Comparison and evaluation of retrospective intermodality brain image registration techniques.

PURPOSE: The primary objective of this study is to perform a blinded evaluation of a group of retrospective image registration techniques using as a gold standard a prospective, marker-based registration method. To ensure blindedness, all retrospective registrations were performed by participants who had no knowledge of the gold standard results until after their results had been submitted. A secondary goal of the project is to evaluate the importance of correcting geometrical distortion in MR images by comparing the retrospective registration error in the rectified images, i.e., those that have had the distortion correction applied, with that of the same images before rectification. METHOD: Image volumes of three modalities (CT, MR, and PET) were obtained from patients undergoing neurosurgery at Vanderbilt University Medical Center on whom bone-implanted fiducial markers were mounted. These volumes had all traces of the markers removed and were provided via the Internet to project collaborators outside Vanderbilt, who then performed retrospective registrations on the volumes, calculating transformations from CT to MR and/ or from PET to MR. These investigators communicated their transformations again via the Internet to Vanderbilt, where the accuracy of each registration was evaluated. In this evaluation, the accuracy is measured at multiple volumes of interest (VOIs), i.e., areas in the brain that would commonly be areas of neurological interest. A VOI is defined in the MR image and its centroid c is determined. Then, the prospective registration is used to obtain the corresponding point c' in CT or PET. To this point, the retrospective registration is then applied, producing c" in MR. Statistics are gathered on the target registration error (TRE), which is the distance between the original point c and its corresponding point c". RESULTS: This article presents statistics on the TRE calculated for each registration technique in this study and provides a brief description of each technique and an estimate of both preparation and execution time needed to perform the registration. CONCLUSION: Our results indicate that retrospective techniques have the potential to produce satisfactory results much of the time, but that visual inspection is necessary to guard against large errors.

Comparison of edge-based and ridge-based registration of CT and MR brain images.

In modern medicine, several different imaging techniques are frequently employed in the study of a single patient. This is useful, since different images show complementary information on the functionality and/or structure of the anatomy examined. This very difference between modalities, however, complicates the problem of proper registration of the images involved, and rules out the most basic approaches--like direct grey value correlation--to achieve registration. The observation that some common structures will always exist is supportive of the statement that registration may be feasible using edges or ridges present in the images. The existence of such structures defined in the binary sense is questionable, however, and their extraction from images requires a segmentation by definition. In this paper we propose to use fuzzy edgeness and ridgeness images, thus avoiding the need for segmentation and using more of the available information from the original images. We will show that such fuzzy images can be used to achieve accurate registration. Several ridgeness and edgeness computing operators were compared. The best registration results were obtained using a gradient magnitude operator.

MR geometric distortion correction for improved frame-based stereotaxic target localization accuracy.

We present a method to correct the geometric distortion caused by field inhomogeneity in MR images of patients wearing MR-compatible stereotaxic frames. Our previously published distortion correction method derives patient-dependent error maps by computing the phase-difference of 3D images acquired at different TEs. The time difference (delta TE = 4.9 ms at 1.5 T) is chosen such that the water and fat signals are in phase. However, delta TE is long enough to permit phase wraps in the difference images for frequency offsets greater than 205 Hz. Phase unwrapping techniques resolve these only for connected structures; therefore, the phase difference for fiducial rods may be off by multiples of 2 pi relative to the head. We remove this uncertainty by using an additional single 2D phase-different image with delta TE = 1 ms (during which time no phase-wraps are typically expected) to determine the correct multiple of 2 pi for each rod. We tested our method in a cadaver and in a patient using CT as the gold standard. Targets in the frame coordinates were chosen from CT and compared with their locations in MR. Localizing errors using MR compared with CT were as large as 3.7 mm before correction and were reduced to less than 1.11 mm after correction.

Registration error quantification of a surface-based multimodality image fusion system.

This paper presents a new reference data set and associated quantification methodology to assess the accuracy of registration of computerized tomography (CT) and magnetic-resonance (MR) images. Also described is a new semiautomatic surface-based system for registering and visualizing CT and MR images. The registration error of the system was determined using a reference data set that was obtained from a cadaver in which rigid fiducial tubes were inserted prior to imaging. Registration error was measured as the distance between an analytic expression for each fiducial tube in one image set and transformed samples of the corresponding tube obtained from the other. Registration was accomplished by first identifying surfaces of similar anatomic structures in each image set. A transformation that best registered these structures was determined using a nonlinear optimization procedure. Even though the root-mean-square (rms) distance at the registered surfaces was similar to that reported by other groups, it was found that rms distances for the tubes were significantly larger than the final rms distances between the registered surfaces. It was also found that minimizing rms distance at the surface did not minimize rms distance for the tubes.

Automatic registration of CT and MR brain images using correlation of geometrical features.

Describes an automated approach to register CT and MR brain images. Differential operators in scale space are applied to each type of image data, so as to produce feature images depicting "ridgeness". The resulting CT and MR feature images show similarities which can be used for matching. No segmentation is needed and the method is devoid of human interaction. The matching is accomplished by hierarchical correlation techniques. Results of 2-D and 3-D matching experiments are presented. The correlation function ensures an accurate match even if the scanned volumes to be matched do not completely overlap, or if some of the features in the images are not similar.

A versatile system for multimodality image fusion.

This paper presents a versatile system for registering and visualizing computed tomography and magnetic resonance images. The system utilizes a semi-automatic, surface-based registration strategy which has proven useful for registering a number of different anatomical structures. A triangular mesh approximates surfaces in one image set while a set of surface points is used as a surface approximation in the other set. A non-linear optimization procedure determines the transformation that minimizes the total sum-squared perpendicular distance between triangles of the mesh and surface points. This system has been used without modification to successfully register images of the brain, spine and calcaneus.

Method for correcting magnetic resonance image distortion for frame-based stereotactic surgery, with preliminary results.

We previously described a technique for correcting patient-specific magnetic field inhomogeneity spatial distortion in magnetic resonance images (MRI), which was not applicable to patients fitted with MRI-compatible stereotactic fiducial frames. Here we describe an improvement to the technique that permits application for these patients. Measurements with a cadaver head show that this method achieves MRI stereotactic localization accuracy of 1 mm.

Integrated presentation of multimodal brain images.

This article discusses the fusion of brain images from multiple modalities as well as the presentation of the integrated image information. The paper has three parts. First, individual brain imaging modalities are compared as regards clinical appreciation, invasiveness, dimensionality, spatial resolution, temporal resolution, and cost. Next, methods to combine multiple images are briefly surveyed and collated by characteristics as accuracy, patient-friendliness, reproducibility, labour-extensiveness, feasibility of retrospective matching, and general applicability. Finally, techniques to display multimodal image information are outlined and examples of the various options for integrated presentation are shown.

Geometry driven multimodality matching of brain images.

Clinical diagnosis, as well as therapy planning and evaluation, are increasingly supported by multimodal images. There are many instances desiring integration of the information obtained by various imaging devices. This paper describes a new approach to match images of different modalities. Differential operators are used in combination with Gaussian blurring to extract geometric features from the images that correspond to similar structures. The resulting 'feature' images may be used with existing matching techniques that minimize the distance between the features in the images to be matched. Our first application of this new approach concerns matching of MRI and CT brain images. The so-called L upsilon upsilon operator produces a ridge-like feature image from which in CT and MRI the center curve of the cranium is easily extracted. First results of this operator's performance in matching tasks are shown. Another promising operator is the 'umbilicity' operator, which is presented in combination with SPECT images.

Spatial integration of multimodal brain images in cerebral infarction.

Different structural as well as functional imaging techniques are becoming increasingly important in the investigation of patients suffering from an ischemic stroke. Available imaging procedures usually provide complementary data, but the images can not easily be compared due to differences in patient positioning, angulation, and slice thickness. We studied the value of spatial integration of images from different modalities in a patient with an ischemic stroke and used skin markers to integrate the obtained information. Computed tomography (CT), magnetic resonance imaging (MRI), 99mTcHMPAO-single photon emission computed tomography (SPECT) and magnetic resonance spectroscopic imaging (MRSI) were performed in a patient, presenting with a right sided hemiparesis caused by an ischemic stroke. Combination of MRI with CT demonstrated that the infarction visible on CT and MRI corresponded in size and volume. Furthermore, structural and functional images could readily be integrated, thus allowing us to obtain accurate information in this stroke patient. Different imaging modalities provide complementary information in the acute phase of cerebral infarction and multimodality matching can be of great value for improvement of our understanding of the pathophysiology and course of ischemic stroke.

Accurate matching of electromagnetic dipole data with CT and MR images.

Interpretation of EEG (electroencephalography) or MEG (magnetoencephalography) derived three-dimensional dipole localizations is hampered by poor visualization. This paper describes a method for combining dipole data with structural image data of the same patient. To ensure high precision this method utilizes external markers that are easy to apply. These markers can achieve subslice accuracy and can even be used to pinpoint reference points outside the scanned volume. Accurate matching is thus provided even in standard imaging protocols employing thick slices and/or large interslice gaps. The results of the matching method are presented in 2D and 3D visualizations. The hybrid images facilitate the interpretation of dipole localizations with respect to the patient's anatomy.