MAC, mitral annular calcification; MS, mitral stenosis; MVD, mitral valve disease.