Machine Learning–Based Three-Dimensional Echocardiographic Quantification of Right Ventricular Size and Function: Validation Against Cardiac Magnetic Resonance

Machine learning, artificial intelligence and mechanical circulatory support: A primer for clinicians

Artificial intelligence (AI) refers to the ability of machines to perform intelligent tasks, and machine learning (ML) is a subset of AI describing the ability of machines to learn independently and make accurate predictions. The application of AI combined with "big data" from the electronic health records, is poised to impact how we take care of patients. In recent years, an expanding body of literature has been published using ML in cardiovascular health care, including mechanical circulatory support (MCS). This primer article provides an overview for clinicians on relevant concepts of ML and AI, reviews predictive modeling concepts in ML and provides contextual reference to how AI is being adapted in the field of MCS. Lastly, it explains how these methods could be incorporated in the practices of medicine to improve patient outcomes.

Prognostic Value of Right Ventricular Ejection Fraction Assessed by 3D Echocardiography in COVID-19 Patients

Background: RVEF (right ventricular ejection fraction) measured by three-dimensional echocardiography (3DE) has been used in evaluating right ventricular (RV) function and can provide useful prognostic information in other various cardiovascular diseases. However, the prognostic value of 3D-RVEF in coronavirus disease 2019 (COVID-19) remains unknown. We aimed to investigate whether 3D-RVEF can predict the mortality of COVID-19 patients.

Methods: A cohort of 128 COVID-19-confirmed patients who had undergone echocardiography were studied. Thirty-one healthy volunteers were also enrolled as controls. COVID-19 patients were divided into three subgroups (general, severe, and critical) according to COVID-19 severity-of-illness. Conventional RV structure and function parameters, RV free wall longitudinal strain (FWLS) and 3D-RVEF were acquired. RVFWLS was measured by two-dimensional speckle tracking echocardiography. RVEF was acquired by 3DE.

Results: Compared with controls, 2D-RVFWLS and 3D-RVEF were both significantly decreased in COVID-19 patients (−27.2 ± 4.4% vs. −22.9 ± 4.8%, P < 0.001; 53.7 ± 4.5% vs. 48.5 ± 5.8%, P < 0.001). Critical patients were more likely to have a higher incidence of acute cardiac injury and acute respiratory distress syndrome (ARDS), and worse prognosis than general and severe patients. The critical patients exhibited larger right-heart chambers, worse RV fractional area change (RVFAC), 2D-RVFWLS, and 3D-RVEF and higher proportion of pulmonary hypertension than general and severe patients. Eighteen patients died during a median follow-up of 91 days. The multivariate Cox regression analysis revealed the acute cardiac injury, ARDS, RVFAC, RVFWLS, and 3D-RVEF were independent predictors of death. 3D-RVEF (chi-square to improve 18.3; P < 0.001), RVFAC (chi-square to improve 4.5; P = 0.034) and 2D-RVFWLS (chi-square to improve 5.1; P = 0.024) all provided additional prognostic value of higher mortality over clinical risk factors. Moreover, the incremental predictive value of 3D-RVEF was significantly (P < 0.05) higher than RVFAC and RVFWLS.

Conclusion: 3D-RVEF was the most robust independent predictor of mortality in COVID-19 patients and provided a higher predictive value over conventional RV function parameters and RVFWLS, which may be helpful to identify COVID-19 patients at a higher risk of death.

A machine learning algorithm supports ultrasound-naïve novices in the acquisition of diagnostic echocardiography loops and provides accurate estimation of LVEF

Left ventricular ejection fraction (LVEF) is the most important parameter in the assessment of cardiac function. A machine-learning algorithm was trained to guide ultrasound-novices to acquire diagnostic echocardiography images. The artificial intelligence (AI) algorithm then estimates LVEF from the captured apical-4-chamber (AP4), apical-2-chamber (AP2), and parasternal-long-axis (PLAX) loops. We sought to test this algorithm by having first-year medical students without previous ultrasound knowledge scan real patients. Nineteen echo-naïve first-year medical students were trained in the basics of echocardiography by a 2.5 h online video tutorial. Each student then scanned three patients with the help of the AI. Image quality was graded according to the American College of Emergency Physicians scale. If rated as diagnostic quality, the AI calculated LVEF from the acquired loops (monoplane and also a "best-LVEF" considering all views acquired in the particular patient). These LVEF calculations were compared to images of the same patients captured and read by three experts (ground-truth LVEF [GT-EF]). The novices acquired diagnostic-quality images in 33/57 (58%), 49/57 (86%), and 39/57 (68%) patients in the PLAX, AP4, and AP2, respectively. At least one of the three views was obtained in 91% of the attempts. We found an excellent agreement between the machine's LVEF calculations from images acquired by the novices with the GT-EF (bias of 3.5% ± 5.6 and r = 0.92, p < 0.001 in the "best-LVEF" algorithm). This pilot study shows first evidence that a machine-learning algorithm can guide ultrasound-novices to acquire diagnostic echo loops and provide an automated LVEF calculation that is in agreement with a human expert.

Novelties in 3D Transthoracic Echocardiography

Cardiovascular imaging is developing at a rapid pace and the newer modalities, in particular three-dimensional echocardiography, allow better analysis of heart structures. Identifying valve lesions and grading their severity represents crucial information and nowadays is strengthened by the introduction of new software, such as transillumination, which provide detailed morphology descriptions. Chambers quantification has never been so rapid and accurate: machine learning algorithms generate automated volume measurements, including left ventricular systolic and diastolic function, which is extremely important for clinical decisions. This review provides an overview of the latest innovations in the echocardiography field, and is helpful by providing a better insight into heart diseases.

Perioperative Course of Three-Dimensional-Derived Right Ventricular Strain in Coronary Artery Bypass Surgery: A Prospective, Observational, Pilot Trial

OBJECTIVES

Few data exist on perioperative three-dimensional-derived right ventricular strain. The authors aimed to describe the perioperative course of three-dimensional-derived right ventricular strain in coronary artery bypass graft (CABG) surgery patients.

DESIGN

Prospective, observational, pilot trial.

SETTING

Single university hospital.

PARTICIPANTS

The study comprised 40 patients with preserved left ventricular and right ventricular (RV) function undergoing isolated on-pump CABG surgery.

INTERVENTIONS

Three-dimensional strain analysis and standard echocardiographic evaluation of RV function were performed preoperatively (T1) and postoperatively (T4) with transthoracic echocardiography (TTE) and intraoperatively before sternotomy (T2) and after sternotomy (T3) with transesophageal echocardiography (TEE). All echocardiographic measurements were performed under stable hemodynamic conditions and predefined fluid management without any vasoactive support.

MEASUREMENTS AND MAIN RESULTS

The measurements of three-dimensional-derived RV free-wall strain (3D-RV FWS) and RV ejection fraction were performed using TomTec 4D RV-Function 2.0 software. Philips QLAB 10.8 was used to analyze tissue velocity of the tricuspid annulus, tricuspid annular systolic excursion, and RV fractional area change. There were no significant differences (median [interquartile range {IQR}]) between preoperative TTE and intraoperative TEE measurements for 3D-RV FWS (T1 v T2: -22.35 [IQR -17.70 to -27.22] v -24.35 [IQR -20.63 to -29.88]; not significant). 3D-RV FWS remained unchanged after sternotomy (T2 v T3: -24.35 [IQR -20.63 to -29.88] v -23.75 [IQR -20.25 to -29.28]; not significant) but deteriorated significantly after CABG (T1 v T4: -22.35 [IQR -17.70 to -27.22] v -18.5 [IQR -16.90 to -21.65]; p = 0.004).

CONCLUSION

In patients undergoing on-pump CABG, 3D-RV FWS values for awake, spontaneously breathing patients measured with TTE and values assessed in patients under general anesthesia with TEE did not significantly differ. Three-dimensional RV FWS did not change after sternotomy but deteriorated after on-pump CABG.

Evaluation of left ventricular diastolic function of patients with coronary heart disease by ultrasound images on bilateral filtering image noise reduction algorithm combined with electrocardiogram

OBJECTIVE

To explore the evaluation of left ventricular diastolic function (LVDF) in patients with coronary heart disease (CHD) using ultrasound images (UI) combined with electrocardiogram (ECG) on bilateral filtering image noise reduction algorithm (BFINRA).

METHODS

A BFINRA was constructed, and 60 subjects who were investigated were divided into a control group (CG) from June 2019 to November 2019 in Taizhou People's Hospital, a myocardial infarction group (MIG), and an angina pectoris group (APG). The patient's LVDF was examined by two-dimensional electrophoresis (2DE) and real-time three-dimensional echocardiography (RT-3DE) combined with ECG. The results showed BFINRA could improve UI quality.

RESULTS

Clinical data indicated there were no substantial differences in age, gender, and fasting blood glucose of all subjects. 2DE examination results showed the left ventricular end-diastolic volume (LVEDV), left ventricular end-systolic volume (LVESV), and early diastolic mitral blood flow velocity / early diastolic mitral annulus velocity (E/E') of MIG were much higher than CG (P<0.05), while the left ventricular ejection fraction (LVEF), E / late diastolic mitral blood flow velocity (E/A) and E' peak value were sharply decreased (P<0.05);LVESV and E/E' of APG were increased dramatically (P<0.05), while E peak, E/A and E' peak were decreased greatly. RT-3DE examination results indicated LVEDV and LVESV of MIG were considerably higher than CG (P<0.05), while LVEF and macrophage resistance factor (MRF) were enormously decreased (P<0.05);LVEDV and LVESV of APG were greatly increased (P<0.05). However, LVEF and MRF were not changed significantly (P>0.05). LVEDV had a remarkable difference (P<0.05), but LVESV and LVEF had no obvious differences (P>0.05). The electrocardiogram results illustrated the increase in QT dispersion (QTd) of MIG and APG was statistically significant (P<0.05) compared with CG, while the negative increase of P-wave terminal force in lead V1 (PTFV1) also had a statistical significance (P<0.05). Correlation analysis revealed that MRF and PTFV1 had positive correlation, while MRF and QTd showed a negative correlation.

CONCLUSION

The combination of UI and ECG could better assess LVDF in CHD patients.

Artificial Intelligence Aids Cardiac Image Quality Assessment for Improving Precision in Strain Measurements

OBJECTIVES

The aim of this study was to develop an artificial intelligence tool to assess echocardiographic image quality objectively.

BACKGROUND

Left ventricular global longitudinal strain (LVGLS) has recently been used to monitor cancer therapeutics-related cardiac dysfunction (CTRCD) but image quality limits its reliability.

METHODS

A DenseNet-121 convolutional neural network was developed for view identification from an athlete's echocardiographic dataset. To prove the concept that classification confidence (CC) can serve as a quality marker, values of longitudinal strain derived from feature tracking of cardiac magnetic resonance (CMR) imaging and strain analysis of echocardiography were compared. The CC was then applied to patients with breast cancer free from CTRCD to investigate the effects of image quality on the reliability of strain analysis.

RESULTS

CC of the apical 4-chamber view (A4C) was significantly correlated with the endocardial border delineation index. CC of A4C >900 significantly predicted a <15% relative difference in longitudinal strain between CMR feature tracking and automated echocardiographic analysis. Echocardiographic studies (n =752) of 102 patients with breast cancer without CTRCD were investigated. The strain analysis showed higher parallel forms, inter-rater, and test-retest reliabilities in patients with CC of A4C >900. During sequential comparisons of automated LVGLS in individual patients, those with CC of A4C >900 had a lower false positive detection rate of CTRCD.

CONCLUSIONS

CC of A4C was associated with the reliability of automated LVGLS and could also potentially be used as a filter to select comparable images from sequential echocardiographic studies in individual patients and reduce the false positive detection rate of CTRCD.

Prognostic Value of the Right Ventricular Ejection Fraction, Assessed by Fully Automated Three-Dimensional Echocardiography: A Direct Comparison of Analyses Using Right Ventricular-Focused Views versus Apical Four-Chamber Views

BACKGROUND

Right ventricular (RV) three-dimensional echocardiographic (3DE) data sets are acquired from either the RV-focused view (RVFV) or the apical four-chamber view (4CV). The prognostic value of 3DE RV ejection fraction (RVEF) was investigated using fully automated RV quantification software, and how measurement values with 3DE data sets from the RVFV compare with those from the 4CV was determined.

METHODS

One hundred seventy-four patients who had undergone both cardiac magnetic resonance (CMR) and 3DE imaging were retrospectively selected. RV 3DE data sets were acquired from both the RVFV and the 4CV and were analyzed separately using fully automated RV quantification software. Primary end points were cardiac events, including cardiac death, heart failure requiring hospitalization, nonfatal myocardial infarction, and ventricular tachyarrhythmia.

RESULTS

The feasibility of RVEF measurements on 3DE imaging from the RVFV and 4CV was 92% and 92%, respectively. There was good correlation (r = 0.83) and small bias (0.3%) between RVEF from the RVFV and that from the 4CV. Similar results were obtained when only data from patients whose echocardiograms had poor image quality in one or both views were analyzed (r = 0.83, bias = 1.7%, n = 78). Although fully automated analysis in both the RVFV and 4CV significantly underestimated RV volumes compared with CMR, neither measurement differed significantly for RVEF compared with CMR. During a median follow-up period of 12.5 months, 21 patients experienced primary end points. RVEF assessed by CMR and 3DE imaging was significantly associated with cardiac events. RVEF using fully automated analysis had a significant association with cardiac events, even in patients with poor image quality (RVFV: hazard ratio, 0.90 [P = .009, n = 44]; 4CV: hazard ratio, 0.90 [P = .009, n = 68]).

CONCLUSIONS

RV 3DE data sets from the RVFV and 4CV yielded similar RVEF values using fully automated software. RVEFs from both approaches had significant association with outcomes. Thus, both provide accurate information regarding RV function and risk for adverse outcomes.

A comparison of artificial intelligence-based algorithms for the identification of patients with depressed right ventricular function from 2-dimentional echocardiography parameters and clinical features

Background

Recognizing low right ventricular (RV) function from 2-dimentiontial echocardiography (2D-ECHO) is challenging when parameters are contradictory. We aim to develop a model to predict low RV function integrating the various 2D-ECHO parameters in reference to cardiac magnetic resonance (CMR)-the gold standard.

Methods

We retrospectively identified patients who underwent a 2D-ECHO and a CMR within 3 months of each other at our institution (American University of Beirut Medical Center). We extracted three parameters (TAPSE, S' and FAC_RV) that are classically used to assess RV function. We have assessed the ability of 2D-ECHO derived parameters and clinical features to predict RV function measured by the gold standard CMR. We compared outcomes from four machine learning algorithms, widely used in the biomedical community to solve classification problems.

Results

One hundred fifty-five patients were identified and included in our study. Average age was 43±17.1 years old and 52/156 (33.3%) were females. According to CMR, 21 patients were identified to have RV dysfunction, with an RVEF of 34.7%±6.4%, as opposed to 54.7%±6.7% in the normal RV population (P<0.0001). The Random Forest model was able to detect low RV function with an AUC =0.80, while general linear regression performed poorly in our population with an AUC of 0.62.

Conclusions

In this study, we trained and validated an ML-based algorithm that could detect low RV function from clinical and 2D-ECHO parameters. The algorithm has two advantages: first, it performed better than general linear regression, and second, it integrated the various 2D-ECHO parameters.

Artificial intelligence: improving the efficiency of cardiovascular imaging

INTRODUCTION

Artificial intelligence (AI) describes the use of computational techniques to mimic human intelligence. In healthcare, this typically involves large medical datasets being used to predict a diagnosis, identify new disease genotypes or phenotypes, or guide treatment strategies. Noninvasive imaging remains a cornerstone for the diagnosis, risk stratification, and management of patients with cardiovascular disease. AI can facilitate every stage of the imaging process, from acquisition and reconstruction, to segmentation, measurement, interpretation, and subsequent clinical pathways.

AREAS COVERED

In this paper, we review state-of-the-art AI techniques and their current applications in cardiac imaging, and discuss the future role of AI as a precision medicine tool.

EXPERT OPINION

Cardiovascular medicine is primed for scalable AI applications which can interpret vast amounts of clinical and imaging data in greater depth than ever before. AI-augmented medical systems have the potential to improve workflow and provide reproducible and objective quantitative results which can inform clinical decisions. In the foreseeable future, AI may work in the background of cardiac image analysis software and routine clinical reporting, automatically collecting data and enabling real-time diagnosis and risk stratification.

Development of novel machine learning model for right ventricular quantification on echocardiography-A multimodality validation study

Purpose

Echocardiography (echo) is widely used for right ventricular (RV) assessment. Current techniques for RV evaluation require additional imaging and manual analysis; machine learning (ML) approaches have the potential to provide efficient, fully automated quantification of RV function.

Methods

An automated ML model was developed to track the tricuspid annulus on echo using a convolutional neural network approach. The model was trained using 7791 image frames, and automated linear and circumferential indices quantifying annular displacement were generated. Automated indices were compared to an independent reference of cardiac magnetic resonance (CMR) defined RV dysfunction (RVEF < 50%).

Results

A total of 101 patients prospectively underwent echo and CMR: Fully automated annular tracking was uniformly successful; analyses entailed minimal processing time (<1 second for all) and no user editing. Findings demonstrate all automated annular shortening indices to be lower among patients with CMR‐quantified RV dysfunction (all P < .001). Magnitude of ML annular displacement decreased stepwise in relation to population‐based tertiles of TAPSE, with similar results when ML analyses were localized to the septal or lateral annulus (all P ≤ .001). Automated segmentation techniques provided good diagnostic performance (AUC 0.69–0.73) in relation to CMR reference and compared to conventional RV indices (TAPSE and S′) with high negative predictive value (NPV 84%–87% vs 83%–88%). Reproducibility was higher for ML algorithm as compared to manual segmentation with zero inter‐ and intra‐observer variability and ICC 1.0 (manual ICC: 0.87–0.91).

Conclusions

This study provides an initial validation of a deep learning system for RV assessment using automated tracking of the tricuspid annulus.

Right ventricular three-dimensional echocardiography: the current status and future perspectives

This review focused on right ventricular (RV) three-dimensional echocardiography (3DE) and discussed the following agenda. First, we summarized the clinical RV anatomy and function-related RV3DE use followed by the explanations about 3DSTE image acquisition, including pitfall. Next, we reviewed the reliability and feasibility of RV volume and RV ejection fraction measurements during the last decade. Besides, we described the techniques that might overcome the dropout images at RV anterior and out tract including the current limitations. Finally, speckle tracking echocardiography by RV3DE and novel RV shape assessment were reviewed. This review will help you get comprehensive information on the current status and future perspectives of RV3DE.

Use of artificial intelligence in imaging in rheumatology - current status and future perspectives

After decades of basic research with many setbacks, artificial intelligence (AI) has recently obtained significant breakthroughs, enabling computer programs to outperform human interpretation of medical images in very specific areas. After this shock wave that probably exceeds the impact of the first AI victory of defeating the world chess champion in 1997, some reflection may be appropriate on the consequences for clinical imaging in rheumatology. In this narrative review, a short explanation is given about the various AI techniques, including 'deep learning', and how these have been applied to rheumatological imaging, focussing on rheumatoid arthritis and systemic sclerosis as examples. By discussing the principle limitations of AI and deep learning, this review aims to give insight into possible future perspectives of AI applications in rheumatology.