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.

Diagnostic performance of transthoracic echocardiography in screening acute type A aortic dissection from ST-segment elevated myocardial infarction

BACKGROUND

When patients with type A acute aortic dissection (TAAAD) present with changes to their ST-segment, diagnostic and treatment delays increase significantly. The performance of transthoracic echocardiography (TTE) screening of TAAAD in patients with ST-segment elevated myocardial infarction (STEMI) is yet to be validated.

METHODS

The diagnostic performance of TTE alone and combined with the aortic dissection risk score (ADRS) in TAAAD was evaluated. In this retrospective study (ChiCTR, No. 2000031291), TTE was reviewed to detect direct/indirect signs of TAAAD. The ADRS of each patient was calculated according to guidelines. Case adjudication was based on advanced imaging and surgery.

RESULTS

Among a total of 442 patients, TAAAD was diagnosed in 146 (33.0%). The presence of direct TTE signs had a sensitivity of 43.0% [95% confidence interval (CI): 35.0% to 52.0%] and specificity of 97.0% (95% CI: 95.0% to 99.0%), and the presence of any TTE sign had a sensitivity of 97.0% (95% CI: 93.0% to 99.0%) and specificity of 78.0% (95% CI: 73.0% to 82.0%) for TAAAD. The additive value of TTE was most evident in patients with low clinical probability for TAAAD (ADRS ≤1). The presence of ADRS ≤1 plus an absence of direct TTE signs for TAAAD rule-out had a sensitivity of 98.4% (95% CI: 96.1% to 99.6%).

CONCLUSIONS

The use of TTE adds value in the screening of TAAAD in STEMI patients. In patients with low clinical probability for TAAAD, direct TTE signs can be used to rapidly identify those who require advanced imaging.

A formalized pilot study

Objective

The study objective is to evaluate the effeteness of an existing educational platform blending didactic presentation and hands-on simulation for university doctoral SRNAs in the area of basic, 4 view identification and performance of transthoracic echocardiography (TTE).

Methods

Following IRB approval, SRNAs were exposed to a pre test to evaluate existing skills, then they were exposed to a graphic rich, live presentation of basic 4 view TTE. The presentation was then followed by hands on simulation and performance of the 4 basic TTE views on live models.

Results

Pretest scores averaged 58% and post tests scores rose to 95%. See Table 1.

Conclusion

Our results support the concept that the existing blended platform is effective to train university SRNAs in basic 4 view, bedside transthoracic echocardiography.

Prosthesis-patient mismatch defined by cardiac computed tomography versus echocardiography after transcatheter aortic valve replacement

BACKGROUNDS

Evaluation of prosthesis-patient mismatch (P-PM) after transcatheter aortic valve replacement (TAVR) by transthoracic echocardiography (TTE) has provided conflicting results regarding its impact on outcomes. Whether post-TAVR computed tomography angiography (CTA) evaluation of P-PM can improve our understanding is unknown. We aimed to evaluate the inter-modality (TTE vs. CTA) agreement, inter-valve platform (balloon-expanding valve [BEV] vs. self-expandable valve [SEV]) differences in P-PM severity, and outcomes related to P-PM after TAVR.

METHODS

We analyzed patients with both CTA and TTE before and after TAVR. Indexed effective orifice area was calculated using two methods: TTE-derived left ventricular outflow tract (LVOT) area from measured diameter and post-TAVR CTA-measured area. Body size specific cut-offs for P-PM severity were used: for body mass index (BMI) ​< ​30 ​kg/m², moderate ​= ​0.66-0.85 ​cm²/m² and severe≤0.65 ​cm²/m²; for BMI ≥30 ​kg/m², moderate ​= ​0.56-0.70 ​cm²/m² and severe≤0.55 ​cm²/m².

RESULTS

A total of 447 patients were included (median age, 83 years; 54% male). The prevalence of P-PM (moderate or severe) was lower with CTA vs. TTE (3.5% vs. 19.5%, p ​< ​0.001). The prevalence of P-PM measured by TTE was more common in BEV compared to SEV (p ​= ​0.002), while CTA assessment showed no difference in P-PM incidence and severity between TAVR platforms (p ​= ​0.40). In multivariable analysis, CTA-defined but not TTE-defined P-PM was associated with mortality after TAVR (HR:3.97; 95%CI,1.55-10.2; p ​= ​0.004). Both CTA-defined and TTE-defined P-PM were associated with the composite of death and heart failure rehospitalization.

CONCLUSION

Although post-TAVR CTA substantially downgraded the prevalence of P-PM compared to TTE, it identified a subset of patients with clinically relevant P-PM which associated with outcomes.