Optimización de la interpretación de ecocardiogramas con la utilización de Machine Learning en el estudio WASE / Optimizing Echocardiograms Interpretation with the Application of Machine Learning in the WASE Study

RESUMEN Introducción: El número creciente de estudios ecocardiográficos y la necesidad de cumplir rigurosamente con las recomendaciones de guías internacionales de cuantificación, ha llevado a que los cardiólogos deban realizar tareas sumamente extensas y repetitivas, como parte de la interpretación y análisis de cantidades de información cada vez más abrumadoras. Novedosas técnicas de machine learning (ML), diseñadas para reconocer imágenes y realizar mediciones en las vistas adecuadas, están siendo cada vez más utilizadas para responder a esta necesidad evidente de automatización de procesos. Objetivos: Nuestro objetivo fue evaluar un modelo alternativo de interpretación y análisis de estudios ecocardiográficos, basado fundamentalmente en la utilización de software de ML, capaz de identificar y clasificar vistas y realizar mediciones estandarizadas de forma automática. Material y métodos: Se utilizaron imágenes obtenidas en 2000 sujetos normales, libres de enfermedad, de los cuales 1800 fueron utilizados para desarrollar los algoritmos de ML y 200 para su validación posterior. Primero, una red neuronal convolucional fue desarrollada para reconocer 18 vistas ecocardiográficas estándar y clasificarlas de acuerdo con 8 grupos (stacks) temáticos. Los resultados de la identificación automática fueron comparados con la clasificación realizada por expertos. Luego, algoritmos de ML fueron desarrollados para medir automáticamente 16 parámetros de eco Doppler de evaluación clínica habitual, los cuales fueron comparados con las mediciones realizadas por un lector experto. Finalmente, comparamos el tiempo necesario para completar el análisis de un estudio ecocardiográfico con la utilización de métodos manuales convencionales, con el tiempo necesario con el empleo del modelo que incorpora ML en la clasificación de imágenes y mediciones ecocardiográficas iniciales. La variabilidad inter e intraobservador también fue analizada. Resultados: La clasificación automática de vistas fue posible en menos de 1 segundo por estudio, con una precisión de 90 % en imágenes 2D y de 94 % en imágenes Doppler. La agrupación de imágenes en stacks tuvo una precisión de 91 %, y fue posible completar dichos grupos con las imágenes necesarias en 99% de los casos. La concordancia con expertos fue excelente, con diferencias similares a las observadas entre dos lectores humanos. La incorporación de ML en la clasificación y medición de imágenes ecocardiográficas redujo un 41 % el tiempo de análisis y demostró menor variabilidad que la metodología de interpretación convencional. Conclusión: La incorporación de técnicas de ML puede mejorar significativamente la reproducibilidad y eficiencia de las interpretaciones y mediciones ecocardiográficas. La implementación de este tipo de tecnologías en la práctica clínica podría resultar en reducción de costos y aumento en la satisfacción del personal médico.

ABSTRACT Background: The growing number of echocardiographic tests and the need for strict adherence to international quantification guidelines have forced cardiologists to perform highly extended and repetitive tasks when interpreting and analyzing increasingly overwhelming amounts of data. Novel machine learning (ML) techniques, designed to identify images and perform measurements at relevant visits, are becoming more common to meet this obvious need for process automation. Objectives: Our objective was to evaluate an alternative model for the interpretation and analysis of echocardiographic tests mostly based on the use of ML software in order to identify and classify views and perform standardized measurements automatically. Methods: Images came from 2000 healthy subjects, 1800 of whom were used to develop ML algorithms and 200 for subsequent validation. First, a convolutional neural network was developed in order to identify 18 standard echocardiographic views and classify them based on 8 thematic groups (stacks). The results of automatic identification were compared to classification by experts. Later, ML algorithms were developed to automatically measure 16 Doppler scan parameters for regular clinical evaluation, which were compared to measurements by an expert reader. Finally, we compared the time required to complete the analysis of an echocardiographic test using conventional manual methods with the time needed when using the ML model to classify images and perform initial echocardiographic measurements. Inter- and intra-observer variability was also analyzed. Results: Automatic view classification was possible in less than 1 second per test, with a 90% accuracy for 2D images and a 94% accuracy for Doppler scan images. Stacking images had a 91% accuracy, and it was possible to complete the groups with any necessary images in 99% of cases. Expert agreement was outstanding, with discrepancies similar to those found between two human readers. Applying ML to echocardiographic imaging classification and measurement reduced time of analysis by 41% and showed lower variability than conventional reading methods. Conclusion: Application of ML techniques may significantly improve reproducibility and efficiency of echocardiographic interpretations and measurements. Using this type of technologies in clinical practice may lead to reduced costs and increased medical staff satisfaction.

Deep learning assisted measurement of echocardiographic left heart parameters: improvement in interobserver variability and workflow efficiency.

Machine learning techniques designed to recognize views and perform measurements are increasingly used to address the need for automation of the interpretation of echocardiographic images. The current study was designed to determine whether a recently developed and validated deep learning (DL) algorithm for automated measurements of echocardiographic parameters of left heart chamber size and function can improve the reproducibility and shorten the analysis time, compared to the conventional methodology. The DL algorithm trained to identify standard views and provide automated measurements of 20 standard parameters, was applied to images obtained in 12 randomly selected echocardiographic studies. The resultant measurements were reviewed and revised as necessary by 10 independent expert readers. The same readers also performed conventional manual measurements, which were averaged and used as the reference standard for the DL-assisted approach with and without the manual revisions. Inter-reader variability was quantified using coefficients of variation, which together with analysis times, were compared between the conventional reads and the DL-assisted approach. The fully automated DL measurements showed good agreement with the reference technique: Bland-Altman biases 0-14% of the measured values. Manual revisions resulted in only minor improvement in accuracy: biases 0-11%. This DL-assisted approach resulted in a 43% decrease in analysis time and less inter-reader variability than the conventional methodology: 2-3 times smaller coefficients of variation. In conclusion, DL-assisted approach to analysis of echocardiographic images can provide accurate left heart measurements with the added benefits of improved reproducibility and time savings, compared to conventional methodology.

A Novel Approach for Semiautomated Three-Dimensional Quantification of Mitral Regurgitant Volume Reflects a More Physiologic Approach to Mitral Regurgitation.

BACKGROUND: Quantification of mitral regurgitation (MR) by echocardiography is integral to assessing lesion severity and entails the integration of multiple Doppler-based parameters. These methods are founded primarily upon the principle of proximal isovelocity surface area (PISA), a two-dimensional (2D) method known to involve several assumptions regarding MR jet characteristics. The authors analyzed the results of a semiautomated method of three-dimensional (3D)-based regurgitant volume (RVol) estimation that accounts for jet behavior throughout the cardiac cycle and compared it with conventional 2D PISA methods for MR quantification. METHODS: A total of 50 patients referred for transesophageal echocardiography for evaluation of primary (n = 25) and secondary (n = 25) MR were included for analysis. Three-dimensional full-volume color data sets were acquired, along with standard 2D methods for PISA calculation. A 3D semiautomated MR flow quantification algorithm was applied offline to calculate 3D RVol, with simultaneous temporal curves generated from the 3D data set. Three-dimensional RVol was compared with 2D RVol. Three-dimensional vena contracta area was also performed in all cases. RESULTS: There was a modest correlation between 2D RVol and 3D RVol (r = 0.60). The semiautomated 3D approach resulted in significantly lower values of RVol compared with 2D PISA. Real-time and dynamic flow curve patterns were used for integral estimates of 3D RVol over the cardiac cycle, with a distinct bimodal pattern in functional MR and a brief and solitary peak in primary MR. CONCLUSIONS: Using a semiautomated 3D software for the quantification of MR allows the simultaneous calculation of 3D RVol with an automated generation of dynamic flow curves characteristic of the underlying MR mechanism. The present flow curve pattern results highlight well-known differences between MR flow dynamics in degenerative MR compared with functional MR.

Prevalence of newly diagnosed sarcoidosis in patients with ventricular arrhythmias: a cardiac magnetic resonance and 18F-FDG cardiac PET study.

Cardiac sarcoidosis (CS) is known to be associated with ventricular tachycardia (VT); however, most investigations to date have focused on patients with known extra-cardiac sarcoidosis. The presence of CS is typically evaluated using 18F-fluorodeoxyglucose (18F-FDG) uptake on cardiac positron emission tomography (PET) or late gadolinium enhancement (LGE) on cardiac magnetic resonance (CMR). In this study, we sought to determine the prevalence of primary CS and the relationship between myocardial 18F-FDG uptake and LGE in patients with VT without known sarcoidosis. We retrospectively identified 67 patients without known sarcoidosis or active ischemic heart disease (i.e. significant ischemic disease that had not been previously revascularized) referred for both CMR and PET for evaluation of VT. Standard cine- and LGE- CMR and cardiac PET protocols were used. Myocardial LGE was defined as signal intensity > 5 SDs above the mean signal intensity of normal myocardium. Cardiac PET images were considered positive if there was focal myocardial 18F-FDG uptake having greater activity than the left ventricular blood pool. 45 patients (67%) had LGE, while only 4 (6%) had myocardial FDG uptake. Nine percent of patients with LGE had FDG-uptake while none without LGE did, and 10% of the cohort had indeterminate FDG uptake presumably from poor dietary preparation. Of those with both FDG uptake and LGE, 3/4 ultimately received a clinical diagnosis of CS. 4.5% of patients without previously known sarcoidosis or active ischemic heart disease presenting with VT have newly diagnosed CS. Detection of CS can be increased using a CMR first approach followed by cardiac PET for patients with non-ischemic LGE.

Validation of non-contrast multiple overlapping thin-slab 4D-flow cardiac magnetic resonance imaging.

BACKGROUND: Cardiac magnetic resonance (CMR) flow quantification is typically performed using 2D phase-contrast (PC) imaging of a plane perpendicular to flow. 3D-PC imaging (4D-flow) allows offline quantification anywhere in a thick slab, but is often limited by suboptimal signal, potentially alleviated by contrast enhancement. We developed a non-contrast 4D-flow sequence, which acquires multiple overlapping thin slabs (MOTS) to minimize signal loss, and hypothesized that it could improve image quality, diagnostic accuracy, and aortic flow measurements compared to non-contrast single-slab approach. METHODS: We prospectively studied 20 patients referred for transesophageal echocardiography (TEE), who underwent CMR (GE, 3 T). 2D-PC images of the aortic valve and three 4D-flow datasets covering the heart were acquired, including single-slab, pre- and post-contrast, and non-contrast MOTS. Each 4D-flow dataset was interpreted blindly for ≥moderate valve disease and compared to TEE. Flow visualization through each valve was scored (0 to 4), and aortic-valve flow measured on each 4D-flow dataset and compared to 2D-PC reference. RESULTS: Diagnostic quality visualization was achieved with the pre- and post-contrast 4D-flow acquisitions in 25% and 100% valves, respectively (scores 0.9 ± 1.1 and 3.8 ± 0.5), and in 58% with the non-contrast MOTS (1.6 ± 1.1). Accuracy of detection of valve disease was 75%, 92% and 82%, respectively. Aortic flow measurements were possible in 53%, 95% and in 89% patients, respectively. The correlation between pre-contrast single-slab measurements and 2D-PC reference was weak (r = 0.21), but improved with both contrast enhancement (r = 0.71) and with MOTS (r = 0.67). CONCLUSIONS: Although non-contrast MOTS 4D-flow improves valve function visualization and diagnostic accuracy, a significant proportion of valves cannot be accurately assessed. However, aortic flow measurements using non-contrast MOTS is feasible and reaches similar accuracy to that of contrast-enhanced 4D-flow.

Refining Severe Tricuspid Regurgitation Definition by Echocardiography with a New Outcomes-Based "Massive" Grade.

BACKGROUND: Current echocardiographic guidelines recommend that tricuspid regurgitation (TR) severity be graded in three categories, following assessment of specific parameters. Findings from recent trials have shown that the severity of TR frequently far exceeds the current definition of severe. We postulated that a grading approach that emphasizes outcomes could be useful to identify patients with severe TR at increased risk of mortality. METHODS: We identified 284 patients with echocardiograms demonstrating severe functional TR, defined as vena contracta (VC) ≥ 0.7 cm. Demographics and mortality data were obtained from the medical records. Patients were divided into study (n = 122 patients with three-dimensional images) and validation (n = 162) cohorts. The VC was measured in both the right ventricular (RV) inflow and apical four-chamber views and averaged. For the study cohort, tricuspid annular, RV end-diastolic (basal, mid, long axis) dimensions, tricuspid leaflet tenting height and area, RV free-wall longitudinal strain, and RV volumes were measured from two- and three-dimensional data sets. A K-partition algorithm was used in the study cohort to derive a mortality-related cutoff VC value, above which TR was termed "massive." The ability of this VC cutoff to identify patients at greater mortality risk was then tested in the validation cohort using Kaplan-Meier survival analysis. RESULTS: In the study cohort, VC > 0.92 cm (massive TR) was optimally associated with worse survival. Tricuspid annular and RV size were larger in the massive group (P < .05), while there were no significant differences in demographics between the TR groups. Importantly, in the independent validation cohort, the above VC cutoff also correlated with increased mortality in the massive group (log-rank P < .05). CONCLUSIONS: Among patients traditionally defined as having severe TR, a subset exists with massive TR, resulting in greater adverse RV remodeling and increased mortality. These patients may derive the greatest benefit from emerging percutaneous therapies.

Stabbed Through The Heart: An Unusual VSD Presentation.

Traumatic ventricular septal defects due to penetrating cardiac injury are uncommon. Transthoracic echocardiography is an essential tool in diagnosis. Options for closure include either surgical or percutaneous repair. We present a case of a trauma-related ventricular septal defect in a young patient that was successfully repaired by using a percutaneous approach. (Level of difficulty: Intermediate.).

Progression of aortic stenosis and echocardiographic criteria for its severity.

AIMS: Guidelines-recommended criteria for identifying severe aortic stenosis (AS) are based on small, homogenous cohorts of patients, leading to potentially inconsistent or missed diagnosis. We used a large cohort of patients with varying degrees of AS to (i) characterize its progression; (ii) evaluate the influence of demographic and echocardiographic variables; and (iii) derive haemodynamically consistent cut-off values. METHODS AND RESULTS: We identified 916 patients with mild to severe AS who had undergone >1 echocardiographic study (N = 2547). For each study, aortic valve area (AVA), peak transaortic velocity (Vmax), and mean pressure gradient (ΔP) were extracted. Annual rates of AVA change were determined by a linear mixed-effects model. To determine the prevalence of inconsistent diagnosis of severe AS, AVA was plotted against ΔP and Vmax, with quadrants defined using guidelines-recommended cut-offs. The rate of AVA change was -0.070 ± 0.003 cm2/year and was more rapid in men than women and in Whites than African Americans. AVA = 1 cm2 corresponded to ΔP = 32 mmHg and Vmax = 3.7 m/s, causing discrepancies in defining severe AS in 480 (19%) and 458 (18%) studies, respectively. Conversely, ΔP = 40 mmHg corresponded to AVA = 0.89 cm2 and Vmax = 4.0 m/s corresponded to AVA = 0.92 cm2, confirming the inconsistency of the guidelines. Notably, discrepancy rate was higher in 206 patients with low flow (SVi < 35 mL/m2): 40% vs. 16% in the remaining patients. CONCLUSION: Our findings demonstrated gender- and race-related differences in AS progression and underscored the need to refine the multiparametric criteria for diagnosis of severe AS to minimize internal inconsistencies, which are high with the current cut-offs and amplified in patients with low stroke volumes.

Prevalence of mitral annular disjunction in patients with mitral valve prolapse and severe regurgitation.

Mitral annular disjunction (MAD) is routinely diagnosed by cardiac imaging, mostly by echocardiography, and shown to be a risk factor for ventricular arrhythmias. While MAD is associated with mitral valve (MV) prolapse (MVP), it is unknown which patients with MAD are at higher risk and which additional imaging features may help identify them. The value of cardiac computed tomography (CCT) for the diagnosis of MAD is unknown. Accordingly, we aimed to: (1) develop a standardized CCT approach to identify MAD in patients with MVP and severe mitral regurgitation (MR); (2) determine its prevalence and identify features that are associated with MAD in this population. We retrospectively studied 90 patients (age 63 ± 12 years) with MVP and severe MR, who had pre-operative CCT (256-slice scanner) of sufficient quality for analysis. The presence and degree of MAD was assessed by rotating the view plane around the MV center to visualize disjunction along the annulus. Additionally, detailed measurements of MV apparatus and left heart chambers were performed. Univariate logistic regression analysis was performed to determine which parameters were associated with MAD. MAD was identified in 18 patients (20%), and it was typically located adjacent to a prolapsed or flail mitral leaflet scallop. Of these patients, 75% had maximum MAD distance > 4.8 mm and 90% > 3.8 mm. Female gender was most strongly associated with MAD (p = 0.04). Additionally, smaller end-diastolic mitral annulus area (p = 0.045) and longer posterior leaflet (p = 0.03) were associated with greater MAD. No association was seen between MAD and left ventricular size and function, left atrial size, and papillary muscle geometry. CCT can be used to readily detect MAD, by taking advantage of the 3D nature of this modality. A significant portion of MVP patients referred for mitral valve repair have MAD. The presence of MAD is associated with female gender, smaller annulus size and greater posterior leaflet length.

Measurement errors in serial echocardiographic assessments of aortic valve stenosis severity.

Transthoracic echocardiography (TTE) evaluation of aortic stenosis (AS) is routinely performed using the continuity equation. Inaccurate measurements of the left ventricular (LV) outflow tract (LVOT) diameter are considered the most common source of error in AS grading. We hypothesized that inconsistency in LVOT velocity time integral (VTI) is an under-recognized cause of AS assessment error. We sought to determine which parameters contribute most towards inconsistencies in AS grading by studying the prevalence of different errors in a historic cohort. We identified patients with mild to severe AS with multiple studies from our database from 1994 to 2018 (n = 988 patients, 2859 studies). Errors were defined when: (1) LVOT diameter changed by > 2 mm, (2) LVOT VTI changed by > 15% without change in LV function from the initial TTE, (3) aortic valve (AV) maximum velocity (Vmax), mean pressure gradient (ΔP) or AV VTI decreased by > 15% without change in LV function from prior study. The most common error was the LVOT VTI measurement with 22% prevalence. LVOT diameter, AV VTI, AV Vmax and AV ΔP measurement caused errors in < 7% studies. Patients with normal LV function and more severe AS were more likely to have LVOT VTI errors (P < 0.05). LVOT VTI is a frequent, under-recognized source of error in assessing AS. Greater attention should be directed toward the proper positioning of the pulsed Doppler sample volume, particularly in patients with higher grades of AS and normal systolic function, to ensure accurate and reproducible assessment of AS.

Hemodynamic impact of coronary stenosis using computed tomography: comparison between noninvasive fractional flow reserve and 3D fusion of coronary angiography with stress myocardial perfusion.

Vasodilator-stress CT perfusion imaging in addition to CT coronary angiography (CTCA) may provide a single-test alternative to nuclear stress testing, commonly used to assess hemodynamic significance of stenosis. Another alternative is fractional flow reserve (FFR) calculated from cardiac CT images. We studied the concordance between these two approaches and their relationship to outcomes. We prospectively studied 150 patients with chest pain, who underwent CTCA and regadenoson CT. CTCA images were interpreted for presence and severity of stenosis. Fused 3D displays of subendocardial X-ray attenuation with coronary arteries were created to detect stress perfusion defects (SPD) in each coronary territory. In patients with stenosis > 25%, CT-FFR was quantified. Significant stenosis was determined by: (1) combination of stenosis > 50% with an SPD, (2) CT-FFR ≤ 0.80. Patients were followed-up for 36 ± 25 months for death, myocardial infarction or revascularization. After excluding patients with normal arteries and technical/quality issues, in final analysis of 76 patients, CTCA depicted stenosis > 70% in 13/224 arteries, 50-70% in 24, and < 50% in 187. CT-FFR ≤ 0.80 was found in 41/224 arteries, and combination of SPD with > 50% stenosis in 31/224 arteries. Inter-technique agreement was 89%. Despite high incidence of abnormal CT-FFR (30/76 patients), only 7 patients experienced adverse outcomes; 6/7 also had SPDs. Only 1/9 patients with CT-FFR ≤ 0.80 but normal perfusion had an event. Fusion of CTCA and stress perfusion can help determine the hemodynamic impact of stenosis in one test, in good agreement with CT-FFR. Adding stress CT perfusion analysis may help risk-stratify patients with abnormal CT-FFR.

Importance of the Left Atrium: More Than a Bystander?

Left atrial size and function parameters are associated with adverse outcomes in multiple disease states, including heart failure with reduced and preserved ejection fraction. Recent data suggest that phasic left atrial function and left atrial stain measurements also hold prognostic information. Three-dimensional echocardiography provides more accurate and reproducible quantification of left atrial volumes than 2-dimensional echocardiography when compared with cardiac magnetic resonance reference standards. Greater accessibility to these advanced imaging techniques allows for the integration of these parameters into routine clinical practice.

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

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

Atrial-focused views improve the accuracy of two-dimensional echocardiographic measurements of the left and right atrial volumes: a contribution to the increase in normal values in the guidelines update.

Current guidelines recommend that the atria be measured in 2D echocardiographic (2DE) apical views using the method-of-disks (MOD) or area-length (AL) technique as an alternative, although no definitive data exists that these are interchangeable. However, standard apical views maximize the long-axis of the left ventricle, rather than the dimensions of the atria, resulting in atrial foreshortening. We hypothesized that the increase in normal values of atrial volumes in the recent guidelines update was driven by data obtained using either the AL technique or dedicated atrial-focused views, which maximize the longitudinal dimension of the atria and thus provide larger volumes than the MOD measurements in standard apical views. We prospectively studied 30 patients (Philips iE33) to compare 2DE measurements of left and right atrial volumes (LAV, RAV) using the MOD and AL techniques in standard and atrial-focused views, against 3D echocardiography (3DE) derived volumes (QLab) as a reference. Compared to standard views, atrial-focused views provided significantly larger MOD volumes for both atria, which were in better agreement with 3DE, as reflected by higher correlation coefficients (LAV: r = 0.95 vs. 0.89; RAV: r = 0.89 vs. 0.84), smaller biases (LAV: -1 ml vs. 7 ml; RAV: 3 ml vs. 7 ml) and tighter limits of agreement. This was also the case for the AL measurements, which were minimally larger than the MOD values (NS) for both atria. In conclusion, atrial-focused views are a more accurate alternative to standard apical views, which provides larger volumes. This finding can explain the increase in the normal values in the recent guidelines update, which was mostly driven by the use of atrial-focused views, rather than by the differences between MOD and AL techniques. This understanding is essential in order to correctly integrate the revised normal values into clinical practice.

Evaluation of apical subtype of hypertrophic cardiomyopathy using cardiac magnetic resonance imaging with gadolinium enhancement.

Apical hypertrophic cardiomyopathy (HC) is an uncommon variant of HC. We sought to characterize cardiac magnetic resonance imaging (MRI) findings among apical HC patients. This was a retrospective review of consecutive patients with a diagnosis of apical HC who underwent cardiac MRI examinations at the Mayo Clinic (Rochester, MN) from August 1999 to October 2011. Clinical and demographic data at the time of cardiac MRI study were abstracted. Cardiac MRI study and 2-dimensional echocardiograms performed within 6 months of the cardiac MRI were reviewed; 96 patients with apical HC underwent cardiac MRI examinations. LV end-diastolic and end-systolic volumes were 130.7 ± 39.1 ml and 44.2 ± 20.9 ml, respectively. Maximum LV thickness was 19 ± 5 mm. Hypertrophy extended beyond the apex into other segments in 57 (59.4%) patients. Obstructive physiology was seen in 12 (12.5%) and was more common in the mixed apical phenotype than the pure apical (19.3 vs 2.6%, p = 0.02). Apical pouches were noted in 39 (40.6%) patients. Late gadolinium enhancement (LGE) was present in 70 (74.5%) patients. LGE was associated with severe symptoms and increased maximal LV wall thickness. In conclusion, cardiac MRI is well suited for studying the apical form of HC because of difficulty imaging the cardiac apex with standard echocardiography. Cardiac MRI is uniquely suited to delineate the presence or absence of an apical pouch and abnormal myocardial LGE that may have implications in the natural history of apical HM. In particular, the presence of abnormal LGE is associated with clinical symptoms and increased wall thickness.