Machine Learning of Multimodal Imaging to Differentiate Constrictive Pericarditis and Restrictive Cardiomyopathy

Machine Learning of Multimodal Imaging to Differentiate Constrictive Pericarditis and Restrictive Cardiomyopathy | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Machine Learning of Multimodal Imaging to Differentiate Constrictive Pericarditis and Restrictive Cardiomyopathy Zhiyun Yang, Xiangbin Meng, Rui Zuo, Hongkai Zhang, Xin Du, Jianzeng Dong, and 2 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-8917150/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 8 You are reading this latest preprint version Abstract Aim: Constrictive pericarditis (CP) and restrictive cardiomyopathy (RCM) require distinct treatments but share overlapping clinical and imaging features, making them difficult to distinguish using single conventional imaging. This study evaluated whether machine learning models integrating multimodal imaging data could improve the differential diagnosis. Methods: CP and RCM patients were included between January 2014 and September 2024 at two hospitals. Demographic, laboratory, echocardiography, computed tomography, magnetic resonance imaging, and cardiac catheterization data were collected. Seven machine learning models were trained and validated using five-fold cross-validation. Model performance was assessed by the area under the receiver-operating characteristic curve (AUC), accuracy, sensitivity, specificity and 95% confidence interval (CI). Results: A total of 156 CP and 91 RCM patients were analyzed. Compared with RCM, CP patients showed a higher ejection fraction (62.9 ± 7.5% vs. 51.9 ± 15.2%, P < 0.001) and more frequent pericardial thickening over 4 mm (66.1% vs. 0%, P < 0.001). The Gaussian SVM of multimodal-imaging achieved the highest AUC of 0.97 (95%CI: 0.93–0.99), accuracy of 85%, sensitivity of 89%, and specificity of 90%, outperforming other SVM models, decision tree, logistic regression and k-nearest-neighbor models. ML based on multimodal imaging data achieved higher diagnostic performance than those based on any single-modality imaging features. Conclusions: A Gaussian SVM integrating multimodal imaging data markedly improves the differential diagnosis between CP and RCM. This model may help reduce misclassification and improve timely interventions for the two diseases. Constrictive pericarditis restrictive cardiomyopathy multimodal imaging support vector machine cardiac magnetic resonance imaging. Figures Figure 1 Figure 2 Figure 3 Introduction Constrictive pericarditis (CP) and restrictive cardiomyopathy (RCM) both present preserved ejection function, impaired left ventricular (LV) relaxation, and normal-sized ventricles with dilated atria on imaging. 1 Despite these similarities, the two conditions require fundamentally different management strategies. CP can be cured by pericardiectomy, whereas RCM derives no surgical benefit and is treated mainly with supportive therapy. Misclassification may therefore expose patients to unnecessary operative risk or delay definitive treatment. 2 Current guidelines recommend a stepwise approach for distinguishing CP from RCM, beginning with transthoracic echocardiography, followed by computed tomography (CT) or cardiac magnetic resonance (CMR), with invasive catheterization reserved for cases where noninvasive findings remain inconclusive. 3,4 Although each modality reflects a distinct aspect of the underlying pathophysiology, none alone provides consistent diagnostic accuracy. Echocardiography is operator-dependent and may yield inconclusive hemodynamic findings, particularly in patients with atrial fibrillation, obesity, or elevated filling pressures, which together account for up to 30% of cases. 5 Although CT can delineate pericardial anatomy and calcification, its diagnostic sensitivity is limited in early-stage or post-radiation constrictive pericarditis, where pericardial thickening may be minimal. Approximately 28% of surgically confirmed CP cases show pericardial thickness ≤ 4 mm. 6,7 CMR provides valuable tissue characterization through late gadolinium enhancement (LGE) and strain analysis, but its accuracy is influenced by vendor-specific techniques, complex post-processing, and limited assessment of ventricular interdependence, a key hemodynamic feature of CP. 8–11 The 2023 American Society of Echocardiography consensus therefore endorses a multimodality imaging strategy when echocardiographic findings are equivocal, acknowledging that no single noninvasive test can reliably distinguish constriction from restriction in all cases. 3 Given these diagnostic limitations, machine learning (ML) offers a promising avenue by integrating diverse quantitative parameters across modalities to uncover patterns not readily discernible to human observers. 12 Prior studies have demonstrated the potential of ML, with deep learning models achieving high accuracy in distinguishing CP from cardiac amyloidosis using echocardiographic imaging features. 13 However, these “black-box” models have limited interpretability and generalizability across the broader RCM spectrum. To date, no study has replicated the clinician’s multimodal reasoning by integrating structured quantitative data from echocardiography, CT, and CMR within a unified and interpretable ML framework. 14,15 In this context, the present study aims to develop and validate ML classifiers that combine multiple imaging modalities to differentiate CP from RCM. The goal is to provide a practical and interpretable noninvasive tool to support clinical decision-making and optimize referral for catheterization or surgery, with the ultimate aim of improving diagnostic precision and patient outcomes. Methods Study population This retrospective study included consecutive CP and RCM patients who underwent multimodal imaging examinations including CT, echocardiography, CMR-FT, and cardiac catheterization at Anzhen Hospital, Capital Medical University, and Peking University Third Hospital, China between January 2014 and September 2024. This study was conducted in accordance with the Declaration of Helsinki and was approved by the Institutional Ethics Committees of Peking University Third Hospital and Beijing Anzhen Hospital, Capital Medical University (Approval No. 2006003x). CP and RCM diagnostic criteria All enrolled CP and RCM patients were in sinus rhythm. Individuals with significant valvular diseases, arrhythmias, prior myocardial infarction, hypertension, or other causes of myocardial hypertrophy or dysfunction were excluded. CP was confirmed either during surgical pericardiectomy or, when the surgery was not performed, by the simultaneous presence of pericardial thickening > 4 mm on CT, echocardiography, or CMR and objective evidence of increased left-to-right ventricular coupling manifested as a respiratory septal shift on echocardiography/CMR or a right-to-left ventricular systolic area index ratio > 1.1 during both inspiration and expiration on cardiac catheterization. 16,17,18 RCM was diagnosed either by 1) tissue biopsy of the heart, kidney, or other organs demonstrating amyloidosis, or by 2) all of the following imaging evidence: interventricular septal thickness > 12 mm, preserved LV ejection fraction (LVEF) > 50%, bi-atrial enlargement, a restrictive filling pattern, and late gadolinium enhancement on CMR indicating myocardial involvement. 19 Multimodal imaging acquisition and processing Each patient followed a standardized imaging protocol (Fig. 1 ): 1) non-contrast chest CT for evaluation of pericardial calcification and wall thickening, 2) transthoracic echocardiography for measurement of LV end-diastolic and end-systolic diameters (LVEDD and LVESD), LVEF, septal thickness, early (E) and late (A) transmitral velocities, mitral-annular early-diastolic velocity (e′), and the derived E/A and E/e′ ratios, 3) right and left heart catheterization, when clinically indicated, for recording atrial, ventricular, and pulmonary artery pressures and for calculating the right-to-left ventricular systolic area index required to confirm CP, and 4) cardiac MRI with long- and short-axis cine steady-state free-precession, rest first-pass perfusion, and late gadolinium enhancement sequences for comprehensive assessment of ventricular function and myocardial tissue characteristics. Image post-processing was conducted on a commercial workstation running CVI42 (version 5.11.3, Circle Cardiovascular Imaging). Endocardial and epicardial borders were delineated in end-diastolic and end-systolic frames, after which LV end-diastolic and end-systolic volumes (LVEDV, LVESV), LVEF, and diastolic myocardial mass were quantified automatically. Three-dimensional feature-tracking analysis was subsequently applied to derive global longitudinal, circumferential, and radial peak strain, peak strain rate, time-to-peak strain, and velocity. All imaging acquisitions were completed before any pericardiectomy or other therapeutic intervention in the CP group. ML for CP and RCM classification The ML workflow consisted of data preprocessing, model training/validation, and performance assessments. Any field with < 20% missingness was imputed by k-nearest-neighbor (KNN) learning during preprocessing. 20 LV global time-strain curves derived from CMR-FT were normalized to the cardiac cycle by referencing successive end-diastolic frames, after which strain ratios for the 0–50%/50–75% and 50–75%/75–100% diastolic intervals were calculated in accordance with a published method. 21 ML was performed using support vector machines (SVMs), decision tree, logistic regression and KNN. SVMs were chosen as the primary supervised algorithm because of their capacity to handle small, high-dimensional data sets. Linear (L-SVM), quadratic (Q-SVM), cubic (C-SVM), and Gaussian-kernel (G-SVM) variants were implemented with the LIBSVM toolbox under Matlab R2017b (MathWorks, Natick, MA, USA). 22 During the training, class labels were paired with predictor matrices, and the optimal hyper-plane separating CP from RCM observations was generated. All ML classifiers were trained on the same dataset using imaging features from echocardiography, MRI or CT as model inputs. Five-fold cross-validation was applied: in each fold, 20% of subjects from both diagnostic categories were reserved for testing and 80% for training. The procedure was repeated across all folds to minimize sampling bias. Mean classification accuracy over the five folds was calculated as the final result. ML was performed with both single- and multiple-modal imaging features as the model inputs. ML performance metrics were derived from the confusion matrix using the following equations: 23 $$\:\text{A}\text{c}\text{c}\text{u}\text{r}\text{a}\text{c}\text{y}=\frac{\text{T}\text{P}\:+\:\text{T}\text{N}}{\text{T}\text{N}\:+\:\text{F}\text{P}\:+\:\text{F}\text{N}\:+\:\text{T}\text{P}}$$ $$\:\text{S}\text{e}\text{n}\text{s}\text{i}\text{t}\text{i}\text{v}\text{i}\text{t}\text{y}=\frac{\text{T}\text{P}}{\text{T}\text{P}\:+\:\text{F}\text{N}}$$ $$\:\text{S}\text{p}\text{e}\text{c}\text{i}\text{f}\text{i}\text{c}\text{i}\text{t}\text{y}=\frac{\text{T}\text{N}}{\text{T}\text{N}\:+\:\text{F}\text{P}}$$ $$\:\text{P}\text{r}\text{e}\text{c}\text{i}\text{s}\text{i}\text{o}\text{n}=\frac{\text{T}\text{P}}{\text{T}\text{P}\:+\:\text{F}\text{P}}$$ where TP is true positive, FP is false positive, FN is false negative, and TN is true negative. Receiver-operating characteristic (ROC) curves were plotted, and the area under curve (AUC) was computed with Matlab Statistics. Statistical analysis Continuous data were expressed as mean ± SD for variables with a normal distribution (e.g., LV septum thickness, LVEF) and as median (interquartile range) for skewed variables (e.g., age, BNP, LVEDD). Categorical data were summarized as count (percentage) (e.g., sex, pericardial calcification). Between-group differences were assessed with the independent-samples t -test for normally distributed continuous variables, and the Mann-Whitney U test for non-normal distributions. Categorical variables were compared with the χ² test or Fisher’s exact test, as appropriate. A two-sided P value < 0.05 was considered statistically significant. Statistical analyses were performed with SPSS 22.0 (IBM, Armonk, NY, USA). Results Clinical characteristics A total of 247 patients were included, with 156 CP and 91 RCM (Fig. 2 ). In the CP group, 98 (63%) were confirmed surgically and 58 (37%) were diagnosed by multimodal imaging comprising CT, echocardiography, CMR and cardiac catheterization. In the RCM group, 51 (56%) had biopsy-proven amyloidosis, and the remaining 40 (44%) were diagnosed by multimodal imaging evidence of diffuse myocardial infiltration, subendocardial late gadolinium enhancement and fibrosis. Table 1 summarizes the patients’ baseline characteristics. The CP group had more females (55.4% vs. 39.0%, P = 0.025) and a lower BNP level (188 (300) ng/ml vs. 788 (1408) ng/ml, P = 0.001). No significant differences in other baseline clinical variables were observed between the two groups, highlighting the difficulty of distinguishing CP from RCM based on clinical characteristics alone. Table 1 Clinical characteristics (n = 247) CP (n = 156) RCM (n = 91) P value Age, y 58.0 (19) 58.0 (14) 0.876 Female, n (%) 85 (55.4%) 36 (39.0%) 0.025 Height, cm 167.3 ± 8.4 164.1 ± 7.8 0.341 Weight, kg 66.8 ± 13.1 63.1 ± 12.0 0.812 BSA, cm/m 2 1.74 ± 0.21 1.68 ± 0.18 0.952 BMI, kg/m 2 23.8 ± 3.5 23.6 ± 4.6 0.413 Heart rate, bpm 85.0 ± 15.8 77.9 ± 12.6 0.091 Atrial fibrillation, n (%) 22 (14%) 9 (10%) 0.090 Hypertension, n (%) 61 (39%) 40 (44%) 0.574 Diabetes mellitus, n (%) 45 (29%) 25 (27%) 0.645 Familial history of CAD, n (%) 17 (11%) 9 (10%) 0.610 Laboratory testing BNP, ng/mL 188 (300) 788 (1408) 0.001 GPT, U/L 17.0 (12) 22.0 (20) 0.411 Creatinine, mg/dL 75.2 (23) 78.1 (41) 0.197 BMI, body mass index; BNP, B-type natriuretic peptide; bpm, beats per minute; BSA, body surface area; CAD, coronary artery disease; CP, constrictive pericarditis; GPT, glutamic pyruvic transaminase; RCM, restrictive cardiomyopathy. Imaging findings from echocardiography, CT, CMR and cardiac catheterization Table 2 summarizes the imaging findings in patients. Echocardiography showed a thinner LV septum (P < 0.001), smaller LVEDD (P = 0.007) and LVESD (P < 0.001), and a higher LVEF (P < 0.001) in the CP group than the RCM group. Chest CT detected pericardial calcification and thickening only in the CP group. On CMR, CP had a lower myocardial mass (P < 0.001) and a higher LVEF (P = 0.013). Myocardial enhancement was observed in most RCM patients (80/91, 87.8%) compared with few CP patients (8/156, 5.4%) (P < 0.001). In CMR-FT, the CP group showed larger peak strain and peak strain rate in the longitudinal, circumferential and radial directions compared with RCM (all P < 0.05). Longitudinal and radial myocardial velocities were also higher in CP (P < 0.05). After normalizing the global time-strain curves, strain ratios were calculated for the early- (0–50%), mid- (50–75%), and late- (75–100%) diastolic intervals. The only significant inter-group difference emerged in the strain ratio of circumferential mid- to late-diastolic segment (50–75%/75–100%), which was lower in CP than in RCM (P = 0.005). Pressure indices from cardiac catheterization were similar in CP and RCM (all P > 0.05). Table 2 Cardiac imaging findings (n = 247) CP (n = 156) RCM (n = 91) P Value Echocardiography LV septum, mm 8.4 ± 1.7 13.1 ± 4.4 < 0.001 LVEDD, mm 39.0 (8) 44.5 (9) 0.007 LVESD, mm 26.0 (6) 32.0 (8) < 0.001 LVEF, % 62.9 ± 7.5 51.9 ± 15.2 < 0.001 E wave, m/s 83.0 ± 21.6 74.6 ± 24.5 0.537 A wave, m/s 49.0 (31) 58.5 (44) 0.519 E/A ratio 1.6 (1) 1.3 (1) 0.119 e’, cm/s 8.5 (1) 7.2 (1) 0.667 E/e′ ratio 9.9 (2) 11.6 (2) 0.237 CT Pericardial calcification, n (%) 95 (60.7%) 0 4 mm, n (%) 104 (66.1%) 0 <0.001 CMR LVEDV, ml 88.5 (35.1) 118.0 (77.1) 0.255 LVESV, ml 41.2 (22.6) 58.5 (66.0) 0.057 LVEF, % 50.6 ± 11.8 42.7 ± 16.1 0.013 Myo Mass, g 64.1 (27) 116.2 (84) <0.001 Myocardial enhancement, n (%) 8(5.4%) 80(87.8%) <0.001 Longitudinal Peak strain (%) -12.2 ± 3.9 -8.3 ± 4.0 0.001 Peak strain rate (1/s) 0.9 (0.6) 0.5 (0.3) <0.001 Time to peak strain (ms) 294.3 (77.9) 271.3 (99.8) 0.395 Velocity (mm/s) -28.8 (28.9) -15.3 (20.9) 0.002 Strain ratio of 0–50%/50–75% diastolic period 3.1 (5.1) 3.6 (4.1) 0.301 Strain ratio of 50–75%/75–100% diastolic period 0.1 (2.3) 0.3 (3.6) 0.629 Circumferential Peak strain (%) -15.2 ± 4.8 -11.6 ± 4.8 <0.001 Peak strain rate (1/s) 1.1 (0.7) 0.8 (0.4) 0.002 Time to peak strain (ms) 266.5 (57.6) 246.0 (62.6) 0.180 Velocity (mm/s) -16.4 (35.7) -14.9 (12.3) 0.932 Strain ratio of 0–50%/50–75% diastolic period 4.5 (8.7) 5.0 (5.9) 0.842 Strain ratio of 50–75%/75–100% diastolic period 0.3 (1.5) 1.6 (7.4) 0.005 Radial Peak strain (%) 26.1 (15.5) 12.8 (16.8) <0.001 Peak strain rate (1/s) -2.1 (2.1) -0.8 (1.2) <0.001 Time to peak strain (ms) 260.2 (60.3) 238.0 (66.1) 0.230 Velocity (mm/s) -26.0 (17.6) -18.4 (11.6) 0.019 Strain ratio of 0–50%/50–75% diastolic period 6.7 (10.3) 5.7 (10.5) 0.879 Strain ratio of 50–75%/75–100% diastolic period 0.3 (1.5) 1.1 (3.8) 0.076 Cardiac catheterization RAP, mmHg 20.4 (6.9) 23.2 (7.0) 0.329 RVEDP, mmHg 20.2 (6.4) 20.0 (6.8) 0.940 PAP, mmHg 20.2 (6.5) 22.1 (7.6) 0.436 PAWP, mmHg 24.0 (4.6) 25.2 (5.6) 0.160 LVEDP, mmHg 23.1 (6.7) 21.2 (6.4) 0.600 Findings supporting CP, No./total No. (%) 6/10 (60.0%) 0/6 0.034 CMR, cardiac magnetic resonance; CMR-FT, cardiac magnetic resonance feature tracking; CP, constrictive pericarditis; CT, computed tomography; E wave, early peak transmitral velocity; e′, mitral annular early-diastolic velocity; E/A ratio, ratio of early to late transmitral velocities; E/e′ ratio, ratio of early transmitral velocity to early mitral annular velocity; LV, left ventricle; LVEDD, left ventricular end-diastolic diameter; LVEDP, left ventricular end-diastolic pressure; LVEDV, left ventricular end-diastolic volume; LVEF, left ventricular ejection fraction; LVESD, left ventricular end-systolic diameter; LVESV, left ventricular end-systolic volume; Myo Mass, myocardial mass (diastolic); PAP, pulmonary artery pressure; PAWP, pulmonary arterial wedge pressure; RAP, right atrial pressure; RCM, restrictive cardiomyopathy; RVEDP, right ventricular end-diastolic pressure. Diagnostic performance of single imaging feature Each imaging-derived metric showing a significant difference (P < 0.05) between CP and RCM in Table 2 was assessed for its ability to distinguish the two conditions. Table 3 summarizes the diagnostic performance for each feature of echocardiography, CT and CMR. Despite the variations in diagnostic performance, most AUCs ranged between 0.60 and 0.80. Notably, late gadolinium enhancement on CMR exhibited the highest diagnostic performance with an AUC of 0.87 (95%CI: 0.80–0.92), sensitivity of 93% and specificity of 81%. It is followed by peak longitudinal strain rate on CMR (AUC = 0.79, 95%CI: 0.69-0. 85, P = 0.001), pericardial thickness (> 4 mm) on CT (AUC = 0.78, 95%CI: 0.73–0.80, P = 0.002) and LV septal thickness on echocardiography (AUC = 0.78, 95%CI: 0.62–0.88, P = 0.002). The distribution of these high-performing parameters across multiple imaging modalities indicates the potential of multimodal imaging integration for more accurate differentiation between CP and RCM. Table 3 Diagnostic performance of single cardiac imaging feature in CP and RCM AUC 95%CI Sensitivity Specificity Accuracy P value Echocardiography LV septum, mm 0.78 0.62–0.88 0.84 0.65 0.75 0.002 LVEDD, mm 0.69 0.51–0.75 0.69 0.57 0.63 0.045 LVESD, mm 0.75 0.67–0.78 0.73 0.63 0.68 0.008 LVEF, % 0.71 0.64–0.74 0.76 0.51 0.64 0.030 CT Pericardial calcification, n (%) 0.75 0.64–0.82 0.50 0.99 0.73 0.008 Pericardial thickness > 4 mm, n (%) 0.78 0.73–0.80 0.57 0.99 0.77 0.002 CMR LVEDV, ml 0.60 0.51–0.70 0.71 0.49 0.60 0.200 LVESV, ml 0.67 0.56–0.74 0.75 0.51 0.64 0.060 LVEF, % 0.69 0.56–0.78 0.66 0.68 0.67 0.040 Myo Mass, g 0.74 0.66–0.78 0.80 0.57 0.69 0.010 Myocardial enhancement, n (%) 0.87 0.80–0.92 0.93 0.81 0.88 < 0.001 Longitudinal Peak strain (%) 0.72 0.61–0.75 0.66 0.60 0.63 0.030 Peak strain rate (1/s) 0.79 0.69-0. 85 0.68 0.81 0.74 0.001 Velocity (mm/s) 0.70 0.66–0.83 0.66 0.64 0.65 0.080 Circumferential Peak strain (%) 0.70 0.60–0.75 0.69 0.57 0.63 0.050 Peak strain rate (1/s) 0.70 0.67–0.74 0.55 0.76 0.64 0.070 Strain ratio of 50–75%/75–100% diastolic period 0.55 0.38–0.68 0.38 0.63 0.52 0.400 Radial Peak strain (%) 0.75 0.73–0.80 0.62 0.74 0.67 0.010 Peak strain rate (1/s) 0.72 0.64–0.82 0.60 0.71 0.64 0.030 Velocity (mm/s) 0.67 0.50–0.79 0.53 0.68 0.59 0.100 AUC, the area under the receiver-operating characteristic curve; CI, confidence interval; CMR, cardiac magnetic resonance; LV, left ventricle; ESD, end-systolic diameter; EF, ejection fraction; Myo Mass, myocardial mass (diastolic). ML for CP and RCM classification ML techniques were utilized to combine multimodal imaging results to achieve more accurate differential diagnosis between CP and RCM. Table 4 and Fig. 3 summarize the diagnostic performance of multimodal imaging and single imaging in the differential diagnosis. Using multimodal imaging, all ML methods based on multimodal imaging achieved an AUC greater than 0.90, with a mean AUC around 0.95 (all P < 0.001). G-SVM demonstrated the highest diagnostic performance with an AUC of 0.97 (95%CI: 0.93–0.99), accuracy of 85%, sensitivity of 89% and specificity of 90% (P < 0.001). In comparison, the single-imaging ML had AUC values below 0.90, with a mean AUC around 0.80. G-SVM remained the best-performing ML technique for echocardiography or CMR alone. KNN tended to show better diagnostic performance than logistic regression and decision tree methods. Table 4. Machine learning for differential diagnosis between CP and RCM Imaging modality Model AUC 95%CI Sensitivity Specificity Accuracy P Value Multimodal imaging SVMs L-SVM 0.94 0.90-0.97 0.82 0.58 0.90 <0.001 Q-SVM 0.93 0.87-0.97 0.75 0.65 0.74 <0.001 C-SVM 0.96 0.88-0.99 0.87 0.76 0.74 <0.001 G-SVM 0.97 0.93-0.99 0.89 0.90 0.85 <0.001 Decision Tree 0.90 0.85-0.92 0.86 0.69 0.85 <0.001 Logistic regression 0.95 0.87-0.99 0.87 0.68 0.80 <0.001 KNN 0.95 0.92-0.98 0.97 0.68 0.87 <0.001 Echocardiography SVMs L-SVM 0.74 0.66-0.82 0.83 0.62 0.82 0.002 Q-SVM 0.82 0.78-0.85 0.75 0.70 0.73 <0.001 C-SVM 0.83 0.79-0.87 0.80 0.37 0.76 <0.001 G-SVM 0.85 0.79-0.91 0.85 0.90 0.81 <0.001 Decision Tree 0.78 0.71-0.85 0.87 0.69 0.80 <0.001 Logistic regression 0.76 0.71-0.81 0.82 0.68 0.76 0.002 KNN 0.85 0.79-0.92 0.88 0.69 0.79 <0.001 CMR SVMs L-SVM 0.82 0.76-0.88 0.84 0.81 0.89 <0.001 Q-SVM 0.81 0.71-0.91 0.75 0.68 0.89 <0.001 C-SVM 0.80 0.75-0.86 0.85 0.67 0.76 <0.001 G-SVM 0.84 0.80-0.88 0.74 0.81 0.89 <0.001 Decision Tree 0.75 0.68-0.83 0.72 0.83 0.89 <0.001 Logistic regression 0.82 0.75-0.88 0.82 0.81 0.89 <0.001 KNN 0.84 0.80-0.88 0.84 0.78 0.88 <0.001 CT SVMs L-SVM 0.75 0.70-0.80 0.82 0.44 0.68 0.007 Q-SVM 0.72 0.72-0.80 0.75 0.83 0.68 0.020 C-SVM 0.75 0.72-0.78 0.80 0.67 0.73 <0.001 G-SVM 0.75 0.72-0.80 0.84 0.80 0.79 <0.001 Decision Tree 0.78 0.71-0.85 0.64 0.74 0.68 0.023 Logistic regression 0.76 0.71-0.81 0.71 0.76 0.73 <0.001 KNN 0.76 0.72-0.80 0.80 0.66 0.81 <0.001 AUC, the area under the receiver-operating characteristic curve; CI, confidence interval; CMR, cardiac magnetic resonance; CMR-FT, cardiac magnetic resonance feature tracking; C-SVM, cubic SVM; G-SVM, Gaussian-SVM; KNN, k-nearest neighbor ; L-SVM, linear SVM; SVM: support vector machine; Q-SVM: quadratic SVM. Comparison with other imaging studies Table 5 shows a comparison of the proposed multimodal G-SVM with other published imaging studies for CP and RCM classification. Echocardiographic studies on annular displacement analysis and tissue Doppler/speckle tracking had a small sample size (n < 100) and reported a low AUC < 0.85. 24,25 Deep learning resulted in an AUC up to 0.97 in a highly selected group of patients with only cardiac amyloidosis, a subtype of RCM. 13 The study used imaging features from transthoracic echocardiography as ML input without providing specific imaging features. CT analysis of pericardial and cardiac abnormalities provided a low AUC < 0.75. 6,16 CMR-FT on LV and LA strain imaging resulted in high AUCs between 0.88–0.95, though heavy postprocessing procedures might have been involved. 19 , 21 , 26 Table 5 Comparison of cardiac imaging for CP vs. RCM differentiation Study (Ref) Population Imaging modality Key methodologies AUC Liu et al. (4) 45 CP , 40 RCM Echo Annular displacement analysis 0.82 Sengupta et al. (7) 30 CP , 30 RCM Tissue Doppler/speckle-tracking 0.85 Chao et al. (10) 112 CP , 98 RCM Deep learning 0.97 Garcia et al. (3) 50 CP , 30 RCM CT Structural evaluation 0.72 Talreja et al. (5) 143 CP Pericardial thickness/calcification assessment 0.75 Amaki et al. (8) 25 CP , 25 RCM CMR-FT Strain 0.89 Yang et al. (16) 50 CP , 45 RCM LV Time-strain curve analysis 0.88 Bo et al. (11) 60 CP , 55 RCM LA strain analysis 0.95 Our Study 156 CP , 91 RCM Multimodal imaging (Echo + CT + CMR) Gaussian SVM with 12 imaging variables 0.97 AUC, area under the receiver-operating characteristic curve; CP, constrictive pericarditis; CMR-FT, cardiac magnetic resonance feature tracking; CT, computed tomography; echo, echocardiography; LA, left atrium; LV, left ventricle; ML, machine learning; RCM, restrictive cardiomyopathy; SVM, support vector machine Discussion Drawing on a ten-year two-center cohort, this study provides the largest assessment to date of multimodal ML for differentiating CP from RCM. A G-SVM trained with routine echocardiography, CT and CMR variables reached an AUC of 0.97 and outperformed every single-modality model as well as the best individual imaging marker. These findings indicate that a compact multimodal feature set can achieve clinically meaningful accuracy when baseline profiles overlap and treatment strategies differ. They also highlight the potential of multimodal-imaging ML to streamline diagnostic workflows and lessen the need for invasive testing. Clinical dilemmas in the differential diagnosis of CP and RCM The differential diagnosis of CP and RCM remains challenging because both conditions present with preserved ejection fraction, impaired left-ventricular relaxation and bi-atrial enlargement, yet they require opposite treatments. Timely recognition of CP is essential because pericardiectomy can restore hemodynamics and yields five-to-seven-year survival above 80% 27 , whereas RCM, particularly when associated with cardiac amyloidosis, lacks disease-modifying therapy and has a median survival below 38 months. 2 Misclassifying RCM as CP exposes patients to unnecessary surgery with reported perioperative mortality close to 50%, 28 while failure to identify operable CP removes the only curative option. Baseline clinical features alone cannot resolve this dilemma. Each single-modality imaging technique provides a limited aspect of the diagnostic assessment. Echocardiography relies on subjective assessment of septal motion and cannot reveal myocardial infiltration. 29,30 CT depicts only static anatomy and may miss functionally constrictive cases with normal pericardial thickness. 7 CMR strain analysis requires labor-intensive post-processing and does not capture real-time hemodynamic information. 31,32 Consequently about 30% of cases remain indeterminate after non-invasive evaluation and proceed to invasive catheterization. 2 These limitations support the development of a ML strategy that integrates multimodal imaging data to improve diagnostic accuracy and guide timely clinical decisions. Multimodal-imaging ML enhances classification accuracy Our analysis indicates that integrating echocardiography, CT and CMR improves the classification of CP and RCM in this cohort. Across SVMs, KNN, decision tree and logistic regression, the multimodal Gaussian SVM achieved the highest performance with an AUC of 0.97 and balanced sensitivity 0.89 and specificity 0.90. Feature-importance analysis showed approximately equal contributions from CMR-derived strain measures, echocardiographic functional indices and CT-based anatomical features. This distribution is consistent with clinical reasoning because tissue characterization, ventricular function and pericardial structure provide complementary information that supports differential diagnosis. The observed accuracy is comparable to reports based on single-modality models using carefully curated echocardiographic datasets 13 or CMR feature tracking parameters. 19,26 However, our evaluation used routine acquisitions from all three modalities in a larger and less selected cohort, which improves applicability to everyday practice. In this dataset, SVMs outperformed decision tree, logistic regression and KNN, likely because margin-based classifiers handle high dimensional predictors in modest samples while maintaining generalization. The radial basis kernel can capture nonlinear relations such as those between pericardial thickness and myocardial strain without excessive parameterization. All models used structured imaging features rather than raw images, which reduced variability introduced by preprocessing. Similar advantages of kernel-based SVMs have been reported in cardiovascular classification tasks. 33 When the same algorithms were trained on any single modality in our dataset, AUCs were consistently lower. This pattern supports a mechanistic interpretation in which echocardiography describes chamber motion and filling, CT depicts pericardial structure and CMR characterizes myocardial tissue. The combination of these selected variables therefore offers a broader view of the underlying pathophysiology than any single technique. In this context, the feature selection step identified clinically meaningful variables and limited redundancy, which likely contributed to efficiency and interpretability. Future directions Prospective validation across diverse multicenter cohorts should be prioritized to address key limitations, including the retrospective study design and predominance of amyloidosis-related RCM cases within our group. 17 , 18 Such validation would better establish generalizability across the broader spectrum of RCM. Incorporating longitudinal clinical endpoints particularly post-pericardiectomy functional recovery and survival outcomes 27 would further substantiate the model’s clinical utility and prognostic relevance. For practical implementation, carefully designed clinician-ML collaborative frameworks 33 , 34 could potentially enhance diagnostic accuracy while preserving essential clinical reasoning and mitigating over-reliance on algorithmic outputs. Future methodological refinements might incorporate adaptive learning mechanisms to accommodate evolving disease phenotypes and emerging diagnostic biomarkers. 35 Limitations This study has several limitations. First, the analysis is retrospective and confined to two centers, and the model was tested only with internal cross-validation, so validation in independent external cohorts and prospective evaluation of clinical benefit are still needed. Second, the cohort is modest in size, and most RCM cases are amyloidosis, which may limit the applicability of the findings to other restrictive phenotypes. Third, the algorithm assumes that echocardiography, CT and CMR are available for every patient, a multimodal protocol that may be difficult to achieve in resource-limited or time-pressured settings. Conclusions This study delineates the clinical and multimodal imaging distinctions between CP and RCM and shows that ML classifiers combining echocardiography, CT and CMR data are a feasible and effective tool for accurate differential diagnosis. Declarations Ethics approval and consent to participate This study was conducted in accordance with the Declaration of Helsinki and was approved by the Institutional Ethics Committee of Peking University Third Hospital (Beijing, China) and the Institutional Ethics Committee of Beijing Anzhen Hospital, Capital Medical University (Beijing, China) (Approval No. 2006003x). Due to the retrospective nature of the study and the use of anonymized clinical data, the requirement for written informed consent was waived by the ethics committees. Consent for publication Not applicable. Availability of data and materials The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request. Competing interests The authors declare no competing financial interests. Funding This work was supported by the Beijing Research Ward Excellence Clinical Study Program (BRWEP2024W014090201); the National Natural Science, Foundation of China (Grant No. 82170381). Authors' contributions Z.Y. analyzed data and drafted the manuscript. J.W. and X.M. were involved in designing the tool, interviewing patients and analyzing data. H.Z. and X.D. collected the data and performed statistical analysis. J.D. and M.C. collected data. R.Z. contributed to the data analysis. H.L. made contributions to interpretation of the data and critically edited the manuscript. All authors contributed to discussion and interpretation of the results, critically reviewed the manuscript, and approved the final version to be submitted. All authors had access to the data and agreed to submit the manuscript for publication. References Nishimura RA, Borlaug BA. Diastology for the clinician. J Cardiol. 2019;73(6):445–52. Hirshfeld JW Jr, Johnston-Cox H. Distinguishing Constrictive Pericarditis From Restrictive Cardiomyopathy-An Ongoing Diagnostic Challenge. JAMA Cardiol. 2022;7(1):13–4. Lloyd JW, Anavekar NS, Oh JK, Miranda WR. Multimodality imaging in differentiating constrictive pericarditis from restrictive cardiomyopathy: a comprehensive overview for clinicians and imagers. J Am Soc Echocardiogr. 2023;36:1254–65. Adler Y, Charron P, Imazio M, et al. 2015 ESC Guidelines for the diagnosis and management of pericardial diseases. Eur Heart J. 2015;36:2921–64. Welch TD, Ling LH, Espinosa RE, et al. Echocardiographic diagnosis of constrictive pericarditis: Mayo Clinic criteria. Circ Cardiovasc Imaging. 2014;7:526–34. Garcia MJ. Constrictive Pericarditis Versus Restrictive Cardiomyopathy? J Am Coll Cardiol. 2016;67(17):2061–76. Talreja DR, Edwards WD, Danielson GK, et al. Constrictive pericarditis in 26 patients with histologically normal pericardial thickness. Circulation. 2003;108(15):1852–7. Amaki M, Savino J, Ain DL, et al. Diagnostic concordance of echocardiography and cardiac magnetic resonance-based tissue tracking for differentiating constrictive pericarditis from restrictive cardiomyopathy. Circ Cardiovasc Imaging. 2014;7(5):819–27. Bo K, Zhao Y, Gao X, et al. Cardiac magnetic resonance feature tracking derived left atrial strain in the diagnosis of patients with constrictive pericarditis and restrictive cardiomyopathy. Heliyon. 2024;10(7):e28768. Syed FF, Schaff HV, Oh JK. Constrictive pericarditis—a curable diastolic heart failure. Nat Rev Cardiol. 2015;12(9):682–90. Ling LH, Oh JK, Schaff HV, et al. Constrictive pericarditis in the modern era: evolving clinical spectrum and impact on outcome after pericardiectomy. Circulation. 1999;100(13):1380–6. Duchateau N, King AP, De Craene M. Machine Learning Approaches for Myocardial Motion and Deformation Analysis. Front Cardiovasc Med. 2019;6:190. Chao C-J, Jeong J, Arsanjani R, et al. Echocardiography-based deep learning model to differentiate constrictive pericarditis and restrictive cardiomyopathy. JACC Cardiovasc Imaging. 2024;17(4):349–60. Tabassian M, Sunderji I, Erdei T, et al. Diagnosis of heart failure with preserved ejection fraction: machine learning of spatiotemporal variations in left ventricular deformation. J Am Soc Echocardiogr. 2018;31(12):1272–84. Sengupta PP, Huang YM, Bansal M, Ashrafi A, Fisher M, Shameer K, Gall W, Dudley JT. Cognitive Machine-Learning Algorithm for Cardiac Imaging: A Pilot Study for Differentiating Constrictive Pericarditis From Restrictive Cardiomyopathy. Circ Cardiovasc Imaging. 2016;9(6). Talreja DR, Nishimura RA, Oh JK, Holmes DR. Constrictive pericarditis in the modern era: novel criteria for diagnosis in the cardiac catheterization laboratory. J Am Coll Cardiol. 2008;51(3):315–9. Kusunose K, Dahiya A, Popovic ZB, Motoki H, Alraies MC, Zurick AO, Bolen MA, Kwon DH, Flamm SD, Klein AL. Biventricular mechanics in constrictive pericarditis comparison with restrictive cardiomyopathy and impact of pericardiectomy. Circ Cardiovasc Imaging. 2013;6(3):399–406. Hurrell DG, Nishimura RA, Higano ST, Appleton CP, Danielson GK, Holmes DR Jr, Tajik AJ. Value of dynamic respiratory changes in left and right ventricular pressures for the diagnosis of constrictive pericarditis. Circulation. 1996;93:2007–13. Amaki M, Savino J, Ain DL, Sanz J, Pedrizzetti G, Kulkarni H, Narula J, Sengupta PP. Diagnostic concordance of echocardiography and cardiac magnetic resonance-based tissue tracking for differentiating constrictive pericarditis from restrictive cardiomyopathy. Circ Cardiovasc Imaging. 2014;7(5):819–27. Tabassian M, Sunderji I, Erdei T, Sanchez-Martinez S, Degiovanni A, Marino P, Fraser AG, D'Hooge J. Diagnosis of Heart Failure With Preserved Ejection Fraction: Machine Learning of Spatiotemporal Variations in Left Ventricular Deformation. J Am Soc Echocardiogr. 2018;31(12):1272–84. Yang Z, Wang H, Chang S, Cui J, Zhou L, Lv Q, He Y, Du X, Dong J, Ma C. Left ventricular strain-curve morphology to distinguish between constrictive pericarditis and restrictive cardiomyopathy. ESC Heart Fail. 2021;8(6):4863–72. Hao PY, Chiang JH, Chen YD. Possibilistic classification by support vector networks. Neural Netw. 2022;149:40–56. Miranda A, Lavrador R, Julio F, Januario C, Castelo-Branco M, Caetano G. Classification of Huntington's disease stage with support vector machines: A study on oculomotor performance. Behav Res Methods. 2016;48(4):1667–77. Liu S, Ren W, Zhang J, Ma C, Yang J, Zhang Y, Guan Z. Incremental Value of the Tissue Motion of Annular Displacement Derived From Speckle-Tracking Echocardiography for Differentiating Chronic Constrictive Pericarditis From Restrictive Cardiomyopathy. J Ultrasound Med. 2018;37(11):2637–45. Sengupta PP, Krishnamoorthy VK, Abhayaratna WP, Korinek J, Belohlavek M, Sundt TM 3rd, Chandrasekaran K, Mookadam F, Seward JB, Tajik AJ, Khandheria BK. Disparate patterns of left ventricular mechanics differentiate constrictive pericarditis from restrictive cardiomyopathy. JACC Cardiovasc Imaging. 2008;1(1):29–38. Bo K, Zhao Y, Gao X, Chen Y, Ren Y, Gao Y, Zhou Z, Wang H, Xu L. Cardiac magnetic resonance feature tracking derived left atrial strain in the diagnosis of patients with constrictive pericarditis and restrictive cardiomyopathy. Heliyon. 2024;10(7):e28768. Welch TD. Constrictive pericarditis: diagnosis, management and clinical outcomes. Heart. 2018;104(9):725–31. Talreja DR, Edwards WD, Danielson GK, Schaff HV, Tajik AJ, Tazelaar HD, Breen JF, Oh JK. Constrictive pericarditis in 26 patients with histologically normal pericardial thickness. Circulation. 2003;108(15):1852–7. Bijnens B, Cikes M, Butakoff C, Sitges M, Crispi F. Myocardial motion and deformation: What does it tell us and how does it relate to function? Fetal Diagn Ther. 2012;32(1–2):5–16. Rajagopalan N, Garcia MJ, Rodriguez L, Murray RD, Hansen CA, Stugaard M, Thomas JD, Klein AL. Comparison of new Doppler echocardiographic methods to differentiate constrictive pericardial heart disease and restrictive cardiomyopathy. Am J Cardiol. 2001;87:86–94. Palka P, Lange A, Donnelly JE, Nihoyannopoulos P. Differentiation between restrictive cardiomyopathy and constrictive pericarditis by early diastolic Doppler myocardial velocity gradient at the posterior wall. Circulation. 2000;102:655–62. Claus P, Omar AMS, Pedrizzetti G, Sengupta PP, Nagel E. Tissue Tracking Technology for Assessing Cardiac Mechanics: Principles, Normal Values, and Clinical Applications. JACC Cardiovasc Imaging. 2015;8(12):1444–60. Cikes M, Sanchez-Martinez S, Claggett B, Duchateau N, Piella G, Butakoff C, Pouleur AC, Knappe D, Biering-Sorensen T, Kutyifa V, Moss A, Stein K, Solomon SD, Bijnens B. Machine learning-based phenogrouping in heart failure to identify responders to cardiac resynchronization therapy. Eur J Heart Fail. 2019;21(1):74–85. Quitadamo LR, Cavrini F, Sbernini L, Riillo F, Bianchi L, Seri S, Saggio G. Support vector machines to detect physiological patterns for EEG and EMG-based human-computer interaction: a review. J Neural Eng. 2017;14(1):011001. Song-Men S. Intelligent Diagnosis Method for New Diseases Based on Fuzzy SVM Incremental Learning. Comput Math Methods Med. 2022;2022:7631271. Additional Declarations No competing interests reported. Supplementary Files Supplementarytable.docx Cite Share Download PDF Status: Under Review Version 1 posted Reviews received at journal 15 May, 2026 Reviewers agreed at journal 07 May, 2026 Reviewers agreed at journal 07 May, 2026 Reviewers invited by journal 29 Apr, 2026 Editor invited by journal 07 Apr, 2026 Editor assigned by journal 21 Feb, 2026 Submission checks completed at journal 21 Feb, 2026 First submitted to journal 19 Feb, 2026 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-8917150","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":635535437,"identity":"6727f4ef-f78c-49bc-a914-75bb0fce4258","order_by":0,"name":"Zhiyun Yang","email":"","orcid":"","institution":"Department of Cardiology and Institute of Vascular Medicine, Peking University Third Hospital, Key Laboratory of Molecular Cardiovascular Science, Ministry of Education, Beijing 100191, China.","correspondingAuthor":false,"prefix":"","firstName":"Zhiyun","middleName":"","lastName":"Yang","suffix":""},{"id":635535438,"identity":"403e0900-f603-40e8-b119-e600e6b56a84","order_by":1,"name":"Xiangbin Meng","email":"","orcid":"","institution":"Pengcheng Laboratory, Shenzhen 518055, China.","correspondingAuthor":false,"prefix":"","firstName":"Xiangbin","middleName":"","lastName":"Meng","suffix":""},{"id":635535439,"identity":"19ec006e-9e45-463c-8976-dfc6ce8af4d8","order_by":2,"name":"Rui Zuo","email":"","orcid":"","institution":"Information Management and Big Data Center, Peking University Third Hospital, Beijing 100191, China.","correspondingAuthor":false,"prefix":"","firstName":"Rui","middleName":"","lastName":"Zuo","suffix":""},{"id":635535440,"identity":"84d7197c-a6e7-4e6b-9e00-ddb0ed2f2f02","order_by":3,"name":"Hongkai Zhang","email":"","orcid":"","institution":"Department of Radiology, Beijing Anzhen Hospital, Capital Medical University, Beijing 100029, China.","correspondingAuthor":false,"prefix":"","firstName":"Hongkai","middleName":"","lastName":"Zhang","suffix":""},{"id":635535441,"identity":"b1e03457-59a2-4cb7-a7f3-cd787d1b30dd","order_by":4,"name":"Xin Du","email":"","orcid":"","institution":"Department of Cardiology, Beijing Anzhen Hospital, Capital Medical University, Beijing 100029, China.","correspondingAuthor":false,"prefix":"","firstName":"Xin","middleName":"","lastName":"Du","suffix":""},{"id":635535442,"identity":"7fcbce01-d73f-4b51-8d7b-1e758b48be0f","order_by":5,"name":"Jianzeng Dong","email":"","orcid":"","institution":"Department of Cardiology, Beijing Anzhen Hospital, Capital Medical University, Beijing 100029, China.","correspondingAuthor":false,"prefix":"","firstName":"Jianzeng","middleName":"","lastName":"Dong","suffix":""},{"id":635535443,"identity":"587905c9-3444-41a3-aee9-deb1474a35f9","order_by":6,"name":"Hongxing Luo","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA5UlEQVRIiWNgGAWjYBACA2YGNjADSDI+ANI8fKRoYTYAaWEjqIWBAa6GTQKqFz8wZ2dge/BxR21in3T7tcqvOXYybAzMDx/dwKPFspmB3XDmmePGbDJnym7LbksGOozN2DgHn8MOM7BJ87Ydk2OTyEm7LbmNGaiFh02aGC08IC3FktvqidZSA7Ql/Rjjx22HidHC2CY5s+2AMdAWZmnGbcd52JgJ+eX84WMSH9vqEufPSH/48ee2ant+9uaHj/FpAcZ5A5A4DMQ8Bsw8IAFmvMrhoA6I2R8w/iBO9SgYBaNgFIwwAAA83T7exG7QvQAAAABJRU5ErkJggg==","orcid":"","institution":"Institute for Surgical Research, Oslo University Hospital, Rikshospitalet, Postboks 4950 Nydalen, 0424 Oslo, Norway.Institute for Surgical Research, Oslo University Hospital, Rikshospitalet","correspondingAuthor":true,"prefix":"","firstName":"Hongxing","middleName":"","lastName":"Luo","suffix":""},{"id":635535444,"identity":"ec05ae6b-6582-4cb7-afd7-2055b66088dc","order_by":7,"name":"Ming Cui","email":"","orcid":"","institution":"Department of Cardiology and Institute of Vascular Medicine, Peking University Third Hospital, Key Laboratory of Molecular Cardiovascular Science, Ministry of Education, Beijing 100191, China.","correspondingAuthor":false,"prefix":"","firstName":"Ming","middleName":"","lastName":"Cui","suffix":""}],"badges":[],"createdAt":"2026-02-19 11:38:29","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-8917150/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-8917150/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":108955671,"identity":"80cc5f3e-1450-4ce8-89c8-80abdb3444e8","added_by":"auto","created_at":"2026-05-11 08:09:42","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":724218,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003e\u003cstrong\u003eMultimodal imaging and clinical data collections.\u003c/strong\u003e\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eBMI, body mass index; BNP, B-type natriuretic peptide; BSA, body surface area; CMR, cardiac magnetic resonance; CMR-FT, cardiac magnetic resonance-feature tracking; CT, computed tomography; GPT, glutamic pyruvic transaminase.\u003c/em\u003e\u003c/p\u003e","description":"","filename":"image1.png","url":"https://assets-eu.researchsquare.com/files/rs-8917150/v1/6d89b4d69af70dce512c3da1.png"},{"id":108955672,"identity":"dc73a0aa-245e-41d6-85e7-0d60262cd25b","added_by":"auto","created_at":"2026-05-11 08:09:42","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":45917,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003e\u003cstrong\u003eStudy population selection flowchart.\u003c/strong\u003e\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eCMR, cardiac magnetic resonance; CP, constrictive pericarditis; CT, computed tomography; Echo, echocardiography; EF, ejection fraction; IVS, interventricular septal thickness; LAE, left atrial enlargement; LGE, late gadolinium enhancement; LV, left ventricle; RAE, right atrial enlargement; RCM, restrictive cardiomyopathy; RV, right ventricle.\u003c/em\u003e\u003c/p\u003e","description":"","filename":"image2.png","url":"https://assets-eu.researchsquare.com/files/rs-8917150/v1/3b4ed01c7156d3ff3b6b4b57.png"},{"id":108956340,"identity":"82a296db-971f-4c97-9824-8e3c5da6075e","added_by":"auto","created_at":"2026-05-11 08:12:56","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":436356,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003e\u003cstrong\u003eMachine Learning Framework for Differentiating Constrictive Pericarditis (CP) and Restrictive Cardiomyopathy (RCM). \u003c/strong\u003e\u003c/em\u003e\u003cem\u003e(A) ML flowchart from feature selection, model training to model evaluation. (B) The ROC curves derived from multimodal imaging of ML models for CP and RCM classification. (C) The ROC curves of Echocardiography of ML models for CP and RCM classification. (D) The ROC curves of CMR of ML models for CP and RCM classification. (E) The ROC curves of CT of ML models for CP and RCM classification.\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eAUC, the area under the receiver-operating characteristic curve; CP, constrictive pericarditis; CMR, cardiac magnetic resonance; CT, computed tomography; G-SVM, Gaussian-SVM; RCM, restrictive cardiomyopathy; ROC, receiver-operating characteristic; SVM, support vector machine.\u003c/em\u003e\u003c/p\u003e","description":"","filename":"image3.png","url":"https://assets-eu.researchsquare.com/files/rs-8917150/v1/19a32eb80efcf5ea8891fdde.png"},{"id":108977958,"identity":"f7f39a1f-fa97-43b2-a6b6-95ca096ec1ba","added_by":"auto","created_at":"2026-05-11 11:33:30","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":2014090,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8917150/v1/2ba8d141-be90-48a7-9ec6-f479d267197b.pdf"},{"id":108955632,"identity":"a0fe63cb-148b-40b1-b349-81cf398d5963","added_by":"auto","created_at":"2026-05-11 08:09:28","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":18567,"visible":true,"origin":"","legend":"","description":"","filename":"Supplementarytable.docx","url":"https://assets-eu.researchsquare.com/files/rs-8917150/v1/c11ddc6c5872142c518b66e0.docx"}],"financialInterests":"No competing interests reported.","formattedTitle":"Machine Learning of Multimodal Imaging to Differentiate Constrictive Pericarditis and Restrictive Cardiomyopathy","fulltext":[{"header":"Introduction","content":"\u003cp\u003eConstrictive pericarditis (CP) and restrictive cardiomyopathy (RCM) both present preserved ejection function, impaired left ventricular (LV) relaxation, and normal-sized ventricles with dilated atria on imaging.\u003csup\u003e\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e\u003c/sup\u003e Despite these similarities, the two conditions require fundamentally different management strategies. CP can be cured by pericardiectomy, whereas RCM derives no surgical benefit and is treated mainly with supportive therapy. Misclassification may therefore expose patients to unnecessary operative risk or delay definitive treatment.\u003csup\u003e\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e\u003c/sup\u003e\u003c/p\u003e \u003cp\u003e Current guidelines recommend a stepwise approach for distinguishing CP from RCM, beginning with transthoracic echocardiography, followed by computed tomography (CT) or cardiac magnetic resonance (CMR), with invasive catheterization reserved for cases where noninvasive findings remain inconclusive. \u003csup\u003e3,4\u003c/sup\u003e Although each modality reflects a distinct aspect of the underlying pathophysiology, none alone provides consistent diagnostic accuracy. Echocardiography is operator-dependent and may yield inconclusive hemodynamic findings, particularly in patients with atrial fibrillation, obesity, or elevated filling pressures, which together account for up to 30% of cases. \u003csup\u003e5\u003c/sup\u003e Although CT can delineate pericardial anatomy and calcification, its diagnostic sensitivity is limited in early-stage or post-radiation constrictive pericarditis, where pericardial thickening may be minimal. Approximately 28% of surgically confirmed CP cases show pericardial thickness\u0026thinsp;\u0026le;\u0026thinsp;4 mm. \u003csup\u003e6,7\u003c/sup\u003e CMR provides valuable tissue characterization through late gadolinium enhancement (LGE) and strain analysis, but its accuracy is influenced by vendor-specific techniques, complex post-processing, and limited assessment of ventricular interdependence, a key hemodynamic feature of CP. \u003csup\u003e8\u0026ndash;11\u003c/sup\u003e The 2023 American Society of Echocardiography consensus therefore endorses a multimodality imaging strategy when echocardiographic findings are equivocal, acknowledging that no single noninvasive test can reliably distinguish constriction from restriction in all cases. \u003csup\u003e3\u003c/sup\u003e\u003c/p\u003e \u003cp\u003eGiven these diagnostic limitations, machine learning (ML) offers a promising avenue by integrating diverse quantitative parameters across modalities to uncover patterns not readily discernible to human observers.\u003csup\u003e\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e\u003c/sup\u003e Prior studies have demonstrated the potential of ML, with deep learning models achieving high accuracy in distinguishing CP from cardiac amyloidosis using echocardiographic imaging features.\u003csup\u003e\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e\u003c/sup\u003e However, these \u0026ldquo;black-box\u0026rdquo; models have limited interpretability and generalizability across the broader RCM spectrum. To date, no study has replicated the clinician\u0026rsquo;s multimodal reasoning by integrating structured quantitative data from echocardiography, CT, and CMR within a unified and interpretable ML framework. \u003csup\u003e14,15\u003c/sup\u003e\u003c/p\u003e \u003cp\u003eIn this context, the present study aims to develop and validate ML classifiers that combine multiple imaging modalities to differentiate CP from RCM. The goal is to provide a practical and interpretable noninvasive tool to support clinical decision-making and optimize referral for catheterization or surgery, with the ultimate aim of improving diagnostic precision and patient outcomes.\u003c/p\u003e"},{"header":"Methods","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003eStudy population\u003c/h2\u003e \u003cp\u003eThis retrospective study included consecutive CP and RCM patients who underwent multimodal imaging examinations including CT, echocardiography, CMR-FT, and cardiac catheterization at Anzhen Hospital, Capital Medical University, and Peking University Third Hospital, China between January 2014 and September 2024. This study was conducted in accordance with the Declaration of Helsinki and was approved by the Institutional Ethics Committees of Peking University Third Hospital and Beijing Anzhen Hospital, Capital Medical University (Approval No. 2006003x).\u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eCP and RCM diagnostic criteria\u003c/h3\u003e\n\u003cp\u003eAll enrolled CP and RCM patients were in sinus rhythm. Individuals with significant valvular diseases, arrhythmias, prior myocardial infarction, hypertension, or other causes of myocardial hypertrophy or dysfunction were excluded.\u003c/p\u003e \u003cp\u003eCP was confirmed either during surgical pericardiectomy or, when the surgery was not performed, by the simultaneous presence of pericardial thickening\u0026thinsp;\u0026gt;\u0026thinsp;4 mm on CT, echocardiography, or CMR and objective evidence of increased left-to-right ventricular coupling manifested as a respiratory septal shift on echocardiography/CMR or a right-to-left ventricular systolic area index ratio\u0026thinsp;\u0026gt;\u0026thinsp;1.1 during both inspiration and expiration on cardiac catheterization. \u003csup\u003e16,17,18\u003c/sup\u003e\u003c/p\u003e \u003cp\u003eRCM was diagnosed either by 1) tissue biopsy of the heart, kidney, or other organs demonstrating amyloidosis, or by 2) all of the following imaging evidence: interventricular septal thickness\u0026thinsp;\u0026gt;\u0026thinsp;12 mm, preserved LV ejection fraction (LVEF)\u0026thinsp;\u0026gt;\u0026thinsp;50%, bi-atrial enlargement, a restrictive filling pattern, and late gadolinium enhancement on CMR indicating myocardial involvement.\u003csup\u003e\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e\u003c/sup\u003e\u003c/p\u003e\n\u003ch3\u003eMultimodal imaging acquisition and processing\u003c/h3\u003e\n\u003cp\u003eEach patient followed a standardized imaging protocol (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e): 1) non-contrast chest CT for evaluation of pericardial calcification and wall thickening, 2) transthoracic echocardiography for measurement of LV end-diastolic and end-systolic diameters (LVEDD and LVESD), LVEF, septal thickness, early (E) and late (A) transmitral velocities, mitral-annular early-diastolic velocity (e\u0026prime;), and the derived E/A and E/e\u0026prime; ratios, 3) right and left heart catheterization, when clinically indicated, for recording atrial, ventricular, and pulmonary artery pressures and for calculating the right-to-left ventricular systolic area index required to confirm CP, and 4) cardiac MRI with long- and short-axis cine steady-state free-precession, rest first-pass perfusion, and late gadolinium enhancement sequences for comprehensive assessment of ventricular function and myocardial tissue characteristics.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eImage post-processing was conducted on a commercial workstation running CVI42 (version 5.11.3, Circle Cardiovascular Imaging). Endocardial and epicardial borders were delineated in end-diastolic and end-systolic frames, after which LV end-diastolic and end-systolic volumes (LVEDV, LVESV), LVEF, and diastolic myocardial mass were quantified automatically. Three-dimensional feature-tracking analysis was subsequently applied to derive global longitudinal, circumferential, and radial peak strain, peak strain rate, time-to-peak strain, and velocity. All imaging acquisitions were completed before any pericardiectomy or other therapeutic intervention in the CP group.\u003c/p\u003e\n\u003ch3\u003eML for CP and RCM classification\u003c/h3\u003e\n\u003cp\u003eThe ML workflow consisted of data preprocessing, model training/validation, and performance assessments. Any field with \u0026lt;\u0026thinsp;20% missingness was imputed by k-nearest-neighbor (KNN) learning during preprocessing.\u003csup\u003e\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e\u003c/sup\u003e LV global time-strain curves derived from CMR-FT were normalized to the cardiac cycle by referencing successive end-diastolic frames, after which strain ratios for the 0\u0026ndash;50%/50\u0026ndash;75% and 50\u0026ndash;75%/75\u0026ndash;100% diastolic intervals were calculated in accordance with a published method. \u003csup\u003e21\u003c/sup\u003e\u003c/p\u003e \u003cp\u003eML was performed using support vector machines (SVMs), decision tree, logistic regression and KNN. SVMs were chosen as the primary supervised algorithm because of their capacity to handle small, high-dimensional data sets. Linear (L-SVM), quadratic (Q-SVM), cubic (C-SVM), and Gaussian-kernel (G-SVM) variants were implemented with the LIBSVM toolbox under Matlab R2017b (MathWorks, Natick, MA, USA).\u003csup\u003e\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e\u003c/sup\u003e During the training, class labels were paired with predictor matrices, and the optimal hyper-plane separating CP from RCM observations was generated. All ML classifiers were trained on the same dataset using imaging features from echocardiography, MRI or CT as model inputs. Five-fold cross-validation was applied: in each fold, 20% of subjects from both diagnostic categories were reserved for testing and 80% for training. The procedure was repeated across all folds to minimize sampling bias. Mean classification accuracy over the five folds was calculated as the final result. ML was performed with both single- and multiple-modal imaging features as the model inputs.\u003c/p\u003e \u003cp\u003eML performance metrics were derived from the confusion matrix using the following equations: \u003csup\u003e23\u003c/sup\u003e\u003cdiv id=\"Equa\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equa\" name=\"EquationSource\"\u003e\n$$\\:\\text{A}\\text{c}\\text{c}\\text{u}\\text{r}\\text{a}\\text{c}\\text{y}=\\frac{\\text{T}\\text{P}\\:+\\:\\text{T}\\text{N}}{\\text{T}\\text{N}\\:+\\:\\text{F}\\text{P}\\:+\\:\\text{F}\\text{N}\\:+\\:\\text{T}\\text{P}}$$\u003c/div\u003e\u003c/div\u003e\u003cdiv id=\"Equb\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equb\" name=\"EquationSource\"\u003e\n$$\\:\\text{S}\\text{e}\\text{n}\\text{s}\\text{i}\\text{t}\\text{i}\\text{v}\\text{i}\\text{t}\\text{y}=\\frac{\\text{T}\\text{P}}{\\text{T}\\text{P}\\:+\\:\\text{F}\\text{N}}$$\u003c/div\u003e\u003c/div\u003e\u003cdiv id=\"Equc\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equc\" name=\"EquationSource\"\u003e\n$$\\:\\text{S}\\text{p}\\text{e}\\text{c}\\text{i}\\text{f}\\text{i}\\text{c}\\text{i}\\text{t}\\text{y}=\\frac{\\text{T}\\text{N}}{\\text{T}\\text{N}\\:+\\:\\text{F}\\text{P}}$$\u003c/div\u003e\u003c/div\u003e\u003cdiv id=\"Equd\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equd\" name=\"EquationSource\"\u003e\n$$\\:\\text{P}\\text{r}\\text{e}\\text{c}\\text{i}\\text{s}\\text{i}\\text{o}\\text{n}=\\frac{\\text{T}\\text{P}}{\\text{T}\\text{P}\\:+\\:\\text{F}\\text{P}}$$\u003c/div\u003e\u003c/div\u003e\u003c/p\u003e \u003cp\u003ewhere TP is true positive, FP is false positive, FN is false negative, and TN is true negative. Receiver-operating characteristic (ROC) curves were plotted, and the area under curve (AUC) was computed with Matlab Statistics.\u003c/p\u003e \u003cdiv id=\"Sec7\" class=\"Section2\"\u003e \u003ch2\u003eStatistical analysis\u003c/h2\u003e \u003cp\u003eContinuous data were expressed as mean\u0026thinsp;\u0026plusmn;\u0026thinsp;SD for variables with a normal distribution (e.g., LV septum thickness, LVEF) and as median (interquartile range) for skewed variables (e.g., age, BNP, LVEDD). Categorical data were summarized as count (percentage) (e.g., sex, pericardial calcification). Between-group differences were assessed with the independent-samples \u003cem\u003et\u003c/em\u003e-test for normally distributed continuous variables, and the Mann-Whitney U test for non-normal distributions. Categorical variables were compared with the χ\u0026sup2; test or Fisher\u0026rsquo;s exact test, as appropriate. A two-sided P value\u0026thinsp;\u0026lt;\u0026thinsp;0.05 was considered statistically significant. Statistical analyses were performed with SPSS 22.0 (IBM, Armonk, NY, USA).\u003c/p\u003e \u003c/div\u003e"},{"header":"Results","content":"\u003cdiv id=\"Sec9\" class=\"Section2\"\u003e \u003ch2\u003eClinical characteristics\u003c/h2\u003e \u003cp\u003eA total of 247 patients were included, with 156 CP and 91 RCM (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e). In the CP group, 98 (63%) were confirmed surgically and 58 (37%) were diagnosed by multimodal imaging comprising CT, echocardiography, CMR and cardiac catheterization. In the RCM group, 51 (56%) had biopsy-proven amyloidosis, and the remaining 40 (44%) were diagnosed by multimodal imaging evidence of diffuse myocardial infiltration, subendocardial late gadolinium enhancement and fibrosis.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eTable\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e summarizes the patients\u0026rsquo; baseline characteristics. The CP group had more females (55.4% \u003cem\u003evs.\u003c/em\u003e 39.0%, P\u0026thinsp;=\u0026thinsp;0.025) and a lower BNP level (188 (300) ng/ml \u003cem\u003evs.\u003c/em\u003e 788 (1408) ng/ml, P\u0026thinsp;=\u0026thinsp;0.001). No significant differences in other baseline clinical variables were observed between the two groups, highlighting the difficulty of distinguishing CP from RCM based on clinical characteristics alone.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eClinical characteristics (n\u0026thinsp;=\u0026thinsp;247)\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"4\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eCP (n\u0026thinsp;=\u0026thinsp;156)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eRCM (n\u0026thinsp;=\u0026thinsp;91)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003e\u003cem\u003eP\u003c/em\u003e value\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAge, y\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e58.0 (19)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e58.0 (14)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.876\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eFemale, n (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e85 (55.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e36 (39.0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.025\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHeight, cm\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e167.3\u0026thinsp;\u0026plusmn;\u0026thinsp;8.4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e164.1\u0026thinsp;\u0026plusmn;\u0026thinsp;7.8\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.341\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eWeight, kg\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e66.8\u0026thinsp;\u0026plusmn;\u0026thinsp;13.1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e63.1\u0026thinsp;\u0026plusmn;\u0026thinsp;12.0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.812\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eBSA, cm/m\u003csup\u003e2\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1.74\u0026thinsp;\u0026plusmn;\u0026thinsp;0.21\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1.68\u0026thinsp;\u0026plusmn;\u0026thinsp;0.18\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.952\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eBMI, kg/m\u003csup\u003e2\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e23.8\u0026thinsp;\u0026plusmn;\u0026thinsp;3.5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e23.6\u0026thinsp;\u0026plusmn;\u0026thinsp;4.6\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.413\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHeart rate, bpm\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e85.0\u0026thinsp;\u0026plusmn;\u0026thinsp;15.8\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e77.9\u0026thinsp;\u0026plusmn;\u0026thinsp;12.6\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.091\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAtrial fibrillation, n (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e22 (14%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e9 (10%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.090\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHypertension, n (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e61 (39%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e40 (44%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.574\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDiabetes mellitus, n (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e45 (29%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e25 (27%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.645\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eFamilial history of CAD, n (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e17 (11%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e9 (10%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.610\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLaboratory testing\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eBNP, ng/mL\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e188 (300)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e788 (1408)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGPT, U/L\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e17.0 (12)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e22.0 (20)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.411\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCreatinine, mg/dL\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e75.2 (23)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e78.1 (41)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.197\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"4\"\u003e\u003cem\u003eBMI, body mass index; BNP, B-type natriuretic peptide; bpm, beats per minute; BSA, body surface area; CAD, coronary artery disease; CP, constrictive pericarditis; GPT, glutamic pyruvic transaminase; RCM, restrictive cardiomyopathy.\u003c/em\u003e\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eImaging findings from echocardiography, CT, CMR and cardiac catheterization\u003c/h3\u003e\n\u003cp\u003eTable\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e summarizes the imaging findings in patients. Echocardiography showed a thinner LV septum (P\u0026thinsp;\u0026lt;\u0026thinsp;0.001), smaller LVEDD (P\u0026thinsp;=\u0026thinsp;0.007) and LVESD (P\u0026thinsp;\u0026lt;\u0026thinsp;0.001), and a higher LVEF (P\u0026thinsp;\u0026lt;\u0026thinsp;0.001) in the CP group than the RCM group. Chest CT detected pericardial calcification and thickening only in the CP group. On CMR, CP had a lower myocardial mass (P\u0026thinsp;\u0026lt;\u0026thinsp;0.001) and a higher LVEF (P\u0026thinsp;=\u0026thinsp;0.013). Myocardial enhancement was observed in most RCM patients (80/91, 87.8%) compared with few CP patients (8/156, 5.4%) (P\u0026thinsp;\u0026lt;\u0026thinsp;0.001). In CMR-FT, the CP group showed larger peak strain and peak strain rate in the longitudinal, circumferential and radial directions compared with RCM (all P\u0026thinsp;\u0026lt;\u0026thinsp;0.05). Longitudinal and radial myocardial velocities were also higher in CP (P\u0026thinsp;\u0026lt;\u0026thinsp;0.05). After normalizing the global time-strain curves, strain ratios were calculated for the early- (0\u0026ndash;50%), mid- (50\u0026ndash;75%), and late- (75\u0026ndash;100%) diastolic intervals. The only significant inter-group difference emerged in the strain ratio of circumferential mid- to late-diastolic segment (50\u0026ndash;75%/75\u0026ndash;100%), which was lower in CP than in RCM (P\u0026thinsp;=\u0026thinsp;0.005). Pressure indices from cardiac catheterization were similar in CP and RCM (all P\u0026thinsp;\u0026gt;\u0026thinsp;0.05).\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab2\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eCardiac imaging findings (n\u0026thinsp;=\u0026thinsp;247)\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"4\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eCP (n\u0026thinsp;=\u0026thinsp;156)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eRCM (n\u0026thinsp;=\u0026thinsp;91)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003e\u003cem\u003eP\u003c/em\u003e Value\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eEchocardiography\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLV septum, mm\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e8.4\u0026thinsp;\u0026plusmn;\u0026thinsp;1.7\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e13.1\u0026thinsp;\u0026plusmn;\u0026thinsp;4.4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLVEDD, mm\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e39.0 (8)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e44.5 (9)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.007\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLVESD, mm\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e26.0 (6)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e32.0 (8)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLVEF, %\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e62.9\u0026thinsp;\u0026plusmn;\u0026thinsp;7.5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e51.9\u0026thinsp;\u0026plusmn;\u0026thinsp;15.2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eE wave, m/s\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e83.0\u0026thinsp;\u0026plusmn;\u0026thinsp;21.6\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e74.6\u0026thinsp;\u0026plusmn;\u0026thinsp;24.5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.537\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eA wave, m/s\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e49.0 (31)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e58.5 (44)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.519\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eE/A ratio\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1.6 (1)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1.3 (1)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.119\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ee\u0026rsquo;, cm/s\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e8.5 (1)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e7.2 (1)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.667\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eE/e\u0026prime; ratio\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e9.9 (2)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e11.6 (2)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.237\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eCT\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePericardial calcification, n (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e95 (60.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e\u0026lt;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePericardial thickness\u0026thinsp;\u0026gt;\u0026thinsp;4 mm, n (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e104 (66.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e\u0026lt;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eCMR\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLVEDV, ml\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e88.5 (35.1)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e118.0 (77.1)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.255\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLVESV, ml\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e41.2 (22.6)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e58.5 (66.0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.057\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLVEF, %\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e50.6\u0026thinsp;\u0026plusmn;\u0026thinsp;11.8\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e42.7\u0026thinsp;\u0026plusmn;\u0026thinsp;16.1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.013\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMyo Mass, g\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e64.1 (27)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e116.2 (84)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e\u0026lt;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMyocardial enhancement, n (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e8(5.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e80(87.8%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e\u0026lt;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eLongitudinal\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePeak strain (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e-12.2\u0026thinsp;\u0026plusmn;\u0026thinsp;3.9\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e-8.3\u0026thinsp;\u0026plusmn;\u0026thinsp;4.0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePeak strain rate (1/s)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.9 (0.6)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.5 (0.3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e\u0026lt;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTime to peak strain (ms)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e294.3 (77.9)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e271.3 (99.8)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.395\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eVelocity (mm/s)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e-28.8 (28.9)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e-15.3 (20.9)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.002\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eStrain ratio of 0\u0026ndash;50%/50\u0026ndash;75% diastolic period\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e3.1 (5.1)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e3.6 (4.1)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.301\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eStrain ratio of 50\u0026ndash;75%/75\u0026ndash;100% diastolic period\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.1 (2.3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.3 (3.6)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.629\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eCircumferential\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePeak strain (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e-15.2\u0026thinsp;\u0026plusmn;\u0026thinsp;4.8\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e-11.6\u0026thinsp;\u0026plusmn;\u0026thinsp;4.8\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e\u0026lt;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePeak strain rate (1/s)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1.1 (0.7)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.8 (0.4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.002\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTime to peak strain (ms)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e266.5 (57.6)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e246.0 (62.6)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.180\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eVelocity (mm/s)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e-16.4 (35.7)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e-14.9 (12.3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.932\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eStrain ratio of 0\u0026ndash;50%/50\u0026ndash;75% diastolic period\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e4.5 (8.7)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e5.0 (5.9)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.842\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eStrain ratio of 50\u0026ndash;75%/75\u0026ndash;100% diastolic period\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.3 (1.5)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1.6 (7.4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.005\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eRadial\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePeak strain (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e26.1 (15.5)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e12.8 (16.8)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e\u0026lt;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePeak strain rate (1/s)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e-2.1 (2.1)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e-0.8 (1.2)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e\u0026lt;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTime to peak strain (ms)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e260.2 (60.3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e238.0 (66.1)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.230\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eVelocity (mm/s)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e-26.0 (17.6)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e-18.4 (11.6)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.019\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eStrain ratio of 0\u0026ndash;50%/50\u0026ndash;75% diastolic period\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e6.7 (10.3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e5.7 (10.5)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.879\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eStrain ratio of 50\u0026ndash;75%/75\u0026ndash;100% diastolic period\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.3 (1.5)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1.1 (3.8)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.076\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eCardiac catheterization\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eRAP, mmHg\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e20.4 (6.9)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e23.2 (7.0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.329\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eRVEDP, mmHg\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e20.2 (6.4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e20.0 (6.8)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.940\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePAP, mmHg\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e20.2 (6.5)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e22.1 (7.6)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.436\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePAWP, mmHg\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e24.0 (4.6)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e25.2 (5.6)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.160\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLVEDP, mmHg\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e23.1 (6.7)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e21.2 (6.4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.600\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eFindings supporting CP, No./total No. (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e6/10 (60.0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0/6\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.034\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"4\"\u003e\u003cem\u003eCMR, cardiac magnetic resonance; CMR-FT, cardiac magnetic resonance feature tracking; CP, constrictive pericarditis; CT, computed tomography; E wave, early peak transmitral velocity; e\u0026prime;, mitral annular early-diastolic velocity; E/A ratio, ratio of early to late transmitral velocities; E/e\u0026prime; ratio, ratio of early transmitral velocity to early mitral annular velocity; LV, left ventricle; LVEDD, left ventricular end-diastolic diameter; LVEDP, left ventricular end-diastolic pressure; LVEDV, left ventricular end-diastolic volume; LVEF, left ventricular ejection fraction; LVESD, left ventricular end-systolic diameter; LVESV, left ventricular end-systolic volume; Myo Mass, myocardial mass (diastolic); PAP, pulmonary artery pressure; PAWP, pulmonary arterial wedge pressure; RAP, right atrial pressure; RCM, restrictive cardiomyopathy; RVEDP, right ventricular end-diastolic pressure.\u003c/em\u003e\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cdiv id=\"Sec11\" class=\"Section2\"\u003e \u003ch2\u003eDiagnostic performance of single imaging feature\u003c/h2\u003e \u003cp\u003eEach imaging-derived metric showing a significant difference (P\u0026thinsp;\u0026lt;\u0026thinsp;0.05) between CP and RCM in Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e was assessed for its ability to distinguish the two conditions. Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e summarizes the diagnostic performance for each feature of echocardiography, CT and CMR. Despite the variations in diagnostic performance, most AUCs ranged between 0.60 and 0.80. Notably, late gadolinium enhancement on CMR exhibited the highest diagnostic performance with an AUC of 0.87 (95%CI: 0.80\u0026ndash;0.92), sensitivity of 93% and specificity of 81%. It is followed by peak longitudinal strain rate on CMR (AUC\u0026thinsp;=\u0026thinsp;0.79, 95%CI: 0.69-0. 85, P\u0026thinsp;=\u0026thinsp;0.001), pericardial thickness (\u0026gt;\u0026thinsp;4 mm) on CT (AUC\u0026thinsp;=\u0026thinsp;0.78, 95%CI: 0.73\u0026ndash;0.80, P\u0026thinsp;=\u0026thinsp;0.002) and LV septal thickness on echocardiography (AUC\u0026thinsp;=\u0026thinsp;0.78, 95%CI: 0.62\u0026ndash;0.88, P\u0026thinsp;=\u0026thinsp;0.002). The distribution of these high-performing parameters across multiple imaging modalities indicates the potential of multimodal imaging integration for more accurate differentiation between CP and RCM.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab3\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 3\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eDiagnostic performance of single cardiac imaging feature in CP and RCM\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"7\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eAUC\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003e95%CI\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eSensitivity\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eSpecificity\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003eAccuracy\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c7\"\u003e \u003cp\u003eP value\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eEchocardiography\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLV septum, mm\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.78\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.62\u0026ndash;0.88\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.84\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.65\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.75\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.002\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLVEDD, mm\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.69\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.51\u0026ndash;0.75\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.69\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.57\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.63\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.045\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLVESD, mm\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.75\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.67\u0026ndash;0.78\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.73\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.63\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.68\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.008\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLVEF, %\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.71\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.64\u0026ndash;0.74\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.76\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.51\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.64\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.030\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eCT\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePericardial calcification, n (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.75\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.64\u0026ndash;0.82\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.50\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.99\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.73\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.008\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePericardial thickness\u0026thinsp;\u0026gt;\u0026thinsp;4 mm, n (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.78\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.73\u0026ndash;0.80\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.57\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.99\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.77\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.002\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eCMR\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLVEDV, ml\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.60\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.51\u0026ndash;0.70\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.71\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.49\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.60\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.200\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLVESV, ml\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.67\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.56\u0026ndash;0.74\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.75\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.51\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.64\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.060\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLVEF, %\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.69\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.56\u0026ndash;0.78\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.66\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.68\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.67\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.040\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMyo Mass, g\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.74\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.66\u0026ndash;0.78\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.80\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.57\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.69\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.010\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMyocardial enhancement, n (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.87\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.80\u0026ndash;0.92\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.93\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.81\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.88\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eLongitudinal\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePeak strain (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.72\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.61\u0026ndash;0.75\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.66\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.60\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.63\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.030\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePeak strain rate (1/s)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.79\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.69-0. 85\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.68\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.81\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.74\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eVelocity (mm/s)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.70\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.66\u0026ndash;0.83\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.66\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.64\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.65\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.080\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eCircumferential\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePeak strain (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.70\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.60\u0026ndash;0.75\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.69\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.57\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.63\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.050\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePeak strain rate (1/s)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.70\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.67\u0026ndash;0.74\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.55\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.76\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.64\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.070\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eStrain ratio of 50\u0026ndash;75%/75\u0026ndash;100% diastolic period\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.55\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.38\u0026ndash;0.68\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.38\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.63\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.52\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.400\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eRadial\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePeak strain (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.75\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.73\u0026ndash;0.80\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.62\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.74\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.67\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.010\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePeak strain rate (1/s)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.72\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.64\u0026ndash;0.82\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.60\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.71\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.64\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.030\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eVelocity (mm/s)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.67\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.50\u0026ndash;0.79\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.53\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.68\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.59\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.100\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"7\"\u003e\u003cem\u003eAUC, the area under the receiver-operating characteristic curve; CI, confidence interval; CMR, cardiac magnetic resonance; LV, left ventricle; ESD, end-systolic diameter; EF, ejection fraction; Myo Mass, myocardial mass (diastolic).\u003c/em\u003e\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec12\" class=\"Section2\"\u003e \u003ch2\u003eML for CP and RCM classification\u003c/h2\u003e \u003cp\u003eML techniques were utilized to combine multimodal imaging results to achieve more accurate differential diagnosis between CP and RCM. Table\u0026nbsp;\u003cspan refid=\"Tab4\" class=\"InternalRef\"\u003e4\u003c/span\u003e and Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e summarize the diagnostic performance of multimodal imaging and single imaging in the differential diagnosis. Using multimodal imaging, all ML methods based on multimodal imaging achieved an AUC greater than 0.90, with a mean AUC around 0.95 (all P\u0026thinsp;\u0026lt;\u0026thinsp;0.001). G-SVM demonstrated the highest diagnostic performance with an AUC of 0.97 (95%CI: 0.93\u0026ndash;0.99), accuracy of 85%, sensitivity of 89% and specificity of 90% (P\u0026thinsp;\u0026lt;\u0026thinsp;0.001). In comparison, the single-imaging ML had AUC values below 0.90, with a mean AUC around 0.80. G-SVM remained the best-performing ML technique for echocardiography or CMR alone. KNN tended to show better diagnostic performance than logistic regression and decision tree methods.\u003c/p\u003e \u003cp\u003e\u003cstrong\u003eTable 4. Machine learning for differential diagnosis between CP and RCM\u003c/strong\u003e\u003c/p\u003e\n\u003cdiv\u003e\n \u003ctable border=\"0\" cellspacing=\"0\" cellpadding=\"0\" width=\"656\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 92px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eImaging modality\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 83px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eModel\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 69px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eAUC\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 82px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e95%CI\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 82px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eSensitivity\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 82px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eSpecificity\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 84px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eAccuracy\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 84px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eP Value\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"8\" valign=\"top\" style=\"width: 92px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eMultimodal imaging\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 83px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eSVMs\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 69px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 82px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 82px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 82px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 84px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 84px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 83px;\"\u003e\n \u003cp\u003e\u0026nbsp; L-SVM\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 69px;\"\u003e\n \u003cp\u003e0.94\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 82px;\"\u003e\n \u003cp\u003e0.90-0.97\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 82px;\"\u003e\n \u003cp\u003e0.82\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 82px;\"\u003e\n \u003cp\u003e0.58\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 84px;\"\u003e\n \u003cp\u003e0.90\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 84px;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 83px;\"\u003e\n \u003cp\u003e\u0026nbsp; Q-SVM\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 69px;\"\u003e\n \u003cp\u003e0.93\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 82px;\"\u003e\n \u003cp\u003e0.87-0.97\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 82px;\"\u003e\n \u003cp\u003e0.75\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 82px;\"\u003e\n \u003cp\u003e0.65\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 84px;\"\u003e\n \u003cp\u003e0.74\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 84px;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 83px;\"\u003e\n \u003cp\u003e\u0026nbsp; C-SVM\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 69px;\"\u003e\n \u003cp\u003e0.96\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 82px;\"\u003e\n \u003cp\u003e0.88-0.99\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 82px;\"\u003e\n \u003cp\u003e0.87\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 82px;\"\u003e\n \u003cp\u003e0.76\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 84px;\"\u003e\n \u003cp\u003e0.74\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 84px;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 83px;\"\u003e\n \u003cp\u003e\u0026nbsp; G-SVM\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 69px;\"\u003e\n \u003cp\u003e0.97\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 82px;\"\u003e\n \u003cp\u003e0.93-0.99\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 82px;\"\u003e\n \u003cp\u003e0.89\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 82px;\"\u003e\n \u003cp\u003e0.90\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 84px;\"\u003e\n \u003cp\u003e0.85\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 84px;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 83px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eDecision Tree\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 69px;\"\u003e\n \u003cp\u003e0.90\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 82px;\"\u003e\n \u003cp\u003e0.85-0.92\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 82px;\"\u003e\n \u003cp\u003e0.86\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 82px;\"\u003e\n \u003cp\u003e0.69\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 84px;\"\u003e\n \u003cp\u003e0.85\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 84px;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 83px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eLogistic regression\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 69px;\"\u003e\n \u003cp\u003e0.95\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 82px;\"\u003e\n \u003cp\u003e0.87-0.99\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 82px;\"\u003e\n \u003cp\u003e0.87\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 82px;\"\u003e\n \u003cp\u003e0.68\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 84px;\"\u003e\n \u003cp\u003e0.80\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 84px;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 83px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eKNN\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 69px;\"\u003e\n \u003cp\u003e0.95\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 82px;\"\u003e\n \u003cp\u003e0.92-0.98\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 82px;\"\u003e\n \u003cp\u003e0.97\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 82px;\"\u003e\n \u003cp\u003e0.68\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 84px;\"\u003e\n \u003cp\u003e0.87\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 84px;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"8\" valign=\"top\" style=\"width: 92px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eEchocardiography\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 83px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eSVMs\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 69px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 82px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 82px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 82px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 84px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 84px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 83px;\"\u003e\n \u003cp\u003e\u0026nbsp; L-SVM\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 69px;\"\u003e\n \u003cp\u003e0.74\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 82px;\"\u003e\n \u003cp\u003e0.66-0.82\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 82px;\"\u003e\n \u003cp\u003e0.83\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 82px;\"\u003e\n \u003cp\u003e0.62\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 84px;\"\u003e\n \u003cp\u003e0.82\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 84px;\"\u003e\n \u003cp\u003e0.002\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 83px;\"\u003e\n \u003cp\u003e\u0026nbsp; Q-SVM\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 69px;\"\u003e\n \u003cp\u003e0.82\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 82px;\"\u003e\n \u003cp\u003e0.78-0.85\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 82px;\"\u003e\n \u003cp\u003e0.75\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 82px;\"\u003e\n \u003cp\u003e0.70\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 84px;\"\u003e\n \u003cp\u003e0.73\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 84px;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 83px;\"\u003e\n \u003cp\u003e\u0026nbsp; C-SVM\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 69px;\"\u003e\n \u003cp\u003e0.83\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 82px;\"\u003e\n \u003cp\u003e0.79-0.87\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 82px;\"\u003e\n \u003cp\u003e0.80\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 82px;\"\u003e\n \u003cp\u003e0.37\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 84px;\"\u003e\n \u003cp\u003e0.76\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 84px;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 83px;\"\u003e\n \u003cp\u003e\u0026nbsp; G-SVM\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 69px;\"\u003e\n \u003cp\u003e0.85\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 82px;\"\u003e\n \u003cp\u003e0.79-0.91\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 82px;\"\u003e\n \u003cp\u003e0.85\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 82px;\"\u003e\n \u003cp\u003e0.90\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 84px;\"\u003e\n \u003cp\u003e0.81\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 84px;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 83px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eDecision Tree\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 69px;\"\u003e\n \u003cp\u003e0.78\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 82px;\"\u003e\n \u003cp\u003e0.71-0.85\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 82px;\"\u003e\n \u003cp\u003e0.87\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 82px;\"\u003e\n \u003cp\u003e0.69\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 84px;\"\u003e\n \u003cp\u003e0.80\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 84px;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 83px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eLogistic regression\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 69px;\"\u003e\n \u003cp\u003e0.76\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 82px;\"\u003e\n \u003cp\u003e0.71-0.81\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 82px;\"\u003e\n \u003cp\u003e0.82\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 82px;\"\u003e\n \u003cp\u003e0.68\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 84px;\"\u003e\n \u003cp\u003e0.76\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 84px;\"\u003e\n \u003cp\u003e0.002\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 83px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eKNN\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 69px;\"\u003e\n \u003cp\u003e0.85\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 82px;\"\u003e\n \u003cp\u003e0.79-0.92\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 82px;\"\u003e\n \u003cp\u003e0.88\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 82px;\"\u003e\n \u003cp\u003e0.69\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 84px;\"\u003e\n \u003cp\u003e0.79\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 84px;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"8\" valign=\"top\" style=\"width: 92px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eCMR\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 83px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eSVMs\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 69px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 82px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 82px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 82px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 84px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 84px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 83px;\"\u003e\n \u003cp\u003e\u0026nbsp; L-SVM\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 69px;\"\u003e\n \u003cp\u003e0.82\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 82px;\"\u003e\n \u003cp\u003e0.76-0.88\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 82px;\"\u003e\n \u003cp\u003e0.84\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 82px;\"\u003e\n \u003cp\u003e0.81\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 84px;\"\u003e\n \u003cp\u003e0.89\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 84px;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 83px;\"\u003e\n \u003cp\u003e\u0026nbsp; Q-SVM\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 69px;\"\u003e\n \u003cp\u003e0.81\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 82px;\"\u003e\n \u003cp\u003e0.71-0.91\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 82px;\"\u003e\n \u003cp\u003e0.75\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 82px;\"\u003e\n \u003cp\u003e0.68\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 84px;\"\u003e\n \u003cp\u003e0.89\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 84px;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 83px;\"\u003e\n \u003cp\u003e\u0026nbsp; C-SVM\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 69px;\"\u003e\n \u003cp\u003e0.80\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 82px;\"\u003e\n \u003cp\u003e0.75-0.86\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 82px;\"\u003e\n \u003cp\u003e0.85\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 82px;\"\u003e\n \u003cp\u003e0.67\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 84px;\"\u003e\n \u003cp\u003e0.76\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 84px;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 83px;\"\u003e\n \u003cp\u003e\u0026nbsp; G-SVM\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 69px;\"\u003e\n \u003cp\u003e0.84\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 82px;\"\u003e\n \u003cp\u003e0.80-0.88\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 82px;\"\u003e\n \u003cp\u003e0.74\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 82px;\"\u003e\n \u003cp\u003e0.81\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 84px;\"\u003e\n \u003cp\u003e0.89\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 84px;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 83px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eDecision Tree\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 69px;\"\u003e\n \u003cp\u003e0.75\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 82px;\"\u003e\n \u003cp\u003e0.68-0.83\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 82px;\"\u003e\n \u003cp\u003e0.72\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 82px;\"\u003e\n \u003cp\u003e0.83\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 84px;\"\u003e\n \u003cp\u003e0.89\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 84px;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 83px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eLogistic regression\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 69px;\"\u003e\n \u003cp\u003e0.82\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 82px;\"\u003e\n \u003cp\u003e0.75-0.88\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 82px;\"\u003e\n \u003cp\u003e0.82\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 82px;\"\u003e\n \u003cp\u003e0.81\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 84px;\"\u003e\n \u003cp\u003e0.89\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 84px;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 83px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eKNN\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 69px;\"\u003e\n \u003cp\u003e0.84\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 82px;\"\u003e\n \u003cp\u003e0.80-0.88\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 82px;\"\u003e\n \u003cp\u003e0.84\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 82px;\"\u003e\n \u003cp\u003e0.78\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 84px;\"\u003e\n \u003cp\u003e0.88\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 84px;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 92px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eCT\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 83px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eSVMs\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 69px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 82px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 82px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 82px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 84px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 84px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 92px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 83px;\"\u003e\n \u003cp\u003e\u0026nbsp; L-SVM\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 69px;\"\u003e\n \u003cp\u003e0.75\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 82px;\"\u003e\n \u003cp\u003e0.70-0.80\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 82px;\"\u003e\n \u003cp\u003e0.82\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 82px;\"\u003e\n \u003cp\u003e0.44\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 84px;\"\u003e\n \u003cp\u003e0.68\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 84px;\"\u003e\n \u003cp\u003e0.007\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 92px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 83px;\"\u003e\n \u003cp\u003e\u0026nbsp; Q-SVM\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 69px;\"\u003e\n \u003cp\u003e0.72\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 82px;\"\u003e\n \u003cp\u003e0.72-0.80\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 82px;\"\u003e\n \u003cp\u003e0.75\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 82px;\"\u003e\n \u003cp\u003e0.83\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 84px;\"\u003e\n \u003cp\u003e0.68\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 84px;\"\u003e\n \u003cp\u003e0.020\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 92px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 83px;\"\u003e\n \u003cp\u003e\u0026nbsp; C-SVM\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 69px;\"\u003e\n \u003cp\u003e0.75\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 82px;\"\u003e\n \u003cp\u003e0.72-0.78\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 82px;\"\u003e\n \u003cp\u003e0.80\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 82px;\"\u003e\n \u003cp\u003e0.67\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 84px;\"\u003e\n \u003cp\u003e0.73\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 84px;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 92px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 83px;\"\u003e\n \u003cp\u003e\u0026nbsp; G-SVM\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 69px;\"\u003e\n \u003cp\u003e0.75\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 82px;\"\u003e\n \u003cp\u003e0.72-0.80\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 82px;\"\u003e\n \u003cp\u003e0.84\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 82px;\"\u003e\n \u003cp\u003e0.80\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 84px;\"\u003e\n \u003cp\u003e0.79\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 84px;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 92px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 83px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eDecision Tree\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 69px;\"\u003e\n \u003cp\u003e0.78\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 82px;\"\u003e\n \u003cp\u003e0.71-0.85\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 82px;\"\u003e\n \u003cp\u003e0.64\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 82px;\"\u003e\n \u003cp\u003e0.74\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 84px;\"\u003e\n \u003cp\u003e0.68\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 84px;\"\u003e\n \u003cp\u003e0.023\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 92px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 83px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eLogistic regression\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 69px;\"\u003e\n \u003cp\u003e0.76\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 82px;\"\u003e\n \u003cp\u003e0.71-0.81\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 82px;\"\u003e\n \u003cp\u003e0.71\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 82px;\"\u003e\n \u003cp\u003e0.76\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 84px;\"\u003e\n \u003cp\u003e0.73\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 84px;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 92px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 83px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eKNN\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 69px;\"\u003e\n \u003cp\u003e0.76\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 82px;\"\u003e\n \u003cp\u003e0.72-0.80\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 82px;\"\u003e\n \u003cp\u003e0.80\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 82px;\"\u003e\n \u003cp\u003e0.66\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 84px;\"\u003e\n \u003cp\u003e0.81\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 84px;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003c/table\u003e\n\u003c/div\u003e\n\u003cp\u003e\u003cem\u003eAUC, the area under the receiver-operating characteristic curve;\u003c/em\u003e \u003cem\u003eCI, confidence interval; CMR, cardiac magnetic resonance;\u003c/em\u003e\u003cem\u003e\u0026nbsp;CMR-FT, cardiac magnetic resonance feature tracking;\u003c/em\u003e\u003cem\u003e\u0026nbsp;C-SVM, cubic SVM; G-SVM, Gaussian-SVM; KNN, k-nearest neighbor ; L-SVM, linear SVM; SVM: support vector machine; Q-SVM: quadratic SVM.\u003c/em\u003e\u003c/p\u003e\u003cdiv id=\"Sec13\" class=\"Section2\"\u003e \u003ch2\u003eComparison with other imaging studies\u003c/h2\u003e \u003cp\u003eTable\u0026nbsp;\u003cspan refid=\"Tab5\" class=\"InternalRef\"\u003e5\u003c/span\u003e shows a comparison of the proposed multimodal G-SVM with other published imaging studies for CP and RCM classification. Echocardiographic studies on annular displacement analysis and tissue Doppler/speckle tracking had a small sample size (n\u0026thinsp;\u0026lt;\u0026thinsp;100) and reported a low AUC\u0026thinsp;\u0026lt;\u0026thinsp;0.85. \u003csup\u003e24,25\u003c/sup\u003e Deep learning resulted in an AUC up to 0.97 in a highly selected group of patients with only cardiac amyloidosis, a subtype of RCM.\u003csup\u003e\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e\u003c/sup\u003e The study used imaging features from transthoracic echocardiography as ML input without providing specific imaging features. CT analysis of pericardial and cardiac abnormalities provided a low AUC\u0026thinsp;\u0026lt;\u0026thinsp;0.75.\u003csup\u003e6,16\u003c/sup\u003e CMR-FT on LV and LA strain imaging resulted in high AUCs between 0.88\u0026ndash;0.95, though heavy postprocessing procedures might have been involved.\u003csup\u003e\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e,\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e,\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e\u003c/sup\u003e\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab5\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 5\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eComparison of cardiac imaging for CP vs. RCM differentiation\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"5\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cem\u003eStudy (Ref)\u003c/em\u003e\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u003cem\u003ePopulation\u003c/em\u003e\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003e\u003cem\u003eImaging modality\u003c/em\u003e\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003e\u003cem\u003eKey methodologies\u003c/em\u003e\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003e\u003cem\u003eAUC\u003c/em\u003e\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cem\u003eLiu et al. (4)\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u003cem\u003e45 CP\u003c/em\u003e,\u003c/p\u003e \u003cp\u003e\u003cem\u003e40 RCM\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\" morerows=\"2\" rowspan=\"3\"\u003e \u003cp\u003e\u003cem\u003eEcho\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e\u003cem\u003eAnnular displacement analysis\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e\u003cem\u003e0.82\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cem\u003eSengupta et al. (7)\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u003cem\u003e30 CP\u003c/em\u003e,\u003c/p\u003e \u003cp\u003e\u003cem\u003e30 RCM\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e\u003cem\u003eTissue Doppler/speckle-tracking\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e\u003cem\u003e0.85\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cem\u003eChao et al. (10)\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u003cem\u003e112 CP\u003c/em\u003e,\u003c/p\u003e \u003cp\u003e\u003cem\u003e98 RCM\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e\u003cem\u003eDeep learning\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e\u003cem\u003e0.97\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cem\u003eGarcia et al. (3)\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u003cem\u003e50 CP\u003c/em\u003e,\u003c/p\u003e \u003cp\u003e\u003cem\u003e30 RCM\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003e\u003cem\u003eCT\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e\u003cem\u003eStructural evaluation\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e\u003cem\u003e0.72\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cem\u003eTalreja et al. (5)\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u003cem\u003e143 CP\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e\u003cem\u003ePericardial thickness/calcification assessment\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e\u003cem\u003e0.75\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cem\u003eAmaki et al. (8)\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u003cem\u003e25 CP\u003c/em\u003e,\u003c/p\u003e \u003cp\u003e\u003cem\u003e25 RCM\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\" morerows=\"2\" rowspan=\"3\"\u003e \u003cp\u003e\u003cem\u003eCMR-FT\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e\u003cem\u003eStrain\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e\u003cem\u003e0.89\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cem\u003eYang et al. (16)\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u003cem\u003e50 CP\u003c/em\u003e,\u003c/p\u003e \u003cp\u003e\u003cem\u003e45 RCM\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e\u003cem\u003eLV Time-strain curve analysis\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e\u003cem\u003e0.88\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cem\u003eBo et al. (11)\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u003cem\u003e60 CP\u003c/em\u003e,\u003c/p\u003e \u003cp\u003e\u003cem\u003e55 RCM\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e\u003cem\u003eLA strain analysis\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e\u003cem\u003e0.95\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eOur Study\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u003cb\u003e156 CP\u003c/b\u003e,\u003c/p\u003e \u003cp\u003e\u003cb\u003e91 RCM\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e\u003cb\u003eMultimodal imaging (Echo\u0026thinsp;+\u0026thinsp;CT\u0026thinsp;+\u0026thinsp;CMR)\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e\u003cb\u003eGaussian SVM with 12 imaging variables\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e\u003cb\u003e0.97\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"5\"\u003e\u003cem\u003eAUC, area under the receiver-operating characteristic curve; CP, constrictive pericarditis; CMR-FT, cardiac magnetic resonance feature tracking; CT, computed tomography; echo, echocardiography; LA, left atrium; LV, left ventricle; ML, machine learning; RCM, restrictive cardiomyopathy; SVM, support vector machine\u003c/em\u003e\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003c/div\u003e"},{"header":"Discussion","content":"\u003cp\u003eDrawing on a ten-year two-center cohort, this study provides the largest assessment to date of multimodal ML for differentiating CP from RCM. A G-SVM trained with routine echocardiography, CT and CMR variables reached an AUC of 0.97 and outperformed every single-modality model as well as the best individual imaging marker. These findings indicate that a compact multimodal feature set can achieve clinically meaningful accuracy when baseline profiles overlap and treatment strategies differ. They also highlight the potential of multimodal-imaging ML to streamline diagnostic workflows and lessen the need for invasive testing.\u003c/p\u003e \u003cdiv id=\"Sec15\" class=\"Section2\"\u003e \u003ch2\u003eClinical dilemmas in the differential diagnosis of CP and RCM\u003c/h2\u003e \u003cp\u003eThe differential diagnosis of CP and RCM remains challenging because both conditions present with preserved ejection fraction, impaired left-ventricular relaxation and bi-atrial enlargement, yet they require opposite treatments. Timely recognition of CP is essential because pericardiectomy can restore hemodynamics and yields five-to-seven-year survival above 80%\u003csup\u003e27\u003c/sup\u003e, whereas RCM, particularly when associated with cardiac amyloidosis, lacks disease-modifying therapy and has a median survival below 38 months.\u003csup\u003e\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e\u003c/sup\u003e Misclassifying RCM as CP exposes patients to unnecessary surgery with reported perioperative mortality close to 50%, \u003csup\u003e28\u003c/sup\u003e while failure to identify operable CP removes the only curative option. Baseline clinical features alone cannot resolve this dilemma.\u003c/p\u003e \u003cp\u003eEach single-modality imaging technique provides a limited aspect of the diagnostic assessment. Echocardiography relies on subjective assessment of septal motion and cannot reveal myocardial infiltration. \u003csup\u003e29,30\u003c/sup\u003e CT depicts only static anatomy and may miss functionally constrictive cases with normal pericardial thickness. \u003csup\u003e7\u003c/sup\u003e CMR strain analysis requires labor-intensive post-processing and does not capture real-time hemodynamic information. \u003csup\u003e31,32\u003c/sup\u003e Consequently about 30% of cases remain indeterminate after non-invasive evaluation and proceed to invasive catheterization. \u003csup\u003e2\u003c/sup\u003e These limitations support the development of a ML strategy that integrates multimodal imaging data to improve diagnostic accuracy and guide timely clinical decisions.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec16\" class=\"Section2\"\u003e \u003ch2\u003eMultimodal-imaging ML enhances classification accuracy\u003c/h2\u003e \u003cp\u003eOur analysis indicates that integrating echocardiography, CT and CMR improves the classification of CP and RCM in this cohort. Across SVMs, KNN, decision tree and logistic regression, the multimodal Gaussian SVM achieved the highest performance with an AUC of 0.97 and balanced sensitivity 0.89 and specificity 0.90. Feature-importance analysis showed approximately equal contributions from CMR-derived strain measures, echocardiographic functional indices and CT-based anatomical features. This distribution is consistent with clinical reasoning because tissue characterization, ventricular function and pericardial structure provide complementary information that supports differential diagnosis.\u003c/p\u003e \u003cp\u003eThe observed accuracy is comparable to reports based on single-modality models using carefully curated echocardiographic datasets\u003csup\u003e\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e\u003c/sup\u003e or CMR feature tracking parameters. \u003csup\u003e19,26\u003c/sup\u003e However, our evaluation used routine acquisitions from all three modalities in a larger and less selected cohort, which improves applicability to everyday practice. In this dataset, SVMs outperformed decision tree, logistic regression and KNN, likely because margin-based classifiers handle high dimensional predictors in modest samples while maintaining generalization. The radial basis kernel can capture nonlinear relations such as those between pericardial thickness and myocardial strain without excessive parameterization. All models used structured imaging features rather than raw images, which reduced variability introduced by preprocessing. Similar advantages of kernel-based SVMs have been reported in cardiovascular classification tasks. \u003csup\u003e33\u003c/sup\u003e\u003c/p\u003e \u003cp\u003eWhen the same algorithms were trained on any single modality in our dataset, AUCs were consistently lower. This pattern supports a mechanistic interpretation in which echocardiography describes chamber motion and filling, CT depicts pericardial structure and CMR characterizes myocardial tissue. The combination of these selected variables therefore offers a broader view of the underlying pathophysiology than any single technique. In this context, the feature selection step identified clinically meaningful variables and limited redundancy, which likely contributed to efficiency and interpretability.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec17\" class=\"Section2\"\u003e \u003ch2\u003eFuture directions\u003c/h2\u003e \u003cp\u003eProspective validation across diverse multicenter cohorts should be prioritized to address key limitations, including the retrospective study design and predominance of amyloidosis-related RCM cases within our group.\u003csup\u003e\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e,\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e\u003c/sup\u003e Such validation would better establish generalizability across the broader spectrum of RCM. Incorporating longitudinal clinical endpoints particularly post-pericardiectomy functional recovery and survival outcomes \u003csup\u003e\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e\u003c/sup\u003e would further substantiate the model\u0026rsquo;s clinical utility and prognostic relevance. For practical implementation, carefully designed clinician-ML collaborative frameworks \u003csup\u003e\u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e,\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e\u003c/sup\u003e could potentially enhance diagnostic accuracy while preserving essential clinical reasoning and mitigating over-reliance on algorithmic outputs. Future methodological refinements might incorporate adaptive learning mechanisms to accommodate evolving disease phenotypes and emerging diagnostic biomarkers. \u003csup\u003e35\u003c/sup\u003e\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec18\" class=\"Section2\"\u003e \u003ch2\u003eLimitations\u003c/h2\u003e \u003cp\u003eThis study has several limitations. First, the analysis is retrospective and confined to two centers, and the model was tested only with internal cross-validation, so validation in independent external cohorts and prospective evaluation of clinical benefit are still needed. Second, the cohort is modest in size, and most RCM cases are amyloidosis, which may limit the applicability of the findings to other restrictive phenotypes. Third, the algorithm assumes that echocardiography, CT and CMR are available for every patient, a multimodal protocol that may be difficult to achieve in resource-limited or time-pressured settings.\u003c/p\u003e \u003c/div\u003e"},{"header":"Conclusions","content":"\u003cp\u003eThis study delineates the clinical and multimodal imaging distinctions between CP and RCM and shows that ML classifiers combining echocardiography, CT and CMR data are a feasible and effective tool for accurate differential diagnosis.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eEthics approval and consent to participate\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis study was conducted in accordance with the Declaration of Helsinki and was approved by the Institutional Ethics Committee of Peking University Third Hospital (Beijing, China) and the Institutional Ethics Committee of Beijing Anzhen Hospital, Capital Medical University (Beijing, China) (Approval No. 2006003x). Due to the retrospective nature of the study and the use of anonymized clinical data, the requirement for written informed consent was waived by the ethics committees.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent for publication\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAvailability of data and materials\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting interests\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors declare no competing financial interests.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis work was supported by the Beijing Research Ward Excellence Clinical Study Program (BRWEP2024W014090201); the National Natural Science, Foundation of China (Grant No. 82170381).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthors' contributions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eZ.Y. analyzed data and drafted the manuscript. J.W. and X.M. were involved in designing the tool, interviewing patients and analyzing data. H.Z. and X.D. collected the data and performed statistical analysis. J.D. and M.C. collected data. R.Z. contributed to the data analysis. H.L. made contributions to interpretation of the data and critically edited the manuscript. All authors contributed to discussion and interpretation of the results, critically reviewed the manuscript, and approved the final version to be submitted. All authors had access to the data and agreed to submit the manuscript for publication.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eNishimura RA, Borlaug BA. Diastology for the clinician. J Cardiol. 2019;73(6):445\u0026ndash;52.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHirshfeld JW Jr, Johnston-Cox H. Distinguishing Constrictive Pericarditis From Restrictive Cardiomyopathy-An Ongoing Diagnostic Challenge. JAMA Cardiol. 2022;7(1):13\u0026ndash;4.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLloyd JW, Anavekar NS, Oh JK, Miranda WR. Multimodality imaging in differentiating constrictive pericarditis from restrictive cardiomyopathy: a comprehensive overview for clinicians and imagers. J Am Soc Echocardiogr. 2023;36:1254\u0026ndash;65.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAdler Y, Charron P, Imazio M, et al. 2015 ESC Guidelines for the diagnosis and management of pericardial diseases. Eur Heart J. 2015;36:2921\u0026ndash;64.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWelch TD, Ling LH, Espinosa RE, et al. Echocardiographic diagnosis of constrictive pericarditis: Mayo Clinic criteria. Circ Cardiovasc Imaging. 2014;7:526\u0026ndash;34.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGarcia MJ. Constrictive Pericarditis Versus Restrictive Cardiomyopathy? J Am Coll Cardiol. 2016;67(17):2061\u0026ndash;76.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTalreja DR, Edwards WD, Danielson GK, et al. Constrictive pericarditis in 26 patients with histologically normal pericardial thickness. Circulation. 2003;108(15):1852\u0026ndash;7.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAmaki M, Savino J, Ain DL, et al. Diagnostic concordance of echocardiography and cardiac magnetic resonance-based tissue tracking for differentiating constrictive pericarditis from restrictive cardiomyopathy. Circ Cardiovasc Imaging. 2014;7(5):819\u0026ndash;27.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBo K, Zhao Y, Gao X, et al. Cardiac magnetic resonance feature tracking derived left atrial strain in the diagnosis of patients with constrictive pericarditis and restrictive cardiomyopathy. Heliyon. 2024;10(7):e28768.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSyed FF, Schaff HV, Oh JK. Constrictive pericarditis\u0026mdash;a curable diastolic heart failure. Nat Rev Cardiol. 2015;12(9):682\u0026ndash;90.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLing LH, Oh JK, Schaff HV, et al. Constrictive pericarditis in the modern era: evolving clinical spectrum and impact on outcome after pericardiectomy. Circulation. 1999;100(13):1380\u0026ndash;6.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDuchateau N, King AP, De Craene M. Machine Learning Approaches for Myocardial Motion and Deformation Analysis. Front Cardiovasc Med. 2019;6:190.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChao C-J, Jeong J, Arsanjani R, et al. Echocardiography-based deep learning model to differentiate constrictive pericarditis and restrictive cardiomyopathy. JACC Cardiovasc Imaging. 2024;17(4):349\u0026ndash;60.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTabassian M, Sunderji I, Erdei T, et al. Diagnosis of heart failure with preserved ejection fraction: machine learning of spatiotemporal variations in left ventricular deformation. J Am Soc Echocardiogr. 2018;31(12):1272\u0026ndash;84.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSengupta PP, Huang YM, Bansal M, Ashrafi A, Fisher M, Shameer K, Gall W, Dudley JT. Cognitive Machine-Learning Algorithm for Cardiac Imaging: A Pilot Study for Differentiating Constrictive Pericarditis From Restrictive Cardiomyopathy. Circ Cardiovasc Imaging. 2016;9(6).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTalreja DR, Nishimura RA, Oh JK, Holmes DR. Constrictive pericarditis in the modern era: novel criteria for diagnosis in the cardiac catheterization laboratory. J Am Coll Cardiol. 2008;51(3):315\u0026ndash;9.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKusunose K, Dahiya A, Popovic ZB, Motoki H, Alraies MC, Zurick AO, Bolen MA, Kwon DH, Flamm SD, Klein AL. Biventricular mechanics in constrictive pericarditis comparison with restrictive cardiomyopathy and impact of pericardiectomy. Circ Cardiovasc Imaging. 2013;6(3):399\u0026ndash;406.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHurrell DG, Nishimura RA, Higano ST, Appleton CP, Danielson GK, Holmes DR Jr, Tajik AJ. Value of dynamic respiratory changes in left and right ventricular pressures for the diagnosis of constrictive pericarditis. Circulation. 1996;93:2007\u0026ndash;13.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAmaki M, Savino J, Ain DL, Sanz J, Pedrizzetti G, Kulkarni H, Narula J, Sengupta PP. Diagnostic concordance of echocardiography and cardiac magnetic resonance-based tissue tracking for differentiating constrictive pericarditis from restrictive cardiomyopathy. Circ Cardiovasc Imaging. 2014;7(5):819\u0026ndash;27.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTabassian M, Sunderji I, Erdei T, Sanchez-Martinez S, Degiovanni A, Marino P, Fraser AG, D'Hooge J. Diagnosis of Heart Failure With Preserved Ejection Fraction: Machine Learning of Spatiotemporal Variations in Left Ventricular Deformation. J Am Soc Echocardiogr. 2018;31(12):1272\u0026ndash;84.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYang Z, Wang H, Chang S, Cui J, Zhou L, Lv Q, He Y, Du X, Dong J, Ma C. Left ventricular strain-curve morphology to distinguish between constrictive pericarditis and restrictive cardiomyopathy. ESC Heart Fail. 2021;8(6):4863\u0026ndash;72.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHao PY, Chiang JH, Chen YD. Possibilistic classification by support vector networks. Neural Netw. 2022;149:40\u0026ndash;56.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMiranda A, Lavrador R, Julio F, Januario C, Castelo-Branco M, Caetano G. Classification of Huntington's disease stage with support vector machines: A study on oculomotor performance. Behav Res Methods. 2016;48(4):1667\u0026ndash;77.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLiu S, Ren W, Zhang J, Ma C, Yang J, Zhang Y, Guan Z. Incremental Value of the Tissue Motion of Annular Displacement Derived From Speckle-Tracking Echocardiography for Differentiating Chronic Constrictive Pericarditis From Restrictive Cardiomyopathy. J Ultrasound Med. 2018;37(11):2637\u0026ndash;45.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSengupta PP, Krishnamoorthy VK, Abhayaratna WP, Korinek J, Belohlavek M, Sundt TM 3rd, Chandrasekaran K, Mookadam F, Seward JB, Tajik AJ, Khandheria BK. Disparate patterns of left ventricular mechanics differentiate constrictive pericarditis from restrictive cardiomyopathy. JACC Cardiovasc Imaging. 2008;1(1):29\u0026ndash;38.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBo K, Zhao Y, Gao X, Chen Y, Ren Y, Gao Y, Zhou Z, Wang H, Xu L. Cardiac magnetic resonance feature tracking derived left atrial strain in the diagnosis of patients with constrictive pericarditis and restrictive cardiomyopathy. Heliyon. 2024;10(7):e28768.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWelch TD. Constrictive pericarditis: diagnosis, management and clinical outcomes. Heart. 2018;104(9):725\u0026ndash;31.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTalreja DR, Edwards WD, Danielson GK, Schaff HV, Tajik AJ, Tazelaar HD, Breen JF, Oh JK. Constrictive pericarditis in 26 patients with histologically normal pericardial thickness. Circulation. 2003;108(15):1852\u0026ndash;7.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBijnens B, Cikes M, Butakoff C, Sitges M, Crispi F. Myocardial motion and deformation: What does it tell us and how does it relate to function? Fetal Diagn Ther. 2012;32(1\u0026ndash;2):5\u0026ndash;16.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRajagopalan N, Garcia MJ, Rodriguez L, Murray RD, Hansen CA, Stugaard M, Thomas JD, Klein AL. Comparison of new Doppler echocardiographic methods to differentiate constrictive pericardial heart disease and restrictive cardiomyopathy. Am J Cardiol. 2001;87:86\u0026ndash;94.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePalka P, Lange A, Donnelly JE, Nihoyannopoulos P. Differentiation between restrictive cardiomyopathy and constrictive pericarditis by early diastolic Doppler myocardial velocity gradient at the posterior wall. Circulation. 2000;102:655\u0026ndash;62.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eClaus P, Omar AMS, Pedrizzetti G, Sengupta PP, Nagel E. Tissue Tracking Technology for Assessing Cardiac Mechanics: Principles, Normal Values, and Clinical Applications. JACC Cardiovasc Imaging. 2015;8(12):1444\u0026ndash;60.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCikes M, Sanchez-Martinez S, Claggett B, Duchateau N, Piella G, Butakoff C, Pouleur AC, Knappe D, Biering-Sorensen T, Kutyifa V, Moss A, Stein K, Solomon SD, Bijnens B. Machine learning-based phenogrouping in heart failure to identify responders to cardiac resynchronization therapy. Eur J Heart Fail. 2019;21(1):74\u0026ndash;85.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eQuitadamo LR, Cavrini F, Sbernini L, Riillo F, Bianchi L, Seri S, Saggio G. Support vector machines to detect physiological patterns for EEG and EMG-based human-computer interaction: a review. J Neural Eng. 2017;14(1):011001.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSong-Men S. Intelligent Diagnosis Method for New Diseases Based on Fuzzy SVM Incremental Learning. Comput Math Methods Med. 2022;2022:7631271.\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"bmc-medical-imaging","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"bmim","sideBox":"Learn more about [BMC Medical Imaging](http://bmcmedimaging.biomedcentral.com/)","snPcode":"","submissionUrl":"https://www.editorialmanager.com/bmim/default.aspx","title":"BMC Medical Imaging","twitterHandle":"BMC_series","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"em","reportingPortfolio":"BMC Series","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"Constrictive pericarditis, restrictive cardiomyopathy, multimodal imaging, support vector machine, cardiac magnetic resonance imaging.","lastPublishedDoi":"10.21203/rs.3.rs-8917150/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-8917150/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003ch2\u003eAim:\u003c/h2\u003e \u003cp\u003eConstrictive pericarditis (CP) and restrictive cardiomyopathy (RCM) require distinct treatments but share overlapping clinical and imaging features, making them difficult to distinguish using single conventional imaging. This study evaluated whether machine learning models integrating multimodal imaging data could improve the differential diagnosis.\u003c/p\u003e\u003ch2\u003eMethods:\u003c/h2\u003e \u003cp\u003eCP and RCM patients were included between January 2014 and September 2024 at two hospitals. Demographic, laboratory, echocardiography, computed tomography, magnetic resonance imaging, and cardiac catheterization data were collected. Seven machine learning models were trained and validated using five-fold cross-validation. Model performance was assessed by the area under the receiver-operating characteristic curve (AUC), accuracy, sensitivity, specificity and 95% confidence interval (CI).\u003c/p\u003e\u003ch2\u003eResults:\u003c/h2\u003e \u003cp\u003eA total of 156 CP and 91 RCM patients were analyzed. Compared with RCM, CP patients showed a higher ejection fraction (62.9\u0026thinsp;\u0026plusmn;\u0026thinsp;7.5% \u003cem\u003evs.\u003c/em\u003e 51.9\u0026thinsp;\u0026plusmn;\u0026thinsp;15.2%, P\u0026thinsp;\u0026lt;\u0026thinsp;0.001) and more frequent pericardial thickening over 4 mm (66.1% \u003cem\u003evs.\u003c/em\u003e 0%, P\u0026thinsp;\u0026lt;\u0026thinsp;0.001). The Gaussian SVM of multimodal-imaging achieved the highest AUC of 0.97 (95%CI: 0.93\u0026ndash;0.99), accuracy of 85%, sensitivity of 89%, and specificity of 90%, outperforming other SVM models, decision tree, logistic regression and k-nearest-neighbor models. ML based on multimodal imaging data achieved higher diagnostic performance than those based on any single-modality imaging features.\u003c/p\u003e\u003ch2\u003eConclusions:\u003c/h2\u003e \u003cp\u003eA Gaussian SVM integrating multimodal imaging data markedly improves the differential diagnosis between CP and RCM. This model may help reduce misclassification and improve timely interventions for the two diseases.\u003c/p\u003e","manuscriptTitle":"Machine Learning of Multimodal Imaging to Differentiate Constrictive Pericarditis and Restrictive Cardiomyopathy","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-05-11 08:08:13","doi":"10.21203/rs.3.rs-8917150/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"editorInvitedReview","content":"","date":"2026-05-15T15:10:58+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"268228838835479909091806447555621718336","date":"2026-05-07T15:23:25+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"289740037783696451079864966172474287113","date":"2026-05-07T08:29:10+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2026-04-29T04:45:30+00:00","index":"","fulltext":""},{"type":"editorInvited","content":"","date":"2026-04-07T05:18:24+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2026-02-21T05:10:47+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2026-02-21T05:10:44+00:00","index":"","fulltext":""},{"type":"submitted","content":"BMC Medical Imaging","date":"2026-02-19T11:28:44+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"bmc-medical-imaging","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"bmim","sideBox":"Learn more about [BMC Medical Imaging](http://bmcmedimaging.biomedcentral.com/)","snPcode":"","submissionUrl":"https://www.editorialmanager.com/bmim/default.aspx","title":"BMC Medical Imaging","twitterHandle":"BMC_series","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"em","reportingPortfolio":"BMC Series","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"f179ed27-8652-46d7-8150-00b5d2c45e84","owner":[],"postedDate":"May 11th, 2026","published":true,"recentEditorialEvents":[{"type":"editorInvitedReview","content":"","date":"2026-05-15T15:10:58+00:00","index":66,"fulltext":""},{"type":"reviewerAgreed","content":"268228838835479909091806447555621718336","date":"2026-05-07T15:23:25+00:00","index":60,"fulltext":""},{"type":"reviewerAgreed","content":"289740037783696451079864966172474287113","date":"2026-05-07T08:29:10+00:00","index":59,"fulltext":""}],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[],"tags":[],"updatedAt":"2026-05-11T08:08:13+00:00","versionOfRecord":[],"versionCreatedAt":"2026-05-11 08:08:13","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-8917150","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-8917150","identity":"rs-8917150","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}