Tests to evaluate left ventricular systolic function

INTRODUCTION — Assessment of left ventricular (LV) systolic function is important for diagnosis, management, follow-up, and prognostic evaluation of patients in a variety of clinical settings. Accurate and reproducible determination of LV systolic function is important given the key role this plays in clinical practice.

INDICATIONS/CLINICAL USE — Indications for evaluation of LV systolic function include the following:

●Signs and symptoms of heart disease.

•Signs or symptoms suggestive of heart disease such as unexplained electrocardiographic abnormality, palpitations, stroke, or peripheral embolic event.

•Signs or symptoms of heart failure (HF). Information on LV systolic function as well as diastolic function, chamber geometry, regional wall motion, and valve function is important for diagnosis and management [1]. Among patients with HF, LV ejection fraction (LVEF) is used to identify categories of HF with preserved ejection fraction (HFpEF; LVEF ≥50 percent), HF with reduced ejection fraction (HFrEF; LVEF ≤40 percent), and HF with mid-range ejection fraction (HFmrEF; LVEF 41 to 49 percent) [1,2]. These categories are important for diagnosis and management of HF. Thus, documentation of LV systolic function, including evaluation of LV ejection fraction, is considered a quality-of-care performance measure in HF [3]. (See "Heart failure: Clinical manifestations and diagnosis in adults".)

•Signs or symptoms of coronary artery disease. Assessment of regional and global LV systolic function is commonly combined with stress testing.

•The presence of ventricular arrhythmias is a common indication for evaluation of LV function and structure as part of an evaluation to determine whether there is a structural cause for the arrhythmia. (See "Sustained monomorphic ventricular tachycardia: Clinical manifestations, diagnosis, and evaluation", section on 'Additional diagnostic evaluation'.)

●Planned or prior exposure to potentially cardiotoxic therapy. Patients undergoing treatment with potentially cardiotoxic therapy require serial evaluation of LV systolic function for early detection of cardiotoxicity, which may affect continued treatment. (See "Clinical manifestations, diagnosis, and treatment of anthracycline-induced cardiotoxicity" and "Cardiotoxicity of cancer chemotherapy agents other than anthracyclines, HER2-targeted agents, and fluoropyrimidines" and "Cardiotoxicity of trastuzumab and other HER2-targeted agents" and "Cardiotoxicity of radiation therapy for breast cancer and other malignancies".)

●Evaluation prior to a procedure for which LV systolic dysfunction may be a risk factor or contraindication. As an example, a transthoracic echocardiogram (TTE) is performed in patients with suspected or increased risk of LV dysfunction prior to renal transplantation. However, a TTE is not generally indicated prior to noncardiac surgery unless there is a specific indication such as a symptom or sign of HF or valve dysfunction. (See "Evaluation of cardiac risk prior to noncardiac surgery", section on 'Resting echocardiography'.)

CHOICE OF TEST — The choice of noninvasive imaging modality for assessment of LV systolic function depends on a number of factors, including test indication, patient-specific factors, and facility-related factors such as availability of expertise and technology for performance of various types of testing. When the primary indication for imaging is assessment of LV systolic function, tests that do not entail radiation or potentially nephrotoxic contrast agents (such as echocardiography [with or without ultrasound contrast] and cardiovascular magnetic resonance (CMR) imaging [without gadolinium contrast]) are generally preferred, particularly when serial imaging is expected.

TTE is the most commonly preferred initial imaging modality, while other imaging modalities are less commonly used as the initial test to assess LV systolic function but may be helpful in specific circumstances.

Unstable patients who require urgent evaluation of LV systolic function require rapid testing, ideally at the bedside and usually by TTE. The urgency of other evaluations, such as assessment of possible acute coronary syndrome, should also be considered.

If additional information is needed, such as identification of structural heart disease (eg, infarction, cardiomyopathy, or congenital heart disease), myocardial ischemia, viability, or fibrosis, then choosing a test that enables comprehensive evaluation including this additional information is ideal from the standpoint of efficiency, patient convenience, and cost. (See "Selecting the optimal cardiac stress test".)

Patients who require serial evaluations of LV systolic function (eg, patients undergoing potentially cardiotoxic chemotherapy) should undergo a noninvasive, minimal risk test that has high reproducibility and accuracy since small changes in measurement may affect management. When available, volumetric (three-dimensional [3D]) methods are preferred for serial evaluation and detection of small changes [4].When serial testing is anticipated, use of a consistent imaging modality (ideally at a single institution) is preferred since studies have shown that measurement of LV ejection fraction by different modalities are not interchangeable [5-7] (see 'Variability' below).

●TTE is the most commonly preferred initial imaging modality due to widespread availability, portability, and negligible risk.

●The following imaging modalities are less commonly used as the initial imaging modality for LV systolic function, but may be helpful in specific circumstances when echocardiography is nondiagnostic (eg, suboptimal image quality or results that conflict with other clinical information) or when specific information from another test is needed:

•Volumetric CMR imaging is considered more accurate and reproducible compared with two-dimensional (2D) transesophageal echocardiography (TEE) and unlike computed tomography (CT) does not entail ionizing radiation. CMR also provides additional information on cardiac structures (including pericardium, myocardium, and valves), myocardial perfusion, viability, and fibrosis that can be useful in certain clinical scenarios such as in identifying HF etiology [8], and is helpful in evaluating congenital heart disease [9]. However, CMR is not as widely available as echocardiography and CT. (See 'Cardiovascular magnetic resonance imaging' below.)

•Radionuclide ventriculography (RVG) may be considered when other tests are unavailable or inadequate, but exposes the patient to 6 to 7 mSv of ionizing radiation [10]. RVG may be particularly useful in patients with significant baseline wall motion abnormalities, distorted geometry [11], or when TTE has poor image quality. Given its high reproducibility, it is helpful when serial evaluations are needed (eg, patients undergoing potentially cardiotoxic chemotherapy). (See 'Radionuclide ventriculography' below.)

•Cardiac CT enables evaluation of the coronary arteries and assessment for pulmonary embolus, in addition to assessment of LV function [12] but exposes the patient to both ionizing radiation (approximately 4 to 10 mSv) and iodinated contrast. Diagnostic accuracy may be compromised in patients with high heart rates or irregular heart rhythms.

•Assessment of LV function is generally not the primary reason for referral for single photon emission CT myocardial perfusion imaging (SPECT-MPI), but SPECT-MPI does provide information on LV function when ordered for the assessment of myocardial ischemia and/or viability [11,13]. (See "Selecting the optimal cardiac stress test" and "Stress testing for the diagnosis of obstructive coronary heart disease".)

•Invasive left ventriculography may be performed at the time of cardiac catheterization. Evaluation of LV function with left ventriculography is generally not the primary reason for cardiac catheterization, as most patients are referred for coronary angiography or to assess other structural heart disease such as valve disease or congenital heart disease. Left ventriculography is performed selectively as needed (eg, when adequate timely noninvasive imaging is not available) [14].

Lastly, nonimaging methods, such as quantification of the ECG signal amplitude, have been proposed to quantify LVEF [15], but the accuracy and applicability of such approaches do not yet allow for their clinical use.

MEASURES OF LEFT VENTRICULAR SYSTOLIC FUNCTION

Left ventricular ejection fraction — LVEF is the most commonly reported clinical metric of global LV systolic function.

However, it is important to recognize that categorization of LV function by LVEF is not based upon etiology or pathophysiology but rather by clinical convention given the prognostic value of LVEF, inclusion of LVEF thresholds as criteria in many clinical trials (including HF trials), and the widespread availability of methods to measure LVEF (particularly transthoracic echocardiography). LVEF is not a robust measure of contractility [16], commonly changes over time [17-19], may be impacted by blood pressure and valvular function, and slightly varies by the method used (eg, echocardiography, CMR, cardiac CT, and radionuclide angiography) . Limits of agreement between modalities for measuring LVEF are wide, and the interobserver and intraobserver variability for measurement of LVEF are substantial (ranging up to nearly 20 percent), as discussed below. In addition, LVEF alone does not incorporate other potentially important metrics such as LV cavity size and whether systolic dysfunction is regional or global. (See 'Variability' below.)

Patients with the same LVEF may display marked and important differences in underlying pathophysiology and prognosis. Despite these limitations, depressed LVEF is an adverse prognostic indicator. (See "Determining the etiology and severity of heart failure or cardiomyopathy" and "Predictors of survival in heart failure with reduced ejection fraction", section on 'Left ventricular ejection fraction' and "Prognosis of heart failure", section on 'Factors affecting mortality rates' and "Treatment and prognosis of heart failure with mildly reduced ejection fraction", section on 'Prognosis'.)

Definition — LVEF is a measure of the percentage of blood ejected during systole in relation to the total end-diastolic volume. The stroke volume (SV) is the difference between end-diastolic and end-systolic LV volumes. LVEF is calculated by dividing the SV by the end-diastolic volume as follows:

A larger ventricle requires a lower EF to achieve the same SV (as compared with a smaller ventricle).

In addition to quantitative calculation of LVEF, semi-quantitative (eg, visual estimates) [20] and qualitative (normal, hyperdynamic, depressed) assessments of LV systolic function have also been described and used clinically (table 1).

Variability — An important limitation of LVEF assessment is the substantial variability in LVEF determinations among various imaging modalities and interpreters. This was illustrated by a study of intermodality variability of LVEF based on data from the international multicenter STICH trial [21]. The study analyzed core lab baseline LVEF data on patients with coronary artery disease and LVEF<35 percent; 1948 had a baseline echocardiogram (only three patients received echo contrast), 774 patients were studied with single photon emission CT (SPECT), and 417 patients were studied with CMR.

●Correlation between LVEF determined by quantitative versus visual echocardiographic methods (r = 0.90 for biplane versus visual) was greater than the correlation of LVEF assessed by different modalities (r = 0.60 for biplane echocardiography and SPECT, r = 0.66 for CMR and SPECT, and r = 0.49 for biplane echocardiography and CMR).

●Bland-Altman plots showed no substantial overestimation or underestimation of LVEF by any modality: Biplane quantitation with echocardiography averaged 2.5 percent higher than CMR, and SPECT was 0.8 percent higher than CMR. Limits of agreement were broad.

●The percentage of LVEF observations that fell within a range of 5 absolute percentage points ranged from 43 to 54 percent between different imaging modalities.

These observations demonstrate a significant limitation of using LVEF thresholds to guide clinical management and suggest that serial follow-up of a given patient may be best accomplished using a single consistent imaging modality and interpretation methodology. When available, quantitative volumetric methods are preferred.

Other measures — Other measures of LV systolic function are less commonly used than LVEF.

●Myocardial velocities, strain, and strain rate are additional parameters of myocardial contractility that can be measured using various techniques, although the most common is speckle tracking echocardiography. Among all myocardial strain parameters, global longitudinal strain [22], reflecting myocardial function as measured from the echocardiographic apical views, is the most used. The term "strain" reflects deformation of a structure and refers to the fractional or percentage change in the structure’s dimension corrected for its original dimension and is calculated as follows:

Strain rate is the rate of this change in deformation, and is calculated as follows:

The myocardial velocity gradient is the difference in velocities between two points of the myocardial wall. Strain and strain rate can be calculated for various myocardial loci in radial, circumferential, and longitudinal directions [23]. Parameters such as strain and strain rate may prove to be more sensitive, reliable, and reproducible than LVEF, though their clinical roles have not been established. (See "Tissue Doppler echocardiography".)

●Fractional shortening (FS) is an older, simplified, linear measure of the percentage LV systolic diameter reduction, usually performed with M-mode or 2D TTE. Normal values for fractional shortening are lower than corresponding LVEF measurements (table 2). It is calculated using linear dimensional measurements of the LV cavity diameter at end-diastole and end-systole as follows:

While calculation of FS is simple, its accuracy in quantifying LV systolic dysfunction is generally limited to settings with global (rather than segmental) LV systolic dysfunction.

●Myocardial contractility, another parameter of LV systolic function, can be measured by calculating rate of pressure rise with time (dP/dt) with the normal range >3200 Hg/s.

●Cardiac output and cardiac index are hemodynamic parameters that are impacted by LV systolic function. Cardiac output and index are calculated from SV and heart rate as follows (see "Pulmonary artery catheterization: Interpretation of hemodynamic values and waveforms in adults", section on 'Calculation of cardiac output'):

ECHOCARDIOGRAPHY — Echocardiography uses ultrasound waves to produce real time images of the heart. A variety of echocardiographic techniques are clinically available:

●2D TTE, with the ultrasound probe oriented in various positions over the chest wall and upper abdomen, enables visualization of the cardiac chambers and myocardium for evaluation of LV size, myocardial wall motion, and wall thickening for qualitative and quantitative assessment of LV systolic function. The use of harmonic imaging and echocardiographic contrast media have been shown to significantly improve image quality for assessment of LV systolic functional parameters, particularly in patients with poor acoustic windows [24-27]. (See "Transthoracic echocardiography: Normal cardiac anatomy and tomographic views" and "Contrast echocardiography: Clinical applications".)

●3D echocardiographic images are generated from the integration of information from multiple 2D imaging planes, allowing for volumetric assessment of LV function. (See "Three-dimensional echocardiography".)

●TEE is similar to TTE except images of the heart are obtained with the ultrasound probe positioned within the esophagus or stomach, providing higher resolution images of the posterior heart structures. (See "Transesophageal echocardiography: Indications, complications, and normal views".) Quantitative TEE assessment of ventricular function is rarely performed.

Indications/clinical use — Patients presenting with symptoms or signs potentially caused by a cardiac disorder, including chest discomfort, shortness of breath, arrhythmias, lightheadedness, or syncope, are commonly referred for echocardiography; in such patients, assessment of LV systolic function often impacts clinical decision making. TTE is often used in the initial and follow-up assessment of ventricular function following acute coronary syndromes. Given its availability and portability, TTE is ideal for the assessment of patients with hypotension or hemodynamic instability of uncertain or suspected cardiac etiology, allowing for real time imaging at the bedside [28].

Function parameters — Global LV systolic function can be evaluated by qualitative visual assessment, estimated semi-quantitative visual grading, or quantitatively. Qualitative visual assessment of LV systolic function is commonly reported as normal, hyperdynamic, or depressed. Depressed LV function is often further characterized by the degree of dysfunction and whether the dysfunction is global or regional. There has been some variation among LV ejection fraction (LVEF) ranges used to define the degree of LV systolic dysfunction [29,30]. The American Society of Echocardiography (ASE) 2015 update on chamber quantification included the following LVEF percentage ranges: mildly depressed (41 to 51 for men and 41 to 53 for women), moderately depressed (30 to 40), and severely depressed (<30). Visual estimation of LVEF represents one of the most common methods used in clinical practice for assessment of LV systolic function, and is typically reported in intervals of 5 or 10 percent, often with a range of values (eg, 35 to 40 percent).

Several methods have been proposed for the quantitative assessment of LVEF. The biplane method of discs (modified Simpson method) using area tracings of the LV cavity is the preferred 2D method for LV volume quantification and measurement of LVEF (image 1) [29]. In comparison with earlier linear methods, the modified Simpson method corrects for shape distortions with fewer geometrical assumptions. 3D echocardiographic imaging is available in some centers as an accurate method for evaluation of LV systolic function without geometric assumptions needed for 2D methods. However, adequate acoustic windows are required for 3D echocardiographic image acquisition. (See "Three-dimensional echocardiography", section on 'Assessment of left ventricular volume and function'.)

Myocardial contractility, a measure of LV systolic function, can be described by measuring dP/dt, or rate of pressure rise in the LV during systole using the continuous wave Doppler spectrum of the mitral regurgitation jet [31].

Newer echocardiographic techniques using tissue Doppler imaging and speckle tracking echocardiography can directly obtain information on myocardial contractility, including myocardial velocities, strain, and strain rate. These methods allow for assessment of the various components of contraction, including radial, longitudinal, and circumferential contraction, allowing for assessment of global and regional systolic function (image 2). Parameters such as global longitudinal strain (GLS) are more sensitive for detection of decreased LV function than traditional measures of LVEF [32-35] and have been recommended as complementary clinical techniques that offer incremental prognostic information over LVEF in selected patients with conditions including cancer-therapy related cardiac dysfunction, HF, valve disease (eg, asymptomatic severe mitral regurgitation), and cardiomyopathy (eg, cardiac amyloidosis) [35]. Based on a meta-analysis of 24 studies involving 2597 subjects, the normal reference range for GLS ranges from 15.9 to 22.1 percent (mean 19.7 percent; 95% CI 20.4-18.9) [36]. (See "Tissue Doppler echocardiography", section on 'Strain and strain rate imaging' and "Clinical manifestations, diagnosis, and treatment of anthracycline-induced cardiotoxicity".)

Regional evaluation of LV function is commonly determined based on qualitative visual assessment of wall thickening and endocardial motion of each myocardial segment visualized in multiple views. Seventeen- or 16-segment models of the heart are recommended (figure 1), with the ASE guidelines recommending the 16-segment model for routine assessment of wall motion due to limited endocardial excursion and thickening at the tip of the apex [37]. ASE guidelines also recommend use of a semiquantitative wall motion score (1-normal or hyperkinetic, 2-hypokinetic [reduced thickening], 3-akinetic [absent or negligible thickening], and 4-dyskinetic [systolic thinning or stretching]) assigned to each segment for calculation of LV wall motion score index as the average of scores of all visualized segments [37]. (See "Overview of stress echocardiography", section on 'Two-dimensional imaging'.)

Hemodynamic information can be obtained with 2D echocardiography using Doppler techniques combined with 2D imaging for measurement of stroke volume, cardiac output, and cardiac index. Accurate measurement of these parameters is largely dependent on the measurement of the LV outflow tract diameter (which assumes a circular LV outflow tract), although serial evaluations by comparing LV outflow tract velocity time integral are not as dependent and thus may allow for detection of subtle changes in these parameters.

Diagnostic performance — Measurement of LV volume by 2D TTE has been shown to be a very accurate technique in experimental studies using balloons and heart models [38-40], including asymmetric heart models, with average mean difference of approximately 10 mL compared with the true volume [39]. Patient studies consistently demonstrate underestimation of LV volumes as measured by TTE compared with volumetric techniques with reported mean differences of up to 70 mL in end-diastolic volume measurements depending on technique [6,41-44]. Sources of error are mainly related to positioning of the transducer probe anterior and superior to the LV apex, leading to tangential imaging of the heart and foreshortened apical views [41,44,45].

Visual estimation of LVEF is the most commonly utilized method for routine clinical evaluation of global LV systolic function, and demonstrates improved correlation with LVEF as determined by radionuclide ventriculography (RVG) [46] and invasive left ventriculography [47], when compared with other, more time consuming quantitative techniques. A mean difference of 1 to 7.5 percent is seen with LVEF measurements when compared with invasive left ventriculography [43,45,48], with wide limits of agreement as high as 28 percent when using Simpson’s biplane method and even higher for LVEF calculation using other 2D and/or linear measurements [6,48]. When compared with other noninvasive techniques, including RVG, CMR, and single photon emission CT myocardial perfusion imaging, there is better correlation in LVEF with mean difference <5 percent, although limits of agreement tend to be wide, ranging from 8 to 27 percent [48-50].

Reproducibility studies show significant variation in visual estimates of LVEF with inter- and intraobserver variability ranging from 15 to 18 percent [11]. Similar variability is seen with quantitative LVEF measurements with differences of 8.5 to 19 percent noted for 2D TTE [11,24,25,43,51], greater for repeat TTE studies or readings by different readers than with repeated measurements from the same study by the same reader [51].

Quantified LVEF using 2D TTE is generally higher when compared with LVEF measured by SPECT-MPI [48,52], lower when compared with invasive left ventriculography [48], and similar when compared with CMR [5,25], although the size and direction of bias between modalities may vary depending upon institutional protocols [53].

The use of an intravenous ultrasound enhancing agent (UEA) improves the accuracy and reproducibility of measurements of LV volumes and LVEF by TTE [24-27], but still underestimates the LV volume [25,26]. Intra- and inter-reader variability in measurement of LVEF ranges from 2 to 10 percent [24,25,54]. Contrast echocardiography has also been shown to improve agreement and decrease variability for assessment for presence and degree of regional wall motion abnormalities with improvement in agreement from 68 to 82 percent for identification of normal segments and 82 to 100 percent for identification of abnormal segments as determined by CMR [55]. (See "Contrast echocardiography: Clinical applications".)

3D TTE provides better accuracy, precision, and reproducibility in measuring LV volumes and modest increase in precision in measuring LVEF compared with 2D TTE. Compared with CMR, 3D TTE consistently underestimates LV volumes, although not LVEF. However, there tends to be substantial variability in the 3D TTE calculations of 30 to 38 mL for volumes and 12 to 21 percent for LVEF, and is worse in patients with poor image quality or larger ventricles [49,56].

TEE provides superior resolution of endocardial borders compared with TTE and can be used to evaluate LV systolic functional parameters. However, TEE significantly underestimates ventricular volumes obtained with other techniques, likely due to foreshortening of the apex [57], which may be an even greater problem than that seen with TTE. Although LVEF by TEE correlated well with LVEF by TTE, TEE often overestimated LVEF in patients with regional LV systolic dysfunction [58]. LVEF by TEE and invasive left ventriculography with mean difference less than 5 percent correlate well but TEE tends to underestimated LV volumes due to foreshortening of the apex [57].

GLS is a more reproducible measurement compared with LVEF, with interobserver relative mean error of 5.4 to 8.6 percent and intraobserver mean error of 4.9 to 7.3 percent [59,60]. LV segmental longitudinal strain measurements have greater variability, particularly when compared across different vendors, although other methods for segmental function assessment, such a strain curve shape analysis or relative comparison between regions, may provide a robust alternative assessment [61].

Machine-learning algorithms for automated volume-independent LVEF calculation have been shown to be feasible [62,63] and may develop as clinically helpful tools.

Advantages — 2D TTE is the most widely used imaging technique for evaluation of LV systolic function. Advantages of TTE include its noninvasive nature, ready availability, low cost, and portability, which makes it especially useful for assessment at bedside in unstable patients. Additionally, TTE offers real time imaging assessment and can be easily repeated without incurring additional risks to the patient, making it good for use in serial follow-up studies.

Limitations and trouble shooting — 2D TTE is highly operator dependent and requires adequate acoustic windows. Patients with obesity, patients with chronic obstructive pulmonary disease, and patients with chest wall abnormalities (eg, pectus or left mastectomy) or limited space between the ribs will often have poor imaging quality. Studies have shown that quantitative assessment of LV function by 2D TTE is suboptimal (endocardium not well visualized in all 16 segments) in up to 20 percent of patients [54]. The use of echo endocardial border definition contrast has been shown to improve the determination of LVEF in some of these patients [54]. (See "Contrast echocardiography: Clinical applications".)

The main limitation with conventional 2D TTE for quantitative assessments is its reliance on geometrical assumptions for measurement of LV volumes and LVEF. Measurements are less accurate in patients with regional systolic dysfunction and/or altered LV geometry, since measurements taken in one location may not be representative of the other regions. In addition, imaging planes that are off axis or foreshortened due to limited acoustic windows from the chest wall may lead to incorrect measurements entirely. In these cases, visual estimation of LVEF may be better than quantitative methods.

The other significant limitation of 2D TTE is the limited accuracy of LVEF quantitation in the presence of regional wall motion abnormalities and/or altered LV geometry given the geometric assumptions required for LVEF calculation. 3D TTE improves the accuracy of quantitative LV function parameters compared with 2D TTE, but has even greater reliance on adequate acoustic windows and sometimes inadequate temporal resolution leading to error. Semi-automated software programs are often used for quantification, but rely on accurate placement of fiduciary points or endocardial border definition by the software or operator. With 3D TTE, image data are usually acquired over several heart beats, and artifacts are notable in patients who have ectopic beats or who cannot sustain a breath hold during image acquisition [64-66]. Real-time methods can minimize some of these artifacts. (See "Three-dimensional echocardiography", section on 'Assessment of left ventricular volume and function'.)

Complications — There are no known complications associated with routine, noncontrast TTE, and there are no well-documented harmful effects during pregnancy. (See "Overview of ultrasound examination in obstetrics and gynecology", section on 'Safety'.)

The use of echocardiographic endocardial border definition contrast agents has been shown to be safe with rare adverse events as discussed separately. Contraindications for use of these contrast agents include pregnancy; known right to left, bidirectional, or transient right to left cardiac shunts; hypersensitivity to perflutren; and hypersensitivity to blood, blood products, or albumin (in the case of Optison only) [67,68]. (See "Contrast echocardiography: Contrast agents, safety, and imaging technique", section on 'Safety'.)

TEE is a moderately-invasive technique that is rarely used for primary LVEF assessment except in patients with poor TTE acoustic windows. Considerations for TEE include the risk for injury to the gastrointestinal tract, posterior pharynx, and respiratory complications due to small reduction in oxygen saturation associated with sedation, bronchospasm, laryngospasm, posterior hematoma in the posterior pharynx, pulmonary edema, atelectasis, and airway obstruction. Complications related to medications administered for sedation and local anesthesia include respiratory depression, hypotension, agitation allergy, anaphylaxis, overdose, and methemoglobinemia. Some of these issues are of particular concern among individuals with sleep apnea, small airway, or class II or III obesity. Quantitative TEE methods for LVEF assessment are rarely used. (See "Transesophageal echocardiography: Indications, complications, and normal views", section on 'Safety of TEE examination'.)

CARDIOVASCULAR MAGNETIC RESONANCE IMAGING — CMR imaging creates anatomic and functional cross-sectional images of the cardiovascular system. CMR offers high spatial and temporal resolution with image acquisition in any desired imaging plane without limitation by chest wall anatomy. The entire heart can be imaged to create full 3D datasets using a fast gradient echo technique (eg, balanced steady state free precession), which offers excellent contrast between the blood pool and LV endocardium without the need for administration of exogenous intravenous contrast (ie, no need for gadolinium) or ionizing radiation.

Gadolinium-based contrast is administered for selected CMR examination (eg, identification of myocardial infarction or fibrosis) but is not needed for LV volume, LVEF, or strain assessment.

Indications/clinical use — CMR imaging provides a wealth of information for a variety of cardiovascular indications [69], in which qualitative and quantitative assessment of LV systolic function parameters are commonly assessed as part of the evaluation. The most common referrals are for evaluation of cardiomyopathy, congenital heart disease, valvular heart disease, and ischemic heart disease [70]. CMR is particularly helpful in evaluating patients with indeterminate or discrepant echocardiography results, as well as in clinical scenarios in which accurate assessment of serial changes in LV function is important. (See "Clinical utility of cardiovascular magnetic resonance imaging".)

Function parameters — Interpretation of CMR images can include qualitative and quantitative assessment of global and regional LV systolic function.

Assessment of LV systolic function is commonly performed using a stack of short-axis cine images covering the entire LV, which are then reviewed and analyzed for quantitative evaluation of ventricular volumes, mass, and LVEF without the need for geometrical assumptions. Similar to other techniques, qualitative visual assessment of LV function is commonly reported as normal, hyperdynamic, or depressed, with degree of dysfunction further characterized (eg, mild, moderate, or severe) and location (global or regional with specification of LV segments). LVEF can be estimated by visual assessment, but more commonly is calculated using manual, semiautomated, or automated methods. The summation of discs method using the short-axis cine images of the LV is the most commonly employed method in which LV endocardial borders are traced on each short-axis image at end-diastole and end-systole to determine LV cavity area, which is then multiplied by the slice interval to determine a volume for each slice. The volumes of slices are summed to determine an LV volume, which is used to calculate LVEF (image 3).

Global myocardial contractility parameters including myocardial velocities, strain, and strain rate can be measured using myocardial tissue tagging, velocity encoding, DENSE, and feature tracking [71-73]. although these techniques have not yet been routinely incorporated into clinical examinations due to post processing limitations.

Regional LV function by CMR is commonly evaluated using cine CMR and either visual estimation and/or quantitative assessment of wall motion and/or wall thickening. Qualitative assessment of regional function demonstrates high concordance with other techniques, including CT, 2D and 3D TTE, invasive left ventriculography with >82 percent agreement for wall motion assessment [49], and low (10 percent) interobserver variability [74,75]. Segmental strain imaging using tagged cine CMR, DENSE, feature-tracking cine CMR, or other CMR techniques allows for measurements of all directions of myocardial contraction, and has been shown to allow a more accurate assessment than measuring wall thickening alone [71,76].

Hemodynamic measures of stroke volume, cardiac output, and cardiac index can be calculated using phase contrast velocity mapping sequences, which function similar to Doppler echocardiography.

Diagnostic performance — CMR is considered a clinical gold standard for the noninvasive assessment of LV volumes and LVEF due to its high accuracy and reproducibility with minimal risk. Experimental studies using latex and wax heart model casts and dynamic phantoms have shown that CMR is a very accurate method for quantification of LV volumes with absolute error <10 mL [77,78]. CMR measurements of LVEF correlate well with other modalities, particularly CT and 3D TTE [49], and correlate less well with 2D TTE [50]. Intraobserver, interobserver, and interstudy variability for CMR parameters of LV systolic function demonstrate very good agreement with variability of <6 percent for LV end-diastolic volumes and <7 percent for LVEF [79-83]. As computer technology and machine-learning algorithms develop, automated analysis of CMR data may become the preferred clinical analysis tool. In a multicenter study, a fully automated machine-learning LVEF algorithm had high accuracy and reproducibility compared with novice and expert cardiologist CMR data analysis, providing the results nearly 200-fold faster [84].

Advantages — Cine CMR produces high resolution images of the cardiac chambers without exposure to ionizing radiation or need for contrast agent. Unlike echocardiography, CMR can produce images of cardiovascular structures without interference from thoracic structures such as the ribs or lung tissue. CMR is also less operator dependent than echocardiography. Cine data acquisition for assessment of LV global and regional systolic function does not require administration of gadolinium contrast. Given the volumetric nature of image acquisition, quantitative evaluations of ventricular volumes for calculation of LVEF are highly accurate and reproducible since they do not require geometric assumptions as is the case with 2D echocardiography and invasive left ventriculography.

Limitations and trouble shooting — The availability of appropriate expertise and technology for CMR is growing, but remains more limited than for echocardiography and radionuclide modalities. Patients must undergo a careful safety check to exclude any CMR incompatible objects or devices prior to scanning. Compared with other cardiovascular tests of LV systolic function, CMR requires relatively more patient cooperation and compliance with remaining still in the scanner. CMR typically requires multiple breath-holds, though free breathing protocols and real time acquisitions are available (albeit with lower temporal and spatial resolution). Claustrophobic patients may be unable to undergo the examination unless they receive antianxiety medications in preparation for the test. Examination times are significantly longer compared with cardiac CT, although new techniques enable shorter examination times compared with standard protocols without sacrificing diagnostic accuracy [85,86]. In addition, patients are isolated from direct care when they are in the scanner, and many medical devices including life support equipment cannot safely enter the scanner room, so this test is generally unsuitable for unstable patients. (See "Patient evaluation for metallic or electrical implants, devices, or foreign bodies before magnetic resonance imaging", section on 'Assessing implants, devices, or foreign bodies for MRI'.)

CMR assessment of LV volumes and LVEF typically require multiple breath holds to acquire images throughout the entire LV. Image quality may be limited in patients who are unable to hold their breath. If the level of breath holding is inconsistent, image acquisition may occur at different levels of the heart with some segments getting missed and others getting imaged twice, which may lead to variability in the calculated volumes and LVEF. Imaging is often performed at end expiration to minimize these differences. Since imaging data are usually acquired over multiple cardiac cycles with electrocardiographic gating, arrhythmias and ectopic beats may degrade image quality. Real-time methods are an alternative in patients with arrhythmias, though spatial resolution is reduced. Variation in data analysis such as the methods of identifying the border between the myocardium and cavity (including how the papillary muscles and trabeculae are traced) and of selecting image slices for inclusion may affect quantitative results and lead to variability among different readers [87,88].

Complications — Complications related to noncontrast CMR are relatively rare, and primarily related to the strong magnetic field and its potential impact on magnetizable materials and electronic devices.

A thorough safety screen is required prior to every CMR scan. The strong, static magnetic fields produced by CMR scanners can cause motion of magnetizable objects and may also affect the performance of implanted devices, such as cardiac pacemakers and defibrillators, although some newer devices have been designed to be MR compatible. Pulsed, gradient, magnetic fields during image acquisition can induce electrical currents in implanted or attached devices and may cause local heating, arrhythmias, and neuromuscular stimulation [89]. Another issue is that unstable patients may be difficult to monitor and access during a scan. Complications related to CMR scanning are discussed in detail separately. (See "Nephrogenic systemic fibrosis/nephrogenic fibrosing dermopathy in advanced kidney disease" and "Patient evaluation for metallic or electrical implants, devices, or foreign bodies before magnetic resonance imaging", section on 'Assessing implants, devices, or foreign bodies for MRI'.)

No harmful effects of CMR during pregnancy have been reported, but data are limited. Decisions regarding use are based upon weighing the potential benefits and risks of imaging. (See "Diagnostic imaging in pregnant and lactating patients", section on 'Magnetic resonance imaging'.)

INVASIVE LEFT VENTRICULOGRAPHY — Invasive left ventriculography is commonly performed in conjunction with coronary angiography or cardiac catheterization performed for other indications. Quantitative assessment of ventricular volumes, LVEF, and regional contraction [90] are often performed using single plane/projection (30 degree right anterior oblique) imaging, although biplane techniques (30 degree right anterior oblique and 60 degree left anterior oblique) are considered more accurate [91,92]. Automatic computerized software is available for calculation of various parameters, including LV volumes and LVEF.

Indications/clinical use — Invasive left ventriculography is indicated when LV function or wall motion is unknown, or a mechanical complication is suspected (ie, in the setting of acute coronary syndrome), and the patient is undergoing coronary angiography/cardiac catheterization prior to any noninvasive assessment of LV function or when an acute change in clinical status suggests that LV function may have also changed. Both single and biplane methods are used, with biplane requiring double the radiation dose (and double the contrast if not performed with biplane system). Given this, the risks and benefits must be weighed for individual patients.

Function parameters — Qualitative and semiquantitative methods are available for assessment of LVEF using rapid, automated digital methods. LV systolic function can be graded qualitatively using visual assessment of cine image loops. LVEF can be determined by visual assessment of cine image loops or quantitatively using geometric models for calculation of LV volumes in end-systole and end-diastole. Automated, semi-automated, or manual tracing of endocardial borders using post-processing software is performed at end-diastole and end-systole, and LV volumes are calculated using an appropriate geometric reference figure such as the prolate ellipsoid [90]. LVEF is calculated based on the end-diastolic and end-systolic volumes using the LVEF formula. Regional function is commonly evaluated using visual assessment of wall motion for each segment using cine loops of each imaging plane. Concomitant hemodynamic LV measurements of LV pressure can be obtained for assessment of circumferential stress, generation of pressure volume curves, and assessment of LV compliance [90].

Diagnostic performance — Invasive left ventriculography has been demonstrated in multiple studies to significantly overestimate LV volumes and LVEF compared with noninvasive techniques [48,49,93], with up to 20 percent difference in agreement for LVEF [49]. The use of biplane left ventriculography has not been shown to consistently provide incremental information compared with single plane left ventriculography, although biplane left ventriculography measurements correlate better with CMR for LVEF, ventricular volumes, and wall motion than single plane left ventriculography [14]. Visual estimation of LVEF by left ventriculography correlates variably to LVEF from TTE, particularly in patients with coronary artery disease [94-96]. Invasive left ventriculography demonstrates high agreement between readers for global LV systolic function, with limits of agreement of 11 percent for LVEF and ±18 to 21 mL for LV volumes [49]. With invasive left ventriculography, there is high interrater agreement for regional wall motion compared with CMR, CT, and echocardiography [49]. Using CMR as the reference standard for per-segment regional function, invasive left ventriculography and 2D and 3D echocardiography are less sensitive and more specific than cardiac CT [49].

Advantages — Invasive left ventriculography was previously considered the gold standard for evaluation of LV systolic function, but it has now largely been supplanted by noninvasive methods. When performed in the setting of a cardiac catheterization procedure, it adds only a few additional minutes with low cost compared with performing an additional noninvasive study. Although limited, concomitant assessment of mitral valve regurgitation and other structures (eg, interventricular shunting) can be obtained.

Limitations and trouble shooting — Invasive left ventriculography is an invasive procedure requiring arterial access, radiation exposure, and iodinated contrast. Average radiation exposure for coronary angiography plus left ventriculography is approximately 5 to 7 mSv [97,98]. The typical volume of iodinated contrast is 35 to 40 mL/view. Two injections are needed for biplane methods if a biplane system is not available.

Invasive left ventriculography is commonly performed in one or two imaging planes, which are used for assessment of LV volume, LVEF, and regional wall motion. Assumptions using geometric models and correction for radiographic magnification are needed for quantification of LV volumes and LVEF. The different ventricular shapes that occur with disease states can make these assumptions invalid. Left ventriculography can only assess a single valvular abnormality (mitral regurgitation) and a single cardiac chamber, unlike other techniques that allow for a more comprehensive evaluation of cardiac structure and function. There can be a high rate of nondiagnostic or suboptimal diagnostic image quality related to a variety of factors [99].

Proper performance of left ventriculography is important to generate high-quality data. Power injections, as opposed to hand injections, are required for adequate opacification of the LV cavity. Multi-sidehole catheters are the safest and most effective as compared with single end-hole catheters. The contrast volume should be sufficient (usually 35 to 40 mL) to completely opacify the ventricle [100] with imaging in appropriate angulated views to best visualize the LV and allow for accurate quantification of LV volumes and LVEF [101]. Common problems include improper measurement of LVEF during a postextrasystolic contraction, catheter-induced mitral regurgitation, and incomplete LV opacification [99].

Complications — Although rare, complications can occur with invasive left ventriculography. The 35 to 40 mL injection of iodinated contrast required for left ventriculography increases the total amount of contrast used for a diagnostic cardiac catheterization. Complications include hypersensitivity reactions to contrast agent. Patients with chronic kidney disease, hypotension, anemia, and HF are at increased risk for developing contrast-induced nephropathy [102]. Patients with elevated end-diastolic pressure may be at increased risk of acute respiratory decompensation [14]. Embolization of air, thrombus, or the catheter tip may occur due to technical errors [103,104]. Patients with LV mural thrombus or aortic valve vegetations are also at risk of systemic embolism [14]. Mechanical stimulation of the ventricular endocardium by the catheter or injection jet can lead to ventricular extra-systoles and ventricular tachycardia. Atrial fibrillation, sustained ventricular tachycardia, and ventricular fibrillation are very rare complications [8,105-107]. Different degrees of heart block may also occur due to the proximity of the left bundle to the LV outflow tract. Very rarely, myocardial and pericardial staining, myocardial rupture, and pericardial tamponade have been described related to improper catheter positioning or use of end-hole catheters [108,109]. Finally, performance of a left ventriculography may increase total radiation exposure of an exam by up to 30 percent [110]. (See "Diagnosis and treatment of an acute reaction to a radiologic contrast agent".)

CARDIAC COMPUTED TOMOGRAPHY — CT uses multiple radiographic images to create cross-sectional images of the body with the aid of a computer. Cardiac CT refers to CT imaging of the heart with multidetector (eg, 64 and higher) CT systems using electrocardiographic (ECG) techniques to synchronize the image acquisition with the cardiac cycle. Contrast-enhanced cardiac CT with volumetric image acquisition covering the entire cardiac cycle is commonly required for assessment of LV function. The 3D datasets can be processed to generate contiguous short-axis cine images that allow for quantitative evaluation of LV volumes, mass, and LVEF using the disc summation method (image 4). 3D techniques using threshold-based segmentation rely on summation of contiguous voxels of a predefined attenuation threshold to determine LV volume and LVEF (image 5).

Indications/clinical use — Patients are commonly referred for cardiac CT imaging for evaluation of the coronary arteries in the assessment of coronary artery disease or coronary artery anomalies [111]. Depending on the specific imaging protocol used, concomitant assessment of LV systolic function can be performed in conjunction with analysis of the coronary arteries. Ventricular wall motion can be assessed by visual evaluation of cine loop displays of multiple cardiac phases. Cardiac CT is considered an appropriate alternate imaging technique for evaluation of LV function in patients following an acute myocardial infarction or in HF who have inadequate images or discrepant information available from other imaging tests [111].

Function parameters — Global LV systolic function can be qualitatively and quantitatively evaluated. Visual assessment of LV global wall motion can be graded qualitatively by viewing cine volume loops encompassing the entire cardiac cycle. Quantitative assessment of LV function, including measurement of LVEF, is performed using Simpson’s method, area length method, or 3D threshold-based segmentation for measurement of end-diastolic and end-systolic volumes.

Diagnostic performance — Cardiac CT measurement of various LV functional parameters demonstrates good correlation and agreement with measurements obtained by other imaging modalities, including CMR, 2D TTE, ECG-gated single photon emission CT, and radionuclide ventriculography [112-119]. Cardiac CT has been shown to have good to excellent intra and interobserver agreement for assessment of LVEF, with excellent correlation with 2D TTE [120]. Compared with TTE, LVEF is overestimated with cardiac CT by an average of 1.4±5.6 percent [120]. Early 4- and 16-slice MDCT studies demonstrated slight overestimation of LV volume and underestimation of LVEF by cardiac CT compared with invasive left ventriculography and CMR [121-125]. With spatial and temporal resolution offered by 64-slice or higher and dual source CT systems, the average difference in global LV function indices between cardiac CT and other imaging techniques narrows [126] and intermethod differences are no longer reproduced [127-130].

3D threshold-based segmentation algorithms systematically underestimate LV volumes and overestimate LVEF when compared with Simpson’s method of discs [131]. Discrepancies between these two techniques may be related to inclusion of papillary muscles or setting of the apical and basal LV boundaries. The automated nature of the 3D region growing approach has been shown to be more accurate and reproducible with minimal interobserver variability compared with 2D-based methods. Interobserver variability with MDCT is reported to be 2 to 8.5 percent for both LV end-diastolic volume and LVEF [125,131-135].

Assessment of regional wall motion abnormalities with cardiac CT demonstrates excellent agreement with other imaging modalities, with 75 to 96 percent agreement with 2D TTE [113,114,120] and 85 to 90 percent agreement with CMR imaging [49,120].

Advantages — Cardiac CT image acquisition is rapid relative to other techniques. Fast automated tools for analysis of segmental and global function have been shown to have high reproducibility and excellent correlation with echocardiography [136], resulting in significant time savings compared with manual approaches [137].

Limitations and trouble shooting — Precise measurement of LVEF requires temporal resolution of 30 to 50 milliseconds per image, especially in patients with high heart rates. The temporal resolution with cardiac CT is still inferior to other techniques, including TTE, invasive contrast left ventriculography, and CMR, which leads to CT overestimation of end systolic volumes, and, as a consequence, underestimation of LVEF. The limited temporal resolution of cardiac CT is the primary limitation in the evaluation for regional wall motion abnormalities, and scanners with higher temporal resolution demonstrate higher agreement compared with other scanners, with 96.7 percent agreement with dual source CT scanners [138].

Depending on the method used, LV volumes will vary based on whether the papillary muscles and/or trabeculations are included in the LV cavity. Variability in the selection of the basal segments for inclusion will also cause variability in volume and LVEF when using the Simpson method. Since ECG gating is required for image reconstruction, cardiac arrhythmias or ectopic beats will degrade image quality, which can reduce the accuracy of the LVEF calculation. Respiratory artifacts due to poor breath holding are another source of error. Finally, iodinated contrast enhancement is needed to distinguish myocardium from the cavity. Poor contrast enhancement of the cavity due to inadequate intravenous access, poor timing of contrast injection and scanning, or obesity may lead to inaccurate distinction of the endocardial cavity interface, particularly with automated methods and inaccurate quantification [139].

Complications — Cardiac CT exposes the patient to radiation associated with the acquisition of the scan, although lower radiation dose protocols have been proposed that limit radiation exposure to less than 5 mSv [119]. Additionally, cardiac CT exposes the patient to intravenous administration of iodinated contrast. The total volume of iodinated contrast administered is usually based on the injection rate and scan duration, and typically is 50 to 120 mL [140]. The risks specific to iodinated contrast induce allergic and nonidiosyncratic reactions. Nonidiosyncratic reactions reflect the physiologic effects of contrast media and direct organ toxicity, such as contrast-induced nephropathy, which is estimated to occur in 2 to 7 percent of all patients undergoing CT scans with contrast. Iodinated contrast agents cross the placenta, although clinical sequelae from brief exposures have not been reported. Breastfeeding is considered reasonable, as less than 1 percent of the contrast dose is excreted into breast milk and only 1 percent of the amount ingested actually absorbed from the infant's gastrointestinal tract, but mothers should be informed of the theoretical risks of toxicity or allergic reaction [141,142]. (See "Diagnosis and treatment of an acute reaction to a radiologic contrast agent" and "Prevention of contrast-associated acute kidney injury related to angiography" and "Diagnostic imaging in pregnant and lactating patients" and "Diagnostic imaging in pregnant and lactating patients", section on 'Use of iodinated contrast materials'.)

RADIONUCLIDE VENTRICULOGRAPHY — Radionuclide ventriculography (RVG) is performed using a radioactive tracer to label the patient’s red blood cell pool and a gamma-ray camera to capture images of the blood circulating through the heart. Other names have been used to describe this type of blood pool imaging, including radionuclide angiography (RNA), radionuclide cine angiography (RNCA), multiple gated cardiac blood pool imaging (MUGA), and equilibrium radionuclide angiography (ERNA).

Assessment of LVEF by RVG is based on the principle that changes in radiotracer count density are proportional to LV volume changes. An RVG scan can be performed as first-pass (immediately following injection) or equilibrium (approximately 20 to 30 minutes after injection) radionuclide angiocardiography. Planar images are usually obtained in three standard cardiac views (anterior, left anterior oblique, and septal) that are reviewed and analyzed to provide quantitative measurements of ventricular volumes (both right and left) and LVEF. Tomographic ERNA using electrocardiographic (ECG)-gated blood pool single photon emission CT (SPECT) allows for evaluation of LV volume and function without any geometric assumptions.

Indications/clinical use — RVG provides a detailed and accurate assessment of LVEF, with the most common indication being the quantification of ventricular volumes and LVEF when this diagnostic information is unreliable or unavailable by other imaging modalities. RVG is one of the main modalities used for baseline and serial assessment of ventricular function in patients undergoing chemotherapy with cardiotoxic agents such as anthracyclines [143]. (See "Cardiotoxicity of cancer chemotherapy agents other than anthracyclines, HER2-targeted agents, and fluoropyrimidines" and "Cardiotoxicity of trastuzumab and other HER2-targeted agents" and "Clinical manifestations, diagnosis, and treatment of anthracycline-induced cardiotoxicity".)

Function parameters — Global LV systolic function can be graded qualitatively using visual assessment of images acquired throughout the cardiac cycle and displayed as a cine image loop. For quantification of LVEF, radioactivity counts within the LV cavity acquired over multiple cardiac cycles are isolated using automated, semi-automated, or manual edge detection post-processing software, often using the left anterior oblique projection, and measured at end-diastole and end-systole. LVEF is calculated using the LVEF formula substituting net counts for volume measurements.

Regional function is commonly evaluated using visual assessment of wall motion for each LV segment visualized by standard views using cine images. Parametric phase and amplitude images and principal components analysis or factor analyses are additional post-processing methods that are commonly used to supplement visual analysis of the cine images [144,145].

Diagnostic performance — RVG is an accurate and highly reproducible [146-150] technique for quantitative assessment of LVEF, particularly when compared with visual estimation with 2D TTE [11]. RVG quantification of LVEF has also been shown to be reliable for the detection of relatively small changes in LVEF [11].

RVG scans for LVEF measurements demonstrate diagnostic performance similar to TTE [151], but are considered less accurate when compared with CMR. Average relative errors in LVEF range from 7 to 22 percent when compared with CMR [152,153]. Intra and inter-reader reproducibility is high, with limits of agreement of 1.8 to 3.6 percent [11]. Identification of regional wall motion abnormalities with RVG is considered fairly accurate with moderate to good agreement with CMR for identifying segments with abnormal wall motion and/or abnormal wall thickening (73 to 88 percent accuracy depending on method) [153].

Advantages — RVG scans provide accurate results that are highly reproducible and capable of detecting subtle changes in cardiac function. Good diagnostic performance is achieved with standard protocols that minimize the risk of user dependent errors in acquiring and interpreting images.

Limitations and trouble shooting — RVG scans require intravenous access for administration of the radioactive tracer. Hence, patients with poor intravenous access may have difficulty with exam performance. In addition, as with other exams that utilize radioactive tracers, there is associated radiation exposure estimated to be approximately 6 to 7 mSv [10].

Unlike other diagnostic tests of LV systolic function, the scope of information provided by RVG scans is generally limited to assessment of ventricular size and function, with limited assessment of valvular and LV myocardial disease. Thus, if additional information on other cardiac structures or characterization of the myocardium is needed, additional testing with other modalities will generally be preferred.

Image acquisition is gated to the patient’s ECG, assumes a regular R-R interval, and is averaged over several cardiac cycles. Therefore, arrhythmias resulting in irregular cardiac cycles, including atrial fibrillation or frequent ectopic beats, will reduce the accuracy of LVEF calculations [144]. However, beat-rejection software is standard with programs performing RVG, such that most arrhythmia beats are removed from analysis and therefore do not interfere with accuracy, unless the arrhythmia is poorly controlled (eg, atrial fibrillation with widely varying R-R intervals).

As with any test, the quality of the study needs to be taken into account when assessing the accuracy of the data. In one study, up to 14 percent of RVG scans were of poor quality [11]. Poor labeling of the red blood cells can lead to poor count rates in the blood pool, which may lead to difficulty identifying the border between cavity and myocardium due to low target to background ratio of tracer counts [66]. Variability in the heart position and rotation in the standard views may lead to misinterpretation of the images. For example, patients with severe obstructive airway disease, pulmonary hypertension, or congenital heart disease may have significant variability in heart position, and if not corrected at the time of image acquisition, this may lead to misinterpretation of the ventricular function and segmental function due to chamber overlap or misidentification of segments or both. In some cases, the chamber overlap may be so severe that the LV cannot be assessed at all.

Complications — The radioactive tracer used in RVG scans is generally safe and excreted through the kidneys within 24 hours. Administration is contraindicated during pregnancy or nursing. As with any test that involves radiation exposure, consideration of the benefit must be weighed against the risk as cumulative exposure over an extended period of time is potentially harmful. (See "Radiation dose and risk of malignancy from cardiovascular imaging", section on 'Balancing risks and benefits'.)

SINGLE-PHOTON EMISSION COMPUTED TOMOGRAPHY MYOCARDIAL PERFUSION IMAGING (SPECT-MPI) — SPECT-MPI is commonly used for assessment of ischemic heart disease. Routine comprehensive analysis of SPECT-MPI generally includes evaluation of LV volumes and LVEF on the reconstructed 3D data sets using commercially available software programs that automatically define the border between the LV myocardium and the LV cavity. (See "Overview of stress radionuclide myocardial perfusion imaging", section on 'SPECT imaging'.)

Indications/clinical use — SPECT-MPI is used for diagnostic evaluation of symptomatic patients with known or suspected coronary artery disease, generally with rest and stress imaging [154]. The addition of regional and global functional parameters to perfusion assessment permits improved discrimination between attenuation artifacts from true perfusion abnormalities and also adds to the prognostic information [155,156].

Function parameters — Global assessment of LV systolic function can be evaluated qualitatively and quantitatively. Visual qualitative assessment of LV function is made using cine loops encompassing the entire cardiac cycle. For quantification of LVEF, volume-based rather than count-based methods are used. The LV cavity borders are identified using automated, semi-automated, or manual contouring of the endocardial border, with generation of time-volume curves over the entire cardiac cycle. End-diastolic and end-systolic LV cavity volumes are measured for calculation of LVEF [38,157].

Regional evaluation of LV systolic function relies on visual assessment of the LV for presence of regional wall motion abnormalities using a six-point system (0 = normal, 1 = mild hypokinesis, 2 = moderate hypokinesis, 3 = severe hypokinesis, 4 = akinesis, and 5 = dyskinesis). Wall motion analysis is performed by visualizing the endocardial edge of the LV, and many commercially available programs allow readers to utilize a 3D representation of the function data. Visual assessment of wall thickening or apparent increase in the brightness of a wall during the cardiac cycle give another parameter that can be used for evaluation of regional function [156]. The degree of wall thickening can be scored with a four-point system (0 = normal to 3 = absent thickening). Quantitative assessment of degree of regional wall motion and wall thickening can be measured using commercially available software, and can augment the visual assessment of regional function [158,159].

Diagnostic performance — There are several different quantitative algorithm programs available for the assessment of LV systolic function that have been validated against other imaging modalities, including 2D and 3D radionuclide ventriculography, CMR, 2D and 3D transthoracic echocardiography, and invasive left ventriculography, depending on the vendor and parameter of interest [156]. Meta-analyses comparing electrocardiographic (ECG)-gated SPECT-MPI for evaluation of LV volumes and LVEF have demonstrated overall high correlation with CMR, with correlation coefficient values of 0.89 to 0.92 for ventricular volumes and 0.87 for LVEF, mean weighted difference in LVEF 3 percent with limits of agreement of -6.2 to 12.2 percent [50,160]. However, substantial differences may be seen on an individual basis, with deviations in LVEF of at least 5 percent in approximately 50 percent of patients, and at least 10 percent in approximately 25 percent of patients when compared with CMR [160].

In general, ECG-gated SPECT-MPI demonstrates good to high reproducibility for visual and automated evaluation of wall motion and wall thickening, for both global and regional measures of function (coefficient values 0.72 to 0.94). Global assessment of LVEF showed the best reproducibility with correlation coefficient values of 0.87 to 0.97 with differences of 3 to 8 percent between measures [159,161-163]. Automated measures of wall motion demonstrated repeatability coefficient <2 mm and wall thickening of <10.4 percent [159]. LV volumes demonstrate reproducibility and repeatability of 7.5 to 10 mL between measures [164].

Advantages — Gated SPECT-MPI can evaluate LV perfusion and function within a single study. The functional status of hypoperfused or normally perfused areas is important for clinical management, making this test an important method for evaluation of ischemic heart disease [165]. Automated image processing for quantification of LV volumes and function have enhanced practical interpretation in the clinical setting and improved reproducibility of the measurements [166,167].

Limitations and trouble shooting — Gated SPECT-MPI scans require intravenous access for administration of the radioactive tracer. There is an associated radiation exposure of 8 to 22 mSv, depending on the dose and radiotracer used and the associated perfusion imaging protocol, although the effective dose is decreasing over time [10]. Gated SPECT-MPI studies provide relatively lower spatial resolution compared with other techniques and are primarily used for evaluation of myocardial perfusion for assessment of coronary heart disease. There is also limited assessment of valvular heart disease. Thus, if additional information on other cardiac structures is needed, additional testing with other modalities will often be required.

Several factors may interfere with interpretation of SPECT images. Attenuation can affect image quality, and each wall location has different attenuation characteristics [168,169]. The presence of hypoperfused regions can lead to decreased tracer counts that limit detection and contouring of myocardial boundaries, which may limit the assessment of ventricular volumes and LVEF [170].

Since imaging data are usually acquired over multiple cardiac cycles with ECG gating, arrhythmias and ectopic beats may degrade image quality. In patients who have perfusion defects from prior infarcts or attenuation artifacts, detection of the myocardial border may be limited, which may lead to inaccurate quantification of LV volumes and LVEF [171]. Patients with small LV cavities may have underestimation of their LV volumes, particularly at end-systolic phases due to relatively poor resolution compared with LV wall thickness and increase in myocardial count density with contraction with resultant overestimation of LVEF due to underestimation of LV end-systolic volume [29,172-175]. The LVEF tends to be overestimated in women due to their smaller LV cavities. Variability in the calculated LVEF may be seen due to differences in the border detection technique with different vendors. Finally, depending on the study protocol, true physiological changes may be seen if the "resting" LVEF is obtained 15 to 60 minutes following the stress portion of the test due to prolonged stunning [176]. This could also be seen as an advantage, as data suggests differences between rest and stress ejection fraction are indicative of high grade stenosis at cardiac catheterization.

Complications — Gated SPECT-MPI scans are considered safe, and complications are most commonly related to the associated stress portion of the examination and long-term risk of radiation exposure. Technetium 99m sestamibi has been rarely associated with allergic and anaphylactic reactions, including angioedema and generalized urticarial.

In general, SPECT-MPI studies are not performed in pregnant patients due to the risk of fetal exposure to radioactivity, and breastfeeding mothers are often instructed to pump and discard breast milk for 12 hours after the examination, although no official recommendations have been published.

As with any test that involves radiation exposure, consideration of the benefit must be weighed against the risk as cumulative exposure over an extended period of time is potentially harmful. (See "Radiation dose and risk of malignancy from cardiovascular imaging", section on 'Balancing risks and benefits'.)

POSITRON EMISSION TOMOGRAPHIC MYOCARDIAL PERFUSION IMAGING (PET-MPI) — Cardiac PET imaging is an alternative to single photon emission CT MPI (SPECT-MPI) when the technology is available. There are several advantages to PET such as high diagnostic accuracy and excellent image quality for the assessment of ischemic heart disease [177]. PET-MPI routine procedures include evaluation of LV volumes and LVEF, commonly at both rest and stress conditions.

Indications/clinical use — PET-MPI is an alternative to SPECT-MPI for the diagnosis and risk stratification of symptomatic patients with known or suspected coronary artery disease, generally with rest and stress pharmacologic imaging [177]. The addition of myocardial blood flow assessment increases the diagnostic accuracy and provides identification of nonepicardial coronary artery disease [178]. Cardiac PET routinely uses either radionuclide line source or CT attenuation correction to reduce the impact of attenuation artifact and improve specificity.

Function parameters — Global and regional assessment of LV systolic function using PET-MPI can be evaluated similarly to SPECT-MPI using both visual interpretation and computerized quantitation. Unlike SPECT-MPI, ventricular function from PET-MPI is routinely captured at both resting and stress conditions. Visual qualitative assessment of LV function is made using cine loops encompassing the entire cardiac cycle, generally 8 to 16 frames per study. For quantification of LVEF, volume-based rather than count-based methods are used. The LV cavity borders are identified using automated, semi-automated, or manual contouring of the endocardial border, with generation of time-volume curves over the entire cardiac cycle. End-diastolic and end-systolic LV cavity volumes are measured for calculation of LVEF [38,157]. Similar software is available for PET-MPI and SPECT-MPI. In addition, the change in LVEF between rest and stress provides diagnostic and prognostic information [179,180]. In one study, gated PET-MPI ejection fraction provided incremental and complimentary information to PET-MPI alone (without gating) in the same patient [181].

Diagnostic performance — Cardiac gated PET-MPI is based upon the same concepts as SPECT-MPI. However, fewer comparisons to other imaging modalities have been performed, and the assumption is that the same data are being provided, particularly at rest. A study found that PET-MPI and SPECT-MPI provided similar information [182].

Advantages — Gated PET-MPI can evaluate LV perfusion and function within a single study at both rest and hyperemic stress conditions. While cardiac PET-MPI is not indicated for evaluation of ventricular function alone, the information provided does have both diagnostic and prognostic value for both ischemic and nonischemic conditions.

Limitations and troubleshooting — Gated PET-MPI LV function assessment provides no additional risk to the patient already undergoing PET-MPI. However, the cardiac PET-MPI study does require administration of a radioactive tracer, with a radiation exposure of 2 to 6 mSv, substantially lower than SPECT-MPI. As this technique is primarily that of measuring perfusion, the spatial resolution does not provide many of the other aspects routinely assessed with other modalities such as echocardiography or CMR imaging.

Since imaging data are usually acquired over multiple cardiac cycles with electrocardiographic gating, arrhythmias and ectopic beats may degrade image quality. A disadvantage of gated PET-MPI is that if an arrhythmia occurs that causes the study to be stopped, the perfusion data are also affected and cannot be repeated due to the short half-life of the tracers.

Complications — Gated PET-MPI scans are considered safe, and complications are most commonly related to the associated stress portion of the examination.

As with any test that involves radiation exposure, consideration of the benefit must be weighed against the risk as cumulative exposure over an extended period of time is potentially harmful. Radiation exposure from gated PET-MPI is substantially lower than a typical SPECT-MPI study. (See "Radiation dose and risk of malignancy from cardiovascular imaging", section on 'Balancing risks and benefits'.)

MANAGEMENT OF DISCREPANT RESULTS — Careful clinical judgment should be exercised when tests evaluating LV systolic function yields unexpected or discrepant results. Possible causes include technical difficulties (eg, suboptimal image quality or image processing), inaccurate interpretation, or development of an unexpected clinical condition. Poor image quality and/or technical difficulties encountered during imaging are the most common reasons for inaccurate measurement of LV systolic function. Repeat imaging of the same or a different method or re-interpretation of the original image data may be required. In addition, when clinical circumstances and conditions are inconsistent with the reported ejection fraction, re-measurement by another method may be required to resolve the discrepancy [183].

SOCIETY GUIDELINE LINKS — Links to society and government-sponsored guidelines from selected countries and regions around the world are provided separately. (See "Society guideline links: Multimodality cardiovascular imaging appropriate use criteria".)

SUMMARY AND RECOMMENDATIONS

●Assessment of left ventricular (LV) systolic function plays an important role in the diagnosis, management, and prognosis of patients with a wide variety of cardiovascular disease. (See 'Indications/clinical use' above.)

●Measurement of LV ejection fraction (LVEF) represents the most commonly reported measure of LV systolic performance. Parameters that directly measure myocardial contractility, such as strain and strain rate, are more sensitive, reliable, and reproducible, with growing clinical applications. (See 'Measures of left ventricular systolic function' above.)

●Two-dimensional transthoracic echocardiography (TTE) is the most commonly used initial imaging modality due to widespread availability, portability, and negligible risk. Other tests are indicated if information obtainable from another test is needed for patient care or if TTE is nondiagnostic. (See 'Choice of test' above.)

●When TTE is not available or is of suboptimal image quality, radionuclide ventriculography or cardiovascular magnetic resonance (CMR) should be considered. (See 'Choice of test' above.)

●Patients who require serial evaluations of LV systolic function should undergo a noninvasive, minimal risk test that has high reproducibility and sensitivity since small changes in measurement may affect management. When serial testing is anticipated, use of a consistent imaging modality is preferred since numerous studies have shown that measurement of LVEF by various modalities are not interchangeable. Volumetric/three-dimensional methods may also be preferred for serial evaluation and detection of small changes. (See 'Choice of test' above.)

●Careful clinical judgement should be exercised when a test evaluating LV systolic function yields unexpected or discrepant results. Possible causes include technical difficulties (eg, suboptimal image quality or image processing), inaccurate interpretation, or development of an unexpected clinical condition. (See 'Management of discrepant results' above.)