Deformable Registration of Brain Tumor Images

Overview

Deformable registration of brain tumor images can make possible the pooling of data from different brain tumor patients into a common stereotaxic space, thereby enabling the construction of statistical brain tumor atlases that are based on collective morphological, functional, and pathological information. Such atlases will act as tools for optimal planning of neuro-surgical operations and therapeutic approaches that deal with brain tumors by statistically linking the functional and structural information provided by multi-modality radiological images to variables such as tumor size, grade, subsequent progression, therapeutic approach and outcome. Therefore the goal of this approach is to develop a deformable registration method of brain tumor images from different subjects into a common stereotaxic space. The direct application of currently available deformable registration methods, including our methods described elsewhere (HAMMER [1]) to adapt a neuroanatomical atlas to a patient's tumor-bearing images is destined to be inaccurate in the vicinity of the tumor, due substantial dissimilarity between the two images. Image dissimilarity arises from topological differences between the atlas and the patient's images, severe deformation in the vicinity of the tumor, tissue death and resorption and effects of edema and tissue infiltration. Our approach to solving this problem [2-4] aims to facilitate the registration process by creating first an atlas image that is as similar as possible to the patient's image. Specifically, it involves the integration of three components:

Example 1

Spatial normalization into the atlas space: Such mappings can facilitate the correlation of treatment parameters with therapy outcome. For example the correlation of tumor recurrence and radiation dose profiles (defined in the atlas) can be studied from a large number of patients with tumors emerging from similar anatomical locations. Also tumor model parameters, e.g. size of initial seed indicating tissue death or amount of infiltration, can be quantified and compared across patients, when they are measured in the same space.

The figure on the left shows an example of spatially normalizing a brain tumor image, shown in (a) with contrast agent and in (b) without contrast agent, into a normal atlas space (shown in (d)), using our method The spatially normalized (registered) patient’s image is shown in (c). The registration causes relaxation of the mass effect due to tumor growth and removal of the inter-subject differences. The warped image in (c) shows a small tumor mass indicating the initial location of tumor appearance (defined in the atlas space). Also the surrounding peri-tumor edema or infiltration, as mapped in the normal atlas, is visible.

Example 2

Mapping an atlas into the patient’s image space: The following figure shows the registration of a normal template (atlas) image to eight patients' images using the proposed method. The skull stripped T1-weighted patient’s image (axial, sagittal, or coronal), is shown on the left, and the corresponding section of the atlas warped with our method is shown on the right for each pair of images, correspondingly.

Contact

Dinggang Shen

People

Dinggang Shen, Evangelia Zacharaki

References

[1] D. Shen and C. Davatzikos, "HAMMER: Hierarchical attribute matching mechanism for elastic registration," IEEE Transactions on Medical Imaging, vol. 21, pp. 1421-1439, November 2002.

[2] E. I. Zacharaki, D. Shen, A. Mohamed, and C. Davatzikos, "Registration of Brain Images with Tumors: Towards the Construction of Statistical Atlases for Therapy Planning," in ISBI, Arlington, VA, 2006.

[3] E. I. Zacharaki, D. Shen, S.-K. Lee, and C. Davatzikos, "A Multiresolution Framework for Deformable Registration of Brain Tumor Images," IEEE Transactions on Medical Imaging, submitted, 2006.

[4] A. Mohamed, E. I. Zacharaki, D. Shen, and C. Davatzikos, "Deformable registration of brain tumor images via a statistical model of tumor-induced deformation," Medical Image Analysis, vol. 10, pp. 752-763, 2006.

[2] C. S. Hogea, F. Abraham, G. Biros, and C. Davatzikos , “Fast Solvers for Soft Tissue Simulation with Application to Construction of Brain Tumor Atlases”, technical report, http://www.seas.upenn.edu/~biros/papers/brain06.pdf

[3] C.S. Hogea, C. Davatzikos and G. Biros, “An image-driven parameter estimation problem for a glioma growth model with mass effects”, submitted for publication, http://www.seas.upenn.edu/~biros/papers/gliomas1d.pdf

[4] C.S. Hogea, C. Davatzikos and G. Biros, “Modeling Glioma Growth and Mass Effect in 3D MR Images of the Brain”, accepted MICCAI07

[5] C.S. Hogea, C. Davatzikos and G. Biros, “Brain-tumor interaction biophysical models for medical image registration”, submitted for publication



Find It

FOR SBIA MEMBERS

SBIA Wiki
Collaborative Documentation
Status of Clusters, etc.

How to get here

3600 Market St.
Suite 380
Philadelphia, PA 19104