Deformable medical image registration: a survey

Analysis of macular OCT images using deformable registration

Optical coherence tomography (OCT) of the macula has become increasingly important in the investigation of retinal pathology. However, deformable image registration, which is used for aligning subjects for pairwise comparisons, population averaging, and atlas label transfer, has not been well-developed and demonstrated on OCT images. In this paper, we present a deformable image registration approach designed specifically for macular OCT images. The approach begins with an initial translation to align the fovea of each subject, followed by a linear rescaling to align the top and bottom retinal boundaries. Finally, the layers within the retina are aligned by a deformable registration using one-dimensional radial basis functions. The algorithm was validated using manual delineations of retinal layers in OCT images from a cohort consisting of healthy controls and patients diagnosed with multiple sclerosis (MS). We show that the algorithm overcomes the shortcomings of existing generic registration methods, which cannot be readily applied to OCT images. A successful deformable image registration algorithm for macular OCT opens up a variety of population based analysis techniques that are regularly used in other imaging modalities, such as spatial normalization, statistical atlas creation, and voxel based morphometry. Examples of these applications are provided to demonstrate the potential benefits such techniques can have on our understanding of retinal disease. In particular, included is a pilot study of localized volumetric changes between healthy controls and MS patients using the proposed registration algorithm.

Diffusion maps for multimodal registration

Multimodal image registration is a difficult task, due to the significant intensity variations between the images. A common approach is to use sophisticated similarity measures, such as mutual information, that are robust to those intensity variations. However, these similarity measures are computationally expensive and, moreover, often fail to capture the geometry and the associated dynamics linked with the images. Another approach is the transformation of the images into a common space where modalities can be directly compared. Within this approach, we propose to register multimodal images by using diffusion maps to describe the geometric and spectral properties of the data. Through diffusion maps, the multimodal data is transformed into a new set of canonical coordinates that reflect its geometry uniformly across modalities, so that meaningful correspondences can be established between them. Images in this new representation can then be registered using a simple Euclidean distance as a similarity measure. Registration accuracy was evaluated on both real and simulated brain images with known ground-truth for both rigid and non-rigid registration. Results showed that the proposed approach achieved higher accuracy than the conventional approach using mutual information.

MSM: a new flexible framework for Multimodal Surface Matching

Surface-based cortical registration methods that are driven by geometrical features, such as folding, provide sub-optimal alignment of many functional areas due to variable correlation between cortical folding patterns and function. This has led to the proposal of new registration methods using features derived from functional and diffusion imaging. However, as yet there is no consensus over the best set of features for optimal alignment of brain function. In this paper we demonstrate the utility of a new Multimodal Surface Matching (MSM) algorithm capable of driving alignment using a wide variety of descriptors of brain architecture, function and connectivity. The versatility of the framework originates from adapting the discrete Markov Random Field (MRF) registration method to surface alignment. This has the benefit of being very flexible in the choice of a similarity measure and relatively insensitive to local minima. The method offers significant flexibility in the choice of feature set, and we demonstrate the advantages of this by performing registrations using univariate descriptors of surface curvature and myelination, multivariate feature sets derived from resting fMRI, and multimodal descriptors of surface curvature and myelination. We compare the results with two state of the art surface registration methods that use geometric features: FreeSurfer and Spherical Demons. In the future, the MSM technique will allow explorations into the best combinations of features and alignment strategies for inter-subject alignment of cortical functional areas for a wide range of neuroimaging data sets.

A projected landmark method for reduction of registration error in image-guided surgery systems

PURPOSE

Image-guided surgery systems are limited by registration error, so practical and effective methods to improve accuracy are necessary. A projection point-based method for reducing the surface registration error in image-guided surgery was developed and tested.

METHODS

Checkerboard patterns are projected on visible surfaces to create projected landmarks over a region of interest. Surface information thus becomes available in the form of point clouds of surface point coordinates with submillimeter resolution. The reconstructed 3D point cloud is registered using iterative closest point (ICP) approximation to a 3D point cloud extracted from preoperative CT images of the same region of interest. The projected landmark surface registration method was compared with two other methods using a facial surface phantom: (a) landmark registration using anatomical features, and (b) surface matching based on an additional 40 surface points.

RESULTS

The mean error for the projected landmark surface registration method was 0.64 mm, which was 47.4 and 35.3 % lower relative to mean errors of the anatomical landmark registration and the surface-matching methods, respectively. After applying the proposed method, using target registration error as a gold standard, the resulting mean error was 1.1 mm or a reduction of 61.2 % compared to the anatomical landmark registration.

CONCLUSION

Optical checkerboard pattern projection onto visible surfaces was used to acquire surface point clouds for image-guided surgery registration. A projected landmark method eliminated the effects of unwanted and overlapping points by acquiring the desired points at specific locations. The results were more accurate than conventional landmark or surface registration.

Image artefact propagation in motion estimation and reconstruction in interventional cardiac C-arm CT

The acquisition of data for cardiac imaging using a C-arm computed tomography system requires several seconds and multiple heartbeats. Hence, incorporation of motion correction in the reconstruction step may improve the resulting image quality. Cardiac motion can be estimated by deformable three-dimensional (3D)/3D registration performed on initial 3D images of different heart phases. This motion information can be used for a motion-compensated reconstruction allowing the use of all acquired data for image reconstruction. However, the result of the registration procedure and hence the estimated deformations are influenced by the quality of the initial 3D images. In this paper, the sensitivity of the 3D/3D registration step to the image quality of the initial images is studied. Different reconstruction algorithms are evaluated for a recently proposed cardiac C-arm CT acquisition protocol. The initial 3D images are all based on retrospective electrocardiogram (ECG)-gated data. ECG-gating of data from a single C-arm rotation provides only a few projections per heart phase for image reconstruction. This view sparsity leads to prominent streak artefacts and a poor signal to noise ratio. Five different initial image reconstructions are evaluated: (1) cone beam filtered-backprojection (FDK), (2) cone beam filtered-backprojection and an additional bilateral filter (FFDK), (3) removal of the shadow of dense objects (catheter, pacing electrode, etc) before reconstruction with a cone beam filtered-backprojection (cathFDK), (4) removal of the shadow of dense objects before reconstruction with a cone beam filtered-backprojection and a bilateral filter (cathFFDK). The last method (5) is an iterative few-view reconstruction (FV), the prior image constrained compressed sensing combined with the improved total variation algorithm. All reconstructions are investigated with respect to the final motion-compensated reconstruction quality. The algorithms were tested on a mathematical phantom data set with and without a catheter and on two porcine models using qualitative and quantitative measures. The quantitative results of the phantom experiments show that if no dense object is present within the scan field of view, the quality of the FDK initial images is sufficient for motion estimation via 3D/3D registration. When a catheter or pacing electrode is present, the shadow of these objects needs to be removed before the initial image reconstruction. An additional bilateral filter shows no major improvement with respect to the final motion-compensated reconstruction quality. The results with respect to image quality of the cathFDK, cathFFDK and FV images are comparable. In conclusion, in terms of computational complexity, the algorithm of choice is the cathFDK algorithm.

Multi-atlas based representations for Alzheimer's disease diagnosis

Brain morphometry based classification from magnetic resonance (MR) acquisitions has been widely investigated in the diagnosis of Alzheimer's disease (AD) and its prodromal stage, i.e., mild cognitive impairment (MCI). In the literature, a morphometric representation of brain structures is obtained by spatial normalization of each image into a common space (i.e., a pre-defined atlas) via non-linear registration, thus the corresponding regions in different brains can be compared. However, representations generated from one single atlas may not be sufficient to reveal the underlying anatomical differences between the groups of disease-affected patients and normal controls (NC). In this article, we propose a different methodology, namely the multi-atlas based morphometry, which measures morphometric representations of the same image in different spaces of multiple atlases. Representations generated from different atlases can thus provide the complementary information to discriminate different groups, and also reduce the negative impacts from registration errors. Specifically, each studied subject is registered to multiple atlases, where adaptive regional features are extracted. Then, all features from different atlases are jointly selected by a correlation and relevance based scheme, followed by final classification with the support vector machine (SVM). We have evaluated the proposed method on 459 subjects (97 AD, 117 progressive-MCI (p-MCI), 117 stable-MCI (s-MCI), and 128 NC) from the Alzheimer's Disease Neuroimaging Initiative (ADNI) database, and achieved 91.64% for AD/NC classification and 72.41% for p-MCI/s-MCI classification. Our results clearly demonstrate that the proposed multi-atlas based method can significantly outperform the previous single-atlas based methods.

Concurrent tumor segmentation and registration with uncertainty-based sparse non-uniform graphs

In this paper, we present a graph-based concurrent brain tumor segmentation and atlas to diseased patient registration framework. Both segmentation and registration problems are modeled using a unified pairwise discrete Markov Random Field model on a sparse grid superimposed to the image domain. Segmentation is addressed based on pattern classification techniques, while registration is performed by maximizing the similarity between volumes and is modular with respect to the matching criterion. The two problems are coupled by relaxing the registration term in the tumor area, corresponding to areas of high classification score and high dissimilarity between volumes. In order to overcome the main shortcomings of discrete approaches regarding appropriate sampling of the solution space as well as important memory requirements, content driven samplings of the discrete displacement set and the sparse grid are considered, based on the local segmentation and registration uncertainties recovered by the min marginal energies. State of the art results on a substantial low-grade glioma database demonstrate the potential of our method, while our proposed approach shows maintained performance and strongly reduced complexity of the model.

Image registration for quantitative parametric response mapping of cancer treatment response

Imaging biomarkers capable of early quantification of tumor response to therapy would provide an opportunity to individualize patient care. Image registration of longitudinal scans provides a method of detecting treatment associated changes within heterogeneous tumors by monitoring alterations in the quantitative value of individual voxels over time, which is unattainable by traditional volumetric-based histogram methods. The concepts involved in the use of image registration for tracking and quantifying breast cancer treatment response using parametric response mapping (PRM), a voxel-based analysis of diffusion-weighted magnetic resonance imaging (DW-MRI) scans, are presented. Application of PRM to breast tumor response detection is described, wherein robust registration solutions for tracking small changes in water diffusivity in breast tumors during therapy are required. Methodologies that employ simulations are presented for measuring expected statistical accuracy of PRM for response assessment. Test-retest clinical scans are used to yield estimates of system noise to indicate significant changes in voxel-based changes in water diffusivity. Overall, registration-based PRM image analysis provides significant opportunities for voxel-based image analysis to provide the required accuracy for early assessment of response to treatment in breast cancer patients receiving neoadjuvant chemotherapy.

The distance discordance metric-a novel approach to quantifying spatial uncertainties in intra- and inter-patient deformable image registration

Previous methods to estimate the inherent accuracy of deformable image registration (DIR) have typically been performed relative to a known ground truth, such as tracking of anatomic landmarks or known deformations in a physical or virtual phantom. In this study, we propose a new approach to estimate the spatial geometric uncertainty of DIR using statistical sampling techniques that can be applied to the resulting deformation vector fields (DVFs) for a given registration. The proposed DIR performance metric, the distance discordance metric (DDM), is based on the variability in the distance between corresponding voxels from different images, which are co-registered to the same voxel at location (X) in an arbitrarily chosen 'reference' image. The DDM value, at location (X) in the reference image, represents the mean dispersion between voxels, when these images are registered to other images in the image set. The method requires at least four registered images to estimate the uncertainty of the DIRs, both for inter- and intra-patient DIR. To validate the proposed method, we generated an image set by deforming a software phantom with known DVFs. The registration error was computed at each voxel in the 'reference' phantom and then compared to DDM, inverse consistency error (ICE), and transitivity error (TE) over the entire phantom. The DDM showed a higher Pearson correlation (Rp) with the actual error (Rp ranged from 0.6 to 0.9) in comparison with ICE and TE (Rp ranged from 0.2 to 0.8). In the resulting spatial DDM map, regions with distinct intensity gradients had a lower discordance and therefore, less variability relative to regions with uniform intensity. Subsequently, we applied DDM for intra-patient DIR in an image set of ten longitudinal computed tomography (CT) scans of one prostate cancer patient and for inter-patient DIR in an image set of ten planning CT scans of different head and neck cancer patients. For both intra- and inter-patient DIR, the spatial DDM map showed large variation over the volume of interest (the pelvis for the prostate patient and the head for the head and neck patients). The highest discordance was observed in the soft tissues, such as the brain, bladder, and rectum, due to higher variability in the registration. The smallest DDM values were observed in the bony structures in the pelvis and the base of the skull. The proposed metric, DDM, provides a quantitative tool to evaluate the performance of DIR when a set of images is available. Therefore, DDM can be used to estimate and visualize the uncertainty of intra- and/or inter-patient DIR based on the variability of the registration rather than the absolute registration error.

SELF-SUPERVISED MRI TISSUE SEGMENTATION BY DISCRIMINATIVE CLUSTERING

The study of brain lesions can benefit from a clear identification of transitions between healthy and pathological tissues, through the analysis of brain imaging data. Current signal processing methods, able to address these issues, often rely on strong prior information. In this article, a new method for tissue segmentation is proposed. It is based on a discriminative strategy, in a self-supervised machine learning approach. This method avoids the use of prior information, which makes it very versatile, and able to cope with different tissue types. It also returns tissue probabilities for each voxel, crucial for a good characterization of the evolution of brain lesions. Simulated as well as real benchmark data were used to validate the accuracy of the method and compare it against other segmentation algorithms.