Faculty Bibliography

The purpose of this paper is to demonstrate the clinical advantages of using semiautomatic volume registration where automatic registration is problematic due to large deformations, small bone anatomy, or extraneous structures. Examples are drawn from clinical cases of MRI/PET breast studies, CT angiography/SPECT cardiac studies, and total wrist arthroplasty. These types of studies should be contrasted with those involving the head, thorax, and pelvis where there is much less deformation and the existence of (some) large bones facilitates automatic matching

PURPOSE: To develop a multislice, first-pass perfusion imaging sequence for increasing the effective dynamic range of the contrast-enhanced blood signal and the contrast-to-noise ratio (CNR) of myocardial wall enhancement. MATERIALS AND METHODS: A hybrid echo-planar imaging (EPI) pulse sequence was modified to acquire data for both the arterial input function (AIF) and the myocardium, using two different saturation-recovery time delays (TDs) and spatial resolutions, after a single saturation pulse. Five healthy subjects were scanned at 3T in three short-axis levels of the heart per heartbeat during passage of a high-dose bolus of contrast agent. The T(1)-weighted signal-time curve of the blood was converted to AIF using empirical conversion tables derived from phantom experiments. RESULTS: In all subjects the calculated AIF was consistently less distorted and higher for the short-TD protocol than for the long-TD protocol (peak concentration: 5.0 +/- 1.0 mM vs. 3.0 +/- 0.6 mM; P < 0.01). A combination of EPI, long TD, high-dose bolus of contrast agent, and 3T imaging yielded relatively strong peak enhancement in the myocardium (CNR = 11.9 +/- 3.3). CONCLUSION: Our dual-imaging approach at 3T seems promising for acquiring both a relatively accurate AIF and a high CNR of myocardial wall enhancement in multiple slices per heartbeat

In this paper, we describe a new approach for reconstructing 3D strains in the myocardium using tagged MR images. We first segment the myocardium using a 3D deformable model driven by image gradients and Gabor filter responses. Tags are automatically detected and tracked as deformable thin plates during systole and early diastole. To keep the tracking results more stable and consistent, we use a combination of gradient information, an intensity probabilistic model, the phase information, and a temporal-spatial smoothness constraint. Based on the tag deformation, we compute a dense displacement in the myocardium around both ventricles. The displacements in x-, y-, and z- directions are calculated separately and are combined to form the final displacement maps. We do not use the information outside the segmented surface of the myocardium to avoid displacement errors caused by noises, artifacts, and correlations between different regions in the myocardium. The strain in the myocardium during the heart cycle is derived from the displacement. This method accepts images of either a tag grid or separate horizontal and vertical tag lines as its input. Experimental results on phantom and real data demonstrate good performance of this method in calculating the myocardial strain

In this paper we present a robust method for segmenting and tracking cardiac contours and tags in 4D cardiac MRI tagged images via spatio-temporal propagation. Our method is based on two main techniques: the Metamorphs segmentation for robust boundary estimation, and the tunable Gabor filter bank for tagging lines enhancement, removal and myocardium tracking. We have developed a prototype system based on the integration of these two techniques, and achieved efficient, robust segmentation and tracking with minimal human interaction

In this paper, we present a novel method for automatically extracting the tagging sheets in tagged cardiac MR images, and tracking their displacement during the heart cycle, using a tunable 3D Gabor filter bank. Tagged MRI is a non-invasive technique for the study of myocardial deformation. We design the 3D Gabor filter bank based on the geometric characteristics of the tagging sheets. The tunable parameters of the Gabor filter bank are used to adapt to the myocardium deformation. The whole 3D image dataset is convolved with each Gabor filter in the filter bank, in the Fourier domain. Then we impose a set of deformable meshes onto the extracted tagging sheets and track them over time. Dynamic estimation of the filter parameters and the mesh internal smoothness are used to help the tracking. Some very encouraging results are shown

In this paper we present an accurate cardiac boundary tracking method for 2D tagged MRI time sequences. This method naturally integrates the motion and the static local appearance features and generates accurate boundary criteria via a boosting approach. We extend the conventional Adaboost classifier into a posterior probability form, which can be embedded in a particle filtering-based shape tracking framework. To make the tracking process more robust and faster, we use a PCA subspace shape representation to constrain the shape variation and lower the dimensionality. We also learn two shape-dynamic models for systole and diastole separately, to predict the shape evolution. Our tracking method incorporates the static appearance, the motion appearance, the shape constraints, and the dynamic prediction in a unified way. The proposed method has been implemented on 50 tagged MRI sequences. The experimental results show the accuracy and robustness of our approach

The pericardium, although seldom the primary cause of systemic illness, can be involved in almost every type of disease. Pericardial involvement may be subtle and escape detection unless specifically sought, or it can overshadow features of the underlying systemic disease. Suspected pericardial disease is usually initially evaluated with echocardiography. However, magnetic resonance imaging can offer additional valuable information. In addition to the excellent resolution and unlimited imaging planes available for visualization of the entire pericardial sac, the wide field of view allows for evaluation of involvement of adjacent cardiac structures. Dynamic functional imaging and tissue characterization with and without contrast can further characterize disease and provide information regarding concomitant myocardial disease and effects on cardiac motion. The treatment of specific pericardial conditions ultimately depends on the underlying disease process. Magnetic resonance imaging can provide useful information to aid in diagnosis, management, and guidance of therapy for pericardial disease

Myocardial perfusion can be estimated, in principle, from first-pass MR images by converting the T(1)-weighted signal-time curves to contrast agent concentration-time curves. Typically, T(1) weighting is achieved by saturating the magnetization with a nonselective radiofrequency (RF) pulse prior to the imaging sequence. The accuracy of the perfusion estimate derived from the single-point T(1)-weighted signal depends on the initial residual longitudinal magnetization (RLM) produced by the saturation pulse. In this study we demonstrate that single-shot, echo-planar imaging can be used to show initial RLM resulting from incomplete saturation due to static magnetic field and RF field inhomogeneities in the heart at 1.5 T. Three saturation pulses, single, composite simple, and composite B(1)-insensitive rotation (BIR-4) were evaluated in phantom and cardiac experiments. The RLM image was calculated by normalizing the saturated image by a proton-density-weighted image. Mean RLM produced by the three saturation pulses was significantly different in noncontrast cardiac imaging (RLM(single) = 0.108 +/- 0.078; RLM(composite) = 0.051 +/- 0.052; RLM(BIR-4) = 0.011 +/- 0.009; P < 0.001; n = 20). Using a BIR-4 pulse to perform saturation of magnetization seems promising for improving the effectiveness and uniformity of T(1) weighting for first-pass perfusion imaging

The right ventricle (RV) of the heart is responsible for pumping blood to the lungs. Its kinematics are not as well understood as that of the left ventricle (LV) due to its thin wall and asymmetric geometry. In this study, the combination of tagged MRI and three-dimensional (3-D) image-processing techniques was used to reconstruct 3-D RV-LV motion and deformation. The reconstructed models were used to quantify the 3-D global and local deformation of the ventricles in a set of normal subjects. When compared with the LV, the RV exhibited a similar twisting pattern, a more longitudinal strain pattern, and a greater amount of displacement