Myocardial texture analysis in idiopathic dilated cardiomyopathy: prediction of contractile reserve on dobutamine echocardiography

Intelligent diagnosis of left ventricular hypertrophy using transthoracic echocardiography videos

PURPOSE

Left ventricular hypertrophy (LVH) is an independent risk factor for cardiovascular events and mortality. Pathological LVH can be caused by various diseases. In this study, we explored the possibility of using time and frequency domain analysis of myocardial radiomics features for patients with LVH in differentiating hypertrophic cardiomyopathy (HCM), hypertensive heart disease (HHD) and uremic cardiomyopathy (UCM) based on transthoracic echocardiography (TTE). This was the first study to explore TTE myocardial time and frequency domain analyses for multiple LVH etiology differentiation.

MATERIALS AND METHODS

We proposed an artificially intelligent diagnosis system based on radiomics techniques for differentiating HCM, HHD and UCM on TTE videos of the apical four-chamber view, which mainly included interventricular septum (IVS) segmentation, feature extraction and classification. We used two independent cohorts, one with 150 patients, including 50 HHD, 50 HCM and 50 UCM, for segmentation training and testing, and another with 149 patients (namely the main cohort), including 50 HHD, 46 HCM and 53 UCM, for classification training and testing after segmentation and feature extraction. Firstly, the U-Net, Residual U-Net (ResUNet) and nnU-Net were trained and tested to segment the IVS on TTE still images in the first cohort. Then the trained model with the best segmentation performance was further used for IVS prediction of ordered TTE images in video sequences in the main cohort. The post-processing was used to eliminate the noisy debris by selecting the maximum connected region and smoothing the edges of the predicted IVS region. Secondly, static radiomics features were extracted from the IVS of ordered TTE images in each video sequence, and subsequently the time and frequency domain features were further extracted from each time series of a static radiomics feature in the video sequence. Finally, the point-wise gated Boltzmann machine (PGBM) was used to learn and fuse the time and frequency domain features, and the support vector machine was used to classify the learned features for LVH diagnosis. The classification was performed with five-fold cross validation.

RESULTS

The ResUNet showed the best segmentation performance, with Dice coefficient, sensitivity, specificity and accuracy of 0.817, 76.3%, 99.6% and 98.6%, respectively. With post-processing, the Dice coefficient, sensitivity, specificity and accuracy of the ResUNet were further improved to 0.839, 77.0%, 99.8%, and 98.8%, respectively. The classification areas under the receiver operating characteristic curves (AUCs) were 0.838 ± 0.049 for HHD vs. HCM, 0.868 ± 0.042 for HCM vs. UCM and 0.701 ± 0.140 for HHD vs. UCM.

CONCLUSION

In this work, we proposed an intelligent identification system for LVH etiology classification based on routine TTE video images with good diagnostic performance. This deep learning method is feasible in automatic TTE images interpretation and expected to assist clinicians in detecting the primary cause of LVH.

Assessment and Prognostic Relevance of Right Ventricular Contractile Reserve in Patients With Severe Pulmonary Hypertension

Background—

This study sought to analyze a new approach to assess exercise-induced pulmonary artery systolic pressure (PASP) increase by means of stress Doppler echocardiography as a possible measure of right ventricular contractile reserve in patients with severe pulmonary hypertension and right heart failure.

Methods and Results—

In this prospective study, patients with invasively diagnosed pulmonary arterial hypertension or inoperable chronic thromboembolic pulmonary hypertension and impaired right ventricular pump function despite a stable targeted pulmonary arterial hypertension medication underwent a broad panel of noninvasive assessments, including stress echocardiography and cardiopulmonary exercise testing. On the basis of the assumption that exercise-induced PASP is a measure of right ventricular contractile reserve, patients were classified into 2 groups according to an exercise-induced PASP increase above or below the median. Patients were followed up for 3.0±1.8 years. Univariate and multivariate analyses were used for factors predicting survival. Of 124 patients, 66 were below the median exercise-induced PASP increase of 30 mm Hg (low PASP), and 58 patients were above the median (high PASP). These groups were not significantly different in terms of medication and resting hemodynamics. Low PASP was associated with a significantly lower 6-minute walking distance, peak o 2 per kilogram, and 1-, 3-, and 4-year survival rates (92%, 69%, and 48%, respectively, versus 96%, 92%, and 89%). In the multivariate Cox model analysis adjusted for age and sex, PASP increase during exercise and peak o 2 per kilogram remained independent prognostic markers (hazard ratio, 2.56 for peak o 2 per kilogram and 2.84 for PASP increase).

Conclusions—

Exercise-induced PASP increase is of high clinical and prognostic relevance in pulmonary hypertension patients and may indicate right ventricular contractile reserve. Stress Doppler echocardiography may be a useful tool for prognostic assessment in pulmonary hypertension patients.

Relation of myocardial histomorphometric features and left ventricular contractile reserve assessed by high-dose dobutamine stress echocardiography in patients with idiopathic dilated cardiomyopathy

BACKGROUND

This study was designed to determine the relationship between histomorphometric features and contractile reserve assessed by high-dose dobutamine stress echocardiography in patients with idiopathic dilated cardiomyopathy.

METHODS

Twenty-four consecutive patients (21 men, aged 43.4+/-8.7 years) with idiopathic dilated cardiomyopathy underwent dobutamine stress echocardiography. Wall motion score index, ejection fraction, cardiac power output and end-systolic pressure/volume ratio were used as indices of left ventricular contractility. Left ventricular endomyocardial biopsy specimens (3-5 per patient) were routinely processed and stained with Masson trichrome, interstitial fibrosis and myocyte diameter were calculated quantitatively.

RESULTS

Myocyte diameter and interstitial fibrosis showed strongest correlation with change in wall motion score index (r=-0.667, p<0.001, and r=-0.567, p=0.004, respectively), followed by change in ejection fraction (r=-0.603, p=0.002, and r=-0.467, p=0.021, respectively). Interstitial fibrosis showed no correlation with change of cardiac power output and end-systolic pressure/volume ratio, whereas myocyte diameter was associated with change of both indices (r=-0.565, p=0.004, and r=-0.455, p=0.025).

CONCLUSIONS

Contractile reserve elicited by high-dose dobutamine is strongly related to the degree of histological disruption in patients with idiopathic dilated cardiomyopathy.

Texture analysis in digitally-acquired echocardiographic images: the effect of JPEG compression and video storage

The analysis of texture in video-stored echocardiographic images is an established method to characterize myocardial pathologies. We investigated whether or not texture parameters calculated from video-stored images and those derived from the joint photographic expert group (JPEG) format compressed data are equivalent to those calculated from uncompressed digital images. Texture parameters were calculated using uncompressed digital data, images stored on videotape, and three forms of compressed digital data (baseline JPEG, JPEG 2000 and lossless JPEG 2000). Video storage heavily affected most texture parameters. Although first-order texture parameters derived from JPEG-compressed images were generally equivalent to those derived from the uncompressed data, several second-order parameters differed significantly. We conclude that texture of video-stored images is not comparable to that of digitally-stored images and that JPEG compression changes important second-order texture parameters. This observation should be taken into account when analyzing texture of modern image data (uncompressed or compressed) and comparing the results with earlier studies utilizing video-stored data.