Automatic apical view classification of echocardiograms using a discriminative learning dictionary.

As part of striving towards fully automatic cardiac functional assessment of echocardiograms, automatic classification of their standard views is essential as a pre-processing stage. The similarity among three of the routinely acquired longitudinal scans: apical two-chamber (A2C), apical four-chamber (A4C) and apical long-axis (ALX), and the noise commonly inherent to these scans - make the classification a challenge. Here we introduce a multi-stage classification algorithm that employs spatio-temporal feature extraction (Cuboid Detector) and supervised dictionary learning (LC-KSVD) approaches to uniquely enhance the automatic recognition and classification accuracy of echocardiograms. The algorithm incorporates both discrimination and labelling information to allow a discriminative and sparse representation of each view. The advantage of the spatio-temporal feature extraction as compared to spatial processing is then validated. A set of 309 clinical clips (103 for each view), were labeled by 2 experts. A subset of 70 clips of each class was used as a training set and the rest as a test set. The recognition accuracies achieved were: 97%, 91% and 97% of A2C, A4C and ALX respectively, with average recognition rate of 95%. Thus, automatic classification of echocardiogram views seems promising, despite the inter-view similarity between the classes and intra-view variability among clips belonging to the same class.

Detection of small subendocardial infarction using speckle tracking echocardiography in a rat model.

BACKGROUND: It is challenging to detect small nontransmural infarcts visually or automatically. As it is important to detect myocardial infarction (MI) at early stages, we tested the hypothesis that small nontransmural MI can be detected using speckle tracking echocardiography (STE) at the acute stage. METHODS: Minimal nontransmural infarcts were induced in 18 rats by causing recurrent ischemia-reperfusion of the left anterior descending (LAD) coronary artery, followed by a 30-min ligation and by reperfusion. A week later, the scar size was measured by histological analysis. Each rat underwent three echocardiography measurements: at baseline, 1 day post-MI, and 1 week post-MI. To measure the peak circumferential strain (CS), peak systolic CS, radial strain (RS), and time-to-peak (TTP) of the CS, short-axis view of the apex was analyzed by a STE program. The TTP was normalized by the duration of the heart cycle to create percent change of heart cycle. RESULTS: Histological analysis after 1 week showed scar size of 4±6% at the anterior wall. At 24 h post-MI, the peak CS, peak systolic CS, and RS were reduced compared to baseline at the anterior wall due to the MI, and at the adjacent segments-the anterior septum and lateral wall, due to stunning (P<.05). However, only the anterior wall, the genuine damaged segment, showed prolonged TTP vs baseline (baseline 36%, 24 h 48%, P<.05). CONCLUSION: The TTP of the CS can distinguish between regions adjacent to MI (stunned or tethered) and MI, even in small nontransmural infarcts.

Optimization-Based Speckle Tracking Algorithm for Left Ventricle Strain Estimation: A Feasibility Study.

Speckle tracking echocardiography (STE) is a widespread method for calculating myocardial strains and estimating left ventricle function. Since echocardiographic clips are corrupted by speckle decorrelation noise, resulting in irregular, nonphysiological tissue displacement fields, smoothing is performed on the displacement data, affecting the strain results. Thus, strain results may depend on the specific implementations of 2-D STE, as well as other systems' characteristics of the various vendors. A novel algorithm (called K-SAD) is introduced, which integrates the physiological constraint of smoothness of the displacement field into an optimization process. Simulated B-mode clips, modeling healthy and abnormal cases, were processed by K-SAD. Peak global and subendocardial longitudinal strains, as well as regional strains, were calculated. In addition, 410 healthy subjects were also processed. The results of K-SAD are compared with those of one of the leading commercial product. K-SAD provides global mid-wall strain values, as well as subendocardial and regional strain values, all in good agreement with the ground-truth-simulated phantom data. K-SAD peak global longitudinal systolic strain values for 410 healthy subjects are quite similar for the different regions: - 17.02 ± 4.02%, - 19.00 ± 3.45%, and - 19.72 ± 5.06% at the basal, mid, and apical regions, respectively. Improved performance under noisy conditions was demonstrated by comparing a subgroup of 40 subjects with the best image quality with the remaining 370 cohort: K-SAD provides statistically similar global and regional results for the two cohorts. Our study indicates that the sensitivity of strain values to speckle noise, caused by the post block-matching weighted smoothing, can be significantly reduced and accuracy enhanced by employing an integrated one-stage, physiologically constrained optimization process.

Left Ventricular Mechanical Property Changes During Acute AV Synchronous Right Ventricular Pacing in Children.

Prolonged RV pacing is recognized as a cause of LV dysfunction due to dyssynchronous activation. There are no specific longitudinal parameters known to help predict RV pacing-induced LV dysfunction. The aim of the study was to assess the acute effects of AV synchronous RV pacing on LV mechanics using echocardiographic speckle tracking. Nineteen children, aged 6-23 years, underwent echocardiographic evaluation prior to and following elective electrophysiology and ablation studies. The subjects were evaluated in sinus rhythm and later with AV synchronous RV pacing at a cycle length of 550 ms with a short AV delay of 80 ms. The echocardiographic clips were analyzed using speckle tracking methods to calculate LV circumferential and longitudinal strain, rotation and twist in all conditions. Acute RV apical pacing decreased LV longitudinal strain from 16.1 ± 3.7% in sinus rhythm to 14.4 ± 3.3% (p = 0.03) and LV base rotation from -8.4° ± 3.6° to -6.4° ± 4.0° (p = 0.04). The circumferential strain, apical rotation and LV twist were not affected. Separate analysis of subjects with no prior preexcitation showed that acute RV pacing caused significant twist reduction, from 15.9° ± 7.6° to 12.1° ± 7.0° (p = 0.02), and decreased longitudinal strain and base rotation. Patients with preexcitation had abnormalities that persisted acutely after ablation. Acute RV apical pacing causes reductions in LV base rotation, longitudinal strain and twist. The recognition of abnormal LV activation patterns may provide longitudinal clues to LV dysfunction in chronically paced patients and potential novel indices of effective CRT interventions to reverse these abnormalities.