Spatially weighted mutual information image registration for image guided radiation therapy.

PURPOSE To develop a new metric for image registration that incorporates the (sub)pixelwise differential importance along spatial location and to demonstrate its application for image guided radiation therapy (IGRT). METHODS It is well known that rigid-body image registration with mutual information is dependent on the size and location of the image subset on which the alignment analysis is based [the designated region of interest (ROI)]. Therefore, careful review and manual adjustments of the resulting registration are frequently necessary. Although there were some investigations of weighted mutual information (WMI), these efforts could not apply the differential importance to a particular spatial location since WMI only applies the weight to the joint histogram space. The authors developed the spatially weighted mutual information (SWMI) metric by incorporating an adaptable weight function with spatial localization into mutual information. SWMI enables the user to apply the selected transform to medically "important" areas such as tumors and critical structures, so SWMI is neither dominated by, nor neglects the neighboring structures. Since SWMI can be utilized with any weight function form, the authors presented two examples of weight functions for IGRT application: A Gaussian-shaped weight function (GW) applied to a user-defined location and a structures-of-interest (SOI) based weight function. An image registration example using a synthesized 2D image is presented to illustrate the efficacy of SWMI. The convergence and feasibility of the registration method as applied to clinical imaging is illustrated by fusing a prostate treatment planning CT with a clinical cone beam CT (CBCT) image set acquired for patient alignment. Forty-one trials are run to test the speed of convergence. The authors also applied SWMI registration using two types of weight functions to two head and neck cases and a prostate case with clinically acquired CBCT/ MVCT image sets. The SWMI registration with a Gaussian weight function (SWMI-GW) was tested between two different imaging modalities: CT and MRI image sets. RESULTS SWMI-GW converges 10% faster than registration using mutual information with an ROI. SWMI-GW as well as SWMI with SOI-based weight function (SWMI-SOI) shows better compensation of the target organ's deformation and neighboring critical organs' deformation. SWMI-GW was also used to successfully fuse MRI and CT images. CONCLUSIONS Rigid-body image registration using our SWMI-GW and SWMI-SOI as cost functions can achieve better registration results in (a) designated image region(s) as well as faster convergence. With the theoretical foundation established, we believe SWMI could be extended to larger clinical testing.

[1]  M. Staring,et al.  A rigidity penalty term for nonrigid registration. , 2007, Medical physics.

[2]  Max A. Viergever,et al.  Mutual-information-based registration of medical images: a survey , 2003, IEEE Transactions on Medical Imaging.

[3]  Frank Chung-Hoon Rhee,et al.  SU‐HH‐AUD C‐08: Spatial Weighted Mutual Information for Image Registration in Image Guided Radiation Therapy , 2008 .

[4]  Max A. Viergever,et al.  Mutual information matching in multiresolution contexts , 2001, Image Vis. Comput..

[5]  Weiguo Lu,et al.  Multi-margin optimization with daily selection (mmods) for image-guided radiotherapy , 2002 .

[6]  Paul A. Viola,et al.  Alignment by Maximization of Mutual Information , 1997, International Journal of Computer Vision.

[7]  Soo-Hyung Kim,et al.  Multimodality Image Registration Using Spatial Procrustes Analysis and Modified Conditional Entropy , 2009, J. Signal Process. Syst..

[8]  D. Hill,et al.  Medical image registration , 2001, Physics in medicine and biology.

[9]  Guy Marchal,et al.  Multimodality image registration by maximization of mutual information , 1997, IEEE Transactions on Medical Imaging.

[10]  L. Xing,et al.  Overview of image-guided radiation therapy. , 2006, Medical dosimetry : official journal of the American Association of Medical Dosimetrists.

[11]  Josien P. W. Pluim,et al.  Evaluation of Optimization Methods for Nonrigid Medical Image Registration Using Mutual Information and B-Splines , 2007, IEEE Transactions on Image Processing.

[12]  J. Kapatoes,et al.  Limited-data image registration for radiotherapy positioning and verification. , 2002, International journal of radiation oncology, biology, physics.

[13]  Daniel Rueckert,et al.  Non-rigid registration using higher-order mutual information , 2000, Medical Imaging.

[14]  Max A. Viergever,et al.  Image registration by maximization of combined mutual information and gradient information , 2000, IEEE Transactions on Medical Imaging.

[15]  Jorge Nocedal,et al.  On the limited memory BFGS method for large scale optimization , 1989, Math. Program..

[16]  D. McLean,et al.  The effect of source to image distance on scattered radiation to the image receptor , 2004, Australasian Physics & Engineering Sciences in Medicine.

[17]  R. Jeraj,et al.  Automatic registration of megavoltage to kilovoltage CT images in helical tomotherapy: an evaluation of the setup verification process for the special case of a rigid head phantom. , 2006, Medical physics.

[18]  J H Siewerdsen,et al.  Compensators for dose and scatter management in cone-beam computed tomography. , 2007, Medical physics.

[19]  David R. Haynor,et al.  PET-CT image registration in the chest using free-form deformations , 2003, IEEE Transactions on Medical Imaging.

[20]  Murray H. Loew,et al.  Weighted and deterministic entropy measure for image registration using mutual information , 1998, Medical Imaging.

[21]  Michael Unser,et al.  Optimization of mutual information for multiresolution image registration , 2000, IEEE Trans. Image Process..

[22]  Lewis D. Griffin,et al.  Zen and the art of medical image registration: correspondence, homology, and quality , 2003, NeuroImage.

[23]  Lei Xing,et al.  Image registration with auto-mapped control volumes , 2006 .

[24]  Zujun Hou,et al.  A Review on MR Image Intensity Inhomogeneity Correction , 2006, Int. J. Biomed. Imaging.

[25]  Edward R. Vrscay,et al.  Mutual Information-Based Methods to Improve Local Region-of-Interest Image Registration , 2005, ICIAR.

[26]  Mark Holden,et al.  A Review of Geometric Transformations for Nonrigid Body Registration , 2008, IEEE Transactions on Medical Imaging.

[27]  Wufan Chen,et al.  Elastic Registration Algorithm of Medical Images Based on Fuzzy Set , 2006, WBIR.

[28]  Guy Marchal,et al.  Automated multi-moda lity image registration based on information theory , 1995 .