Evaluation of ventricular dysfunction using semi-automatic longitudinal strain analysis of four-chamber cine MR imaging

Assessment of Bi-Ventricular and Bi-Atrial Areas Using Four-Chamber Cine Cardiovascular Magnetic Resonance Imaging: Fully Automated Segmentation with a U-Net Convolutional Neural Network

Four-chamber (4CH) cine cardiovascular magnetic resonance imaging (CMR) facilitates simultaneous evaluation of cardiac chambers; however, manual segmentation is time-consuming and subjective in practice. We evaluated deep learning based on a U-Net convolutional neural network (CNN) for fully automated segmentation of the four cardiac chambers using 4CH cine CMR. Cine CMR datasets from patients were randomly assigned for training (1400 images from 70 patients), validation (600 images from 30 patients), and testing (1000 images from 50 patients). We validated manual and automated segmentation based on the U-Net CNN using the dice similarity coefficient (DSC) and Spearman’s rank correlation coefficient (ρ); p < 0.05 was statistically significant. The overall median DSC showed high similarity (0.89). Automated segmentation correlated strongly with manual segmentation in all chambers—the left and right ventricles, and the left and right atria (end-diastolic area: ρ = 0.88, 0.76, 0.92, and 0.87; end-systolic area: ρ = 0.81, 0.81, 0.92, and 0.83, respectively; p < 0.01). The area under the curve for the left ventricle, left atrium, right ventricle, and right atrium showed high scores (0.96, 0.99, 0.88, and 0.96, respectively). Fully automated segmentation could facilitate simultaneous evaluation and detection of enlargement of the four cardiac chambers without any time-consuming analysis.

Balloon pulmonary angioplasty improves right atrial reservoir and conduit functions in chronic thromboembolic pulmonary hypertension

AIMS

Right atrial (RA) function largely contributes to the maintenance of right ventricular (RV) function. This study investigated the effect of balloon pulmonary angioplasty (BPA) on RA functions in patients with inoperable chronic thromboembolic pulmonary hypertension (CTEPH) using cardiac magnetic resonance imaging (CMRI).

METHODS AND RESULTS

CMRI and RV catheterization were performed before BPA sessions and at the follow-up periods in 29 CTEPH patients. Reservoir [RA longitudinal strain (RA-LS)], passive conduit [RA early LS rate (LSR)], and active (RA late LSR) phases were assessed by using cine CMRI and a feature-tracking algorithm. The relationships between the changes in RA functions and in brain natriuretic peptide (BNP) were evaluated in both the dilated and non-dilated RA groups. RA-LS (32.4% vs. 42.7%), RA LSR (6.3% vs. 8.3%), and RA early LSR (-2.3% vs. -4.3%) were improved after BPA, whereas no significant change was seen in RA late LSR. The changes in RA peak LS and in RA early LSR were significantly correlated with the changes in BNP (ΔRA-LS: r = -0.63, ΔRA-early LSR: r = 0.65) and pulmonary vascular resistance (PVR) (ΔRA-LS: r = -0.69, ΔRA-early LSR: r = 0.66) in the nondilated RA group.

CONCLUSION

The RA reservoir and passive conduit functions were impaired in inoperable CTEPH, whereas RA active function was preserved. BPA markedly reversed these impaired functions. The improvements in RA reservoir and conduit functions were significantly correlated with the changes in BNP levels and PVR in CTEPH patients with normal RA sizes.

Feature-Tracking MRI Fractal Analysis of Right Ventricular Remodeling in Adults with Congenitally Corrected Transposition of the Great Arteries

Purpose

To assess a recently available technique for quantification of right ventricular (RV) trabeculae that is based on fractal analysis performed by using cardiac MRI feature tracking, in patients with congenitally corrected transposition of the great arteries (cc-TGA).

Materials and Methods

A total of 19 patients (eight men, 11 women; mean age, 35 years ± 10 [standard deviation]) with consecutive cc-TGA who underwent cardiac MRI were enrolled in the study. For analysis, patients were divided into two groups: six patients (four men, two women; mean age, 34 years ± 14) with an end-systolic RV volume index higher than 72 mL/m² (indicative of adverse RV remodeling) and 13 patients (four men, nine women; mean age, 36 years ± 9) in whom this index was lower than or equal to 72 mL/m² (indicative of adapted RV). The following outcomes were quantified in the midsection of the RV: fractional fractal dimension (FD) and diastolic FD, circumferential strain, and radial strain. Receiver operating characteristic (ROC) analysis was performed to determine the cutoff FD values for the detection of adverse RV remodeling. Correlations among fractional FD, diastolic FD, circumferential strain, and radial strain were calculated by using Pearson correlation coefficient (r) analysis.

Results

The following ROC values were identified for fractional and diastolic FD: cutoff, 0.09 and 1.39, respectively; area under the ROC curve, 0.95 and 0.68, respectively; sensitivity, 1.00 and 0.33, respectively; and specificity, 0.92 and 1.00, respectively. Fractional FD correlated with circumferential strain and radial strain (r = -0.70 and 0.69, respectively; P < .01), as did diastolic FD (r = 0.37 and -0.38, respectively; P < .05).

Conclusion

The fractional FD derived from cardiac MRI feature-tracking analysis correlates with adverse RV remodeling, including a changed strain pattern and trabeculae, in patients with cc-TGA.© RSNA, 2019.

Clinical usefulness of right ventricular 3D area strain in the assessment of treatment effects of balloon pulmonary angioplasty in chronic thromboembolic pulmonary hypertension: comparison with 2D feature-tracking MRI

OBJECTIVES

To evaluate the usefulness of right ventricular (RV) area strain analysis via cardiac MRI (CMRI) as a tool for assessing the treatment effects of balloon pulmonary angioplasty (BPA) in inoperable chronic thromboembolic pulmonary hypertension (CTEPH), RV area strain was compared to two-dimensional (2D) strain with feature-tracking MRI (FTMRI) before and after BPA.

METHODS

We retrospectively analyzed 21 CTEPH patients who underwent BPA. End-systolic global area strain (GAS), longitudinal strain (LS), circumferential strain (CS), and radial strain (RS) were measured before and after BPA. Changes in GAS and RV ejection fraction (RVEF) values after BPA were defined as ΔGAS and ΔRVEF. Receiver operating characteristic (ROC) analyses were performed to determine the optimal cutoff of the strain at after BPA for detection of improved patients with decreased mean pulmonary artery pressure (mPAP) less than 30 mmHg and increased RVEF more than 50%.

RESULTS

ROC analysis revealed the optimal cutoffs of strains (GAS, LS, CS, and RS) for identifying improved patients with mPAP < 30 mmHg (cutoff (%) = - 41.2, - 13.8, - 16.7, and 14.4: area under the curve, 0.75, 0.56, 0.65, and 0.75) and patients with RVEF > 50% (cutoff (%) = - 37.2, - 29.5, - 2.9, and 14.4: area under the curve, 0.81, 0.60, 0.56, and 0.56).

CONCLUSIONS

Area strain analysis via CMRI may be a more useful tool for assessing the treatment effects of BPA in patients with CTEPH than 2D strains with FTMRI.

KEY POINTS

• Area strain values can detect improvement of right ventricular (RV) pressure and function after balloon pulmonary angioplasty (BPA) equally or more accurately than two-dimensional strains. • Area strain analysis is a useful analytical method that reflects improvements in complex RV myocardial deformation by BPA. • Area strain analysis is a robust method with reproducibility equivalent to that of 2D strain analysis.

Abnormal right atrial performance in repaired tetralogy of Fallot: A CMR feature tracking analysis

BACKGROUND

We hypothesized that right atrial (RA) performance is abnormal in repaired tetralogy of Fallot (TOF).

METHODS

TOF patients were prospectively enrolled for cardiovascular magnetic resonance (CMR), echocardiography and exercise stress following a standardized 14-center protocol. Peak RA longitudinal strain (RALS) and right ventricular longitudinal strain (RVLS) were measured using CMR feature tracking (FT) and correlated to RA and RV end diastolic volumes (EDVi) and ejection fraction (EF).

RESULTS

The cohort had 311 subjects: 171 TOF (94 male, age 18.2±8years) and 140 healthy controls (69 male, 16.4±11years). RAEDVi, RALS, RVEDVi, RVLS, RAEF, and RVEF in TOF were 60.8±17.1ml/m², 13.6±5.7%, 120.3±30.3ml/m², 12.3±4.2%, 32.5±9.9% and 51.2±8.4% and differed from respective indices in controls: 51.7±15.7ml/m², 27±10.1%, 74±19.0ml/m², 18.5±5.3%, 54±8% and 62.5±5.5% (p<0.001). RAEDVi and RALS correlated with RVLS (p=0.004, <0.001, r=0.2,0.3). RAEDVi was higher in older TOF, while RALS did not increase with age. RAEDVi but not RALS correlated with RV systolic pressure(r=0.2, 0). Neither RAEDVi nor RALS was associated with tricuspid regurgitation grade or peak oxygen uptake (r=0.1, 0). Positive correlation was observed for RVEDVi with RAEDVi (p=0.035, r=0.2) and a trend toward negative correlation with RALS (p=0.09, r=0.1).

CONCLUSION

RALS, RAEDVi and RAEF are abnormal in TOF. Reduced RALS indicates decreased RA reservoir function. Because they correlate with other functional RV indices, these abnormalities may represent RA diastolic burden from chronic RV dysfunction in TOF. The young cohort age might explain the absence of RALS correlation to tricuspid regurgitation and peak oxygen uptake.