Assessment of right ventricular size and function from cardiovascular magnetic resonance images using artificial intelligence

Evaluation of the Truncated Cone-Rhomboid Pyramid Formula for Simplified Right Ventricular Quantification: A Cardiac Magnetic Resonance Study

Background/Objective: Cardiac magnetic resonance (CMR) is the reference method for right ventricular (RV) volume and function analysis, but time-consuming manual segmentation and corrections of imperfect automatic segmentations are needed. This study sought to evaluate the applicability of an echocardiographically established truncated cone-rhomboid pyramid formula (CPF) for simplified RV quantification using CMR. Methods: A total of 70 consecutive patients assigned to RV analysis using CMR were included. As standard method, the manual contouring of RV-short axis planes was performed for the measurement of end-diastolic volume (EDV) and end-systolic volume (ESV). Additionally, two linear measurements in four-chamber views were obtained in systole and diastole: basal diameters at the level of tricuspid valve (Dd and Ds) and baso-apical lengths from the center of tricuspid valve to the RV apex (Ld and Ls) were measured for the calculation of RV-EDV = 1.21 × Dd2 × Ld and RV-ESV = 1.21 × Ds 2 × Ls using CPF. Results: RV volumes using CPF were slightly higher than those using standard CMR analysis (RV-EDV index: 86.2 ± 29.4 mL/m2 and RV-ESV index: 51.5 ± 22.5 mL/m2 vs. RV-EDV index: 81.7 ± 24.1 mL/m2 and RV-ESV index: 44.5 ± 23.2 mL/m2) and RV-EF was lower (RV-EF: 41.1 ± 13.5% vs. 48.4 ± 13.7%). Both methods had a strong correlation of RV volumes (ΔRV-EDV index = -4.5 ± 19.0 mL/m2; r = 0.765, p < 0.0001; ΔRV-ESV index = -7.0 ± 14.4 mL/m2; r = 0.801, p < 0.0001). Conclusions: Calculations of RV volumes and function using CPF assuming the geometrical model of a truncated cone-rhomboid pyramid anatomy of RV is feasible, with a strong correlation to measurements using standard CMR analysis, and only two systolic and diastolic linear measurements in four-chamber views are needed.

CT and MR imaging of patients with a dilated right ventricle due to congenital causes and their treatment

A variety of both acquired and congenital conditions can significantly affect the right ventricle, with a variety of potential origins that can have substantial clinical ramifications. These conditions can range from the impact of diseases like pulmonary arterial hypertension and ischaemic heart disease to valvular deficiencies resulting in heart failure. Moreover, the right ventricle response to factors like abnormal loading conditions, and its subsequent clinical effects, are influenced by factors such as age, disease progression, potential interventions, and their immediate and long-term clinical outcomes. Therefore, a readily available and reproducible non-invasive imaging assessment can aid in diagnosing the underlying condition of a dilated right ventricle, track its evolution, and help devising the most appropriate treatment strategy and optimal timing for its implementation throughout the patient's life.In this review, our primary focus will be on the non-invasive imaging with CT and MR of an enlarged right ventricle resulting from congenital causes and their treatment.

Anatomical Prior-Based Automatic Segmentation for Cardiac Substructures from Computed Tomography Images

Cardiac substructure segmentation is a prerequisite for cardiac diagnosis and treatment, providing a basis for accurate calculation, modeling, and analysis of the entire cardiac structure. CT (computed tomography) imaging can be used for a noninvasive qualitative and quantitative evaluation of the cardiac anatomy and function. Cardiac substructures have diverse grayscales, fuzzy boundaries, irregular shapes, and variable locations. We designed a deep learning-based framework to improve the accuracy of the automatic segmentation of cardiac substructures. This framework integrates cardiac anatomical knowledge; it uses prior knowledge of the location, shape, and scale of cardiac substructures and separately processes the structures of different scales. Through two successive segmentation steps with a coarse-to-fine cascaded network, the more easily segmented substructures were coarsely segmented first; then, the more difficult substructures were finely segmented. The coarse segmentation result was used as prior information and combined with the original image as the input for the model. Anatomical knowledge of the large-scale substructures was embedded into the fine segmentation network to guide and train the small-scale substructures, achieving efficient and accurate segmentation of ten cardiac substructures. Sixty cardiac CT images and ten substructures manually delineated by experienced radiologists were retrospectively collected; the model was evaluated using the DSC (Dice similarity coefficient), Recall, Precision, and the Hausdorff distance. Compared with current mainstream segmentation models, our approach demonstrated significantly higher segmentation accuracy, with accurate segmentation of ten substructures of different shapes and sizes, indicating that the segmentation framework fused with prior anatomical knowledge has superior segmentation performance and can better segment small targets in multi-target segmentation tasks.

Emerging Roles for Artificial Intelligence in Heart Failure Imaging

Artificial intelligence (AI) applications are expanding in cardiac imaging. AI research has shown promise in workflow optimization, disease diagnosis, and integration of clinical and imaging data to predict patient outcomes. The diagnostic and prognostic paradigm of heart failure is heavily reliant on cardiac imaging. As AI becomes increasingly validated and integrated into clinical practice, AI influence on heart failure management will grow. This review discusses areas of current research and potential clinical applications in AI as applied to heart failure cardiac imaging.

Assessment of Right Ventricle Function and Tricuspid Regurgitation in Heart Failure: Current Advances in Diagnosis and Imaging

Right ventricular (RV) systolic dysfunction increases mortality among heart failure patients, and therefore, accurate diagnosis and monitoring is paramount. RV anatomy and function are complex, usually requiring a combination of imaging modalities to completely quantitate volumes and function. Tricuspid regurgitation usually occurs with RV dysfunction, and quantifying this valvular lesion also may require multiple imaging modalities. Echocardiography is the first-line imaging tool for identifying RV dysfunction, with cardiac MRI and cardiac computed tomography adding valuable additional information.

Multimodality Imaging of Right Heart Function: JACC Scientific Statement

Right ventricular (RV) size and function assessed by multimodality imaging are associated with outcomes in a variety of cardiovascular diseases. Understanding RV anatomy and physiology is essential in appreciating the strengths and weaknesses of current imaging methods and gives these measurements greater context. The adaptation of the right ventricle to different types and severity of stress, particularly over time, is specific to the cardiovascular disease process. Multimodality imaging parameters, which determine outcomes, reflect the ability to image the initial and longitudinal RV response to stress. This paper will review the standard and novel imaging methods for assessing RV function and the impact of these parameters on outcomes in specific disease states.

Deep learning can yield clinically useful right ventricular segmentations faster than fully manual analysis

Right ventricular (RV) volumes are commonly obtained through time-consuming manual delineations of cardiac magnetic resonance (CMR) images. Deep learning-based methods can generate RV delineations, but few studies have assessed their ability to accelerate clinical practice. Therefore, we aimed to develop a clinical pipeline for deep learning-based RV delineations and validate its ability to reduce the manual delineation time. Quality-controlled delineations in short-axis CMR scans from 1114 subjects were used for development. Time reduction was assessed by two observers using 50 additional clinical scans. Automated delineations were subjectively rated as (A) sufficient for clinical use, or as needing (B) minor or (C) major corrections. Times were measured for manual corrections of delineations rated as B or C, and for fully manual delineations on all 50 scans. Fifty-eight % of automated delineations were rated as A, 42% as B, and none as C. The average time was 6 min for a fully manual delineation, 2 s for an automated delineation, and 2 min for a minor correction, yielding a time reduction of 87%. The deep learning-based pipeline could substantially reduce the time needed to manually obtain clinically applicable delineations, indicating ability to yield right ventricular assessments faster than fully manual analysis in clinical practice. However, these results may not generalize to clinics using other RV delineation guidelines.

Free-breathing cardiovascular cine magnetic resonance imaging using compressed-sensing and retrospective motion correction: accurate assessment of biventricular volume at 3T

PURPOSE

We applied a combination of compressed-sensing (CS) and retrospective motion correction to free-breathing cine magnetic resonance (MR) (FBCS cine MoCo). We validated FBCS cine MoCo by comparing it with breath-hold (BH) conventional cine MR.

MATERIALS AND METHODS

Thirty-five volunteers underwent both FBCS cine MoCo and BH conventional cine MR imaging. Twelve consecutive short-axis cine images were obtained. We compared the examination time, image quality and biventricular volumetric assessments between the two cine MR.

RESULTS

FBCS cine MoCo required a significantly shorter examination time than BH conventional cine (135 s [110-143 s] vs. 198 s [186-349 s], p < 0.001). The image quality scores were not significantly different between the two techniques (End-diastole: FBCS cine MoCo; 4.7 ± 0.5 vs. BH conventional cine; 4.6 ± 0.6; p = 0.77, End-systole: FBCS cine MoCo; 4.5 ± 0.5 vs. BH conventional cine; 4.5 ± 0.6; p = 0.52). No significant differences were observed in all biventricular volumetric assessments between the two techniques. The mean differences with 95% confidence interval (CI), based on Bland-Altman analysis, were - 0.3 mL (- 8.2 - 7.5 mL) for LVEDV, 0.2 mL (- 5.6 - 5.9 mL) for LVESV, - 0.5 mL (- 6.3 - 5.2 mL) for LVSV, - 0.3% (- 3.5 - 3.0%) for LVEF, - 0.1 g (- 8.5 - 8.3 g) for LVED mass, 1.4 mL (- 15.5 - 18.3 mL) for RVEDV, 2.1 mL (- 11.2 - 15.3 mL) for RVESV, - 0.6 mL (- 9.7 - 8.4 mL) for RVSV, - 1.0% (- 6.5 - 4.6%) for RVEF.

CONCLUSION

FBCS cine MoCo can potentially replace multiple BH conventional cine MR and improve the clinical utility of cine MR.

Papillary and Trabecular Muscles Have Substantial Impact on Quantification of Left Ventricle in Patients with Hypertrophic Obstructive Cardiomyopathy

Patients with obstructive hypertrophic cardiomyopathy (HOCM) have large papillary and trabecular muscles (PTMs), which are myocardial tissue. PTMs are usually excluded from the myocardium and included in the left ventricular (LV) cavity when determining LV mass (LVM) and volumes using cardiac magnetic resonance (CMR). This conventional method may result in large distortion of LVM and other indices. We investigated 74 patients with HOCM undergoing CMR imaging. LV short-axis cine images were obtained. LV contours were drawn using two different methods: (1) the conventional method, where PTMs were included in the LV cavity; and (2) the mask method, which includes the TPMs in the LV myocardium. The LV end-diastolic volume (LV-EDV), LV end-systolic volume (LV-ESV), LV ejection fraction (LVEF), and the LVM were then calculated. Fasting NT-proBNP and CK-MB levels were measured with ELISA. In patients with HOCM, mass of PTMs (MOPTM) was 47.9 ± 18.7 g, which represented 26.9% of total LVM. Inclusion of PTMs with the mask method resulted in significantly greater LVM and LVM index (both p < 0.0001) in comparison with those measured with the conventional method. In addition, the mask method produced a significant decrease in LV-EDV and LV-ESV. LVEF was significantly increased with the mask method (64.3 ± 7.9% vs. 77.2 ± 7.1%, p < 0.0001). MOPTM was positively correlated with BMI, septal wall thickness, LVM, LV-EDV, and LV-ESV. LVEF was inversely correlated with MOPTM. In addition, MOPTM correlated positively with NT-proBNP (r = 0.265, p = 0.039) and CK-MB (r = 0.356, p = 0.002). In conclusion, inclusion of PTMs in the myocardium has a substantial impact on quantification of the LVM, LV-EDV, LV-ESV, and LVEF in patients with HOCM. The effects of the PTMs in women was greater than that in men. Furthermore, the MOPTM was positively associated with NT-proBNP and CK-MB. The PTMs might be included in the myocardium when measuring the LV volumes and mass of patients with HOCM. At present, the clinical and prognostic meaning and relevance of the PTMs is not clear and should be further studied.

The Applications of Artificial Intelligence in Cardiovascular Magnetic Resonance-A Comprehensive Review

Cardiovascular disease remains an integral field on which new research in both the biomedical and technological fields is based, as it remains the leading cause of mortality and morbidity worldwide. However, despite the progress of cardiac imaging techniques, the heart remains a challenging organ to study. Artificial intelligence (AI) has emerged as one of the major innovations in the field of diagnostic imaging, with a dramatic impact on cardiovascular magnetic resonance imaging (CMR). AI will be increasingly present in the medical world, with strong potential for greater diagnostic efficiency and accuracy. Regarding the use of AI in image acquisition and reconstruction, the main role was to reduce the time of image acquisition and analysis, one of the biggest challenges concerning magnetic resonance; moreover, it has been seen to play a role in the automatic correction of artifacts. The use of these techniques in image segmentation has allowed automatic and accurate quantification of the volumes and masses of the left and right ventricles, with occasional need for manual correction. Furthermore, AI can be a useful tool to directly help the clinician in the diagnosis and derivation of prognostic information of cardiovascular diseases. This review addresses the applications and future prospects of AI in CMR imaging, from image acquisition and reconstruction to image segmentation, tissue characterization, diagnostic evaluation, and prognostication.