Martin Baiker

Atlas-based whole-body segmentation of mice from low-contrast Micro-CT data

This paper presents a fully automated method for atlas-based whole-body segmentation in non-contrast-enhanced Micro-CT data of mice. The position and posture of mice in such studies may vary to a large extent, complicating data comparison in cross-sectional and follow-up studies. Moreover, Micro-CT typically yields only poor soft-tissue contrast for abdominal organs. To overcome these challenges, we propose a method that divides the problem into an atlas constrained registration based on high-contrast organs in Micro-CT (skeleton, lungs and skin), and a soft tissue approximation step for low-contrast organs. We first present a modification of the MOBY mouse atlas (Segars et al., 2004) by partitioning the skeleton into individual bones, by adding anatomically realistic joint types and by defining a hierarchical atlas tree description. The individual bones as well as the lungs of this adapted MOBY atlas are then registered one by one traversing the model tree hierarchy. To this end, we employ the Iterative Closest Point method and constrain the Degrees of Freedom of the local registration, dependent on the joint type and motion range. This atlas-based strategy renders the method highly robust to exceptionally large postural differences among scans and to moderate pathological bone deformations. The skin of the torso is registered by employing a novel method for matching distributions of geodesic distances locally, constrained by the registered skeleton. Because of the absence of image contrast between abdominal organs, they are interpolated from the atlas to the subject domain using Thin-Plate-Spline approximation, defined by correspondences on the already established registration of high-contrast structures (bones, lungs and skin). We extensively evaluate the proposed registration method, using 26 non-contrast-enhanced Micro-CT datasets of mice, and the skin registration and organ interpolation, using contrast-enhanced Micro-CT datasets of 15 mice. The posture and shape varied significantly among the animals and the data was acquired in vivo. After registration, the mean Euclidean distance was less than two voxel dimensions for the skeleton and the lungs respectively and less than one voxel dimension for the skin. Dice coefficients of volume overlap between manually segmented and interpolated skeleton and organs vary between 0.47+/-0.08 for the kidneys and 0.73+/-0.04 for the brain. These experiments demonstrate the methods effectiveness for overcoming exceptionally large variations in posture, yielding acceptable approximation accuracy even in the absence of soft-tissue contrast in in vivo Micro-CT data without requiring user initialization.

Articulated Whole-Body Atlases for Small Animal Image Analysis: Construction and Applications

PurposeUsing three publicly available small-animal atlases (Sprague–Dawley rat, MOBY, and Digimouse), we built three articulated atlases and present several applications in the scope of molecular imaging.ProceduresMajor bones/bone groups were manually segmented for each atlas skeleton. Then, a kinematic model for each atlas was built: each joint position was identified and the corresponding degrees of freedom were specified.ResultsThe articulated atlases enable automated registration into a common coordinate frame of multimodal small-animal imaging data. This eliminates the postural variability (e.g., of the head, back, and front limbs) that occurs in different time steps and due to modality differences and nonstandardized acquisition protocols.ConclusionsThe articulated atlas proves to be a useful tool for multimodality image combination, follow-up studies, and image processing in the scope of molecular imaging. The proposed models were made publicly available.

Articulated Planar Reformation for Change Visualization in Small Animal Imaging

The analysis of multi-timepoint whole-body small animal CT data is greatly complicated by the varying posture of the subject at different timepoints. Due to these variations, correctly relating and comparing corresponding regions of interest is challenging.In addition, occlusion may prevent effective visualization of these regions of interest. To address these problems, we have developed a method that fully automatically maps the data to a standardized layout of sub-volumes, based on an articulated atlas registration.We have dubbed this process articulated planar reformation, or APR. A sub-volume can be interactively selected for closer inspection and can be compared with the corresponding sub-volume at the other timepoints, employing a number of different comparative visualization approaches. We provide an additional tool that highlights possibly interesting areas based on the change of bone density between timepoints. Furthermore we allow visualization of the local registration error, to give an indication of the accuracy of the registration. We have evaluated our approach on a case that exhibits cancer-induced bone resorption.

FULLY AUTOMATED WHOLE-BODY REGISTRATION IN MICE USING AN ARTICULATED SKELETON ATLAS

In this paper, we propose a fully automated articulated registration approach for whole-body 3D data of mice. The method is based on a hierarchical anatomical model of the skeletal system where we specified position and degrees of freedom for each joint. Model fitting is performed by traversing a hierarchical part-tree, which enables a coarse-to-fine registration from the inner articulations outwards. The method was tested on 12 micro-CT volumes, giving accurate alignment of the skeletal structures in all cases.

Automated registration of whole-body follow-up MicroCT data of mice

In vivo MicroCT imaging of disease models at multiple time points is of great importance for preclinical oncological research, to monitor disease progression. However, the great postural variability between animals in the imaging device complicates data comparison. In this paper we propose a method for automated registration of whole-body MicroCT follow-up datasets of mice. First, we register the skeleton, the lungs and the skin of an articulated animal atlas (Segars et al. 2004) to MicroCT datasets, yielding point correspondence of these structures over all time points. This correspondence is then used to regularize an intensity-based B-spline registration. This two step approach combines the robustness of model-based registration with the high accuracy of intensity-based registration. We demonstrate our approach using challenging whole-body in vivo follow-up MicroCT data and obtain subvoxel accuracy for the skeleton and the skin, based on the Euclidean surface distance. The method is computationally efficient and enables high resolution whole-body registration in approximately 17 minutes with unoptimized code, mostly executed single-threaded.

Organ approximation in μCT data with low soft tissue contrast using an articulated whole-body atlas

In this article, we present an approach for organ approximation in low contrast muCT data of mice using a whole-body mouse atlas (Segars et al. [1]). Starting from a set of landmarks on bone and joint locations, further correspondences are derived on surface representations of the lung by atlas-based registration and on the skin by employing a local geodesic shape context. Subsequently, landmarks on the skeleton, the lung and the skin are used to constrain a Thin-Plate-Spline (TPS) based mapping of major organs from the atlas to the subject domain. The feasibility of the method has been tested by means of 26 CT mouse datasets and a different whole-body mouse atlas (Digimouse [2]). Proper mapping of the lung and the skin as well as major organs could be achieved in all cases yielding a mean Euclidean distance between surface nodes of 0.42 plusmn 0.068 mm for the lung and 0.34 plusmn 0.036 mm for the skin. The performance of the organ interpolation has been assessed on basis of manual segmentations of two CT datasets of mice with injected contrast agent and the Digimouse. The calculated dice indices of volume overlap show significant improvement compared to earlier studies.

2D/3D registration of micro-CT data to multi-view photographs based on a 3D distance map

In this work we present a method for registration of a CT-derived mouse skin surface to two or more 2D, geometrically calibrated, photographs of the same animal using a similarity transformation model. We show that by using a 3D distance map, which is reconstructed from the animal skin silhouettes in the 2D photographs, and by penalizing large angle differences between distance map gradients and CT-based skin surface normals, we are able to construct a registration criterion that is robust to silhouette outliers and yields accurate results for synthetic and real data (mean skin surface distance 0.12mm and 1.35mm respectively).

Atlas-based organ & bone approximation for ex-vivo μMRI mouse data: A pilot study

In this paper we propose a novel semi-automated atlas-based approach for organ and bone approximation for micro-Magnetic Resonance Imaging (μMRI) data of mice. Based on a set of 18 manually indicated landmarks at specific joint & bone locations, individual atlas bones (pelvis, limb bones and sternum) are mapped to the target in a first step and a sparse set of corresponding landmarks on a skin surface representation is determined in a second step. Subsequently, this sparse set on the skin is used to derive a dense set of correspondences relying on matching spectra of local geodesic distances. Finally, determined by the skin correspondence, a Thin-Plate-Spline (TPS) approximation of major organs (heart, lungs, liver, spleen, stomach, kidneys) is performed. The method was tested using 3 µMRI mouse datasets and the MOBY atlas. The performance of the organ approximation algorithm was estimated using manual segmentations of 6 organs for each MRI dataset and calculating Dice indices of organ-volume overlap for each dataset and the atlas. The obtained results indicate excellent fitting of heart and kidneys and moderate fitting of spleen, lungs, liver and stomach. These initial results are satisfactory and comparable to other organ mapping studies using different approaches and μComputed Tomography (CT) mouse data.

Segmentation and Visual Analysis of Whole-Body Mouse Skeleton microSPECT

Whole-body SPECT small animal imaging is used to study cancer, and plays an important role in the development of new drugs. Comparing and exploring whole-body datasets can be a difficult and time-consuming task due to the inherent heterogeneity of the data (high volume/throughput, multi-modality, postural and positioning variability). The goal of this study was to provide a method to align and compare side-by-side multiple whole-body skeleton SPECT datasets in a common reference, thus eliminating acquisition variability that exists between the subjects in cross-sectional and multi-modal studies. Six whole-body SPECT/CT datasets of BALB/c mice injected with bone targeting tracers 99mTc-methylene diphosphonate (99mTc-MDP) and 99mTc-hydroxymethane diphosphonate (99mTc-HDP) were used to evaluate the proposed method. An articulated version of the MOBY whole-body mouse atlas was used as a common reference. Its individual bones were registered one-by-one to the skeleton extracted from the acquired SPECT data following an anatomical hierarchical tree. Sequential registration was used while constraining the local degrees of freedom (DoFs) of each bone in accordance to the type of joint and its range of motion. The Articulated Planar Reformation (APR) algorithm was applied to the segmented data for side-by-side change visualization and comparison of data. To quantitatively evaluate the proposed algorithm, bone segmentations of extracted skeletons from the correspondent CT datasets were used. Euclidean point to surface distances between each dataset and the MOBY atlas were calculated. The obtained results indicate that after registration, the mean Euclidean distance decreased from 11.5±12.1 to 2.6±2.1 voxels. The proposed approach yielded satisfactory segmentation results with minimal user intervention. It proved to be robust for “incomplete” data (large chunks of skeleton missing) and for an intuitive exploration and comparison of multi-modal SPECT/CT cross-sectional mouse data.

Atlas-based articulated skeleton segmentation of μSPECT mouse data

In this paper we propose an automated articulated atlas-based approach for bone segmentation in whole-body μSPECT data of mice, obtained by injecting the 99mTc-methylene diphosphonate (99mTc - MDP). This is a difficult task, since SPECT data is usually noisy and low resolution, and the skeleton image is incomplete with several portions missing (e.g.: in limbs and skull). For this purpose the articulated version of the MOBY atlas skeleton with a correspondent hierarchical tree description is used. Iterative Closest Point registration is used, while constraining the local degrees of freedom (DoFs) in accordance to the type of joint and its range of motion. The method was tested using 3 whole-body μSPECT mouse datasets acquired using a NanoSPECT/CT scanner for small animals and the MOBY atlas. To evaluate the proposed algorithm, manual bone segmentations of extracted skeletons from the correspondent CT datasets were used. Euclidean point to surface distances for each dataset and the MOBY atlas were calculated. The obtained results indicate that after registration, the mean Euclidean distance descreased from 8.37 ± 8.70 to 2.27 ± 2.06 voxels. The results were presented using a novel method for change visualization in small animal imaging (Articulated Planar Reformation).