J.H.C. Reiber

3-D active appearance models: segmentation of cardiac MR and ultrasound images

A model-based method for three-dimensional image segmentation was developed and its performance assessed in segmentation of volumetric cardiac magnetic resonance (MR) images and echocardiographic temporal image sequences. Comprehensive design of a three-dimensional (3-D) active appearance model (AAM) is reported for the first time as an involved extension of the AAM framework introduced by Cootes et al. The models behavior is learned from manually traced segmentation examples during an automated training stage. Information about shape and image appearance of the cardiac structures is contained in a single model. This ensures a spatially and/or temporally consistent segmentation of three-dimensional cardiac images. The clinical potential of the 3-D AAM is demonstrated in short-axis cardiac MR images and four-chamber echocardiographic sequences. The methods performance was assessed by comparison with manually identified independent standards in 56 clinical MR and 64 clinical echo image sequences. The AAM method showed good agreement with the independent standard using quantitative indexes of border positioning errors, endo- and epicardial volumes, and left ventricular mass. In MR, the endocardial volumes, epicardial volumes, and left ventricular wall mass correlation coefficients between manual and AAM were R/sup 2/=0.94,0.97,0.82, respectively. For echocardiographic analysis, the area correlation was R/sup 2/=0.79. The AAM method shows high promise for successful application to MR and echocardiographic image analysis in a clinical setting.

Magnetic resonance imaging during dobutamine stress for detection and localization of coronary artery disease. Quantitative wall motion analysis using a modification of the centerline method.

Quantitative measurement of wall motion is essential to assess objectively the functional significance of coronary artery disease. We developed a quantitative wall thickening analysis on stress magnetic resonance images. This study was designed to assess the clinical value of magnetic resonance imaging (MRI) during dobutamine stress for detection and localization of myocardial ischemia in patients with suspected coronary artery disease. Methods and ResultsThirty-nine consecutive patients with clinically suspected coronary artery disease referred for coronary arteriography and 10 normal volunteers underwent gradient- echo MRI at rest and during peak dobutamine stress (infusion rate, 20 μg· kg−1· min−2). MRI was performed in the short-axis plane at four adjacent levels. Display in a cine loop provided a qualitative impression of regional wall motion (cine MRI). A modification of the centerline method was applied for quantitative wall motion analysis by means of calculation of percent systolic wall thickening. Short-axis cine MRI images were analyzed at 100 equally spaced chords constructed perpendicular to a centerline drawn midway between the end-diastolic and end-systolic contours. Dobutamine MRI was considered positive for coronary artery disease if the percent systolic wall thickening of more than four adjacent chords was < 2 SD below the mean values obtained from the normal volunteers. The overall sensitivity of dobutamine MRI for the detection of significant coronary artery disease (diameter stenosis ≥ 50%) was 91% (30 of 33), specificity was 80% (5 of 6), and accuracy was 90% (35 of 39). The sensitivity for identifying one-vessel disease was 88% (15 of 17), for twovessel disease 91% (10 of 11), and for three-vessel disease 100% (5 of 5). The sensitivity for detection of individual coronary artery lesions was 75% for the left anterior descending coronary artery, 87% for the right coronary artery, and 63% for the left circumflex coronary artery. ConclusionsDobutamine MRI clearly identifies wall motion abnormalities by quantitative analysis using a modification of the centerline method. Dobutamine MRI is an accurate method for detection and localization of myocardial ischemia and may emerge as a new noninvasive approach for evaluation of patients with known or suspected coronary artery disease.

Comparison between manual and semiautomated analysis of left ventricular volume parameters from short-axis MR images

PURPOSE The goal of this study was to evaluate a newly developed semiautomated contour detection algorithm for the quantitative analysis of cardiovascular MRI. METHOD Left ventricular function parameters derived from automatically detected endocardial and epicardial contours were compared with results derived from manually traced contours in short-axis multislice GRE MRI studies of 10 normal volunteers and 10 infarct patients. RESULTS Compared with manual image analysis, the semiautomated method resulted in the following systematic and random differences (auto-manual; mean +/- SD): end-diastolic volume: -5.5 +/- 9.7 ml; end-systolic volume: -3.6 +/- 6.5 ml; ejection fraction: 1.7 +/- 4.1%; left ventricular mass: 7.3 +/- 20.6 g. Total analysis time for a complete study was reduced from 3-4 h for the manual analysis to < 20 min using semiautomated contour detection. CONCLUSION Global left ventricular function parameters can be obtained with a high degree of accuracy and precision using the present semiautomated contour detection algorithm.

Diagnostic accuracy of 320-row multidetector computed tomography coronary angiography in the non-invasive evaluation of significant coronary artery disease.

AIMS Multidetector computed tomography coronary angiography (CTA) has emerged as a feasible imaging modality for non-invasive assessment of coronary artery disease (CAD). Recently, 320-row CTA systems were introduced, with 16 cm anatomical coverage, allowing image acquisition of the entire heart within a single heart beat. The aim of the present study was to assess the diagnostic accuracy of 320-row CTA in patients with known or suspected CAD. METHODS AND RESULTS A total of 64 patients (34 male, mean age 61 +/- 16 years) underwent CTA and invasive coronary angiography. All CTA scans were evaluated for the presence of obstructive coronary stenosis by a blinded expert, and results were compared with quantitative coronary angiography. Four patients were excluded from initial analysis due to non-diagnostic image quality. Sensitivity, specificity, and positive and negative predictive values to detect > or =50% luminal narrowing on a patient basis were 100, 88, 92, and 100%, respectively. Moreover, sensitivity, specificity, and positive and negative predictive values to detect > or =70% luminal narrowing on a patient basis were 94, 95, 88, and 98%, respectively. With inclusion of non-diagnostic imaging studies, sensitivity, specificity, and positive and negative predictive values to detect > or =50% luminal narrowing on a patient basis were 100, 81, 88, and 100%, respectively. CONCLUSION The current study shows that 320-row CTA allows accurate non-invasive assessment of significant CAD.

A multiresolution image segmentation technique based on pyramidal segmentation and fuzzy clustering

In this paper, an unsupervised image segmentation technique is presented, which combines pyramidal image segmentation with the fuzzy c-means clustering algorithm. Each layer of the pyramid is split into a number of regions by a root labeling technique, and then fuzzy c-means is used to merge the regions of the layer with the highest image resolution. A cluster validity functional is used to find the optimal number of objects automatically. Segmentation of a number of synthetic as well as clinical images is illustrated and two fully automatic segmentation approaches are evaluated, which determine the left ventricular volume (LV) in 140 cardiovascular magnetic resonance (MR) images. First fuzzy c-means is applied without pyramids. In the second approach the regions generated by pyramidal segmentation are merged by fuzzy c-means. The correlation coefficients of manually and automatically defined LV lumen of all 140 and 20 end-diastolic images were equal to 0.86 and 0.79, respectively, when images were segmented with fuzzy c-means alone. These coefficients increased to 0.90 and 0.93 when the pyramidal segmentation was combined with fuzzy c-means. This method can be applied to any dimensional representation and at any resolution level of an image series. The evaluation study shows good performance in detecting LV lumen in MR images.

Automated measurement of volume flow in the ascending aorta using MR velocity maps : Evaluation of inter- and intraobserver variability in healthy volunteers

PURPOSE An automated contour detection algorithm was developed for the objective and reproducible quantitative analysis of velocity-encoded MR studies of the ascending aorta. METHOD The only user interaction required is the manual definition of a center point inside the cross-section of the aorta in one of the available images. The automated contour detection algorithm detects an initial model contour in this image and subsequently corrects for motion and deformation of the aortic cross-section in each of the acquired images over the complete cardiac cycle using dynamic programming techniques. Integrating the flow velocity values for each pixel within the detected contour results in an instantaneous flow value. Next, by integrating the instantaneous flow values for each acquired phase over the complete cardiac cycle, left ventricular stroke volume measurement could be obtained. The results of the automated method were compared with results derived from manually traced contours in MR studies from 11 healthy volunteers. RESULTS An excellent agreement in stroke volume measurements was observed: signed difference 0.61+/-1.15%. Inter- and intraobserver variabilities were <2% for both manual and automated image analysis methods. Manual tracing of contours required on the order of 10 min; the analysis time for automated contour detection was <6 s/study. CONCLUSION The present contour detection allows fast and reliable left ventricular stroke volume measurements from aortic flow studies using velocity-encoded MR studies in healthy volunteers. Further study is required to assess the accuracy and reproducibility of the algorithm in patients with aortic and aortic valve disease.

Head-to-head comparison of coronary plaque evaluation between multislice computed tomography and intravascular ultrasound radiofrequency data analysis.

OBJECTIVES The purpose of this study was to perform a head-to-head comparison of plaque observations with multislice computed tomography (MSCT) to virtual histology intravascular ultrasound (VH IVUS). BACKGROUND The VH IVUS allows in vivo coronary plaque characterization with high spatial resolution. Noninvasively, plaques may be evaluated with MSCT, but limited data are available. METHODS A total of 50 patients underwent 64-slice MSCT followed by VH IVUS. The Agatston score was evaluated on MSCT in coronary segments where IVUS was performed. Plaques were classified on MSCT as noncalcified, mixed, and calcified. Four plaque components (fibrotic, fibro-fatty, and necrotic core tissues and dense calcium) were identified on VH IVUS, and the presence of thin-cap fibroatheroma was evaluated. RESULTS A moderate correlation was observed between the Agatston score and calcium volume on VH IVUS (r = 0.69, p < 0.0001). In total, 168 coronary plaques were evaluated (48 [29%] noncalcified, 71 [42%] mixed, 49 [29%] calcified). As compared with calcified plaques, noncalcified plaques contained more fibrotic (60.90 +/- 9.21% vs. 54.60 +/- 8.33%, p = 0.001) and fibro-fatty tissues (28.11 +/- 13.03% vs. 21.37 +/- 9.75%, p = 0.006) on VH IVUS. Mixed and calcified plaques contained more dense calcium (7.61 +/- 8.94% vs. 2.68 +/- 3.01%, p = 0.001; 10.18 +/- 6.71% vs. 2.68 +/- 3.01%, p < 0.0001, respectively). Thin-cap fibroatheromas were most frequently observed in mixed plaques as compared with noncalcified and calcified plaques (32%, 13%, 8%, p = 0.002, respectively). CONCLUSIONS A good correlation was observed between calcium quantification on MSCT and VH IVUS. In addition, plaque classification on MSCT paralleled relative plaque composition on VH IVUS, although VH IVUS provided more precise plaque characterization. Mixed plaques on MSCT were associated with high-risk features on VH IVUS.

A 3-D Active Shape Model Driven by Fuzzy Inference: Application to Cardiac CT and MR

Manual quantitative analysis of cardiac left ventricular function using multislice CT and MR is arduous because of the large data volume. In this paper, we present a 3-D active shape model (ASM) for semiautomatic segmentation of cardiac CT and MR volumes, without the requirement of retraining the underlying statistical shape model. A fuzzy c-means based fuzzy inference system was incorporated into the model. Thus, relative gray-level differences instead of absolute gray values were used for classification of 3-D regions of interest (ROIs), removing the necessity of training different models for different modalities/acquisition protocols. The 3-D ASM was evaluated using 25 CT and 15 MR datasets. Automatically generated contours were compared to expert contours in 100 locations. For CT, 82.4% of epicardial contours and 74.1% of endocardial contours had a maximum error of 5 mm along 95% of the contour arc length. For MR, those numbers were 93.2% (epicardium) and 91.4% (endocardium). Volume regression analysis revealed good linear correlations between manual and semiautomatic volumes, r 2 ges 0.98. This study shows that the fuzzy inference 3-D ASM is a robust promising instrument for semiautomatic cardiac left ventricle segmentation. Without retraining its statistical shape component, it is applicable to routinely acquired CT and MR studies.

Fully Automated Motion Correction in First-Pass Myocardial Perfusion MR Image Sequences

This paper presents a novel method for registration of cardiac perfusion magnetic resonance imaging (MRI). The presented method is capable of automatically registering perfusion data, using independent component analysis (ICA) to extract physiologically relevant features together with their time-intensity behavior. A time-varying reference image mimicking intensity changes in the data of interest is computed based on the results of that ICA. This reference image is used in a two-pass registration framework. Qualitative and quantitative validation of the method is carried out using 46 clinical quality, short-axis, perfusion MR datasets comprising 100 images each. Despite varying image quality and motion patterns in the evaluation set, validation of the method showed a reduction of the average right ventricle (LV) motion from 1.26plusmn0.87 to 0.64plusmn0.46 pixels. Time-intensity curves are also improved after registration with an average error reduced from 2.65plusmn7.89% to 0.87plusmn3.88% between registered data and manual gold standard. Comparison of clinically relevant parameters computed using registered data and the manual gold standard show a good agreement. Additional tests with a simulated free-breathing protocol showed robustness against considerable deviations from a standard breathing protocol. We conclude that this fully automatic ICA-based method shows an accuracy, a robustness and a computation speed adequate for use in a clinical environment.

Correlation between annual change in health status and computer tomography derived lung density in subjects with α1-antitrypsin deficiency

Background: There is increasing recognition that questionnaires of health status and lung density measurements are more sensitive tools for assessing progression of emphysema than forced expiratory volume in 1 second (FEV1) and transfer coefficient (Kco). A study was undertaken to investigate prospectively the correlation between annual change in health status and computer tomography (CT) derived lung density in subjects with α1-antitrypsin deficiency. Methods: Twenty two patients of mean (SD) age 40.7 (9.2) years with ZZ type α1-antitrypsin deficiency were investigated at baseline and 30 months later by FEV1 and Kco, St George Respiratory Questionnaire (SGRQ), and by a spiral CT scan of the chest. CT data of chest images were analysed using software designed for automated lung contour detection and lung density measurements. The density data were corrected for changes in inspiration levels. Results: Changes in lung density, expressed as 15th percentile point or relative area below –950 HU, correlated well with changes in health status (SGRQ total score): R = −0.56, p = 0.007 or R = 0.6, p  = 0.003. Neither changes in health status nor changes in lung density correlated significantly with changes in FEV1 or changes in Kco. Conclusions: The SGRQ total score (which is a global measure in COPD) and lung density (a specific measure of emphysema) are sensitive to deterioration in patients with α1-antitrypsin deficiency. This finding may facilitate future studies with new drugs specific for emphysema, a frequently occurring component of COPD.