Evaluation of a New Method for Automated Detection of Left Ventricular Boundaries in Time Series of Magnetic Resonance Images Using an Active Appearance Motion Model

Abstract

The purpose of this study was the evaluation of a computer algorithm for the automated detection of endocardial and epicardial boundaries of the left ventricle in time series of short-axis magnetic resonance images based on an Active Appearance Motion Model (AAMM). In 20 short-axis MR examinations, manual contours were defined in multiple temporal frames (from end-diastole to end-systole) in multiple slices from base to apex. Using a leave-one-out procedure, the image data and contours were used to build 20 different AAMMs giving a statistical description of the ventricular shape, gray value appearance, and cardiac motion patterns in the training set. Automated contour detection was performed by iteratively deforming the AAMM within statistically allowed limits until an optimal match was found between the deformed AAMM and the underlying image data of the left-out subject. Global ventricular function results derived from automatically detected contours were compared with results obtained from manually traced boundaries. The AAMM contour detection method was successful in 17 of 20 studies. The three failures were excluded from further statistical analysis. Automated contour detection resulted in small, but statistically nonsignificant, underestimations of ventricular volumes and mass: differences for end-diastolic volume were 0.3%+/-12.0%, for end-systolic volume 2.0%+/-23.4% and for left ventricular myocardial mass 0.73%+/-14.9% (mean+/-SD). An excellent agreement was observed in the ejection fraction: difference of 0.1%+/-6.7%. In conclusion, the presented fully automated contour detection method provides assessment of quantitative global function that is comparable to manual analysis.

Automated frame‐by‐frame endocardial border detection from cardiac magnetic resonance images for quantitative assessment of left ventricular function: Validation and clinical feasibility

“…A majority of techniques described in the literature in this context are based on the use of image intensity gradients, and thus may not necessarily be ideally suited for CMR images, because gradients in these images may not be strong enough to allow accurate endocardial border detection and because the algorithms may be dependent on image quality and the specific pulse sequence used for imaging (23). There are different types of algorithms for boundary detection from CMR images, including approaches based on deformable models (24,25), active shape models (26,27), active appearance models (28,29), and expectation maximization method (30). However, these methods usually require extensive manual tracing for building the model database or for the definition of the training set, thus limiting their clinical application.…”

Segmentation of the left ventricle in short‐axis sequences by combining deformation flow and optical flow

“…However, the accuracy of a statistical method is determined by its manually defined training set, which requires significant effort and includes variability due to human error. Statistical methods also have poor performance on images from patients with diseases that are not represented in the training set [9].…”

“…Since many methods have been proposed to solve this challenging and clinically important problem, we divide the state-of-the-art literatures roughly into three categories: (i) spatial domain methods [2][3][4][5][6], (ii) statistical methods [7][8][9][10][11][12][13], (iii) time domain tracking methods [14][15][16][17][18][19][20][21][22]. Many spatial domain methods utilise an intensity threshold to identify the left ventricular cavity from images which have welldefined differences in pixel intensity between the blood pool and the myocardium.…”

A New Approach for Automatic Identification of the Ventricular Boundary in Cine Magnetic Resonance Images

“…Statistical methods also have poor performance on images from patients with diseases that are not represented in the training set. 12 Because of heart's characteristics of motion, tracking methods are perhaps the most efficient methods for segmenting the ventricular endocardium from MRI. These methods "track" a known boundary in the first image frame to subsequent frames based on calculated differences between the images.…”

“…It has been studied for decades yet remains a challenging and clinically important problem. We divide the state of art literatures into three categories: (1) spatial domain methods; [1][2][3][4][5][6] (2) statistical methods; [7][8][9][10][11][12][13][14][15][16][17][18][19][20][21][22][23][24][25] and (3) time domain tracking methods. [26][27][28][29] Many spatial domain intensity methods utilize a global threshold to accurately identify the ventricular cavity from images which have well-defined differences in pixel intensity between the blood pool and the myocardium.…”