Automatic assessment of cardiac function from short-axis MRI: procedure and clinical evaluation

A systematic review of multi-slice and multi-frame descriptors in cardiac MRI exams

Computer-Aided Diagnosis systems have been developed to help medical professional in their decision making routines towards a more accurate diagnosis. These systems process medical exams such as Magnetic Resonance (MRI) in order to quantify meaningful features. These can be used with similarity-measuring techniques in a Content-Based Image Retrieval context, or inputted into a machine learning classifier in order to support early disease detection. For cardiac MRIs, single slice descriptors have been proposed in the two-dimensional domain, shape descriptors have been proposed in the three-dimensional domain, and previous reviews have mapped these two descriptor categories. Nonetheless, no systematic review on these descriptors have looked at full cardiac MRI images sets. We have reviewed the literature by searching for descriptors that consider the whole slice set (multi-slice) or frames (multi-frame) in cardiac MRI exams. We discuss descriptors and techniques, the datasets that were used, and the different evaluation metrics. Finally, we highlight literature gaps and research opportunities.

Ensemble of 2D Residual Neural Networks Integrated with Atrous Spatial Pyramid Pooling Module for Myocardium Segmentation of Left Ventricle Cardiac MRI

Cardiac disease diagnosis and identification is problematic mostly by inaccurate segmentation of the cardiac left ventricle (LV). Besides, LV segmentation is challenging since it involves complex and variable cardiac structures in terms of components and the intricacy of time-based crescendos. In addition, full segmentation and quantification of the LV myocardium border is even more challenging because of different shapes and sizes of the myocardium border zone. The foremost purpose of this research is to design a precise automatic segmentation technique employing deep learning models for the myocardium border using cardiac magnetic resonance imaging (MRI). The ASPP module (Atrous Spatial Pyramid Pooling) was integrated with a proposed 2D-residual neural network for segmentation of the myocardium border using a cardiac MRI dataset. Further, the ensemble technique based on a majority voting ensemble method was used to blend the results of recent deep learning models on different set of hyperparameters. The proposed model produced an 85.43% dice score on validation samples and 98.23% on training samples and provided excellent performance compared to recent deep learning models. The myocardium border was successfully segmented across diverse subject slices with different shapes, sizes and contrast using the proposed deep learning ensemble models. The proposed model can be employed for automatic detection and segmentation of the myocardium border for precise quantification of reflow, myocardial infarction, myocarditis, and h cardiomyopathy (HCM) for clinical applications.

Left Ventricle Quantification Challenge: A Comprehensive Comparison and Evaluation of Segmentation and Regression for Mid-Ventricular Short-Axis Cardiac MR Data

Automatic quantification of the left ventricle (LV) from cardiac magnetic resonance (CMR) images plays an important role in making the diagnosis procedure efficient, reliable, and alleviating the laborious reading work for physicians. Considerable efforts have been devoted to LV quantification using different strategies that include segmentation-based (SG) methods and the recent direct regression (DR) methods. Although both SG and DR methods have obtained great success for the task, a systematic platform to benchmark them remains absent because of differences in label information during model learning. In this paper, we conducted an unbiased evaluation and comparison of cardiac LV quantification methods that were submitted to the Left Ventricle Quantification (LVQuan) challenge, which was held in conjunction with the Statistical Atlases and Computational Modeling of the Heart (STACOM) workshop at the MICCAI 2018. The challenge was targeted at the quantification of 1) areas of LV cavity and myocardium, 2) dimensions of the LV cavity, 3) regional wall thicknesses (RWT), and 4) the cardiac phase, from mid-ventricle short-axis CMR images. First, we constructed a public quantification dataset Cardiac-DIG with ground truth labels for both the myocardium mask and these quantification targets across the entire cardiac cycle. Then, the key techniques employed by each submission were described. Next, quantitative validation of these submissions were conducted with the constructed dataset. The evaluation results revealed that both SG and DR methods can offer good LV quantification performance, even though DR methods do not require densely labeled masks for supervision. Among the 12 submissions, the DR method LDAMT offered the best performance, with a mean estimation error of 301 mm 2 for the two areas, 2.15 mm for the cavity dimensions, 2.03 mm for RWTs, and a 9.5% error rate for the cardiac phase classification. Three of the SG methods also delivered comparable performances. Finally, we discussed the advantages and disadvantages of SG and DR methods, as well as the unsolved problems in automatic cardiac quantification for clinical practice applications.

Development and clinical validation of a hybrid method for semiautomated left ventricle endocardial and epicardial boundary extraction on cine-magnetic resonance images

We describe a hybrid method for left ventricle (LV) endocardial and epicardial segmentation on cardiac magnetic resonance (CMR) images requiring minimal operator intervention. Endocardium extraction results from the union of three independent estimations based on adaptive thresholding, region growing, and active contour with Chan-Vese energy function. Epicardium segmentation relies on conditional morphological dilation of the endocardial mask followed by active contour optimization. The proposed method was first evaluated using an open access database of 18 CMR for which expert manual contouring was available. The method was further validated on a retrospective cohort of 29 patients, who underwent CMR with expert manual segmentation. Regarding the open access database, similarity (Dice index) between hybrid and expert segmentations was good for end-diastolic (ED) endocardium (0.92), end-systolic (ES) endocardium (0.88), and ED epicardium (0.92). As for derived LV parameters, concordance (Lin's coefficient) was good for ED volume (0.91), ES volume (0.93), ejection fraction (EF; 0.89), and fair for myocardial mass (MM; 0.74). Regarding the retrospective patient study, concordance between expert and hybrid estimations was excellent for ED volume (0.95), ES volume (0.96), good for EF (0.86), and fair for MM (0.71). Hybrid segmentation resulted in small biases ([Formula: see text] for ED volume, [Formula: see text] for ES volume, [Formula: see text] for EF, and [Formula: see text] for MM) with little clinical relevance and acceptable for routine practice. The quickness and robustness of the proposed hybrid method and its ability to provide LV volumes, functions, and masses highly concordant with those given by expert segmentation support its pertinence for routine clinical use.

Fast segmentation of the left ventricle in cardiac MRI using dynamic programming

BACKGROUND AND OBJECTIVE

The segmentation of the left ventricle (LV) in cardiac magnetic resonance imaging is a necessary step for the analysis and diagnosis of cardiac function. In most clinical setups, this step is still manually performed by cardiologists, which is time-consuming and laborious. This paper proposes a fast system for the segmentation of the LV that significantly reduces human intervention.

METHODS

A dynamic programming approach is used to obtain the border of the LV. Using very simple assumptions about the expected shape and location of the segmentation, this system is able to deal with many of the challenges associated with this problem. The system was evaluated on two public datasets: one with 33 patients, comprising a total of 660 magnetic resonance volumes and another with 45 patients, comprising a total of 90 volumes. Quantitative evaluation of the segmentation accuracy and computational complexity was performed.

RESULTS

The proposed system is able to segment a whole volume in 1.5 seconds and achieves an average Dice similarity coefficient of 86.0% and an average perpendicular distance of 2.4 mm, which compares favorably with other state-of-the-art methods.

CONCLUSIONS

A system for the segmentation of the left ventricle in cardiac magnetic resonance imaging is proposed. It is a fast framework that significantly reduces the amount of time and work required of cardiologists.

Comparative evaluation of active contour model extensions for automated cardiac MR image segmentation by regional error assessment

OBJECTIVE

In the field of cardiac MR image segmentation, active contour models, or snakes have been extensively used, owing to their promising results and to the numerous extensions proposed to improve their performance. This paper explores a methodology for evaluating cardiac MR image segmentation algorithms, which assesses the distance between computer-generated and the observer's hand-outlined boundaries. This metric was applied to various external force extensions of the traditional snake, since no systematic comparison has been performed.

MATERIALS AND METHODS

Cardiac MRI from six patients were analyzed. Imaging was performed on a 1.5 T MR scanner with ECG-gated balanced steady-state free precession (b-SSFP) sequences. Segmentation performances were established for traditional snake, gradient vector flow snake, standard- and guided- pressure force-based snake. The use of a pre-treatment with non-linear anisotropic filtering was also compared to non-filtered images.

RESULTS

Agreement between manual and segmentation algorithms was satisfactory for ejection fraction for every segmentation scheme. However end-systolic and end-diastolic volumes were systematically underestimated.

CONCLUSION

The developed regional error metric provided a more rigorous evaluation of the segmentation schemes in comparison to the classical derived parameters based on left ventricle volume estimation, usually used in functional cardiac MR studies. These derived parameters can furthermore mask local segmentation errors.

Automatic quantitative left ventricular analysis of cine MR images by using three-dimensional information for contour detection

The purpose of this study was to evaluate an automatic boundary detection algorithm of the left ventricle on magnetic resonance (MR) short-axis images with the essential restriction of no manual corrections. The study comprised 13 patients (nine men, four women) and 12 healthy volunteers (11 men, one woman), and institutional review board approval and informed consent were obtained. The outline of the left ventricle was indicated manually on horizontal and vertical long-axis MR images. The calculated intersection points with the short-axis MR images were the basis of the automatic contour detection. Automatically derived volumes correlated highly with manually derived (short axis-based) volumes (R2 = 0.98); ejection fraction (EF) and mass showed a correlation of 0.95 and 0.93, respectively. Automatic contour detection reduced interobserver variability to 0.1 mL for endocardial end-diastolic and end-systolic volumes, 1.1 mL for epicardial end-diastolic and end-systolic volumes, 0.02% for EF, and 1.1 g for mass. Thus, the algorithm enabled highly reproducible left ventricular parameters to be obtained.

Evaluation of biventricular ejection fraction with ECG-gated 16-slice CT: preliminary findings in acute pulmonary embolism in comparison with radionuclide ventriculography

This study aimed to assess the feasibility of cardiac global function evaluation during a whole-chest multi-slice CT (MSCT) acquisition in patients referred for suspicion of pulmonary embolism (PE), and to compare the results with planar equilibrium radionuclide ventriculography (ERNA). Ten consecutive haemodynamically stable patients (six female, four male; mean age 69.7 years; heart rate 65-99 bpm) with suspicion of PE underwent an MSCT and ERNA within a 6 h period. CT acquisition was performed after contrast medium injection by using 16x1.5 mm collimation and retrospective ECG gating. Left ventricular (LVEF) and right ventricular (RVEF) ejection fractions were calculated using dedicated three-dimensional software. Relationships between measurements obtained with MSCT and ERNA were assessed using linear regression analysis and reliability of MSCT was assessed with intra-class correlation coefficient. Bland-Altman analysis was performed to calculate limits of agreement between MSCT and ERNA. MSCT was performed successfully in ten patients with a mean acquisition time of 16.5+/-2.8 s. Functional cardiac evaluation was possible on CT for all patients except for one due to poor opacification of right ventricle. Linear regression analysis showed a good correlation between MSCT and ERNA for the LVEF (R=0.91) and the RVEF (R=0.89) measurements. Intra-class correlation was superior for LVEF (0.92) than for the RVEF (0.68). Bland-Altman plots demonstrated that MSCT substantially overestimated the ERNA RVEF. Morphological CT data demonstrated PE in four of ten of patients and alternative diagnoses in five of ten patients. Our study reveals that MSCT with retrospective ECG gating may provide in one modality a morphological and a functional cardiopulmonary evaluation. Comparison with ERNA demonstrated a good correlation for both ventricular ejection fractions.

MR imaging assessment of cardiac function

Magnetic resonance (MR) imaging is an accurate and reproducible technique for assessment of ventricular function. Although echocardiography is the mainstay for evaluation of cardiac function, dobutamine stress MR imaging has been shown to be as safe as echocardiography for patients with coronary artery disease and more accurate in patients with suboptimal echocardiographic image quality. This article reviews MR imaging techniques, methods of pharmacologic stress, and clinical applications for assessment of cardiac function, primarily left ventricular function.

Myocardial delineation via registration in a polar coordinate system1

RATIONALE AND OBJECTIVES

Cardiovascular disease is the number one cause of premature death in the western world. Analysis of cardiac function provides clinically useful diagnostic and prognostic information; however, manual analysis of function via delineation is prohibitively time consuming. This article describes a technique for analysis of dynamic magnetic resonance images of the left ventricle using a non-rigid registration algorithm. A manually delineated contour of a single phase was propagated through the dynamic sequence.

MATERIALS AND METHODS

Short-axis cine magnetic resonance images were resampled into polar coordinates before all the time frames were aligned using a non-rigid registration algorithm. The technique was tested on 10 patient data sets, a total of 1,052 images were analyzed.

RESULTS

Results of this approach were investigated and compared with manual delineation at all phases in the cardiac cycle, and with registration performed in a Cartesian coordinate system. The results correlated very well with manually delineated contours.

CONCLUSION

A novel approach to the registration and subsequent delineation of cardiac magnetic resonance images has been introduced. For the endocardium, the polar resampling technique correlated well with manual delineation, and better than for images registered without radial resampling in a Cartesian coordinate system. For the epicardium, the difference was not as apparent with both techniques correlating well.

MR Imaging of the heart with cine true fast imaging with steady-state precession: influence of spatial and temporal resolutions on left ventricular functional parameters

The influence of changes in spatial and temporal resolutions on functional parameters in the left ventricle (LV) were investigated with magnetic resonance (MR) imaging with a modified true fast imaging with steady-state precession, or FISP, two-dimensional sequence that provided temporal resolution of 21-90 msec and spatial resolution of 1-3 mm. MR imaging in the heart was performed in 15 healthy volunteers. A decrease in LV functional parameters was observed with reduced spatial and temporal resolutions. The influence of temporal resolution was more relevant.

Magnetic resonance imaging demonstrates improved regional systolic wall motion and thickening and myocardial perfusion of myocardial territories treated by laser myocardial revascularization

OBJECTIVES

This study was designed to investigate the use of magnetic resonance (MR) functional and perfusion imaging to evaluate laser myocardial revascularization (LMR).

BACKGROUND

Most clinical studies of LMR have shown improvements in angina class and exercise capacity, with minimal or absent improvements in myocardial perfusion and function.

METHODS

Fifteen patients who underwent percutaneous Biosense-guided holmium:yttrium aluminum garnet LMR to areas of viable but ischemic myocardium were followed clinically and underwent functional and perfusion MRI at baseline, 30 days and 6 months.

RESULTS

The mean age was 64 +/- 11 years; four patients were women. The ejection fraction was 47.4 +/- 14.0%. Angina class at baseline was 3.4 +/- 0.6 and improved to 2.5 +/- 1.4 at six months (p = 0.054). Exercise time at baseline was 298 +/- 97 s and increased to 350 +/- 95 s at 30 days and 365 +/- 79 s at six months, p = 0.04. There were no significant changes in nuclear perfusion scans. Although MR determined that resting radial motion and thickening of the target wall were significantly less than normal at baseline (p < 0.001), they improved significantly during follow-up (wall thickening: baseline, 30.6 +/- 11.7%; day 30, 41.2 +/- 13.3% and day 180, 44.2 +/- 11.9%, p = 0.01). The size of the underperfused myocardial area was 14.5 +/- 5.4% at baseline and was reduced to 6.3 +/- 2.8% at 30 days and 7.7 +/- 3.7% at 6 months (p < 0.001).

CONCLUSIONS

This small phase I, open-label, uncontrolled study of MR functional and perfusion imaging in patients undergoing Biosense-guided LMR suggests a beneficial effect of this treatment strategy on myocardial function and perfusion. The efficacy of Biosense-guided LMR is being evaluated in a large phase II, randomized, blinded placebo-controlled trial with an MRI substudy (DIRECT).

Steady-state free precession magnetic resonance imaging of the heart: comparison with segmented k-space gradient-echo imaging

Steady-state free precession imaging is a promising technique for cardiac magnetic resonance imaging (MRI), as it provides improved blood/myocardial contrast in shorter acquisition times compared with conventional gradient-echo acquisition. The better contrast could improve observer agreement and automatic detection of cardiac contours for volumetric assessment of the ventricles, but measurements might differ from those obtained using conventional methods. We compared volumetric measurements, observer variabilities, and automatic contour detection between a steady-state free precession imaging sequence (BFFE = balanced fast field echo) and segmented k-space gradient-echo acquisition (TFE = turbo field echo) in 41 subjects. With BFFE, significantly higher end-diastolic and end-systolic volumes and lower wall thickness, ventricular mass, ejection fraction, and wall motion were observed (P < 0.0001), while interobserver variabilities were lower and automatic contour detection of endocardial contours was more successful. We conclude that the improved image quality of BFFE reduces the observer-dependence of volumetric measurements of the left ventricle (LV) but results in significantly different values in comparison to TFE measurements.

Cardiac phase contrast gradient echo MRI: characterization of abnormal left ventricular wall motion in patients with ischemic heart disease

PURPOSE

In our patient study, we examined the clinical usefulness of phase contrast velocity mapping for the detection and characterization of localized abnormalities of left ventricular motion.

MATERIALS AND METHODS

Velocity encoding is based on the fact that motion in the presence of a magnetic field gradient causes a change of the phase of the MRI signal that is proportional to the velocity of tissue motion. Left ventricular motion was characterized by parameters describing rotation and contraction/dilatation, respectively. We examined 34 patients with localized abnormalities of left ventricular motion due to ischemic heart disease.

RESULTS

Three patients could not be sufficiently evaluated due to technical problems including varying positions of the heart during successive breathhold periods. In 27 of the remaining 31 patients, MRI could demonstrate abnormal radial velocities that corresponded fully or partly with perfusion deficits in single photon emission computed tomography or positron emission tomography. The abnormalities were most pronounced in early diastole. Rotational velocities did not show any regional changes.

CONCLUSION

Our study showed that our technique is suitable for the detection and characterization of localized abnormalities of left ventricular motion in patients with ischemic heart disease.

Intracoronary basic fibroblast growth factor (FGF-2) in patients with severe ischemic heart disease: results of a phase I open-label dose escalation study

OBJECTIVES

Evaluate the safety, tolerability and preliminary efficacy of intracoronary (IC) basic fibroblast growth factor (bFGF, FGF-2).

BACKGROUND

FGF-2 is a heparin-binding growth factor capable of inducing functionally significant angiogenesis in animal models of myocardial ischemia.

METHODS

Phase I, open-label dose-escalation study of FGF-2 administered as a single 20-min infusion in patients with ischemic heart disease not amenable to treatment with CABG or PTCA.

RESULTS

Fifty-two patients enrolled in this study received IC FGF-2 (0.33 to 48 microg/kg). Hypotension was dose-dependent and dose-limiting, with 36 microg/kg being the maximally tolerated dose. Four patients died and four patients had non-Q-wave myocardial infarctions. Laboratory parameters and retinal examinations showed mild and mainly transient changes during the 6-month follow-up. There was an improvement in quality of life as assessed by Seattle Angina Questionnaire and improvement in exercise tolerance as assessed by treadmill exercise testing (510+/-24 s at baseline, 561+/-26 s at day 29 [p = 0.023], 609+/-26 s at day 57 (p < 0.001), and 633+/-24 s at day 180 (p < 0.001), overall p < 0.001). Magnetic resonance (MR) imaging showed increased regional wall thickening (baseline: 34+/-1.7%, day 29: 38.7+/-1.9% [p = 0.006], day 57: 41.4+/-1.9% [p < 0.001], and day 180: 42.0+/-2.3% [p < 0.001], overall p = 0.001) and a reduction in the extent of the ischemic area at all time points compared with baseline.

CONCLUSIONS

Intracoronary administration of rFGF-2 appears safe and is well tolerated over a 100-fold dose range (0.33 to 0.36 microk/kg). Preliminary evidence of efficacy is tempered by the open-label uncontrolled design of the study.

Quantification of global and regional ventricular function in cardiac magnetic resonance imaging

One of the strong assets of cardiac magnetic resonance (CMR) is its ability to assess myocardial anatomy, structure, function, flow, and perfusion within a single examination. Quantification of global and regional function from magnetic resonance imaging (MRI) studies was shown to be accurate and reproducible in experimental and clinical research studies. With the advent of high-performance MRI scanners and newly developed pulse sequences, image acquisition times have been reduced considerably in recent years. However, the clinical use of CMR remains limited for various reasons. Among these limitations is that the amount of images obtained in a typical cardiac examination is so large that visual and especially quantitative image analysis is tedious and time consuming. There is an urgent need for optimized dedicated software tools featuring highly automated contour detection and optimized display capabilities to present the quantitative results to the physician in an orderly fashion, thus facilitating clinical decision making. This article focuses on the state of the art in CMR postprocessing techniques for quantitative assessment of global and regional function.

Assessment of cardiac function: magnetic resonance and computed tomography

A complete cardiac study requires both anatomic and physiologic evaluation. Cardiac function can be evaluated noninvasively by magnetic resonance imaging (MRI)or ultrafast computed tomography (CT). MRI allows for evaluation of cardiac function by cine gradient echo imaging of the ventricles and flow analysis across cardiac valves and the great vessels. Cine gradient echo imaging is useful for evaluation of cardiac wall motion, ventricular volumes and ventricular mass. Flow analysis allows for measurement of velocity and flow during the cardiac cycle that reflects cardiac function. Ultrafast CT allows for measurement of cardiac indices similar to that provided by gradient echo imaging of the ventricles.

The role of cardiovascular magnetic resonance in heart failure

Cardiovascular Magnetic Resonance (CMR) is an accepted gold standard for non-invasive, accurate, and reproducible assessment of cardiac mass and function. The interest in its use for viability, myocardial perfusion and coronary artery imaging is also widespread and growing rapidly as the hardware and expertise becomes available in more centres, and the scans themselves become more cost effective. In patients with heart failure, accurate and reproducible serial assessment of remodelling is of prognostic importance and the lack of exposure to ionizing radiation is helpful. The concept of an integrated approach to heart failure and its complications using CMR is fast becoming a reality, and this will be tested widely in the coming few years, with the new generation of dedicated CMR scanners.

A multicenter validation of an active contour-based left ventricular analysis technique

Quantitative analysis of functional cardiac magnetic resonance (MR) images has been limited by the lack of well-validated, semiautomatic, methods for rapid analysis. We describe the evaluation of a DICOM-compatible PC-based parallel-processing tool, for cardiac magnetic resonance analysis (CAMRA), which supports semiautomatic image mensuration using an active contour model-based algorithm. The CAMRA software was used to analyze data from 12 patients in a multicenter acquisition and analysis trial to compare semiautomatic contour detection with manual planimetry of the left ventricular endocardium from short-axis, breath-held, cine gradient-echo images. There was excellent agreement between the manual and semiautomatic measurements of global left ventricular function, with no significant (P = 0.32) difference in the determination of ejection fraction (-0.9 +/- 3.1% [mean difference +/- 1 standard deviation]). There was no significant interobserver difference in the semiautomatically measured ejection fraction. Additionally, a single observer completed the analysis on data from 30 patients and found no significant (P = 0.05) difference in the determination of ejection fraction (-1.3 +/- 3.5% [mean difference +/- 1 standard deviation]). The CAMRA software demonstrates the capability for the reproducible evaluation of global left ventricular function in cardiac patients, with adequate interobserver reproducibility for use in multicenter trials.