Use of artificial intelligence by novice users to enable accurate point-of-care echocardiographic assessment of left ventricular ejection fraction

Introduction

Point-of-care ultrasound (POCUS) is now universal in the hands of non-experts. However, the results are usually binary (normal/abnormal) but not necessarily specific. Artificial intelligence (AI) is currently used by non-experts in different imaging modalities including echocardiography to aid in diagnosis and decision making.

Aim

To prospectively evaluate whether medical students trained in POCUS and using an AI based assessment tool can accurately evaluate the left ventricular ejection fraction (LVEF) of patients hospitalized in the cardiology department.

Methods

Nine medical students were trained in a 6-hrs session that included lectures and hands-on practice. Participants used a hand-held ultrasound machine (VScan Extend) equipped with LVivoEF, an AI-based tool that automatically evaluates LVEF. The clips were assessed for LVEF by three methods: visually by the students, students plus the AI-based tool, and experienced cardiologists. All LVEF measurements were compared with the gold-standard (a formal echocardiogram completed within 24-hrs with the Simpson method and LVEF eyeball assessment by two blinded fellowship-trained echocardiographers).

Results

The study included 88 patients, (aged 58.3±16.3 yrs, mean BMI 28.3±4.4 kg/m2). Comparing LVEF (continuous values) reported by medical students' visual evaluation, medical students plus AI, and cardiologists vs. the echocardiographers values, revealed Pearson correlations of 0.32 (p=0.003), 0.67 (p<0.0001), and 0.73 (p<0.0001), respectively. The agreements between these three evaluation methods and the echocardiographers, assessed at different LVEF categorical thresholds (at 50% and 40%), using kappa coefficient are presented (Figure 1).

Conclusion

Medical student use of an AI-based tool with a hand-held ultrasound device can improve their LVEF visual assessment to a level of experienced cardiologists. In addition, the use of AI enabled achieving a moderate to substantial inter-rater reliability with echocardiographers' evaluation. This AI tool can be successfully utilized as a decision support tool for POCUS LVEF evaluation by non-experts.

Funding Acknowledgement

Type of funding sources: None.

Validation of an artificial intelligence-based tool for automated evaluation of RV size and function from ultrasound examinations

Funding Acknowledgements

Type of funding sources: Private company. Main funding source(s): Study sponsored by DiA Imaging Analysis

Background

 The assessment of the right ventricular (RV) function is an essential part of every transthoracic echocardiographic examination. It plays an important role in the diagnosis and management of many diseases and conditions. In covid-19 patients, there is evidence showing that enlarged RV and abnormal Free Wall Strain (FWS) correlate with high mortality. Assessment of the RV in the Point of Care settings can assist the differential diagnosis and provide important clinical information in early stages, close to the onset of symptoms, however, it is done qualitatively by visual estimation and thus subjective and requires high level of expertise. The presented study aims to validate the use of a novel artificial intelligence tool (LVivo RV by DiA imaging analysis) that uses deep learning and image processing algorithms to automatically evaluate RV size and function from apical 4 chamber (4CH) focused or modified ultrasound clips.

Methods

 A retrospective study of 100 patients who underwent routine echocardiographic examination was conducted in three medical centers in US and Israel. The apical 4CH focused or modified clips were interpreted by experts to evaluate RV size and function. The End Diastolic Area (EDA) measured manually by two sonographers and the FWS measured by semi-automated Velocity Vector Imaging software (VVI by Siemens) were compared to the same parameters obtained automatically by the LVivo RV, using linear regression and Bland Altman analysis.

Results

 100 cases were included, 39% females and 61% males. Mean age was 64.7 [19-92]. Mean BMI was 28.1 [6.8-17.6]. 74% had pulmonary hypertension and 19% had lung disease. 36% were considered abnormal by their fractional area change values and 64% were considered normal. Three cases were non-interpretable by the physicians thus excluded from the analysis. The LVivo RV was able to process 99% of the cases. Excellent correlation was obtained for EDA between the sonographers" average of manual measurements and the automated EDA by LVivo RV with r = 0.92 (p < 0.0001, 95%CI 0.88-0.94). The bias and limits of agreement for EDA were 0.87 ± 5.76cm^2. For FWS, 4 cases were manually excluded due to insufficient image quality. The Bland Altman analysis for FWS showed small bias and limits of agreement of 0.7 ± 12.2%. A very good correlation of r= 0.78, (95% CI 0.69-0.85) was found, indicating good compatibility between the methods. The specificity and sensitivity for FWS were 80% and 77% respectively, using an optimal cutoff value of -16%, and the overall agreement was 79%.

Conclusions

 The performance of LVivo RV demonstrated a very good agreement with manual and semi-automated quantitative methods for RV assessment. LVivo RV provides fast, accurate, objective and reproducible results and has the potential to be used at the Point of Care settings as a powerful tool for RV size and function evaluation.

Abstract Figure.

Elemental and magnetic sensitive imaging using x-ray excited luminescence microscopy

We demonstrate the potential of x-ray excited luminescence microscopy for full-field elemental and magnetic sensitive imaging using a commercially available optical microscope, mounted on preexisting synchrotron radiation (SR) beamline end stations. The principal components of the instrument will be described. Bench top measurements indicate that a resolution of 1 μm or better is possible; this value was degraded in practice due to vibrations and/or drift in the end station and associated manipulator. X-ray energy dependent measurements performed on model solar cell materials and lithographically patterned magnetic thin film structures reveal clear elemental and magnetic signatures. The merits of the apparatus will be discussed in terms of conventional SR imaging techniques.

The local effect of magnetic impurities on superconductivity in CoxNbSe2 and MnxNbSe2 single crystals

We investigate the effect of individual atomic impurities on the superconducting state that they are embedded in. Using low temperature scanning tunneling microscopy and spectroscopy we could identify Co and Mn atoms in the Co(x)NbSe(2) and Mn(x)NbSe(2) single crystals and observe the influence on the local electronic density of states (LDOS) at 0.4 K. We find that Co is in the weak scattering limit. In this case the LDOS is quite homogeneous on the sample surface, despite the number of defects, and retains sharp coherent superconducting peaks. This is in strong contrast to the effects of Mn impurities, which locally destroy superconductivity. In this case the LDOS shows a strong enhancement of spectral weight inside the superconducting gap even far from the Mn atoms. Moreover, two impurity bound states are found within the superconducting gap at E/Δ(0) = 0.18 and 0.36 at locations close to defects.

Direct observation of geometrical phase transitions in mesoscopic superconductors by scanning tunneling microscopy

Using scanning tunneling microscopy, we mapped the distribution of the local density of states in a single crystal superconductor heterostructure with an array of submicron normal metal islands. We observe the coexistence of strongly interacting multiquanta vortex lattice with interstitial Abrikosov vortices. The newly formed composite magnetic flux structure undergoes a series of phase transitions between different topological configuration states. The vortex configuration states are strongly dependent on the number of flux quanta and the nanoscale confinement architecture of the mesoscopic superconductor. Here, we present images of vortex phase transitions due to confinement effects when the number of magnetic flux quanta in the system changes. The vortex dynamics in these systems could serve as a model for behavior of confined many-body systems when the number of particles changes.

[Surgical repair of myxomatous mitral valve]

Myxomatous mitral valve disease is now the most common cause of mitral regurgitation in the western world. Repair of the leaking valve has become standard surgical procedure during the past 2 decades. Between 1993-1999 we performed 113 repairs of the mitral valve. In 25 patients the etiology was myxomatous degeneration (no mortality). Long-term clinical results depend on patients' functional class at surgery. Based on this fact, and the good surgical results, it is recommended to refer such patients even with severe mitral incompetence for surgery at an early stage, even if symptoms are minimal.

Effect of surgical repair of secundum-type atrial septal defect on right atrial, right ventricular, and left ventricular volumes in adults

Surgical repair of atrial septal defect in adults reduces right ventricular and right atrial diameters and volumes, and improves left ventricular filling.

Prevalence and significance of left ventricular outflow gradient during dobutamine echocardiography

AIMS

This study investigated the clinical and physiological significance of the dynamic left ventricle outflow gradient observed in some patients during dobutamine stress echocardiography.

METHODS

Three hundred and ninety-four consecutive patients completed dobutamine stress echocardiography using Doppler echocardiography to assess the presence of myocardial ischaemia and left ventricular outflow gradient. The prevalence of left ventricular outflow gradient was evaluated and correlated with echocardiographic and clinical findings. Fifteen patients with left ventricular outflow gradient during dobutamine infusion underwent exercise echocardiography for appearance of left ventricular outflow gradient.

RESULTS

Sixty-nine of 394 (17.5%) patients developed a left ventricular outflow gradient of more than 36 mmHg. In nine of them (13%) the anterior mitral valve leaflet had a systolic anterior motion. In 60 of the 69 patients (87%) there was a dynamic obstruction at the level of the papillary muscles. The mean intracavitary gradient was 75.4 (range 36-175) mmHg. There was no correlation between the presence or absence of a dobutamine stress echocardiography-induced left ventricle outflow gradient and chest pain or shortness of breath. In patients who developed a left ventricular outflow gradient ischaemic wall motion abnormalities occurred at a significantly lower frequency during dobutamine stress echocardiography (2.9 vs 16.4% P<0.001). None of the 15 patients who underwent exercise echocardiography developed significant left ventricular outflow gradient.

CONCLUSION

Left ventricular outflow gradient occurs occasionally during dobutamine stress echocardiography examination. Its presence is of no physiological or clinical significance.

Diagnostic value of 12-lead electrocardiogram during dobutamine echocardiographic studies

BACKGROUND

The diagnostic value of 12-lead electrocardiography during dobutamine stress echocardiography (DSE) is not well documented.

METHODS AND RESULTS

We reviewed the records of 116 patients referred for DSE and coronary angiography, 52 of whom were excluded because of abnormal ST segment or inadequate DSE. Of the analyzed 65 patients, 42 had angiographic evidence of significant coronary disease, 41 had evidence of ischemia according to the echocardiographic criteria, and 30 had ST changes during the study. DSE had sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) of 88%, 81%, 90%, and 78%, respectively. Twelve-lead electrocardiography had sensitivity, specificity, PPV, and NPV of 52%, 64%, 72%, and 41%, respectively. NPV increased to 92% in patients with negative DSE and negative ST changes. PPV increased to 95% if both DSE and 12-lead electrocardiographic ischemic changes were observed.

CONCLUSIONS

Twelve-lead electrocardiography has an incremental diagnostic value when used during DSE.

Comparison of exercise hemodynamics among nonstented aortic bioprostheses, mechanical valves, and normal native aortic valves

BACKGROUND

While aortic valve prostheses are known to perform well at rest, few studies have examined them under stress. We compared stress hemodynamics of mechanical valves and nonstented porcine valves in the aortic position to that of normal native aortic valves.

METHODS

Dobutamine echocardiography was used to assess mean and peak gradients and effective orifice area index (EOAI) at rest and exercise in patients with the Toronto Stentless Porcine Valve (SPV) (n = 13, mean implant size 25.7 mm), Sorin Bicarbon mechanical valve (SOR) (n = 11, mean implant size 24.5 mm), and patients with normal native aortic valves (NOR) (n = 10). Dobutamine infusion was started at 5 micron/kg per minute, and increased by increments of 5 micron/kg per minute until the target heart rate was achieved or until a maximal dose of 40 micron/kg per minute.

RESULTS

At rest and exercise, respectively, cardiac output (L/min) was 5.2 and 10.4 for Toronto SPV; 7.4 and 13.5 for SOR; and 4.6 and 11.2 for NOR. Measured EOAI (cm2) was 1.1+/-0.2 and 1.15+/-0.2 for TORONTO SPV; 1.60+/-0.3 and 1.58+/-0.3 for SOR; and 1.45+/-0.2 and 1.46+/-0.2 for NOR. Mean gradients (mmHg) were 5.48+/-1.1 and 5.83+/-0.9 for TORONTO SPV; 5.26+/-0.8 and 11.3+/-1.8 for SOR; and 1.54+/-0.4 and 2.18+/-0.7 for NOR. Peak gradients (mmHg) were 11.9+/-2.0 and 21.0+/-3.7 for TORONTO SPV; 10.79+/-1.7 and 25.9+/-3.4 for SOR; and 2.38+/-0.9 and 6.1+/-2.3 for NOR.

CONCLUSIONS

Although the mechanical group (SOR) had larger measured EOAI, the greater increase in gradients with exercise in this group suggests that the TORONTO SPV is less obstructive to flow.

Valvular dysfunction in antiphospholipid syndrome: prevalence, clinical features, and treatment

Valvular abnormalities develop in 36% and 35% of patients with primary antiphospholipid syndrome (PAPS) and with systemic lupus erythematosus (SLE) respectively, and in 48% of patients with SLE and antiphospholipid antibodies (aPL). Valvulopathy includes leaflet thickening, vegetations, regurgitation, and stenosis. A literature survey shows that significant morbidity from valvular dysfunction, mostly mitral regurgitation leading to congestive heart failure, occurs in 4% and 6% of SLE and PAPS patients, respectively. The pathogenesis of valvulopathy may involve interaction of aPL with antigens on the valve surface, resulting in valvulitis. Current therapy includes symptomatic measures and valve replacement. A novel approach for symptomatic antiphospholipid syndrome (APS) related valvulopathy involves treatment with systemic corticosteroid. We describe four such patients and their dramatic clinical and hemodynamic response to treatment with prednisone when symptomatic measures failed.

Severe hypotension induced by combination of dobutamine and dipyridamole

Pharmacologic testing with dobutamine or dipyridamole in conjunction with echocardiography has become an accepted method for diagnosis of coronary artery disease (CAD). The sensitivity of dobutamine echo ranges from 68 to 86%, and of dipyridamole from 53 to 69% for diagnosis of CAD. Our purpose was to investigate whether the addition of dipyridamole to dobutamine, which may improve the test sensitivity, is safe. Ten patients with low probability of CAD underwent dobutamine echo; 5 were control patients and 5 patients had low dose dipyridamole added at the maximal dose of dobutamine. Four of the latter patients had severe hypotension, while no hypotension was observed in control patients. Our findings suggest that this combination of dobutamine and dipyridamole can be hazardous and should not be used in patients with suspected CAD.

A prospective randomized trial of intravenous magnesium versus intravenous propranolol in acute myocardial infarction

A total of 266 patients entered into a study comparing the effect of intravenous magnesium and propranolol following acute myocardial infarction. Of these, 97 were able to receive either drug and were therefore randomized into the magnesium (n = 51) or propranolol group (n = 46). 88 patients were unable to receive propranolol and formed a third group (NR) while a further 81 patients could not receive either drug and formed a fourth group (N). The study showed that intravenous magnesium was as effective in preventing potentially lethal arrhythmias as propranolol and could be given to some 70 per cent of such patients whereas propranolol could only be given to 36 per cent.

Long-term follow-up after acute myocardial infarction in patients randomized to treatment with intravenous magnesium or intravenous propranolol in the acute phase

Ninety-five patients with acute myocardial infarction were followed up for 6 months to 3 years (mean 25.4 months) in a preliminary study to compare the effects of intravenous magnesium (49 patients) with that of intravenous propranolol (44 patients) given immediately after admission to the intensive care unit. There were four cardiac deaths in the propranolol group and no deaths in the magnesium group (P < 0.046) and 27 per cent of patients who received propranolol subsequently developed cardiac failure as opposed to 12 per cent of those who had received magnesium (P < 0.04). Intravenous magnesium given in the early stages of myocardial infarction reduces the subsequent cardiac death rate possibly by reducing infarct size.

Progression and regression of coronary artery disease in one year. Quantitative angiographic assessment in patients with stable angina pectoris

The rate of progression of coronary artery stenoses (CAS) is not clear. Spontaneous regression may also occur. Seventy-one CAS in 25 patients who were enrolled in a study of the effects of chromium on CAS were analysed. Coronary angiography was performed in multiple views and patients randomised to chromium or placebo treatment. Videodensitometric quantitative analysis was performed using a Vanguard XR70 Analyzer. After 1 year all patients were recatheterised. Corresponding frames from identical views were analysed. CAS were assessed with the observers blinded to the initial study results. No differences were found between chromium or placebo and the results have been combined. There was no overall progression of CAS as assessed by % area stenosis (p = 0.65), % diameter stenosis (p = 0.19), stenotic area (p = 0.87), or stenotic diameter (p = 0.99). However, 20% of individual lesions progressed, while 10% regressed, and 70% remained the same. These changes must be taken into account in studies of interventions which may modify the course of coronary atherosclerosis, and if coronary by-pass surgery is to be performed with a 1 year delay after angiography.

Reproducibility of measurements of coronary narrowings by videodensitometry and by digital calipers

Computer-assisted videodensitometry has been shown to be a reliable and reproducible method of measuring absolute and relative coronary narrowings. Using a commercially available analyzer (Vanguard XR70) we confirmed intra- and interobserver reproducibilities in 34 narrowings in 9 patients. Analyses were performed on normal area and diameter, stenotic area and diameter, percent area stenosis and percent diameter stenosis. For all 6 analyses, excellent intra- and interobserver correlations were found (r = 0.93-0.98), with slopes close to 1 and intercepts close to zero. Caliper measurements (Mitutoyo Digimatic) of the same lesions by the same observers showed good inter- and intraobserver reproducibility for percent diameter stenosis (r = 0.90 and 0.86), with mean interobserver difference of 1.67 +/- (SD) 6.4% and intraobserver difference of 2.97 +/- (SD) 7.9%. However, less good correlations were found between caliper and videodensitometric measurements of percent diameter stenosis; r = 0.61 and 0.76 for the two observers. These data suggest that videodensitometry is a highly reproducible quantitative angiographic method, suitable for documenting changes in the severity of coronary artery lesions, both spontaneous or related to interventions. Caliper measurements do not provide the same degree of accuracy, but they have acceptable reproducibility in measuring diameter stenosis. As such, they are also suitable for assessing changes in severity of coronary artery lesions in individual patients.

Use of pulmonary artery catheters in patients with acute myocardial infarction. Analysis of experience in 5,841 patients in the SPRINT Registry. SPRINT Study Group

This study analyzes the use of PAC in a registry comprising 5,841 hospitalized patients with AMI. A total of 371 patients received PAC. In-hospital mortality was higher in patients with CHF who received PAC, while there was no difference in patients with cardiogenic shock or persistent hypotension. Mortality in patients receiving PAC was higher irrespective of the presence or absence of "pump failure." A separate analysis of discharge summaries of 364 patients with CHF showed that PAC was used more frequently in sicker patients and that when severity of CHF was assessed, no difference in mortality was found in patients with mild or moderate CHF. We conclude that while a higher in-hospital mortality is found in patients receiving PAC, this excess is likely related to difference in severity of CHF, which had not been assessed in every individual. It is unlikely that PAC increases mortality.

Reproducibility of measurements of coronary narrowings by videodensitometry: unreliability of single view measurements

Computer-assisted videodensitometry has been shown to be a reliable and reproducible method of measuring absolute and relative coronary narrowings. Using a commercially available analyzer (Vanguard XR70) we confirmed the intra- and interobserver reproducibilities in 34 narrowings in 9 patients. Analyses were performed on normal area and diameter, stenotic area and diameter, percent area stenosis and percent diameter stenosis. For all 6 analyses, excellent intra- and interobserver correlations were found (r = 0.93-0.98), with slopes close to 1 and intercepts close to zero. In a separate study of 16 lesions in 11 patients, each lesion was analyzed in both the RAO and LAO planes. Correlation between the measurements was reasonable with r = 0.76 for the stenotic area and r = 0.75 for the absolute diameter stenosis. However, with suboptimal slopes (0.54 and 0.63 respectively), actual differences between measurements in the two planes were often unacceptably large. These data suggest that videodensitometry is a highly reproducible quantitative angiographic method; however, single view analyses are inadequate for comparative studies.

Acquired obstruction of the ventricular septal defect in tetralogy of Fallot

The interventricular septal defect in patients with tetralogy of Fallot is, in most cases, a nonrestrictive malalignment type of defect. Rare examples of a restrictive defect have been described. We report a case of tetralogy of Fallot with pulmonic atresia in which the interventricular septal defect was found to be nonrestrictive at birth. The obstruction of the defect developed gradually, until, at the age of 6 years, a gradient of 55 mm Hg across the defect was found at repeat cardiac catheterization. Echocardiography proved that the acquired obstruction was due to excessive tricuspid valve tissue, while Doppler studies suggested a similar gradient between the right ventricle and the ascending aorta. This rare case should alert the clinician to the possibility of an acquired obstruction of the ventricular septal defect, with suprasystemic right ventricular pressure, in cases of tetralogy of Fallot, even if an early study suggested a nonobstructive defect. The diagnosis of this situation is of importance because of the poor prognosis which it carries.

Influence of chronic diuretic therapy on serum, lymphocyte and erythrocyte potassium, magnesium and calcium concentrations

Serum, lymphocyte and erythrocyte potassium, magnesium and calcium concentrations were measured in 31 patients with congestive cardiac failure and 14 patients with mild noncomplicated hypertension, who had been receiving either furosemide or chlorothiazide (with or without potassium supplementation) or a combination of hydrochlorothiazide and amiloride for more than 6 months. Lymphocyte potassium concentrations (pmol/100 cells +/- SE) were as follows: controls 18.1 +/- 1.5, furosemide 14.1 +/- 0.9 (p less than 0.001), furosemide + potassium 12.3 +/- 0.7 (p less than 0.001), chlorothiazide 13.1 +/- 1.0 (p less than 0.001) and hydrochlorothiazide + amiloride 18.6 +/- 0.7 (p = NS). There was a statistically significant relationship between the number of months the patients had been on diuretics and their lymphocyte potassium concentrations. Serum electrolytes, apart from the group receiving chlorothiazide who showed a significant fall in serum K, were unchanged.

Effects of enalapril on lymphocyte sodium, potassium, magnesium and calcium levels in patients with severe congestive heart failure

Fifteen patients (median age 73 years) with severe congestive heart failure were treated with Enalapril for a total of 12 weeks with a significant improvement in their right atrial pressures and in their functional state. Renal function, serum potassium, magnesium and calcium levels were unchanged. Lymphocyte sodium, potassium and calcium levels were generally lower than control values throughout the study but these differences were only statistically significant early in the study. Lymphocyte magnesium levels were unchanged. These findings are in contrast to those previously reported in the literature for such patients treated with conventional diuretics.

Lymphocyte potassium and magnesium concentrations as prognostic factors after acute myocardial infarction

215 patients admitted with acute myocardial infarction were prospectively evaluated for prognostic factors. When lymphocyte potassium concentration increased by up to twofold of normal, a high lymphocyte magnesium concentration was associated with a good prognosis. However, when the increase in lymphocyte potassium concentration was greater than twofold, the presence of a high lymphocyte magnesium made no difference to mortality. Age, the female sex, anterior wall infarction, high serum enzymes, heart failure, diabetes mellitus in females and serious arrhythmias were all bad prognostic signs.

Magnesium in the prevention of lethal arrhythmias in acute myocardial infarction

Seven of 48 patients (14.6%) with acute myocardial infarction who were given 2.4 g of magnesium sulfate as a single intravenous dose had potentially lethal arrhythmias during the first 24 hours after admission, whereas 16 (34.8%) of 46 patients receiving placebo had similar arrhythmias. In addition, 14 of these 16 patients in the placebo group had their first arrhythmia (in the intensive coronary-care unit) within two hours after the start of the study, whereas in the magnesium-treated group, there were no such arrhythmias until some four hours later. The higher the lymphocyte potassium concentration, the greater the reduction in the incidence of arrhythmias. Serum magnesium levels increased by 16.5% and lymphocyte magnesium concentrations by 72% in the magnesium treated group. Intravenous magnesium reduces the incidence of serious arrhythmias after acute myocardial infarction.

Echocardiographic diagnosis of anatomically corrected malposition of the great arteries

In anatomically corrected malposition of the great arteries, the great arteries are abnormally related to one another and to the ventricles but arise from their normal ventricles. We present here a case in which this diagnosis was made by echocardiography and confirmed by cardiac catheterization.

Familial membranous subaortic stenosis

Familial occurrence of membranous subaortic stenosis (MSS) is described in three families. The defect was found in 2 siblings in two of these families, and in 3 siblings of the third family. The importance of early diagnosis and treatment of MSS is emphasized. We suggest early evaluation of first-degree relatives of patients with MSS for the possibility of this defect.