Assessment of right ventricular size and function from cardiovascular magnetic resonance images using artificial intelligence.

BACKGROUND: Theoretically, artificial intelligence can provide an accurate automatic solution to measure right ventricular (RV) ejection fraction (RVEF) from cardiovascular magnetic resonance (CMR) images, despite the complex RV geometry. However, in our recent study, commercially available deep learning (DL) algorithms for RVEF quantification performed poorly in some patients. The current study was designed to test the hypothesis that quantification of RV function could be improved in these patients by using more diverse CMR datasets in addition to domain-specific quantitative performance evaluation metrics during the cross-validation phase of DL algorithm development. METHODS: We identified 100 patients from our prior study who had the largest differences between manually measured and automated RVEF values. Automated RVEF measurements were performed using the original version of the algorithm (DL1), an updated version (DL2) developed from a dataset that included a wider range of RV pathology and validated using multiple domain-specific quantitative performance evaluation metrics, and conventional methodology performed by a core laboratory (CORE). Each of the DL-RVEF approaches was compared against CORE-RVEF reference values using linear regression and Bland-Altman analyses. Additionally, RVEF values were classified into 3 categories: ≤ 35%, 35-50%, and ≥ 50%. Agreement between RVEF classifications made by the DL approaches and the CORE measurements was tested. RESULTS: CORE-RVEF and DL-RVEFs were obtained in all patients (feasibility of 100%). DL2-RVEF correlated with CORE-RVEF better than DL1-RVEF (r = 0.87 vs. r = 0.42), with narrower limits of agreement. As a result, DL2 algorithm also showed increasing accuracy from 0.53 to 0.80 for categorizing RV function. CONCLUSIONS: The use of a new DL algorithm cross-validated on a dataset with a wide range of RV pathology using multiple domain-specific metrics resulted in a considerable improvement in the accuracy of automated RVEF measurements. This improvement was demonstrated in patients whose images were the most challenging and resulted in the largest RVEF errors. These findings underscore the critical importance of this strategy in the development of DL approaches for automated CMR measurements.

Impact of Wideband Late Gadolinium Enhancement Cardiac Magnetic Resonance Imaging on Device-Related Artifacts in Different Implantable Cardioverter-Defibrillator Types.

BACKGROUND: Late gadolinium enhancement (LGE) imaging in patients with implantable cardioverter-defibrillators (ICD) is limited by device-related artifacts (DRA). The use of wideband (WB) LGE protocols improves LGE images, but their efficacy with different ICD types is not well known. PURPOSE: To assess the effects of WB LGE imaging on DRA in different non-MR conditional ICD subtypes. STUDY TYPE: Retrospective. POPULATION: A total of 113 patients undergoing cardiac magnetic resonance imaging with three ICD subtypes: transvenous (TV-ICD, N = 48), cardiac-resynchronization therapy device (CRT-D, N = 48), and subcutaneous (S-ICD, N = 17). FIELD STRENGTH/SEQUENCE: 5 T scanner, standard LGE, and WB LGE imaging with a phase-sensitive inversion recovery segmented gradient echo sequence. ASSESSMENT: DRA burden was defined as the number of artifact-positive short-axis LGE slices as percentage of the total number of short-axis slices covering the left ventricle from based to apex, and was determined for WB and standard LGE studies for each patient. Additionally, artifact area on each slice was quantified. STATISTICAL TESTS: Shapiro-Wilks, Kruskal-Wallis analysis of variance, Dunn tests with Bonferroni correction, and Mann-Whitney U-test. RESULTS: In patients with TV-ICD, DRA burden was significantly reduced and nearly eliminated with WB LGE compared to standard LGE imaging (median [interquartile range]: 0 [0-7]% vs. 18 [0-50]%, P < 0.05), but WB imaging had less of an impact on DRA in the CRT-D (8 [0-23]% vs. 16 [0-45]%, p = 0.12) and S-ICD (60 [15-71]% vs. 67 [50-92]%, P = 0.09) patients. Residual DRA was significantly greater (P < 0.05) for S-ICD compared to other device types with WB LGE imaging, despite the generators of all three ICD types having similar proximity to the heart. The area of S-ICD associated DRA was smaller with WB LGE (P < 0.001) than with standard LGE imaging and the artifacts had different characteristics (dark signal void instead of a bright hyperenhancement artifact). DATA CONCLUSION: Although WB LGE imaging reduced the burden of DRA caused by S-ICD, the residual artifact was greater than that observed with TV-ICD and CRT-D devices. Further developments are needed to better resolve S-ICD artifacts. LEVEL OF EVIDENCE: 1 TECHNICAL EFFICACY: STAGE: 5.

Myocardial strain analysis of the right ventricle: comparison of different cardiovascular magnetic resonance and echocardiographic techniques.

BACKGROUND: Right ventricular (RV) strain is a useful predictor of prognosis in various cardiovascular diseases, including those traditionally believed to impact only the left ventricle. We aimed to determine inter-modality and inter-technique agreement in RV longitudinal strain (LS) measurements between currently available cardiovascular magnetic resonance (CMR) and echocardiographic techniques, as well as their reproducibility and the impact of layer-specific strain measurements. METHODS: RV-LS was determined in 62 patients using 2D speckle tracking echocardiography (STE, Epsilon) and two CMR techniques: feature tracking (FT) and strain-encoding (SENC), and in 17 healthy subjects using FT and SENC only. Measurements included global and free-wall LS (GLS, FWLS). Inter-technique agreement was assessed using linear regression and Bland-Altman analysis. Reproducibility was quantified using intraclass correlation (ICC) and coefficients of variation (CoV). RESULTS: We found similar moderate agreement between both CMR techniques and STE in patients: r = 0.57-0.63 for SENC; r = 0.50-0.62 for FT. The correlation between SENC and STE was better for GLS (r = 0.63) than for FWLS (r = 0.57). Conversely, the correlation between FT and STE was higher for FWLS (r = 0.60-0.62) than GLS (r = 0.50-0.54). FT-midmyocardial strain correlated better with SENC and STE than FT-subendocardial strain. The agreement between SENC and FT was fair (r = 0.36-0.41, bias: - 6.4 to - 10.4%) in the entire study group. All techniques except FT showed excellent reproducibility (ICC: 0.62-0.96, CoV: 0.04-0.30). CONCLUSIONS: We found only moderate inter-modality agreement with STE in RV-LS for both FT and SENC and poor agreement when comparing between the CMR techniques. Different modalities and techniques should not be used interchangeably to determine and monitor RV strain.

Echocardiography and cardiovascular magnetic resonance based evaluation of myocardial strain and relationship with late gadolinium enhancement.

OBJECTIVES: We sought to: (1) determine the agreement in cardiovascular magnetic resonance (CMR) and speckle tracking echocardiography (STE) derived strain measurements, (2) compare their reproducibility, (3) determine which approach is best related to CMR late gadolinium enhancement (LGE). BACKGROUND: While STE-derived strain is routinely used to assess left ventricular (LV) function, CMR strain measurements are not yet standardized. Strain can be measured using dedicated pulse sequences (strain-encoding, SENC), or post-processing of cine images (feature tracking, FT). It is unclear whether these measurements are interchangeable, and whether strain can be used as an alternative to LGE. METHODS: Fifty patients underwent 2D echocardiography and 1.5 T CMR. Global longitudinal strain (GLS) was measured by STE (Epsilon), FT (NeoSoft) and SENC (Myocardial Solutions) and circumferential strain (GCS) by FT and SENC. RESULTS: GLS showed good inter-modality agreement (r-values: 0.71-0.75), small biases (< 1%) but considerable limits of agreement (- 7 to 8%). The agreement between the CMR techniques was better for GLS than GCS (r = 0.81 vs 0.67; smaller bias). Repeated measurements showed low intra- and inter-observer variability for both GLS and GCS (intraclass correlations 0.86-0.99; coefficients of variation 3-13%). LGE was present in 22 (44%) of patients. Both SENC- and FT-derived GLS and GCS were associated with LGE, while STE-GLS was not. Irrespective of CMR technique, this association was stronger for GCS (AUC 0.77-0.78) than GLS (AUC 0.67-0.72) and STE-GLS (AUC = 0.58). CONCLUSION: There is good inter-technique agreement in strain measurements, which were highly reproducible, irrespective of modality or analysis technique. GCS may better reflect the presence of underlying LGE than GLS.

A histopathologic schema to quantify the burden of cardiac amyloidosis: Relationship with survival and echocardiographic parameters.

BACKGROUND: Despite routine use of echocardiographic parameters to evaluate the severity of cardiac amyloidosis (CA), this methodology has not been well validated. We developed a histopathologic schema for quantifying CA burden and evaluated its relationship with clinical outcomes. Additionally, echocardiographic parameters were tested as potential noninvasive indices of CA burden. METHODS: We retrospectively studied 59 patients with CA (17 light chain, 42 transthyretin) who underwent endomyocardial biopsies. Light microscopy with staining was used to categorize CA burden as mild-to-moderate (<50%) or high (≥50%). Kaplan-Meier survival analysis was performed for the two groups. In 34 patients with good-quality echocardiograms, we measured left ventricular volumes, ejection fraction (EF), interventricular septal thickness (IVSt), posterior wall thickness (PWt), LV mass, lateral e'-velocity, and global longitudinal strain (GLS). These parameters were compared between the two groups. RESULTS: Thirty-five patients had mild-to-moderate and 24 severe amyloid burden. Kaplan-Meier curves demonstrated a trend toward worse mortality with high CA burden, which was more common and associated with higher mortality specifically in transthyretin-type patients. Echocardiography-derived IVSt, PWt, and LV mass were directly related to CA burden, while LV EF, e'-velocity, and GLS magnitude were inversely related to CA burden. CONCLUSIONS: Our findings provided a signal that CA burden is a clinically important entity with potentially valuable prognostic information. Echocardiographic parameters of LV anatomy and function correlate with histopathologic burden of CA, which is inversely related to survival. Further studies are needed to determine whether these parameters could be used as imaging biomarkers of treatment-related changes in CA burden.

Automated, machine learning-based, 3D echocardiographic quantification of left ventricular mass.

BACKGROUND: Although 3D echocardiography (3DE) circumvents many limitations of 2D echocardiography by allowing direct measurements of left ventricular (LV) mass, it is seldom used in clinical practice due to time-consuming analysis. A recently developed 3DE machine learning (ML) approach allows automated determination of LV mass. We aimed to evaluate the accuracy of this new approach by comparing it to cardiac magnetic resonance (CMR) reference and to conventional 3DE volumetric analysis. METHODS: We prospectively studied 23 patients who underwent 3DE (Philips EPIQ) and CMR imaging on the same day. Single-beat wide-angle 3D datasets of the left ventricle were acquired. LV mass was quantified using the new automated software (Philips HeartModel) with manual corrections when necessary and using conventional volumetric analysis (TomTec). CMR analysis was performed by manual slice-by-slice tracing of LV endo- and epicardial boundaries. Reproducibility of the ML approach was assessed using repeated measurements and quantified by intra-class correlation (ICC) and coefficients of variation (CoV). RESULTS: Automated LV mass measurements were feasible in 20 patients (87%). The results were similar to CMR-derived values (Bland-Altman bias 5 g, limits of agreement ±37 g) and also to the conventional 3DE analysis (bias 7 g, ±27 g). Processing time was considerably shorter: 1.02 ± 0.24 minutes (CMR: 2.20 ± 0.13 minutes; TomTec: 2.36 ± 0.09 minutes), although manual corrections were performed in most patients. Repeated measurements showed high reproducibility: ICC = 0.99; CoV = 4 ± 5%. CONCLUSIONS: 3D Echocardiography analysis of LV mass using novel ML-based algorithm is feasible, fast, and accurate and may thus facilitate the incorporation of 3DE measurements of LV mass into clinical practice.

Residual native left ventricular function optimization using quantitative 3D echocardiographic assessment of rotational mechanics in patients with left ventricular assist devices.

Preservation of native left ventricular (LV) function in patients supported with LV assist device (LVAD) may be beneficial to attain optimal hemodynamics and enhance potential recovery. Currently, LVAD speed optimization is based on hemodynamic parameters, without considering residual native LV function. We hypothesized that alternatively, LV rotational mechanics can be quantified by 3D echocardiography (3DE), and may help preserve native LV function while optimizing LVAD speed. The goal of this study was to test the feasibility of quantifying the effects of LVAD implantation on LV rotational mechanics and to determine whether conventional speed optimization maximally preserves native LV function. We studied 55 patients with LVADs, who underwent 3DE imaging and quantitative analysis of LV twist. Thirty patients were studied before and after LVAD implantation. The remaining 25 patients were studied during hemodynamic ramp studies. The pump speed at which LV twist was maximal was compared with the hemodynamics-based optimal speed. LV twist decreased following LVAD implantation from 4.2 ± 2.7 to 2.3 ± 1.9° (P < 0.01), reflecting the constricting effects on native function. With lower pump speeds, no significant changes were noted in LV twist, which peaked at a higher speed. In 11/25 (44%) patients, the conventional hemodynamic/2DE methodology and 3DE assessment of maximal residual function did not indicate the same optimal conditions, suggesting that a higher pump speed would have better preserved native function. In conclusion, quantitative 3DE analysis of LV rotational mechanics provides information, which together with hemodynamics may help select optimal pump speed, while maximally preserving native LV function.

Reproducibility and experience dependence of echocardiographic indices of left ventricular function: Side-by-side comparison of global longitudinal strain and ejection fraction.

BACKGROUND: Although left ventricular (LV) ejection fraction (EF) and global longitudinal strain (GLS) are recommended by the current echocardiographic chamber quantification guidelines, these measurements are not performed routinely. Because EF measurements rely on manual tracing of LV boundaries, and are subject to inter-reader variability and experience dependence, we hypothesized that semiautomated GLS measurements using speckle tracking would be more reproducible and less experience-dependent. METHODS: Images from 30 patients were analyzed to obtain biplane EF using manual tracing. GLS was measured in three long-axis views using EchoInsight software (Epsilon Imaging) that automatically detects LV endocardial boundary, which is edited manually as necessary and is then automatically tracked throughout the cardiac cycle. All measurements were performed by an expert echocardiographer and three first-year cardiology fellows. RESULTS: Semiautomated GLS analysis showed excellent correlation (r=.98) and small bias (-1.0±13% of measured value) between the experienced and less experienced readers, superior to EF (r=.91, bias 7.3±16%). Also, in repeated measurements, GLS showed higher intra-class correlation (ICC=.98) than EF (ICC=.89). Additionally, GLS analysis required ~1 minute per patient, while biplane EF measurements took twice as long. CONCLUSIONS: Semiautomated GLS measurements are fast, less experience-dependent, and more reproducible than conventional EF measurements. This is probably because, irrespective of experience, the readers' choice of boundary position varies less when asked to refine the automated detection than to draw borders without initial clues. This technique may facilitate the workflow of a busy laboratory and make a step forward toward incorporating quantitative analysis into everyday echocardiography practice.

3D echocardiographic analysis of aortic annulus for transcatheter aortic valve replacement using novel aortic valve quantification software: Comparison with computed tomography.

BACKGROUND: With the increasing use of transcatheter aortic valve replacement (TAVR) in patients with aortic stenosis (AS), computed tomography (CT) remains the standard for annulus sizing. However, 3D transesophageal echocardiography (TEE) has been an alternative in patients with contraindications to CT. We sought to (1) test the feasibility, accuracy, and reproducibility of prototype 3DTEE analysis software (Philips) for aortic annular measurements and (2) compare the new approach to the existing echocardiographic techniques. METHODS: We prospectively studied 52 patients who underwent gated contrast CT, procedural 3DTEE, and TAVR. 3DTEE images were analyzed using novel semi-automated software designed for 3D measurements of the aortic root, which uses multiplanar reconstruction, similar to CT analysis. Aortic annulus measurements included area, perimeter, and diameter calculations from these measurements. The results were compared to CT-derived values. Additionally, 3D echocardiographic measurements (3D planimetry and mitral valve analysis software adapted for the aortic valve) were also compared to the CT reference values. RESULTS: 3DTEE image quality was sufficient in 90% of patients for aortic annulus measurements using the new software, which were in good agreement with CT (r-values: .89-.91) and small (<4%) inter-modality nonsignificant biases. Repeated measurements showed <10% measurements variability. The new 3D analysis was the more accurate and reproducible of the existing echocardiographic techniques. CONCLUSIONS: Novel semi-automated 3DTEE analysis software can accurately measure aortic annulus in patients with severe AS undergoing TAVR, in better agreement with CT than the existing methodology. Accordingly, intra-procedural TEE could potentially replace CT in patients where CT carries significant risk.

Screening for Outflow Cannula Malfunction of Left Ventricular Assist Devices (LVADs) With the Use of Doppler Echocardiography: New LVAD-Specific Reference Values for Contemporary Devices.

BACKGROUND: Echocardiographic assessment of left ventricular assist devices (LVADs) is used as a screening tool to evaluate the integrity and mechanics of the pump and circuit. We aimed to 1) establish the normal range and upper reference limit of peak velocity of the outflow cannula for the modern era of LVADs and 2) assess the clinical performance of the currently cited and newly proposed reference limits in patients with continuous-flow LVADs as a screening tool for cannula malfunction. METHODS: LVAD outflow peak CW velocities were measured with the use of Doppler transthoracic echocardiography (TTE) in 57 patients with LVADs (44 with Heartmate II (HM2), 13 with Heartware (HW)). The average velocity and the upper and lower normal reference limits (defined as ±2 standard deviations from the mean) for each LVAD type was calculated. The upper reference limit was then used as a screening threshold for cannula malfunction. RESULTS: The average outflow cannula peak velocity for the normal HM2 cohort was 1.86 ± 0.44 m/s with upper and lower reference limits of 2.73 m/s and 0.98 m/s, respectively. The average outflow cannula peak velocity for the normal HW cohort was 2.36 ± 0.53 m/s with upper and lower reference limits of 3.42 m/s and 1.3 m/s, respectively, which was significantly higher than the HM2 cohort (P = .004). CONCLUSIONS: In both HM2 and HW LVADs, the average peak outflow velocity and reference limit for the normal population, as measured by Doppler TTE, was markedly higher than the currently used LVAD reference limits of 2 m/s and are significantly different between devices. Patients with peak outflow velocities above our upper reference limits should be evaluated for LVAD outflow cannula malfunction.

Echocardiographic Diagnosis of Acute Pulmonary Embolism in Patients with McConnell's Sign.

BACKGROUND: "McConnell's sign" (McCS), described as hypo- or akinesis of the right ventricular (RV) free wall with preservation of the apex, is associated with acute pulmonary embolism (aPE). However, the sensitivity of McCS for the detection of aPE is limited. We sought to evaluate in patients with McCS, whether echocardiographic parameters of global and regional RV function could differentiate between patients with and without aPE. METHODS: We reviewed echocardiograms of 81 patients with McCS, who underwent CT or V/Q studies for suspected PE, and 40 normal controls (NL). Echocardiograms were analyzed to measure pulmonary artery systolic pressure (PASP), tricuspid regurgitation (TR) by vena contracta width, conventional indices of RV function, and speckle tracking-derived longitudinal free wall strain. ROC analysis was performed to evaluate the diagnostic accuracy of these parameters for diagnosis of aPE. RESULTS: Fifty-five of eighty-one (68%) had PE (McCS + PE), while 26 of 81 (32%) did not (McCS - PE). Compared to NL, global and segmental RV strain were lower in patients with McCS, contrary to the notion of normal apical function. In McCS + PE, compared to McCS - PE: (1) PASP, fractional area change and TR were significantly lower; (2) strain magnitude was significantly lower globally and in basal and apical segments. Individual parameters had similar diagnostic accuracy by ROC analysis, which further improved by combining parameters. In McCS - PE, 69% of patients had pulmonary hypertension (PH). CONCLUSIONS: McCS and aPE are not synonymous. RV free wall strain may aid in differential diagnosis of patients with McCS evaluated for aPE. Specifically, McCS should prompt an inquiry for evidence of PH, which would indicate that aPE is less likely.

Right Heart Involvement in Patients with Sarcoidosis.

The left ventricle (LV) is affected in 20-25% of patients with sarcoidosis and its involvement is associated with morbidity and mortality. However, effects of sarcoidosis on the right ventricle (RV) are not well documented. Our aims were to investigate the prevalence of RV dysfunction in patients with sarcoidosis and determine whether it is predominantly associated with direct cardiac involvement, severity of lung disease, or pulmonary hypertension (PH). We identified 50 patients with biopsy-proven extra-cardiac sarcoidosis and preserved LV function, who underwent echocardiography, pulmonary function (PF) testing, and cardiovascular magnetic resonance. RV function was quantified by free wall longitudinal strain. Tricuspid valve Doppler and estimated right atrial pressure were used to estimate systolic pulmonary artery pressure. Myocardial late gadolinium enhancement was considered diagnostic for cardiac sarcoidosis and assumed to involve both ventricles. Of the 50 patients, 28 (56%) had RV dysfunction, 4 with poorly defined PF status. Of the remaining 24 patients, 16 (67%) had lung disease, 8 (33%) had PH, and 10 (42%) had LV involvement. Ten patients had greater than one of these findings, and 4 had all 3. In contrast, in 4/24 patients (17%), RV dysfunction could not be explained by these mechanisms, despite severely reduced RV strain. In conclusion, RV dysfunction is common in patients with sarcoidosis and is usually associated with either direct LV involvement, lung disease, or PH, but may occur in the absence of these mechanisms, suggesting the possibility of isolated RV involvement and underscoring the need for imaging protocols that would include RV strain analysis.

Improved detection of myocardial damage in sarcoidosis using longitudinal strain in patients with preserved left ventricular ejection fraction.

BACKGROUND: Cardiac infiltration is an important cause of death in sarcoidosis. Transthoracic echocardiography (TTE) has limited sensitivity for the detection of cardiac sarcoidosis (CS). Late gadolinium enhancement (LGE) cardiovascular magnetic resonance (CMR) is used to diagnose CS but has limitations of cost and availability. We sought to determine whether TTE-derived global longitudinal strain (GLS) may be used to identify individuals with CS, despite preserved left ventricular ejection fraction (LVEF), and whether abnormal GLS is associated with major cardiovascular events (MCE). METHODS: We studied 31 patients with biopsy-proven extra-cardiac sarcoidosis, LVEF>50% and LGE on CMR (CS+ group), and 31 patients without LGE (CS- group), matched by age, sex, and severity of lung disease. GLS was measured using vendor-independent speckle tracking software. Parameters of left and right ventricular systolic and diastolic function were also studied. Receiver-operating characteristic curves were used to identify GLS cutoff for CS detection, and Kaplan-Meier plots to determine the ability of GLS to predict MCE. RESULTS: LGE was associated with reduced GLS (-19.6±1.9% in CS- vs -14.7±2.4% in CS+, P<.01) and with reduced E/A ratio (1.1±0.3 vs 0.9±0.3, respectively, P =.01). No differences were noted in other TTE parameters. GLS magnitude inversely correlated with LGE burden (r=-.59). GLS cutoff of -17% showed sensitivity and specificity 94% for detecting CS. Patients who experienced MCE had worse GLS than those who did not (-13.4±0.9% vs -17.7±0.4%, P=.0003). CONCLUSIONS: CS is associated with significantly reduced GLS in the presence of preserved LVEF. GLS measurements may become part of the TTE study performed to screen for CS.

Right ventricular strain in pulmonary arterial hypertension: a 2D echocardiography and cardiac magnetic resonance study.

BACKGROUND: Right ventricular (RV) strain is a potentially useful prognostic marker in patients with pulmonary arterial hypertension (PAH). However, published reports regarding the accuracy of two-dimensional echocardiography (2DE)-derived RV strain against an independent reference in this patient population are limited. The aims of this study were: (1) to study the relationship between 2DE RV longitudinal strain and cardiovascular magnetic resonance (CMR)-derived RV ejection fraction (RVEF) in patients with PAH; (2) to compare 2DE-derived and CMR-derived RV longitudinal strain in these patients; and (3) to determine the reproducibility of these measurements. METHODS: Thirty patients with PAH underwent 2DE and CMR imaging within a 2-hour time period. 2DE RV longitudinal strain was measured from a focused RV apical four-chamber view using speckle tracking software. CMR RV longitudinal strain was measured from short-axis slices acquired using fast-strain-encoded sequence. Global peak systolic RV longitudinal strain was calculated for both 2DE and CMR. RESULTS: RV longitudinal strain using 2DE software correlated well with CMR-derived RVEF (R = 0.69, P = 0.0006). There was moderate agreement when comparing 2DE to CMR RV longitudinal strain (R = 0.74, P = 0.0002; bias -1%, limits of agreement -9 to 7%). Inter-observer variability and intra-observer variability for RV longitudinal strain were lower for 2DE than CMR. CONCLUSIONS: RV longitudinal strain by 2DE provides a good alternative for CMR-derived RVEF in patients with PAH. The moderate agreement in strain measurements between 2DE and CMR suggests that further software improvements are needed before these measurements can be used interchangeably in clinical practice.

Echocardiographic Detection of Regional Wall Motion Abnormalities Using Artificial Intelligence Compared to Human Readers.

BACKGROUND: Although regional wall motion abnormality (RWMA) detection is foundational to transthoracic echocardiography, current methods are prone to interobserver variability. We aimed to develop a deep learning (DL) model for RWMA assessment and compare it to expert and novice readers. METHODS: We used 15,746 transthoracic echocardiography studies-including 25,529 apical videos-which were split into training, validation, and test datasets. A convolutional neural network was trained and validated using apical 2-, 3-, and 4-chamber videos to predict the presence of RWMA in 7 regions defined by coronary perfusion territories, using the ground truth derived from clinical transthoracic echocardiography reports. Within the test cohort, DL model accuracy was compared to 6 expert and 3 novice readers using F1 score evaluation, with the ground truth of RWMA defined by expert readers. Significance between the DL model and novices was assessed using the permutation test. RESULTS: Within the test cohort, the DL model accurately identified any RWMA with an area under the curve of 0.96 (0.92-0.98). The mean F1 scores of the experts and the DL model were numerically similar for 6 of 7 regions: anterior (86 vs 84), anterolateral (80 vs 74), inferolateral (83 vs 87), inferoseptal (86 vs 86), apical (88 vs 87), inferior (79 vs 81), and any RWMA (90 vs 94), respectively, while in the anteroseptal region, the F1 score of the DL model was lower than the experts (75 vs 89). Using F1 scores, the DL model outperformed both novices 1 (P = .002) and 2 (P = .02) for the detection of any RWMA. CONCLUSIONS: Deep learning provides accurate detection of RWMA, which was comparable to experts and outperformed a majority of novices. Deep learning may improve the efficiency of RWMA assessment and serve as a teaching tool for novices.

Limitations of apical sparing pattern in cardiac amyloidosis: a multicentre echocardiographic study.

AIMS: Although impaired left ventricular (LV) global longitudinal strain (GLS) with apical sparing is a feature of cardiac amyloidosis (CA), its diagnostic accuracy has varied across studies. We aimed to determine the ability of apical sparing ratio (ASR) and most common echocardiographic parameters to differentiate patients with confirmed CA from those with clinical and/or echocardiographic suspicion of CA but with this diagnosis ruled out. METHODS AND RESULTS: We identified 544 patients with confirmed CA and 200 controls (CTRLs) as defined above (CTRL patients). Measurements from transthoracic echocardiograms were performed using artificial intelligence software (Us2.AI, Singapore) and audited by an experienced echocardiographer. Receiver operating characteristic curve analysis was used to evaluate the diagnostic performance and optimal cut-offs for the differentiation of CA patients from CTRL patients. Additionally, a group of 174 healthy subjects (healthy CTRL) was included to provide insight on how patients and healthy CTRLs differed echocardiographically. LV GLS was more impaired (-13.9 ± 4.6% vs. -15.9 ± 2.7%, P < 0.0005), and ASR was higher (2.4 ± 1.2 vs. 1.7 ± 0.9, P < 0.0005) in the CA group vs. CTRL patients. Relative wall thickness and ASR were the most accurate parameters for differentiating CA from CTRL patients [area under the curve (AUC): 0.77 and 0.74, respectively]. However, even with the optimal cut-off of 1.67, ASR was only 72% sensitive and 66% specific for CA, indicating the presence of apical sparing in 32% of CTRL patients and even in 6% healthy subjects. CONCLUSION: Apical sparing did not prove to be a CA-specific biomarker for accurate identification of CA, when compared with clinically similar CTRLs with no CA.

Prognostic value of normal regadenoson stress perfusion cardiovascular magnetic resonance.

BACKGROUND: Regadenoson is a vasodilator stress agent that selectively activates the A2A receptor. Compared to adenosine, regadenoson is easier to administer and results in fewer side effects. Although extensively studied in patients undergoing nuclear perfusion imaging (MPI), its use for perfusion cardiovascular magnetic resonance (CMR) is not well described. The aim of this study was to determine the prognostic value of a normal regadenoson perfusion CMR in patients with known or suspected coronary artery disease. METHODS: Patients with known or suspected coronary artery disease were prospectively enrolled to receive perfusion CMR (Philips 1.5 T) with regadenoson. Three short-axis slices of the left ventricle (LV) were obtained during first pass of contrast using a hybrid GRE-EPI pulse sequence (0.075 mmol/kg Gadolinium-DTPA-BMA at 4 ml/sec). Imaging was performed 1 minute after injection of regadenoson (0.4 mg) and repeated 15 minutes after reversal of hyperemia with aminophylline (125 mg). Perfusion defects were documented if they persisted for ≥ 2 frames after peak enhancement of the LV cavity. CMR was considered abnormal if there was a resting wall motion abnormality, decreased LVEF (<40%), presence of LGE, or the presence of a perfusion defect during hyperemia. All patients were followed for a minimum of 1 year for major adverse cardiovascular event (MACE) defined as coronary revascularization, non-fatal myocardial infarction, and cardiovascular death. RESULTS: 149 patients were included in the final analysis. Perfusion defects were noted in 43/149 (29%) patients; 59/149 (40%) had any abnormality on CMR. During the mean follow-up period of 24 ± 9 months, 17/149 (11.4%) patients experienced MACE. The separation in the survival distributions for those with perfusion defects and those without perfusion defects was highly significant (log-rank p = 0.0001). When the absence of perfusion defects was added to the absence of other resting CMR abnormalities, the negative predictive value improved from 96% to 99%. CONCLUSION: Regadenoson perfusion CMR provides high confidence for excellent prognosis in patients with normal perfusion.

How Often Does Apical Sparing of Longitudinal Strain Indicate the Presence of Cardiac Amyloidosis?

Echocardiographic diagnosis of cardiac amyloidosis (CA) is frequently suggested by the presence of a left ventricular (LV) apical sparing pattern (ASP) on longitudinal strain (LS) assessment, the so-called "cherry on top" pattern, defined by strain magnitude preserved exclusively at the apex. However, it is unclear how frequently this strain pattern truly represents CA. This study aimed to evaluate the predictive value of ASP in the diagnosis of CA. We retrospectively identified consecutive adult patients who had the following studies performed within an 18-month period: (1) transthoracic echocardiogram and (2) either (a) cardiac magnetic resonance imaging, (b) Technetium-Pyrophosphate (PYP) imaging, or (c) endomyocardial biopsy. LS was retrospectively measured in the apical 4-, 3-, and 2-chamber views in patients who had adequate noncontrast images (n = 466). An apical sparing ratio (ASR) was calculated as (average apical strain)/[(average basal strain) + (average midventricular strain)]. Patients with ASR ≥1 were evaluated for the presence/absence of CA, using established criteria. Basic LV parameters were also measured. A total of 33 patients (7.1%) had ASP. Nine of these patients (27%) had "confirmed" CA, 2 (6.1%) "highly probable" CA, 1 (3.0%) "possible" CA, and 21 (64%) no evidence of CA. When comparing patients with and without confirmed CA, there were no significant differences in ASR, average global LS, ejection fraction, or LV mass. Patients with confirmed CA were older (76 ± 9 vs 59 ± 18 years, p = 0.01) and had thicker posterior wall (15 ± 3 vs 11 ± 3 mm, p = 0.004) with a trend toward thicker septal wall (15 ± 2 vs 12 ± 4 mm, p = 0.05). In conclusion, the presence of ASP on LS represents confirmed or highly probable CA in only 1/3 of patients and is more likely to indicate true CA in older patients with increased LV wall thickness. Although a larger, prospective study is needed to confirm these findings, 1/3 should be considered as a large diagnostic yield that justifies further testing, given the poor outcomes associated with CA diagnosis.

Impact of wideband cardiac magnetic resonance on diagnosis, decision-making and outcomes in patients with implantable cardioverter defibrillators.

AIMS: Although myocardial scar assessment using late gadolinium enhancement (LGE) cardiac magnetic resonance (CMR) imaging is frequently indicated for patients with implantable cardioverter defibrillators (ICDs), metal artefact can degrade image quality. With the new wideband technique designed to mitigate device related artefact, CMR is increasingly used in this population. However, the common clinical indications for CMR referral and impact on clinical decision-making and prognosis are not well defined. Our study was designed to address these knowledge gaps. METHODS AND RESULTS: One hundred seventy-nine consecutive patients with an ICD (age 59 ± 13 years, 75% male) underwent CMR using cine and wideband pulse sequences for LGE imaging. Electronic medical records were reviewed to determine the reason for CMR referral, whether there was a change in clinical decision-making, and occurrence of major adverse cardiac events (MACEs). Referral indication was the most common evaluation of ventricular tachycardia (VT) substrate (n = 114, 64%), followed by cardiomyopathy (n = 53, 30%). Overall, CMR resulted in a new or changed diagnosis in 64 (36%) patients and impacted clinical management in 51 (28%). The effect on management change was highest in patients presenting with VT. A total of 77 patients (43%) experienced MACE during the follow-up period (median 1.7 years), including 65 in patients with evidence of LGE. Kaplan-Meier analysis showed that ICD patients with LGE had worse outcomes than those without LGE (P = 0.006). CONCLUSION: The clinical yield from LGE CMR is high and provides management changing and meaningful prognostic information in a significant proportion of patients with ICDs.

Noninvasive assessment of left ventricular end-diastolic pressure using machine learning-derived phasic left atrial strain.

AIMS: While transthoracic echocardiography (TTE) assessment of left ventricular end-diastolic pressure (LVEDP) is critically important, the current paradigm is subject to error and indeterminate classification. Recently, peak left atrial strain (LAS) was found to be associated with LVEDP. We aimed to test the hypothesis that integration of the entire LAS time curve into a single parameter could improve the accuracy of peak LAS in the noninvasive assessment of LVEDP with TTE. METHODS AND RESULTS: We retrospectively identified 294 patients who underwent left heart catheterization and TTE within 24 h. LAS curves were trained using machine learning (100 patients) to detect LVEDP ≥ 15 mmHg, yielding the novel parameter LAS index (LASi). The accuracy of LASi was subsequently validated (194 patients), side by side with peak LAS and ASE/EACVI guidelines, against invasive filling pressures. Within the validation cohort, invasive LVEDP was elevated in 116 (59.8%) patients. The overall accuracy of LASi, peak LAS, and American Society of Echocardiography/European Association for Cardiovascular Imaging (ASE/EACVI) algorithm was 79, 75, and 76%, respectively (excluding 37 patients with indeterminate diastolic function by ASE/EACVI guidelines). When the number of LASi indeterminates (defined by near-zero LASi values) was matched to the ASE/EACVI guidelines (n = 37), the accuracy of LASi improved to 87%. Importantly, among the 37 patients with ASE/EACVI-indeterminate diastolic function, LASi had an accuracy of 81%, compared with 76% for peak LAS. CONCLUSION: LASi allows the detection of elevated LVEDP using invasive measurements as a reference, at least as accurately as peak LAS and current diastolic function guideline algorithm, with the advantage of no indeterminate classifications in patients with measurable LAS.