An automatic technique for finding and localizing externally attached markers in CT and MR volume images of the head

Automatic patient registration for port placement in minimally invasive endoscopic surgery

Optimal port placement is a delicate issue in minimally invasive endoscopic surgery, particularly in robotically assisted surgery. A good choice of the instruments' and endoscope's ports can avoid time-consuming consecutive new port placement. We present a novel method to intuitively and precisely plan the port placement. The patient is registered to its pre-operative CT by just moving the endoscope around fiducials, which are attached to the patient's thorax and are visible in its CT. Their 3D positions are automatically reconstructed. Without prior time-consuming segmentation, the pre-operative CT volume is directly rendered with respect to the endoscope or instruments. This enables the simulation of a camera flight through the patient's interior along the instruments' axes to easily validate possible ports.

MRI/PET nonrigid breast-image registration using skin fiducial markers

We propose a finite-element method (FEM) deformable breast model that does not require elastic breast data for nonrigid PET/MRI breast image registration. The model is applicable only if the stress conditions in the imaged breast are virtually the same in PET and MRI. Under these conditions, the observed intermodality displacements are solely due the imaging/reconstruction process. Similar stress conditions are assured by use of an MRI breast-antenna replica for breast support during PET, and use of the same positioning. The tetrahedral volume and triangular surface elements are used to construct the FEM mesh from the MRI image. Our model requires a number of fiducial skin markers (FSM) visible in PET and MRI. The displacement vectors of FSMs are measured followed by the dense displacement field estimation by first distributing the displacement, vectors linearly over the breast surface and then distributing them throughout the volume. Finally, the floating MRI image is warped to a fixed PET image, by using an appropriate shape function in the interpolation from mesh nodes to voxels. We tested our model on an elastic breast phantom with simulated internal lesions and on a small number of patients imaged, with FMS using PET and MRI. Using simulated lesions (in phantom) and real lesions (in patients) visible in both PET and MRI, we established that the target registration error (TRE) is below two pet voxels.

Hybrid navigation interface for orthopedic and trauma surgery

Several visualization methods for intraoperative navigation systems were proposed in the past. In standard slice based navigation, three dimensional imaging data is visualized on a two dimensional user interface in the surgery room. Another technology is the in-situ visualization i.e. the superimposition of imaging data directly into the view of the surgeon, spatially registered with the patient. Thus, the three dimensional information is represented on a three dimensional interface. We created a hybrid navigation interface combining an augmented reality visualization system, which is based on a stereoscopic head mounted display, with a standard two dimensional navigation interface. Using an experimental setup, trauma surgeons performed a drilling task using the standard slice based navigation system, different visualization modes of an augmented reality system, and the combination of both. The integration of a standard slice based navigation interface into an augmented reality visualization overcomes the shortcomings of both systems.

Minimally invasive, image-guided, facial-recess approach to the middle ear: demonstration of the concept of percutaneous cochlear access in vitro

HYPOTHESIS

Image-guided surgery will permit accurate access to the middle ear via the facial recess using a single drill hole from the lateral aspect of the mastoid cortex.

BACKGROUND

The widespread use of image-guided methods in otologic surgery has been limited by the need for a system that achieves the necessary level of accuracy with an easy-to-use, noninvasive fiducial marker system. We have developed and recently reported such a system (accuracy within the temporal bone = 0.76 +/- 0.23 mm; n = 234 measurements). With this system, image-guided otologic surgery is feasible.

METHODS

Skulls (n = 2) were fitted with a dental bite-block affixed fiducial frame and scanned by computed tomography using standard temporal-bone algorithms. The frame was removed and replaced with an infrared emitter used to track the skull during dissection. Tracking was accomplished using an infrared tracker and commercially available software. Using this system in conjunction with a tracked otologic drill, the middle ear was approached via the facial recess using a single drill hole from the lateral aspect of the mastoid cortex. The path of the drill was verified by subsequently performing a traditional temporal bone dissection, preserving the tunnel of bone through which the drill pass had been made.

RESULTS

An accurate approach to the middle ear via the facial recess was achieved without violating the canal of the facial nerve, the horizontal semicircular canal, or the external auditory canal.

CONCLUSIONS

Image-guided otologic surgery provides access to the cochlea via the facial recess in a minimally invasive, percutaneous fashion. While the present study was confined to in vitro demonstration, these exciting results warrant in vivo testing, which may lead to clinically applicable access.

In vivo three-dimensional wrist motion analysis using magnetic resonance imaging and volume-based registration

This study represents a new attempt to non-invasively analyze three-dimensional motions of the wrist in vivo. A volume-based registration method using magnetic resonance imaging (MRI) was developed to avoid radiation exposure. The primary aim was to evaluate the accuracy of volume-based registration and compare it with surface-based registration. The secondary aim was to evaluate contributions of the scaphoid and lunate to global wrist motion during flexion-extension motion (FEM), radio-ulnar deviation (RUD) and radial-extension/ulnoflexion, "dart-throwing" motion (DTM) in the right wrists of 12 healthy volunteers. Volume-based registration displayed a mean rotation error of 1.29 degrees +/-1.03 degrees and a mean translation error of 0.21+/-0.25 mm and was significantly more accurate than surface-based registration in rotation. Different patterns of contribution of the scaphoid and lunate were identified for FEM, RUD, and DTM. The scaphoid contributes predominantly in the radiocarpal joint during FEM, in the midcarpal joint during RUD and almost equally between these joints during DTM. The lunate contributes almost equally in both joints during FEM and predominantly in the midcarpal joint during RUD and DTM.

Development of a three-dimensional freehand endorectal ultrasound system for use in rectal cancer imaging

The initial study reporting the accuracy of an optically tracked endorectal ultrasound (TERUS) probe for the purpose of improving the staging of rectal cancer is presented here. In this work we describe the need for a more accurate ERUS system and why the incorporation of image guidance makes this goal feasible. A rectal phantom was constructed with five targets placed in positions where tumors normally occur. The locations of these targets were found using two different imaging modalities, CT and ultrasound, and the target registration error (TRE) between these two image sets was calculated. The average TRE of 33 image captures of the five targets using TERUS was 2.1 mm. This is a promising outcome because the desired tumor margins for rectal cancer are on the order of centimeters. These preliminary results support the proof of concept for a TERUS system that should improve ultrasound imaging in rectal cancer.

In vitro assessment of image-guided otologic surgery: submillimeter accuracy within the region of the temporal bone

OBJECTIVES

Application of image-guided surgery to otology has been limited by the need for submillimeter accuracy via a fiducial system that is easily usable (noninvasive and nonobstructive).

METHODS

A dental bite-block was fitted with a rigid frame with 7 fiducial markers surrounding each external ear. The temporal bones of 3 cadaveric skulls were removed and replaced with surgical targets arranged in a bull's-eye pattern about the centroid of each temporal bone. The surgical targets were identified both within CT scans and in physical space using an infrared optical tracking system. The difference between positions in CT space versus physical space was calculated as target registration error.

RESULTS

A total of 234 independent target registration errors were calculated. Mean +/- standard deviation = 0.73 mm +/- 0.25 mm.

CONCLUSIONS

These findings show that image-guided otologic surgery with submillimeter accuracy is achievable with a minimally invasive fiducial frame. Significance In vivo validation of the system is ongoing. With such validation, this system may facilitate clinically applicable image-guided otologic surgery.

EBM RATING

A.

Accuracy of customized miniature stereotactic platforms

In this study, a new system was evaluated for implanting deep-brain stimulators based on a one-piece platform for each trajectory customized from a preoperative planning image. During surgery, the platform is attached to skull-implanted posts that extend through the scalp. The platform acts as a miniature stereotactic frame to provide guidance for parallel cannulas as they are advanced through a burr hole to the target. Accuracy is determined from a postoperative CT. For each implantation, the distance between the position observed in the postoperative image and the position calculated relative to the platform from the preoperative image is our measure of error. Because this measure incorporates the surgical error of electrode anchoring, brain shift between preoperative and postoperative scanning, and error in the measurement of the position of the electrode in CT, it will tend to overestimate the true error. The mean error was 2.8 mm for 20 implantations. These data reflect favorably the accuracy of this system when compared with others.

Intensity-based 2D-3D spine image registration incorporating a single fiducial marker

RATIONALE AND OBJECTIVES

The two-dimensional (2D)-three dimensional (3D) registration of a computed tomography image to one or more x-ray projection images has a number of image-guided therapy applications. In general, fiducial marker-based methods are fast, accurate, and robust, but marker implantation is not always possible, often is considered too invasive to be clinically acceptable, and entails risk. There also is the unresolved issue of whether it is acceptable to leave markers permanently implanted. Intensity-based registration methods do not require the use of markers and can be automated because such geometric features as points and surfaces do not need to be segmented from the images. However, for spine images, intensity-based methods are susceptible to local optima in the cost function and thus need initial transformations that are close to the correct transformation.

MATERIALS AND METHODS

In this report, we propose a hybrid similarity measure for 2D-3D registration that is a weighted combination of an intensity-based similarity measure (mutual information) and a point-based measure using one fiducial marker. We evaluate its registration accuracy and robustness by using gold-standard clinical spine image data from four patients.

RESULTS

Mean registration errors for successful registrations for the four patients were 1.3 and 1.1 mm for the intensity-based and hybrid similarity measures, respectively. Whereas the percentage of successful intensity-based registrations (registration error < 2.5 mm) decreased rapidly as the initial transformation got further from the correct transformation, the incorporation of a single marker produced successful registrations more than 99% of the time independent of the initial transformation.

CONCLUSION

The use of one fiducial marker reduces 2D-3D spine image registration error slightly and improves robustness substantially. The findings are potentially relevant for image-guided therapy. If one marker is sufficient to obtain clinically acceptable registration accuracy and robustness, as the preliminary results using the proposed hybrid similarity measure suggest, the marker can be placed on a spinous process, which could be accomplished without penetrating muscle or using fluoroscopic guidance, and such a marker could be removed relatively easily.

EEG source imaging

OBJECTIVE

Electroencephalography (EEG) is an important tool for studying the temporal dynamics of the human brain's large-scale neuronal circuits. However, most EEG applications fail to capitalize on all of the data's available information, particularly that concerning the location of active sources in the brain. Localizing the sources of a given scalp measurement is only achieved by solving the so-called inverse problem. By introducing reasonable a priori constraints, the inverse problem can be solved and the most probable sources in the brain at every moment in time can be accurately localized.

METHODS AND RESULTS

Here, we review the different EEG source localization procedures applied during the last two decades. Additionally, we detail the importance of those procedures preceding and following source estimation that are intimately linked to a successful, reliable result. We discuss (1) the number and positioning of electrodes, (2) the varieties of inverse solution models and algorithms, (3) the integration of EEG source estimations with MRI data, (4) the integration of time and frequency in source imaging, and (5) the statistical analysis of inverse solution results.

CONCLUSIONS AND SIGNIFICANCE

We show that modern EEG source imaging simultaneously details the temporal and spatial dimensions of brain activity, making it an important and affordable tool to study the properties of cerebral, neural networks in cognitive and clinical neurosciences.

The influence of head frame distortions on stereotactic localization and targeting

A strong attachment of a stereotactic head frame to the patient's skull may cause distortions of the head frame. The aim of this work was to identify possible distortions of the head frame, to measure the degree of distortion occurring in clinical practice and to investigate its influence on stereotactic localization and targeting. A model to describe and quantify the distortion of the Riechert-Mundinger (RM) head frame was developed. Distortions were classified as (a) bending and (b) changes from the circular ring shape. Ring shape changes were derived from stereotactic CT scans and frame bending was determined from intraoperative stereotactic x-ray images of patients with implanted 125I-seeds acting as landmarks. From the examined patient data frame bending was determined to be 0.74 mm+/-0.32 mm and 1.30 mm in maximum. If a CT-localizer with a top ring is used, frame bending has no influence on stereotactic CT-localization. In stereotactic x-ray localization, frame bending leads to an overestimation of the z-coordinate by 0.37 mm+/-0.16 mm on average and by 0.65 mm in maximum. The accuracy of patient positioning in radiosurgery is not affected by frame bending. But in stereotactic surgery with an RM aiming bow trajectory displacements are expected. These displacements were estimated to be 0.36 mm+/-0.16 mm (max. 0.74 mm) at the target point and 0.65 mm+/-0.30 mm (max. 1.31 mm) at the entry point level. Changes from the circularring shape are small and do not compromise the accuracy of stereotactic targeting and localization. The accuracy of CT-localization was found to be close to the resolution limit due to voxel size. Our findings for frame bending of the RM frame could be validated by statistical analysis and by comparison with an independent patient examination. The results depend on the stereotactic system and details of the localizers and instruments and also reflect our clinical practice. Therefore, a generalization is not possible. Preliminary experience with a new MR-compatible RM head frame made of ceramics shows no frame distortions as with the conventional frame made of an Al-Cu-Mg alloy.

MR-guided stereotactic neurosurgery—comparison of fiducial-based and anatomical landmark transformation approaches

For application in magnetic resonance (MR) guided stereotactic neurosurgery, two methods for transformation of MR-image coordinates in stereotactic, frame-based coordinates exist: the direct stereotactic fiducial-based transformation method and the indirect anatomical landmark method. In contrast to direct stereotactic MR transformation, indirect transformation is based on anatomical landmark coregistration of stereotactic computerized tomography and non-stereotactic MR images. In a patient study, both transformation methods have been investigated with visual inspection and mutual information analysis. Comparison was done for our standard imaging protocol, including t2-weighted spin-echo as well as contrast enhanced t1-weighted gradient-echo imaging. For t2-weighted spin-echo imaging, both methods showed almost similar and satisfying performance with a small, but significant advantage for fiducial-based transformation. In contrast, for t1-weighted gradient-echo imaging with more geometric distortions due to field inhomogenities and gradient nonlinearity than t2-weighted spin-echo imaging, mainly caused by a reduced bandwidth per pixel, anatomical landmark transformation delivered markedly better results. Here, fiducial-based transformation yielded results which are intolerable for stereotactic neurosurgery. Mean Euclidian distances between both transformation methods were 0.96 mm for t2-weighted spin-echo and 1.67 mm for t1-weighted gradient-echo imaging. Maximum deviations were 1.72 mm and 3.06 mm, respectively.

Phantom validation of coregistration of PET and CT for image-guided radiotherapy

Radiotherapy treatment planning integrating positron emission tomography (PET) and computerized tomography (CT) is rapidly gaining acceptance in the clinical setting. Although hybrid systems are available, often the planning CT is acquired on a dedicated system separate from the PET scanner. A limiting factor to using PET data becomes the accuracy of the CT/PET registration. In this work, we use phantom and patient validation to demonstrate a general method for assessing the accuracy of CT/PET image registration and apply it to two multi-modality image registration programs. An IAEA (International Atomic Energy Association) brain phantom and an anthropomorphic head phantom were used. Internal volumes and externally mounted fiducial markers were filled with CT contrast and 18F-fluorodeoxyglucose (FDG). CT, PET emission, and PET transmission images were acquired and registered using two different image registration algorithms. CT/PET Fusion (GE Medical Systems, Milwaukee, WI) is commercially available and uses a semi-automated initial step followed by manual adjustment. Automatic Mutual Information-based Registration (AMIR), developed at our institution, is fully automated and exhibits no variation between repeated registrations. Registration was performed using distinct phantom structures; assessment of accuracy was determined from registration of the calculated centroids of a set of fiducial markers. By comparing structure-based registration with fiducial-based registration, target registration error (TRE) was computed at each point in a three-dimensional (3D) grid that spans the image volume. Identical methods were also applied to patient data to assess CT/PET registration accuracy. Accuracy was calculated as the mean with standard deviation of the TRE for every point in the 3D grid. Overall TRE values for the IAEA brain phantom are: CT/PET Fusion = 1.71 +/- 0.62 mm, AMIR = 1.13 +/- 0.53 mm; overall TRE values for the anthropomorphic head phantom are: CT/PET Fusion = 1.66 +/- 0.53 mm, AMIR = 1.15 +/- 0.48 mm. Precision (repeatability by a single user) measured for CT/PET Fusion: IAEA phantom = 1.59 +/- 0.67 mm and anthropomorphic head phantom = 1.63 +/- 0.52 mm. (AMIR has exact precision and so no measurements are necessary.) One sample patient demonstrated the following accuracy results: CT/PET Fusion = 3.89 +/- 1.61 mm, AMIR = 2.86 +/- 0.60 mm. Semi-automatic and automatic image registration methods may be used to facilitate incorporation of PET data into radiotherapy treatment planning in relatively rigid anatomic sites, such as head and neck. The overall accuracies in phantom and patient images are < 2 mm and < 4 mm, respectively, using either registration algorithm. Registration accuracy may decrease, however, as distance from the initial registration points (CT/PET fusion) or center of the image (AMIR) increases. Additional information provided by PET may improve dose coverage to active tumor subregions and hence tumor control. This study shows that the accuracy obtained by image registration with these two methods is well suited for image-guided radiotherapy.

Automated fiducial marker detection for patient registration in image-guided neurosurgery

OBJECTIVE

The registration of applied fiducial markers within the preoperative data is often left to the surgeon, who has to identify and tag the center of each marker. This is both time-consuming and a potential source of error. For this reason, the development of an automated procedure was desirable. In this study, we have investigated the accuracy of a software algorithm for detecting fiducial markers within the navigation data set. The influence of adjustable values for accuracy and threshold on the sensitivity and specificity of the detection process, as well as the time gain, was investigated.

PATIENTS AND METHODS

One hundred MP-RAGE MRI data sets of patients with different pathologies who were scheduled for image-guided surgery were used in this study. A total of 591 applied fiducial markers were to be detected using the algorithm of the software VVPlanning 1.3 (BrainLAB, Heimstetten, Germany) on a Pentium II standard PC. The size value of a marker in the y-direction is called "accuracy" and depends on the slice thickness. "Threshold" describes the gray level above which the algorithm starts searching for pixel clusters. The threshold value was changed stepwise on the basis of a constant "accuracy" value. The "accuracy" value was changed on the basis of that threshold value at which all markers were detected correctly.

RESULTS

The time needed for automatic detection varied between 12 s and 25 s. An optimum value for adjustable marker size was found to be 1.1 mm, with 8 undetected markers (1.35%) and 7 additionally detected structures (1.18%) out of 591. The mean gray level (Threshold) for all data sets above which marker detection was correct was 248.9. The automatic detection of markers was good for higher gray levels, with 11 missed markers (1.86%). Starting the algorithm at lower gray levels led to a decreased incidence of missed markers (0.17%), but increased the incidence of additionally detected structures to 27.92%.

CONCLUSION

The automatic marker-detection algorithm is a robust, fast and objective instrument for reliable fiducial marker registration when used with optimum settings for both threshold and accuracy.

A modified method to perform the frameless biopsy

Neuronavigation is increasingly being used to assist in stereotactic neurosurgery due to its frameless property. In this study, we developed and assessed a modified method of performing stereotactic brain biopsies by combining the use of the Fisher stereotactic biopsy instrument, that was fixed on universal quick-lock holder, under infrared guidance of the BrainLab VectorVision Neuronavigation system. Eighteen patients received a frameless stereotactic procedure in this study, including 5 cases of brain biopsy, 2 cases of abscess aspiration, 10 cases of hematoma aspiration and one case of Ommaya reservoir implantation. All cases were on target and successful. In this paper, we present our technique, discuss the advantage and disadvantages of the method and review the literature.

An imaging co-registration system using novel non-invasive and non-radioactive external markers

We present a system of image co-registration and its validation in phantom and volunteer studies. The system co-registered images via six novel non-invasive and non-radioactive external markers. The fiducial markers were attached with sponge bases on the skin surface of the phantom and the volunteers in a non-collinear and non-coplanar array. The subjects were scanned with a 1.5-T magnetic resonance (MR) imager using 2D spin-echo T1-weighted (SE) and 3D spoiled gradient recalled pulse sequences (SPGR) and with a positron emission tomography (PET) scanner for transmission imaging (TI) and emission imaging (EI). The sponge bases created radiolucent gaps with good contrast between the fiducial markers and skin surface. They made the markers visible with clear edge boundaries on both PET and MR images. The images to be registered were rescaled, interpolated, reformatted and followed by point-to-point registration for coordinate determination and the estimation of geometrical transformation and fiducial registration errors (FREs) via the fiducial markers. The images formed four matched pairs of inter-modality (SE-TI, SPGR-TI, SE-EI and SPGR-EI) and two pairs of intra-modality (SE-SPGR, TI-EI) imaging for direct co-registration. The parameters for direct co-registration of SE-TI and SPRG-TI were subsequently used as a bridge and applied for indirect co-registration of SE with EI (SE-EI(TI)) and SPGR with EI (SPGR-EI(TI)), respectively. The overall FREs of the phantom were, respectively, 1.50 mm for inter-modality and 1.10 mm for intra-modality direct co-registration. Those of volunteers were, respectively, 1.79 mm for inter-modality and 1.21 mm for intra-modality direct co-registration. For indirect co-registration, the overall FREs of the phantom were 2.53 mm (SE-EI(TI)) and 2.28 (SPGR-EI(TI)) mm; those of volunteers were 2.84 mm (SE-EI(TI)) and 2.81 mm (SPGR-EI(TI)). The errors of direct co-registration were smaller than those of indirect co-registration; the errors of phantom studies, MR-EI and SPGR-PET were smaller than those of the volunteer studies, MR-TI and SE-PET, respectively (all P<0.01, paired-difference test). In conclusion, motion artefacts, imaging blurring and spatial resolution of imaging remained the key factors affecting the accuracy of co-registration. High-accuracy indirect co-registration is provided by using non-invasive and non-radioactive external fiducial markers. The errors were less than 3 mm for both phantom and volunteer studies. The system is applicable for imaging co-registration of inter-modality non-dual imaging, inter-modality multi-tracer imaging and intra-modality multiple parameter images in clinical practice.

Computer-aided navigation in neurosurgery

The article comprises three main parts: a historical review on navigation, the mathematical basics for calculation and the clinical applications of navigation devices. Main historical steps are described from the first idea till the realisation of the frame-based and frameless navigation devices including robots. In particular the idea of robots can be traced back to the Iliad of Homer, the first testimony of European literature over 2500 years ago. In the second part the mathematical calculation of the mapping between the navigation and the image space is demonstrated, including different registration modalities and error estimations. The error of the navigation has to be divided into the technical error of the device calculating its own position in space, the registration error due to inaccuracies in the calculation of the transformation matrix between the navigation and the image space, and the application error caused additionally by anatomical shift of the brain structures during operation. In the third part the main clinical fields of application in modern neurosurgery are demonstrated, such as localisation of small intracranial lesions, skull-base surgery, intracerebral biopsies, intracranial endoscopy, functional neurosurgery and spinal navigation. At the end of the article some possible objections to navigation-aided surgery are discussed.

Design and implementation of a PC-based image-guided surgical system

In interactive, image-guided surgery, current physical space position in the operating room is displayed on various sets of medical images used for surgical navigation. We have developed a PC-based surgical guidance system (ORION) which synchronously displays surgical position on up to four image sets and updates them in real time. There are three essential components which must be developed for this system: (1) accurately tracked instruments; (2) accurate registration techniques to map physical space to image space; and (3) methods to display and update the image sets on a computer monitor. For each of these components, we have developed a set of dynamic link libraries in MS Visual C++ 6.0 supporting various hardware tools and software techniques. Surgical instruments are tracked in physical space using an active optical tracking system. Several of the different registration algorithms were developed with a library of robust math kernel functions, and the accuracy of all registration techniques was thoroughly investigated. Our display was developed using the Win32 API for windows management and tomographic visualization, a frame grabber for live video capture, and OpenGL for visualization of surface renderings. We have begun to use this current implementation of our system for several surgical procedures, including open and minimally invasive liver surgery.

Automatic MR volume registration and its evaluation for the pelvis and prostate

A three-dimensional (3D) mutual information registration method was created and used to register MRI volumes of the pelvis and prostate. It had special features to improve robustness. First, it used a multi-resolution approach and performed registration from low to high resolution. Second, it used two similarity measures, correlation coefficient at lower resolutions and mutual information at full resolution, because of their particular advantages. Third, we created a method to avoid local minima by restarting the registration with randomly perturbed parameters. The criterion for restarting was a correlation coefficient below an empirically determined threshold. Experiments determined the accuracy of registration under conditions found in potential applications in prostate cancer diagnosis, staging, treatment and interventional MRI (iMRI) guided therapies. Images were acquired in the diagnostic (supine) and treatment position (supine with legs raised). Images were also acquired as a function of bladder filling and the time interval between imaging sessions. Overall studies on three patients and three healthy volunteers, when both volumes in a pair were obtained in the diagnostic position under comparable conditions, bony landmarks and prostate 3D centroids were aligned within 1.6 +/- 0.2 mm and 1.4 +/- 0.2 mm, respectively, values only slightly larger than a voxel. Analysis suggests that actual errors are smaller because of the uncertainty in landmark localization and prostate segmentation. Between the diagnostic and treatment positions, bony landmarks continued to register well, but prostate centroids moved towards the posterior 2.8-3.4 mm. Manual cropping to remove voxels in the legs was necessary to register these images. In conclusion, automatic, rigid body registration is probably sufficiently accurate for many applications in prostate cancer. For potential iMRI-guided treatments, the small prostate displacement between the diagnostic and treatment positions can probably be avoided by acquiring volumes in similar positions and by reducing bladder and rectal volumes.

Centroid-based maximum intensity projections

PURPOSE

Maximum intensity projection (MIP) is a three-dimensional visualization technique for tomographic angiograms. While conventional MIPs display contralateral vascular anatomy, this study uses centroid calculations to remove this information. It is necessary to provide accurate, unambiguous vessel depiction and identical projections regardless of slice orientation.

METHOD

A mathematical model was formed using parameters from clinical images of the head. The vessel widths from the resulting projections were measured and compared with the model. To test the consistency of the projection process, a clinical image set was reformatted and projections of the same view were compared.

RESULTS

The vessel widths were smaller than in the model while varying interpolation and noise. Similar projection views were generated for all slice orientations, but some misalignment was present.

CONCLUSION

Vessel width is affected by the ray's path length and interpolation method. Some slight misalignment is present because the reformatting process alters the centroid calculations.

Automated CT marker segmentation for image registration in radionuclide therapy

In this paper a novel, automated CT marker segmentation technique for image registration is described. The technique, which is based on analysing each CT slice contour individually, treats the cross sections of the external markers as protrusions of the slice contour. Knowledge-based criteria, using the shape and dimensions of the markers, are defined to enable marker identification and segmentation. Following segmentation, the three-dimensional (3D) markers' centroids are localized using an intensity-weighted algorithm. Finally, image registration is performed using a least-squares fit algorithm. The technique was applied to both simulated and patient studies. The patients were undergoing 131I-mIBG radionuclide therapy with each study comprising several 99mTc single photon emission computed tomography (SPECT) scans and one CT marker scan. The mean residual 3D registration errors (+/- 1 SD) computed for the simulated and patient studies were 1.8 +/- 0.3 mm and 4.3 +/- 0.5 mm respectively.

Localization of intravascular devices with paramagnetic markers in MR images

Magnetic resonance imaging (MRI) offers potential advantages over conventional X-ray techniques for guiding and evaluating intravascular interventions. The development of methods to safely and robustly localize and track devices under MRI guidance is mandatory to enable automatic scan plane adaptation so as to exploit the three-dimensional imaging capabilities of the MRI scanner. With regard to the issue of radiofrequency-induced heating, passive approaches to catheter tracking are inherently safe. These techniques visualize intravascular devices by exploiting the susceptibility artifacts associated with the devices. To promote conspicuity, the devices are equipped with paramagnetic markers. This paper introduces a method to enable automatic localization of devices by its ability to recognize markers in two-dimensional MR images. The method requires a coarse segmentation of the vasculature of interest, and consists of two steps. First, it performs a series of postprocessing operations including calculation of the winding number image and of the Laplacian image to detect marker candidates in the image. Second, the device is localized by matching the detected pattern of candidates to the known distance template of the device markers. Results of an animal experiment and of a clinical application are demonstrated. Validation in phantom experiments shows that the method is able to localize the device in 95% of the cases.

The influence of electrode location errors on EEG dipole source localization with a realistic head model

OBJECTIVES

Inaccurate information about the electrode locations on the scalp will introduce errors in electroencephalogram dipole source localization results. The present study uses computer simulations to evaluate such errors in a realistic head model and in the context of noise.

METHODS

A realistic head model was constructed from magnetic resonance imaging scans and 29 electrodes placed on the head according to the 10-20 International System. Twenty sets of electrode displacements, with a mean value of 5 mm, were generated and 200 single dipoles evenly located in the brain were used as test sources. The boundary element method was employed for the forward calculation and dipole fitting was carried out at different noise levels.

RESULTS

For a noise-free signal, the source localization error due to electrode misplacement is about 5 mm, whereas it is about 2 mm for normal noisy signals.

CONCLUSIONS

For realistic head models, dipole estimation error due to electrode misplacement is negligible compared with errors caused by noise.

Depth-buffer targeting for spatially accurate 3-D visualization of medical images

During interactive image-guided surgery (IIGS), a surgeon uses data from medical images to help guide the surgical procedure. At Vanderbilt University, an IIGS software system called Orion has been developed which is capable of displaying up to four 512 x 512 images and the current surgical position using an active optical tracking system. Orion is capable of displaying data from any tomographic image volume and from any NTSC video image. An additional display module has been implemented to display three-dimensional information as well as the tomographic slices. This provides the surgeon with valuable anatomical information that is not readily obtained from the tomographic slices alone. Before the surgery, a set of rendered images is created, each with a different angular view of the tomographic volume in order to surround the site of surgical interest. The major objectives of the display module are to display the appropriate rendered image from the set, identify the current probe position on the selected image, and provide an indication of distance between the probe and the physical point of the anatomy indicated on the image. This can provide the surgeon with vital information such as distance to blood vessels, tumors, or other critical structures.

Boundary-based warping of brain MR images

The goal of this work was to develop a warping technique for mapping a brain image to another image or atlas data, with minimum user interaction and independent of gray level information. We have developed and tested three different methods for warping magnetic resonance (MR) brain images. We utilize a deformable contour to extract and warp the boundaries of the two images. A mesh-grid coordinate system is constructed for each brain, by applying a distance transformation to the resulting contours, and scaling. In the first method (MGC), the first image is mapped to the second image based on a one-to-one mapping between different layers defined by the mesh-grid. In the second method (IDW), the corresponding pixels in the two images are found using the above mesh-grid system and a local inverse-distance weights interpolation. In the third proposed method (TSB), a subset of grid points is used for finding the parameters of a spline transformation, which defines the global warping. The warping methods were applied to clinical MR consisting of diffusion-weighted and T2-weighted images of the human brain. The IDW and TSB methods were superior in ranking of diagnostic quality of the warped MR images to the MGC (P < 0.01) as defined by a neuroradiologist. The deformable contour warping produced excellent diagnostic quality for the diffusion-weighted images coregistered and warped to T2 weighted images. J. Magn. Reson. Imaging 2000;12:417-429.

Surface registration for use in interactive, image-guided liver surgery

OBJECTIVE

Liver surgery is difficult because of limited external landmarks, significant vascularity, and inexact definition of intra-hepatic anatomy. Intra-operative ultrasound (IOUS) has been widely used in an attempt to overcome these difficulties, but is limited by its two-dimensional nature, inter-user variability, and image obliteration with ablative or resectional techniques. Because the anatomy of the liver and intra-operative removal of hepatic ligaments make intrinsic or extrinsic point-based registration impractical, we have implemented a surface registration technique to map physical space into CT image space, and have tested the accuracy of this method on an anatomical liver phantom with embedded tumor targets.

MATERIALS AND METHODS

Liver phantoms were created from anatomically correct molds with "tumors" embedded within the substance of the liver. Helical CT scans were performed with 3-mm slices. Using an optically active position sensor, the surface of the liver was digitized according to anatomical segments. A surface registration was performed and RMS errors of the locations of internal tumors are presented as verification. An initial point-based marker registration was performed and considered the "gold standard" for error measurement.

RESULTS

Errors for surface registration were 2.9 mm for the entire surface and 2.8 mm for embedded targets.

CONCLUSION

This is an initial study considering the use of surface registration for the purpose of physical-to-image registration in the area of liver surgery.

Automatic localization of EEG electrode markers within 3D MR data

The electrical activity of the brain can be monitored using ElectroEncephaloGraphy (EEG). From the positions of the EEG electrodes, it is possible to localize focal brain activity. Thereby, the accuracy of the localization strongly depends on the accuracy with which the positions of the electrodes can be determined. In this work, we present an automatic, simple, and accurate scheme that detects EEG electrode markers from 3D MR data of the human head.

EEG dipole source localization using artificial neural networks

Localization of focal electrical activity in the brain using dipole source analysis of the electroencephalogram (EEG), is usually performed by iteratively determining the location and orientation of the dipole source, until optimal correspondence is reached between the dipole source and the measured potential distribution on the head. In this paper, we investigate the use of feed-forward layered artificial neural networks (ANNs) to replace the iterative localization procedure, in order to decrease the calculation time. The localization accuracy of the ANN approach is studied within spherical and realistic head models. Additionally, we investigate the robustness of both the iterative and the ANN approach by observing the influence on the localization error of both noise in the scalp potentials and scalp electrode mislocalizations. Finally, after choosing the ANN structure and size that provides a good trade off between low localization errors and short computation times, we compare the calculation times involved with both the iterative and ANN methods. An average localization error of about 3.5 mm is obtained for both spherical and realistic head models. Moreover, the ANN localization approach appears to be robust to noise and electrode mislocations. In comparison with the iterative localization, the ANN provides a major speed-up of dipole source localization. We conclude that an artificial neural network is a very suitable alternative for iterative dipole source localization in applications where large numbers of dipole localizations have to be performed, provided that an increase of the localization errors by a few millimetres is acceptable.

Frameless stereotaxy with scalp-applied fiducial markers for brain biopsy procedures: experience in 218 cases

OBJECT

The goal of this study was to develop and assess the use and limitations of performing brain biopsy procedures by using image-guided surgical navigation systems (SNSs; that is, frameless stereotactic systems) with scalp-applied fiducial markers.

METHODS

Two hundred eighteen percutaneous brain biopsies were performed in 213 patients by using a frameless stereotactic SNS that operated with either sonic or optical digitizer technology and scalp-applied fiducial markers for the purpose of registering image space with operating room space. Common neurosurgical and stereotactic instrumentation was adapted for use with a localizing wand, and recently developed target and trajectory guidance software was used. Eight (3.7%) of the 218 biopsy specimens were nondiagnostic; five of these (2.4%) were obtained during procedures in 208 supratentorial lesions and three were obtained during procedures in 10 infratentorial lesions (30%; p < 0.001). Complications related to the biopsy procedure occurred in eight patients (seven of whom had supratentorial lesions and one of whom had an infratentorial lesion, p > 0.25). Five complications were intracerebral hemorrhages (two of which required craniotomy), two were infections, and one was wound breakdown after instillation of intratumoral carmustine following biopsy. There were only three cases of sustained morbidity, and there were two deaths and one delayed deterioration due to disease progression. Two surgeons performed the majority of the procedures (193 cases). The three surgeons who performed more than 10 biopsies had complication rates lower than 5%, whereas two of the remaining four surgeons had complication rates greater than 10% (p = 0.15). Twenty-three additional procedures were performed in conjunction with the biopsies: nine brachytherapies; five computer-assisted endoscopies; four cyst aspirations; two instillations of carmustine; two placements of Ommaya reservoirs; and one craniotomy.

CONCLUSIONS

Brain biopsy procedures in which guidance is provided by a frameless stereotactic SNS with scalp-applied fiducial markers represents a safe and effective alternative to frame-based stereotactic procedures for supratentorial lesions. There were comparable low rates of morbidity and a high degree of diagnostic success. Strategies for performing posterior fossa biopsies are suggested.

The role of image-guided technology in the surgical planning and resection of gliomas

Today's image-guided techniques provide the surgeon with new tools to plan and execute surgery on gliomas. When coupled with good judgment, they can extend what is safely operable and may maximize the extent of surgical resection. Intraoperative imaging such as ultrasonography or MRI may expand the utility of these systems in the foreseeable future.

Retrospective intermodality registration techniques for images of the head: surface-based versus volume-based

The primary objective of this study is to perform a blinded evaluation of two groups of retrospective image registration techniques, using as a gold standard a prospective marker-based registration method, and to compare the performance of one group with the other. These techniques have already been evaluated individually [27]. In this paper, however, we find that by grouping the techniques as volume based or surface based, we can make some interesting conclusions which were not visible in the earlier study. In order to ensure blindness, all retrospective registrations were performed by participants who had no knowledge of the gold-standard results until after their results had been submitted. Image volumes of three modalities: X-ray computed tomography (CT), magnetic resonance (MR), and positron emission tomography (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 (VOI's). Our results indicate that the volume-based techniques in this study tended to give substantially more accurate and reliable results than the surface-based ones for the CT-to-MR registration tasks, and slightly more accurate results for the PET-to-MR tasks. Analysis of these results revealed that the rotational component of error was more pronounced for the surface-based group. It was also apparent that all of the registration techniques we examined have the potential to produce satisfactory results much of the time, but that visual inspection is necessary to guard against large errors.

Predicting error in rigid-body point-based registration

Guidance systems designed for neurosurgery, hip surgery, and spine surgery, and for approaches to other anatomy that is relatively rigid can use rigid-body transformations to accomplish image registration. These systems often rely on point-based registration to determine the transformation, and many such systems use attached fiducial markers to establish accurate fiducial points for the registration, the points being established by some fiducial localization process. Accuracy is important to these systems, as is knowledge of the level of that accuracy. An advantage of marker-based systems, particularly those in which the markers are bone-implanted, is that registration error depends only on the fiducial localization error (FLE) and is thus to a large extent independent of the particular object being registered. Thus, it should be possible to predict the clinical accuracy of marker-based systems on the basis of experimental measurements made with phantoms or previous patients. This paper presents two new expressions for estimating registration accuracy of such systems and points out a danger in using a traditional measure of registration accuracy. The new expressions represent fundamental theoretical results with regard to the relationship between localization error and registration error in rigid-body, point-based registration. Rigid-body, point-based registration is achieved by finding the rigid transformation that minimizes "fiducial registration error" (FRE), which is the root mean square distance between homologous fiducials after registration. Closed form solutions have been known since 1966. The expected value (FRE2) depends on the number N of fiducials and expected squared value of FLE, (FLE-2, but in 1979 it was shown that (FRE2) is approximately independent of the fiducial configuration C. The importance of this surprising result seems not yet to have been appreciated by the registration community: Poor registrations caused by poor fiducial configurations may appear to be good due to a small FRE value. A more critical and direct measure of registration error is the "target registration error" (TRE), which is the distance between homologous points other than the centroids of fiducials. Efforts to characterize its behavior have been made since 1989. Published numerical simulations have shown that (TRE2) is roughly proportional to (FLE2)/N and, unlike (FRE2), does depend in some way on C. Thus, FRE, which is often used as feedback to the surgeon using a point-based guidance system, is in fact an unreliable indicator of registration-accuracy. In this work we derive approximate expressions for (TRE2), and for the expected squared alignment error of an individual fiducial. We validate both approximations through numerical simulations. The former expression can be used to provide reliable feedback to the surgeon during surgery and to guide the placement of markers before surgery, or at least to warn the surgeon of potentially dangerous fiducial placements; the latter expression leads to a surprising conclusion: Expected registration accuracy (TRE) is worst near the fiducials that are most closely aligned! This revelation should be of particular concern to surgeons who may at present be relying on fiducial alignment as an indicator of the accuracy of their point-based guidance systems.

Surface-based registration of CT images to physical space for image-guided surgery of the spine: a sensitivity study

This paper presents a method designed to register preoperative computed tomography (CT) images to vertebral surface points acquired intraoperatively from ultrasound (US) images or via a tracked probe. It also presents a comparison of the registration accuracy achievable with surface points acquired from the entire posterior surface of the vertebra to the accuracy achievable with points acquired only from the spinous process and central laminar regions. Using a marker-based method as a reference, this work shows that submillimetric registration accuracy can be obtained even when a small portion of the posterior vertebral surface is used for registration. It also shows that when selected surface patches are used, CT slice thickness is not a critical parameter in the registration process. Furthermore, the paper includes qualitative results of registering vertebral surface points in US images to multiple CT slices. The method has been tested with US points and physical points on a plastic spine phantom and with simulated data on a patient CT scan.

Validation of object-induced MR distortion correction for frameless stereotactic neurosurgery

Spatial fidelity is a paramount issue in image guided neurosurgery. Until recently, three-dimensional computed tomography (3D CT) has been the primary modality because it provides fast volume capture with pixel level (1 mm) accuracy. While three-dimensional magnetic resonance (3D MR) images provide superior anatomic information, published image capture protocols are time consuming and result in scanner- and object-induced magnetic field inhomogeneities which raise inaccuracy above pixel size. Using available scanner calibration software, a volumetric algorithm to correct for object-based geometric distortion, and a Fast Low Angle SHot (FLASH) 3D MR-scan protocol, we were able to reduce mean CT to MR skin-adhesed fiducial marker registration error from 1.36 to 1.09 mm. After dropping the worst one or two of six fiducial markers, mean registration error dropped to 0.62 mm (subpixel accuracy). Three dimensional object-induced error maps present highest 3D MR spatial infidelity at the tissue interfaces (skin/air, scalp/skull) where frameless stereotactic fiducial markers are commonly applied. The algorithm produced similar results in two patient 3D MR-scans.

Registration of head CT images to physical space using a weighted combination of points and surfaces

Most previously reported registration techniques that align three-dimensional image volumes by matching geometrical features such as points or surfaces use a single type of feature. We recently reported a hybrid registration technique that uses a weighted combination of multiple geometrical feature shapes. In this study we use the weighted geometrical feature (WGF) algorithm to register computed tomography (CT) images of the head to physical space using the skin surface only, the bone surface only, and various weighted combinations of these surfaces and one fiducial point (centroid of a bone-implanted marker). We use data acquired from 12 patients that underwent temporal lobe craniotomies for the resection of cerebral lesions. We evaluate and compare the accuracy of the registrations obtained using these various approaches by using as a reference gold standard the registration obtained using three bone-implanted markers. The results demonstrate that a combination of geometrical features can improve the accuracy of CT-to-physical space registration. Point-based registration requires a minimum of three noncolinear points. The position of a bone-implanted marker can be determined much more accurately than that of a skin-affixed marker or an anatomic landmark. A major disadvantage of using bone-implanted markers is that an invasive procedure is required to implant each marker. By combining surface information, the WGF algorithm allows registration to be performed using only one or two such markers. One important finding is that the use of a single very accurate point (a bone-implanted marker) allows very accurate surface-based registration to be achieved using very few surface points. Finally, the WGF algorithm, which not only allows the combination of multiple types of geometrical information but also handles point-based and surface-based registration as degenerate cases, could form the foundation of a "flexible" surgical navigation system that allows the surgeon to use what he considers the method most appropriate for an individual clinical situation.

Measurement of intraoperative brain surface deformation under a craniotomy

OBJECTIVE

Several causes of spatial inaccuracies in image-guided surgery have been carefully studied and documented for several systems. These include error in identifying the external features used for registration, geometrical distortion in the preoperative images, and error in tracking the surgical instruments. Another potentially important source of error is brain deformation between the time of imaging and the time of surgery or during surgery. In this study, we measured the deformation of the dura and brain surfaces between the time of imaging and the start of surgical resection for 21 patients.

METHODS

All patients underwent intraoperative functional mapping, allowing us to measure brain surface motion at two times that were separated by nearly an hour after opening the dura but before performing resection. The positions of the dura and brain surfaces were recorded and transformed to the coordinate space of a preoperative magnetic resonance image, using the Acustar surgical navigation system (manufactured by Johnson & Johnson Professional, Inc., Randolph, MA) (the Acustar trademark and associated intellectual property rights are now owned by Picker International, Highland Heights, OH). This system performs image registration with bone-implanted markers and tracks a surgical probe by optical triangulation.

RESULTS

The mean displacements of the dura and the first and second brain surfaces were 1.2, 4.4, and 5.6 mm, respectively, with corresponding mean volume reductions under the craniotomy of 6, 22, and 29 cc. The maximum displacement was greater than 10 mm in approximately one-third of the patients for the first brain surface measurement and one-half of the patients for the second. In all cases, the direction of brain shift corresponded to a "sinking" of the brain intraoperatively, compared with its preoperative position. Analysis of the measurement error revealed that its magnitude was approximately 1 to 2 mm. We observed two different patterns of the brain surface deformation field, depending on the inclination of the craniotomy with respect to gravity. Separate measurements of brain deformation within the closed cranium caused by changes in patient head orientation with respect to gravity suggested that less than 1 mm of the brain shift recorded intraoperatively could have resulted from the change in patient orientation between the time of imaging and the time of surgery.

CONCLUSION

These results suggest that intraoperative brain deformation is an important source of error that needs to be considered when using surgical navigation systems.

Visual assessment of the accuracy of retrospective registration of MR and CT images of the brain

In a previous study we demonstrated that automatic retrospective registration algorithms can frequently register magnetic resonance (MR) and computed tomography (CT) images of the brain with an accuracy of better than 2 mm, but in that same study we found that such algorithms sometimes fail, leading to errors of 6 mm or more. Before these algorithms can be used routinely in the clinic, methods must be provided for distinguishing between registration solutions that are clinically satisfactory and those that are not. One approach is to rely on a human observer to inspect the registration results and reject images that have been registered with insufficient accuracy. In this paper, we present a methodology for evaluating the efficacy of the visual assessment of registration accuracy. Since the clinical requirements for level of registration accuracy are likely to be application dependent, we have evaluated the accuracy of the observer's estimate relative to six thresholds: 1-6 mm. The performance of the observers was evaluated relative to the registration solution obtained using external fiducial markers that are screwed into the patient's skull and that are visible in both MR and CT images. This fiducial marker system provides the gold standard for our study. Its accuracy is shown to be approximately 0.5 mm. Two experienced, blinded observers viewed five pairs of clinical MR and CT brain images, each of which had each been misregistered with respect to the gold standard solution. Fourteen misregistrations were assessed for each image pair with misregistration errors distributed between 0 and 10 mm with approximate uniformity. For each misregistered image pair each observer estimated the registration error (in millimeters) at each of five locations distributed around the head using each of three assessment methods. These estimated errors were compared with the errors as measured by the gold standard to determine agreement relative to each of the six thresholds, where agreement means that the two errors lie on the same side of the threshold. The effect of error in the gold standard itself is taken into account in the analysis of the assessment methods. The results were analyzed by means of the Kappa statistic, the agreement rate, and the area of receiver-operating-characteristic (ROC) curves. No assessment performed well at 1 mm, but all methods performed well at 2 mm and higher. For these five thresholds, two methods agreed with the standard at least 80% of the time and exhibited mean ROC areas greater than 0.84. One of these same methods exhibited Kappa statistics that indicated good agreement relative to chance (Kappa > 0.6) between the pooled observers and the standard for these same five thresholds. Further analysis demonstrates that the results depend strongly on the choice of the distribution of misregistration errors presented to the observers.

Effects of acceleration, jerk, and field inhomogeneities on vessel positions in magnetic resonance angiography

Blood flow and magnetic field inhomogeneities lead to distortions in MR angiography (MRA) images that present added risk for stereotactic neurosurgical applications. These effects are demonstrated in an MRA image of a model of cerebrovasculature. Analysis of the effects of velocity, acceleration, jerk, and field inhomogeneities on vessel position is presented; results are used to predict vessel shifts for several cerebral blood vessels. The actual encoded position for flowing spins is shown to be a moment-weighted average position. Maximum shift of 3.11 mm was reduced to 0.05 mm when velocity compensation was added. Velocity compensation applied specifically in the phase-encoding direction reduces flow-dependent shifts to the point that they can be safely ignored even if acceleration and jerk are present. Those prescribing and using MRA images for stereotactic applications must be aware of whether compensation is actually applied along the phase-encoding axis when a flow-compensated sequence is used.

Correcting scaling errors in tomographic images using a nine degree of freedom registration algorithm

PURPOSE

Clinical imaging systems, especially MR scanners, frequently have errors of a few percent in their voxel dimensions. We evaluate a nine degree of freedom registration algorithm that maximizes mutual information for determining scaling errors. We evaluate it by registering MR and CT images for each of five patients (patient scaling) and by registering MR images of a phantom to a computer model of the phantom (phantom scaling).

METHOD

Each scaling method was validated using bone-implanted markers localized in the patient images and also intraoperatively. The root mean square residual in the alignment of the fiducial markers [fiducial registration error (FRE)] was determined without scale correction, with patient scaling, and with phantom scaling.

RESULTS

Each scaling method significantly reduced the average FRE (p < 0.05) for MR to CT registration and for MR to physical space registration, indicating that voxel scaling errors were reduced. The greater reduction in scaling errors was achieved using the phantom scaling method.

CONCLUSION

We have demonstrated that a nine degree of freedom registration algorithm that maximizes mutual information can significantly reduce scaling errors in MR.

Registration of head volume images using implantable fiducial markers

In this paper, we describe an extrinsic-point-based, interactive image-guided neurosurgical system designed at Vanderbilt University, Nashville, TN, as part of a collaborative effort among the Departments of Neurological Surgery, Computer Science, and Biomedical Engineering. Multimodal image-to-image (II) and image-to-physical (IP) registration is accomplished using implantable markers. Physical space tracking is accomplished with optical triangulation. We investigate the theoretical accuracy of point-based registration using numerical simulations, the experimental accuracy of our system using data obtained with a phantom, and the clinical accuracy of our system using data acquired in a prospective clinical trial by six neurosurgeons at four medical centers from 158 patients undergoing craniotomies to resect cerebral lesions. We can determine the position of our markers with an error of approximately 0.4 mm in X-ray computed tomography (CT) and magnetic resonance (MR) images and 0.3 mm in physical space. The theoretical registration error using four such markers distributed around the head in a configuration that is clinically practical is approximately 0.5-0.6 mm. The mean CT-physical registration error for the phantom experiments is 0.5 mm and for the clinical data obtained with rigid head fixation during scanning is 0.7 mm. The mean CT-MR registration error for the clinical data obtained without rigid head fixation during scanning is 1.4 mm, which is the highest mean error that we observed. These theoretical and experimental findings indicate that this system is an accurate navigational aid that can provide real-time feedback to the surgeon about anatomical structures encountered in the surgical field.

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.

Registration of 3-D images using weighted geometrical features

The authors present a weighted geometrical feature (WGF) registration algorithm. Its efficacy is demonstrated by combining points and a surface. The technique is an extension of Besl and McKay's (1992) iterative closest point (ICP) algorithm. The authors use the WGF algorithm to register X-ray computed tomography (CT) and T2-weighted magnetic resonance (MR) volume head images acquired from eleven patients that underwent craniotomies in a neurosurgical clinical trial. Each patient had five external markers attached to transcutaneous posts screwed into the outer table of the skull. The authors define registration error as the distance between positions of corresponding markers that are not used for registration. The CT and MR images are registered using fiducial paints (marker positions) only, a surface only, and various weighted combinations of points and a surface. The CT surface is derived from contours corresponding to the inner surface of the skull. The MR surface is derived from contours corresponding to the cerebrospinal fluid (CSF)-dura interface. Registration using points and a surface is found to be significantly more accurate then registration using only points or a surface.