Feasibility, repeatability, and reproducibility of contemporary diastolic parameters and classification

Diastolic function assessment with four-dimensional flow cardiovascular magnetic resonance using automatic deep learning E/A ratio analysis

BACKGROUND

Diastolic left ventricular (LV) dysfunction is a powerful contributor to the symptoms and prognosis of patients with heart failure. In patients with depressed LV systolic function, the E/A ratio, the ratio between the peak early (E) and the peak late (A) transmitral flow velocity, is the first step to defining the grade of diastolic dysfunction. Doppler echocardiography (echo) is the preferred imaging technique for diastolic function assessment, while cardiovascular magnetic resonance (CMR) is less established as a method. Previous four-dimensional (4D) Flow-based studies have looked at the E/A ratio proximal to the mitral valve, requiring manual interaction. In this study, we compare an automated, deep learning-based and two semi-automated approaches for 4D Flow CMR-based E/A ratio assessment to conventional, gold-standard echo-based methods.

METHODS

Ninety-seven subjects with chronic ischemic heart disease underwent a cardiac echo followed by CMR investigation. 4D Flow-based E/A ratio values were computed using three different approaches; two semi-automated, assessing the E/A ratio by measuring the inflow velocity (MVvel) and the inflow volume (MVflow) at the mitral valve plane, and one fully automated, creating a full LV segmentation using a deep learning-based method with which the E/A ratio could be assessed without constraint to the mitral plane (LVvel).

RESULTS

MVvel, MVflow, and LVvel E/A ratios were strongly associated with echocardiographically derived E/A ratio (R2 = 0.60, 0.58, 0.72). LVvel peak E and A showed moderate association to Echo peak E and A, while MVvel values were weakly associated. MVvel and MVflow EA ratios were very strongly associated with LVvel (R2 = 0.84, 0.86). MVvel peak E was moderately associated with LVvel, while peak A showed a strong association (R2 = 0.26, 0.57).

CONCLUSION

Peak E, peak A, and E/A ratio are integral to the assessment of diastolic dysfunction and may expand the utility of CMR studies in patients with cardiovascular disease. While underestimation of absolute peak E and A velocities was noted, the E/A ratio measured with all three 4D Flow methods was strongly associated with the gold standard Doppler echocardiography. The automatic, deep learning-based method performed best, with the most favorable runtime of ∼40 seconds. As both semi-automatic methods associated very strongly to LVvel, they could be employed as an alternative for estimation of E/A ratio.

Reducing echocardiographic examination time through routine use of fully automated software: a comparative study of measurement and report creation time

BACKGROUND

Manual interpretation of echocardiographic data is time-consuming and operator-dependent. With the advent of artificial intelligence (AI), there is a growing interest in its potential to streamline echocardiographic interpretation and reduce variability. This study aimed to compare the time taken for measurements by AI to that by human experts after converting the acquired dynamic images into DICOM data.

METHODS

Twenty-three consecutive patients were examined by a single operator, with varying image quality and different medical conditions. Echocardiographic parameters were independently evaluated by human expert using the manual method and the fully automated US2.ai software. The automated processes facilitated by the US2.ai software encompass real-time processing of 2D and Doppler data, measurement of clinically important variables (such as LV function and geometry), automated parameter assessment, and report generation with findings and comments aligned with guidelines. We assessed the duration required for echocardiographic measurements and report creation.

RESULTS

The AI significantly reduced the measurement time compared to the manual method (159 ± 66 vs. 325 ± 94 s, p < 0.01). In the report creation step, AI was also significantly faster compared to the manual method (71 ± 39 vs. 429 ± 128 s, p < 0.01). The incorporation of AI into echocardiographic analysis led to a 70% reduction in measurement and report creation time compared to manual methods. In cases with fair or poor image quality, AI required more corrections and extended measurement time than in cases of good image quality. Report creation time was longer in cases with increased report complexity due to human confirmation of AI-generated findings.

CONCLUSIONS

This fully automated software has the potential to serve as an efficient tool for echocardiographic analysis, offering results that enhance clinical workflow by providing rapid, zero-click reports, thereby adding significant value.

Poor Agreement Between Preoperative Transthoracic Echocardiography and Intraoperative Transesophageal Echocardiography for Grading Diastolic Dysfunction

BACKGROUND

Guidelines for the evaluation and grading of diastolic dysfunction are available for transthoracic echocardiography (TTE). Transesophageal echocardiography (TEE) is used for this purpose intraoperatively but the level of agreement between these 2 imaging modalities for grading diastolic dysfunction is unknown. We assessed agreement between awake preoperative TTE and intraoperative TEE for grading diastolic dysfunction.

METHODS

In 98 patients undergoing cardiac surgery, key Doppler measurements were obtained using TTE and TEE at the following time points: TTE before anesthesia induction (TTEawake), TTE following anesthesia induction (TTEanesth), and TEE following anesthesia induction (TEEanesth). The primary endpoint was grade of diastolic dysfunction categorized by a simplified algorithm, and measured by TTEawake and TEEanesth, for which the weighted κ statistic assessed observed agreement beyond chance. Secondary endpoints were peak early diastolic lateral mitral annular tissue velocity (e'lat) and the ratio of peak early diastolic mitral inflow velocity (E) to e'lat (E/e'lat), measured by TTEawake and TEEanesth, were compared using Bland-Altman limits of agreement.

RESULTS

Disagreement in grading diastolic dysfunction by ≥1 grade occurred in 43 (54%) of 79 patients and by ≥2 grades in 8 (10%) patients with paired measurements for analysis, yielding a weighted κ of 0.35 (95% confidence interval [CI], 0.19-0.51) for the observed level of agreement beyond chance. Bland-Altman analysis of paired data for e'lat and E/e'lat demonstrated a mean difference (95% CI) of 0.51 (-0.06 to 1.09) and 0.70 (0.07-1.34), respectively, for measurements made by TTEawake compared to TEEanesth. The percentage (95% CI) of paired measurements for e'lat and E/e'lat that lay outside the [-2, +2] study-specified boundary of acceptable agreement was 36% (27%-48%) and 39% (29%-51%), respectively. Results were generally robust to sensitivity analyses, including comparing measurements between TTEawake and TTEanesth, between TTEanesth and TEEanesth, and after regrading diastolic dysfunction by the American Society of Echocardiography (ASE)/European Association of CardioVascular Imaging (EACVI) algorithm.

CONCLUSIONS

There was poor agreement between TTEawake and TEEanesth for grading diastolic dysfunction by a simplified algorithm, with disagreement by ≥1 grade in 54% and by ≥2 grades in 10% of the evaluable cohort. Future studies, including comparing the prognostic utility of TTEawake and TEEanesth for clinically important adverse outcomes that may be a consequence of diastolic dysfunction, are needed to understand whether this disagreement reflects random variability in Doppler variables, misclassification by the changed technique and physiological conditions of intraoperative TEE, or the accurate detection of a clinically relevant change in diastolic dysfunction.

Prognostic value of myocardial performance index in individuals with type 1 and type 2 diabetes: Thousand&1 and Thousand&2 studies

AIMS

Cardiovascular disease (CVD) is the leading cause of mortality and morbidity in type 1 (T1D) and type 2 diabetes (T2D). Despite diabetes affects the myocardium, risk prediction models do not include myocardial function parameters. Myocardial performance index (MPI) reflects left ventricular function. The prognostic value of MPI has not been evaluated in large-scale diabetes populations.

METHODS AND RESULTS

We evaluated two prospective cohort studies: Thousand&1 (1093 individuals with T1D) and Thousand&2 (1030 individuals with T2D). Clinical data, including echocardiography, were collected at baseline. We collected follow-up data from national registries. We defined major adverse cardiovascular events (MACE) as incident events of hospital admission for acute coronary syndrome, heart failure, stroke, or all-cause mortality. For included individuals (56% male, 54 ± 15 years, MPI 0.51 ± 0.1, 63% T1D), follow-up was 100% after median of 5.3 years (range: 4.8-6.3). MPI was associated with MACE (HR 1.2, 95%CI 1.0-1.3, P = 0.012, per 0.10-unit increase) and heart failure (HR 1.3, 95%CI 1.1-1.6, P = 0.005, per 0.10-unit increase) after adjusting for clinical and echocardiographic variables. MPI predicted MACE and heart failure better in T1D than T2D (P = 0.031 for interaction). MPI added discriminatory power to the Steno T1 Risk Engine, based on clinical characteristics, in predicting MACE [area under the curve (AUC) from 0.77 to 0.79, P = 0.030] and heart failure (AUC from 0.77 to 0.83, P = 0.009) in T1D.

CONCLUSION

MPI is independently associated with MACE and heart failure in T1D but not T2D and improves prediction in T1D. Echocardiographic assessment in T1D may enhance risk prediction.

Automated Echocardiographic Detection of Heart Failure With Preserved Ejection Fraction Using Artificial Intelligence

Background

Detection of heart failure with preserved ejection fraction (HFpEF) involves integration of multiple imaging and clinical features which are often discordant or indeterminate.

Objectives

The authors applied artificial intelligence (AI) to analyze a single apical 4-chamber transthoracic echocardiogram video clip to detect HFpEF.

Methods

A 3-dimensional convolutional neural network was developed and trained on apical 4-chamber video clips to classify patients with HFpEF (diagnosis of heart failure, ejection fraction ≥50%, and echocardiographic evidence of increased filling pressure; cases) vs without HFpEF (ejection fraction ≥50%, no diagnosis of heart failure, normal filling pressure; controls). Model outputs were classified as HFpEF, no HFpEF, or nondiagnostic (high uncertainty). Performance was assessed in an independent multisite data set and compared to previously validated clinical scores.

Results

Training and validation included 2,971 cases and 3,785 controls (validation holdout, 16.8% patients), and demonstrated excellent discrimination (area under receiver-operating characteristic curve: 0.97 [95% CI: 0.96-0.97] and 0.95 [95% CI: 0.93-0.96] in training and validation, respectively). In independent testing (646 cases, 638 controls), 94 (7.3%) were nondiagnostic; sensitivity (87.8%; 95% CI: 84.5%-90.9%) and specificity (81.9%; 95% CI: 78.2%-85.6%) were maintained in clinically relevant subgroups, with high repeatability and reproducibility. Of 701 and 776 indeterminate outputs from the Heart Failure Association-Pretest Assessment, Echocardiographic and Natriuretic Peptide Score, Functional Testing (HFA-PEFF), and Final Etiology and Heavy, Hypertensive, Atrial Fibrillation, Pulmonary Hypertension, Elder, and Filling Pressure (H2FPEF) scores, the AI HFpEF model correctly reclassified 73.5% and 73.6%, respectively. During follow-up (median: 2.3 [IQR: 0.5-5.6] years), 444 (34.6%) patients died; mortality was higher in patients classified as HFpEF by AI (HR: 1.9 [95% CI: 1.5-2.4]).

Conclusions

An AI HFpEF model based on a single, routinely acquired echocardiographic video demonstrated excellent discrimination of patients with vs without HFpEF, more often than clinical scores, and identified patients with higher mortality.

Isolated diastolic dysfunction is associated with increased mortality in critically ill patients

PURPOSE

Left ventricular (LV) diastolic dysfunction is important in critically ill patients, but prevalence and impact on mortality is not well studied. We classified intensive care patients with normal left ventricular function according to current diastolic guidelines and explored associations with mortality.

MATERIAL AND METHODS

Echocardiography was performed within 24 h of intensive care admission. Patients with reduced LV ejection fraction, regional wall motion abnormality, or a history of cardiac disease were excluded. Patients were classified according to the 2016 EACVI guidelines, Recommendations for the Evaluation of LV Diastolic Function by Echocardiography.

RESULTS

Out of 218 patients, 162 (74%) had normal diastolic function, 21 (10%) had diastolic dysfunction, and 35 (17%) had indeterminate diastolic function. Diastolic dysfunction were more common in female patients, older patients and associated with sepsis, respiratory and cardiovascular comorbidity as well as higher SAPS Score. In a risk-adjusted logistic regression model, patients with indeterminate diastolic dysfunction (OR 4.3 [1.6-11.4], p = 0.004) or diastolic dysfunction (OR 5.1 [1.6-16.5], p = 0.006) had an increased risk of death at 90 days compared to patients with normal diastolic function.

CONCLUSION

Isolated diastolic dysfunction, assessed by a multi-parameter approach, is common in critically ill patients and is associated with mortality.

TRIAL REGISTRATION

Secondary analysis of data from a single-center prospective observational study focused on systolic dysfunction in intensive care unit patients (Clinical Trials ID: NCT03787810.

A formal validation of a deep learning-based automated workflow for the interpretation of the echocardiogram

This study compares a deep learning interpretation of 23 echocardiographic parameters-including cardiac volumes, ejection fraction, and Doppler measurements-with three repeated measurements by core lab sonographers. The primary outcome metric, the individual equivalence coefficient (IEC), compares the disagreement between deep learning and human readers relative to the disagreement among human readers. The pre-determined non-inferiority criterion is 0.25 for the upper bound of the 95% confidence interval. Among 602 anonymised echocardiographic studies from 600 people (421 with heart failure, 179 controls, 69% women), the point estimates of IEC are all <0 and the upper bound of the 95% confidence intervals below 0.25, indicating that the disagreement between the deep learning and human measures is lower than the disagreement among three core lab readers. These results highlight the potential of deep learning algorithms to improve efficiency and reduce the costs of echocardiography.

Predictive Value of Monocyte Chemoattractant Protein-1 in the Development of Diastolic Dysfunction in Patients with Psoriatic Arthritis

We aimed to evaluate the diagnostic accuracy of the proinflammatory monocyte chemotactic protein-1 (MCP-1) in the diagnosis of asymptomatic diastolic dysfunction (DD) in patients with psoriatic arthritis (PsA). The disease activity in psoriatic arthritis (DAPSA) was determined using clinical and laboratory parameters, and echocardiography was performed to estimate DD. Serum MCP-1 concentrations were elevated in PsA patients with DD diagnosed with ultrasound (median (25th percentile, 75th percentile): 366.6 pg/mL (283, 407.1 pg/mL) vs. 277.5 pg/mL (223.5, 319.1 pg/mL) in controls; $P<0.0017$ ). PsA patients with serum MCP-1 concentration higher than the cut-off value of 347.6 pg/mL had a 7.74-fold higher chance of developing DD than PsA patients with lower serum MCP-1 concentrations (controls), with a specificity of 86.36% and sensitivity of 55%, as verified using ultrasound. The group with MCP-1 concentrations above the cut-off value also showed a higher late peak diastolic mitral inflow velocity, A-wave value ( $P=0.000005$ ), E/E ${}^{\prime }$ ratio ( $P=0.00005$ ), and a lower E/A ratio ( $P=0.000002$ ), peak systolic left atrial reservoir strain, SA value ( $P=0.0066$ ), early peak diastolic displacement of the mitral septal annulus, E ${}^{\prime }$ wave value ( $P=0.003$ ), than controls. Systolic blood pressure ( $P=0.01$ ), LDL cholesterol concentration ( $P=0.012$ ), glucose concentration ( $P=0.011$ ), and DAPSA ( $P=0.0000$ ) increased in the PsA group with higher MCP-1 concentrations, although there were no differences in comorbidities and therapy between the groups compared. Thus, the serum MCP-1 concentration was a significant and independent prognostic indicator for asymptomatic DD in PsA patients ( $\text{area}\text{under}\text{the}\text{curve}=0.730$ , $P=0.001$ ). The DAPSA score in PsA patients might indicate the need for echocardiography and adjustment of anti-inflammatory treatment in terms of DD prevention.