Registration of MR and CT images for skull base surgery using point-like anatomical features

Meta-Learning Initializations for Interactive Medical Image Registration

We present a meta-learning framework for interactive medical image registration. Our proposed framework comprises three components: a learning-based medical image registration algorithm, a form of user interaction that refines registration at inference, and a meta-learning protocol that learns a rapidly adaptable network initialization. This paper describes a specific algorithm that implements the registration, interaction and meta-learning protocol for our exemplar clinical application: registration of magnetic resonance (MR) imaging to interactively acquired, sparsely-sampled transrectal ultrasound (TRUS) images. Our approach obtains comparable registration error (4.26 mm) to the best-performing non-interactive learning-based 3D-to-3D method (3.97 mm) while requiring only a fraction of the data, and occurring in real-time during acquisition. Applying sparsely sampled data to non-interactive methods yields higher registration errors (6.26 mm), demonstrating the effectiveness of interactive MR-TRUS registration, which may be applied intraoperatively given the real-time nature of the adaptation process.

Dual-Mode Tumor Imaging Using Probes That Are Responsive to Hypoxia-Induced Pathological Conditions

Hypoxia in solid tumors is associated with poor prognosis, increased aggressiveness, and strong resistance to therapeutics, making accurate monitoring of hypoxia important. Several imaging modalities have been used to study hypoxia, but each modality has inherent limitations. The use of a second modality can compensate for the limitations and validate the results of any single imaging modality. In this review, we describe dual-mode imaging systems for the detection of hypoxia that have been reported since the start of the 21st century. First, we provide a brief overview of the hallmarks of hypoxia used for imaging and the imaging modalities used to detect hypoxia, including optical imaging, ultrasound imaging, photoacoustic imaging, single-photon emission tomography, X-ray computed tomography, positron emission tomography, Cerenkov radiation energy transfer imaging, magnetic resonance imaging, electron paramagnetic resonance imaging, magnetic particle imaging, and surface-enhanced Raman spectroscopy, and mass spectrometric imaging. These overviews are followed by examples of hypoxia-relevant imaging using a mixture of probes for complementary single-mode imaging techniques. Then, we describe dual-mode molecular switches that are responsive in multiple imaging modalities to at least one hypoxia-induced pathological change. Finally, we offer future perspectives toward dual-mode imaging of hypoxia and hypoxia-induced pathophysiological changes in tumor microenvironments.

Application of Surface Landmarks Combined with Image-Guided Sinus Location in the Retrosigmoid Approach and their Clinic-Image Relationship Analysis

Objectives: During craniotomy for cerebellopontine angle (CPA) lesions, the exact exposure of the margin of the venous sinuses complex remains an essential but risky part of the procedure. Here, we revealed the exact position of the asterion and sinus complex by combining preoperative image information and intraoperative cranial landmarks, and analyzed their clinic-image relationship. Methods: Ninety-four patients who underwent removal of vestibular schwannoma (VS) through retrosigmoid craniotomies were enrolled in the series. To determine the exact location of the sigmoid sinus and the transverse sinus and sigmoid sinus junction (TSSJ), we used preoperative images, such as computed tomography (CT) and/or magnetic resonance imaging (MRI) combined with intraoperative anatomical landmarks. The distance between the asterion and the sigmoid sinus was measured using MRI T1 sequences with gadolinium and/or the CT bone window. Results: In 94 cases of retrosigmoid craniotomies, the asterion lay an average of 12.71 mm on the posterior to the body surface projection to the TSSJ. Intraoperative cranial surface landmarks were used in combination with preoperative image information to identify the distance from the asterion to the sigmoid sinus at the transverse sinus level, allowing for an appropriate initial burr-hole (the margin of the TSSJ). Conclusions: By combining intraoperative anatomical landmarks and preoperative image information, the margin of the TSSJ, in particular, the inferior margin of the transverse sinus, can be well and thoroughly identified in the retrosigmoid approach.

Evaluation of CT-MR image registration methodologies for 3D preoperative planning of forearm surgeries

Computerized surgical planning for forearm procedures that considers both soft and bony tissue, requires alignment of preoperatively acquired computed tomography (CT) and magnetic resonance (MR) images by image registration. Normalized mutual information (NMI) registration techniques have been researched to improve efficiency and to eliminate the user dependency associated with manual alignment. While successfully applied in various medical fields, the application of NMI registration to images of the forearm, for which the relative pose of the radius and ulna likely differs between CT and MR acquisitions, is yet to be described. To enable the alignment of CT and MR forearm data, we propose an NMI-based registration pipeline, which allows manual steering of the registration algorithm to the desired image subregion and is, thus, applicable to the forearm. Successive automated registration is proposed to enable planning incorporating multiple target anatomical structures such as the radius and ulna. With respect to gold-standard manual registration, the proposed registration methodology achieved mean accuracies of 0.08 ± 0.09 mm (0.01-0.41 mm range) in comparison with 0.28 ± 0.23 mm (0.03-0.99 mm range) associated with a landmark-based registration when tested on 40 patient data sets. Application of the proposed registration pipeline required less than 10 minutes on average compared with 20 minutes required by the landmark-based registration. The clinical feasibility and relevance of the method were tested on two different clinical applications, a forearm tumor resection and radioulnar joint instability analysis, obtaining accurate and robust CT-MR image alignment for both cases.

Image-Guided Surgery and Intraoperative Imaging in Rhinology: Clinical Update and Current State of the Art

INTRODUCTION

Image-guided surgery (IGS) has gained widespread acceptance in otorhinolaryngology for its applications in sinus and skull base surgery. Although the core concepts of IGS have not changed, advances in image guidance technology, including the incorporation of intraoperative imaging, have the potential to enhance surgical education, allow for more rigorous preoperative planning, and aid in more complete surgery with improved outcomes.

OBJECTIVES

Provide a clinical update regarding the use of image guidance and intraoperative imaging in the field of rhinology and endoscopic skull base surgery with a focus on current state of the art technologies.

METHODS

English-language studies published in PubMed, Cochrane, and Embase were searched for articles relating to image-guided sinus surgery, skull base surgery, and intraoperative imaging. Relevant studies were reviewed and critical appraisals were included in this clinical update, highlighting current state of the art advances.

CONCLUSIONS

As image guidance and intraoperative imaging systems have advanced, their applications in sinus and skull base surgery have expanded. Both technologies offer invaluable real-time feedback on the status and progress of surgery, and thus may help to improve the completeness of surgery and overall outcomes. Recent advances such as augmented and virtual reality offer a window into the future of IGS. Future advancements should aim to enhance the surgeon's operative experience by improving user satisfaction and ultimately lead to better surgical results.

Intermodality feature fusion combining unenhanced computed tomography and ferumoxytol-enhanced magnetic resonance angiography for patient-specific vascular mapping in renal impairment

OBJECTIVE

The purpose of this study was to establish the feasibility of fusing complementary, high-contrast features from unenhanced computed tomography (CT) and ferumoxytol-enhanced magnetic resonance angiography (FE-MRA) for preprocedural vascular mapping in patients with renal impairment.

METHODS

In this Institutional Review Board-approved and Health Insurance Portability and Accountability Act-compliant study, 15 consecutive patients underwent both FE-MRA and unenhanced CT scanning, and the complementary high-contrast features from both modalities were fused to form an integrated, multifeature image. Source images from CT and MRA were segmented and registered. To validate the accuracy, precision, and concordance of fused images to source images, unambiguous landmarks, such as wires from implantable medical devices or indwelling catheters, were marked on three-dimensional (3D) models of the respective modalities, followed by rigid co-registration, interactive fusion, and fine adjustment. We then compared the positional offsets using pacing wires or catheters in the source FE-MRA (defined as points of interest [POIs]) and fused images (n = 5 patients, n = 247 points). Points within 3D image space were referenced to the respective modalities: x (right-left), y (anterior-posterior), and z (cranial-caudal). The respective 3D orthogonal reference axes from both image sets were aligned, such that with perfect registration, a given point would have the same (x, y, z) component values in both sets. The 3D offsets (Δx mm, Δy mm, Δz mm) for each of the corresponding POIs represent nonconcordance between the source FE-MRA and fused images. The offsets were compared using concordance correlation coefficients. Interobserver agreement was assessed using intraclass correlation coefficients and Bland-Altman analyses.

RESULTS

Thirteen patients (aged 76 ± 12 years; seven female) with aortic valve stenosis and chronic kidney disease and two patients with thoracoabdominal vascular aneurysms and chronic kidney disease underwent FE-MRA for preprocedural vascular assessment, and unenhanced CT examinations were available in all patients. No ferumoxytol-related adverse events occurred. There were 247 matched POIs evaluated on the source FE-MRA and fused images. In patients with implantable medical devices, the mean offsets in spatial position were 0.31 ± 0.51 mm (ρ = 0.99; Cb = 1; 95% confidence interval [CI], 0.99-0.99) for Δx, 0.27 ± 0.69 mm (ρ = 0.99; Cb = 0.99; 95% CI, 0.99-0.99) for Δy, and 0.20 ± 0.59 mm (ρ = 1; Cb = 1; 95% CI, 0.99-1.00) for Δz. Interobserver agreement was excellent (intraclass correlation coefficient, >0.99). The mean difference in offset between readers was 1.5 mm.

CONCLUSIONS

Accurate 3D feature fusion is feasible, combining luminal information from FE-MRA with vessel wall information on unenhanced CT. This framework holds promise for combining the complementary strengths of magnetic resonance imaging and CT to generate information-rich, multifeature composite vascular images while avoiding the respective risks and limitations of both modalities.

Accurate Geo-Referencing of Trees with No or Inaccurate Terrestrial Location Devices

Accurate and precise location of trees from data acquired under-the-canopy is challenging and time-consuming. However, current forestry practices would benefit tremendously from the knowledge of tree coordinates, particularly when the information used to position them is acquired with inexpensive sensors. Therefore, the objective of our study is to geo-reference trees using point clouds created from the images acquired below canopy. We developed a procedure that uses the coordinates of the trees seen from above canopy to position the same trees seen below canopy. To geo-reference the trees from above canopy we captured images with an unmanned aerial vehicle. We reconstructed the trunk with photogrammetric point clouds built with a structure–from–motion procedure from images recorded in a circular pattern at multiple locations throughout the stand. We matched the trees segmented from below canopy with the trees extracted from above canopy using a non-rigid point-matching algorithm. To ensure accuracy, we reduced the number of matching trees by dividing the trees segmented from above using a grid with 50 m cells. Our procedure was implemented on a 7.1 ha Douglas-fir stand from Oregon USA. The proposed procedure is relatively fast, as approximately 600 trees were mapped in approximately 1 min. The procedure is sensitive to the point density, directly impacting tree location, as differences larger than 2 m between the coordinates of the tree top and the bottom part of the stem could lead to matching errors larger than 1 m. Furthermore, the larger the number of trees to be matched the higher the accuracy is, which could allow for misalignment errors larger than 2 m between the locations of the trees segmented from above and below.

Oxygen-Enhanced and Dynamic Contrast-Enhanced Optoacoustic Tomography Provide Surrogate Biomarkers of Tumor Vascular Function, Hypoxia, and Necrosis

Measuring the functional status of tumor vasculature, including blood flow fluctuations and changes in oxygenation, is important in cancer staging and therapy monitoring. Current clinically approved imaging modalities suffer long procedure times and limited spatiotemporal resolution. Optoacoustic tomography (OT) is an emerging clinical imaging modality that may overcome these challenges. By acquiring data at multiple wavelengths, OT can interrogate hemoglobin concentration and oxygenation directly and resolve contributions from injected contrast agents. In this study, we tested whether two dynamic OT techniques, oxygen-enhanced (OE) and dynamic contrast-enhanced (DCE)-OT, could provide surrogate biomarkers of tumor vascular function, hypoxia, and necrosis. We found that vascular maturity led to changes in vascular function that affected tumor perfusion, modulating the DCE-OT signal. Perfusion in turn regulated oxygen availability, driving the OE-OT signal. In particular, we demonstrate for the first time a strong per-tumor and spatial correlation between imaging biomarkers derived from these in vivo techniques and tumor hypoxia quantified ex vivo Our findings indicate that OT may offer a significant advantage for localized imaging of tumor response to vascular-targeted therapies when compared with existing clinical DCE methods.Significance: Imaging biomarkers derived from optoacoustic tomography can be used as surrogate measures of tumor perfusion and hypoxia, potentially yielding rapid, multiparametric, and noninvasive cancer staging and therapeutic response monitoring in the clinic.Graphical Abstract: http://cancerres.aacrjournals.org/content/canres/78/20/5980/F1.large.jpg Cancer Res; 78(20); 5980-91. ©2018 AACR.

Discrete Modules and Mesoscale Functional Circuits for Thermal Nociception within Primate S1 Cortex

This study addresses one long-standing question of whether functional separations are preserved for somatosensory modalities of touch, heat, and cold nociception within primate primary somatosensory (S1) cortex. This information is critical for understanding how the nature of pain is represented in the primate brain. Using a combination of submillimeter-resolution fMRI and microelectrode local field potential (LFP) and spike recordings, we identified spatially segregated cortical zones for processing touch and nociceptive heat and cold stimuli in somatotopically appropriate areas 3a, 3b, 1, and 2 of S1 in male monkeys. The distances between zones were comparable (∼3.4 mm) across stimulus modalities (heat, cold, and tactile), indicating the existence of uniform, modality-specific modules. Stimulus-evoked LFP maps validated the fMRI maps in areas 3b and 1. Isolation of heat and cold nociceptive neurons from the fMRI zones confirmed the validity of using fMRI to probe nociceptive regions and circuits. Resting-state fMRI analysis revealed distinct intrinsic functional circuits among functionally related zones. We discovered distinct modular structures and networks for thermal nociception within S1 cortex, a finding that has significant implications for studying chronic pain syndromes and guiding the selection of neuromodulation targets for chronic pain management.SIGNIFICANCE STATEMENT Primate S1 subregions contain discrete heat and cold nociceptive modules. Modules with the same properties exhibit strong functional connection. Nociceptive fMRI response coincides with LFP and spike activities of nociceptive neurons. Functional separation of heat and cold pain is retained within primate S1 cortex.

Optimal Alignment of Structures for Finite and Periodic Systems

Finding the optimal alignment between two structures is important for identifying the minimum root-mean-square distance (RMSD) between them and as a starting point for calculating pathways. Most current algorithms for aligning structures are stochastic, scale exponentially with the size of structure, and the performance can be unreliable. We present two complementary methods for aligning structures corresponding to isolated clusters of atoms and to condensed matter described by a periodic cubic supercell. The first method (Go-PERMDIST), a branch and bound algorithm, locates the global minimum RMSD deterministically in polynomial time. The run time increases for larger RMSDs. The second method (FASTOVERLAP) is a heuristic algorithm that aligns structures by finding the global maximum kernel correlation between them using fast Fourier transforms (FFTs) and fast SO(3) transforms (SOFTs). For periodic systems, FASTOVERLAP scales with the square of the number of identical atoms in the system, reliably finds the best alignment between structures that are not too distant, and shows significantly better performance than existing algorithms. The expected run time for Go-PERMDIST is longer than FASTOVERLAP for periodic systems. For finite clusters, the FASTOVERLAP algorithm is competitive with existing algorithms. The expected run time for Go-PERMDIST to find the global RMSD between two structures deterministically is generally longer than for existing stochastic algorithms. However, with an earlier exit condition, Go-PERMDIST exhibits similar or better performance.

CT-MR Image Fusion for the Management of Skull Base Lesions

OBJECTIVE: Image fusion software creates composite images from computed tomography (CT) and magnetic resonance (MR) images.

STUDY DESIGN AND SETTING: CT-MR fusion studies between August 2004 to July 2005 were reviewed. CT scan images (1-mm axial) were obtained on a multi-detector CT scanner. MR images (1-mm axial) were obtained with a T1-weighted, volume acquisition technique. CT-MR fusion images were created on the Instatrak 3500 Plus or CBYON Suite.

RESULTS: A total of 25 patients had 27 CT-MR image fusion studies. CT-MR fusion accuracy was 2 mm or better. During preoperative planning and intraoperative navigation, CT-MR fusion facilitated the delineation of the anatomic relationships between the lesion and the skull base.

CONCLUSIONS: CT-MR image fusion yields composite images that combine features of each component imaging modality. Image fusion, when combined with surgical navigation, further enhances potential skull base applications for minimally invasive endoscopic surgery.

Evaluation of 3-dimensional superimposition techniques on various skeletal structures of the head using surface models

Objectives

To test the applicability, accuracy, precision, and reproducibility of various 3D superimposition techniques for radiographic data, transformed to triangulated surface data.

Methods

Five superimposition techniques (3P: three-point registration; AC: anterior cranial base; AC + F: anterior cranial base + foramen magnum; BZ: both zygomatic arches; 1Z: one zygomatic arch) were tested using eight pairs of pre-existing CT data (pre- and post-treatment). These were obtained from non-growing orthodontic patients treated with rapid maxillary expansion. All datasets were superimposed by three operators independently, who repeated the whole procedure one month later. Accuracy was assessed by the distance (D) between superimposed datasets on three form-stable anatomical areas, located on the anterior cranial base and the foramen magnum. Precision and reproducibility were assessed using the distances between models at four specific landmarks. Non parametric multivariate models and Bland-Altman difference plots were used for analyses.

Results

There was no difference among operators or between time points on the accuracy of each superimposition technique (p>0.05). The AC + F technique was the most accurate (D<0.17 mm), as expected, followed by AC and BZ superimpositions that presented similar level of accuracy (D<0.5 mm). 3P and 1Z were the least accurate superimpositions (0.79<D<1.76 mm, p<0.005). Although there was no difference among operators or between time points on the precision of each superimposition technique (p>0.05), the detected structural changes differed significantly between different techniques (p<0.05). Bland-Altman difference plots showed that BZ superimposition was comparable to AC, though it presented slightly higher random error.

Conclusions

Superimposition of 3D datasets using surface models created from voxel data can provide accurate, precise, and reproducible results, offering also high efficiency and increased post-processing capabilities. In the present study population, the BZ superimposition was comparable to AC, with the added advantage of being applicable to scans with a smaller field of view.

Parallel functional reorganizations of somatosensory areas 3b and 1, and S2 following spinal cord injury in squirrel monkeys

Multiple somatosensory cortices of adult primates reorganize following spinal cord injury, but little is known about the temporal dynamics and inter-areal differences of the reorganization. Using longitudinal high-resolution fMRI in combination with microelectrode recordings and tracer histology, we previously illustrated a two-phase dynamic spatial reorganization of digit representations in area 3b within weeks after a unilateral lesion of the dorsal column in squirrel monkeys (Chen et al., 2012). Here we report that higher-order area 1 and secondary somatosensory cortex (S2) underwent similar spatial reorganizations, which were characterized by shifted and expanded digit activations at week 4 after lesion, which then shifted back and contracted by week 8. In addition, the responsiveness of areas 3b and 1, and S2, as measured by the magnitude of the BOLD signal change to tactile stimuli, was reduced markedly at 4 weeks and then recovered to ~50% of the prelesion level at 8 weeks, a time when behavioral recovery was complete, as assessed by successful food retrieval rates. Across animals, the extents of spatial reorganizations and changes in cortical responsiveness and activation sizes in all three areas were correlated with the degree of afferent disruption. In summary, our data show that more severe afferent disruption was associated with greater cortical plasticity and behavioral impairment. Reorganization that occurred in area 3b, area 1, and S2 were similar across most measures.

Radio-guided sentinel lymph node identification by lymphoscintigraphy fused with an anatomical vector profile: clinical applications

OBJECTIVE

To develop a method to fuse lymphoscintigraphic images with an adaptable anatomical vector profile and to evaluate its role in the clinical practice.

METHODS

We used Adobe Illustrator CS6 to create different vector profiles, we fused those profiles, using Adobe Photoshop CS6, with lymphoscintigraphic images of the patient. We processed 197 lymphoscintigraphies performed in patients with cutaneous melanomas, breast cancer or delayed lymph drainage.

RESULTS

Our models can be adapted to every patient attitude or position and contain different levels of anatomical details ranging from external body profiles to the internal anatomical structures like bones, muscles, vessels, and lymph nodes. If needed, more new anatomical details can be added and embedded in the profile without redrawing them, saving a lot of time. Details can also be easily hidden, allowing the physician to view only relevant information and structures. Fusion times are about 85 s. The diagnostic confidence of the observers increased significantly. The validation process showed a slight shift (mean 4.9 mm).

CONCLUSIONS

We have created a new, practical, inexpensive digital technique based on commercial software for fusing lymphoscintigraphic images with built-in anatomical reference profiles. It is easily reproducible and does not alter the original scintigraphic image. Our method allows a more meaningful interpretation of lymphoscintigraphies, an easier recognition of the anatomical site and better lymph node dissection planning.

Dynamic reorganization of digit representations in somatosensory cortex of nonhuman primates after spinal cord injury

Somatosensory cortices of adult primates reactivate over time after sensory loss. The time course and the neural mechanisms underlying the cortical reactivation are not well understood. Here we report that longitudinal high-resolution functional magnetic resonance imaging (fMRI) studies on anesthetized squirrel monkeys revealed dynamic reorganizations of digit activations in area 3b, within 2 months after severely disrupting afferent inputs by dorsal column section. We found that digit regions in which inputs were severely disrupted exhibited fMRI tactile responses. Reorganization was characterized by an early moving away phase and a late returning phase, as indicated by spatial shifts of individual digit activation centers in relation to the pre-lesion activation sites. Subsequent optical imaging studies confirmed fMRI activations, and dense microelectrode penetrations identified weak neuronal activity at the reactivated sites. Activation zones detected by fMRI and optical imaging were significantly larger in input-deprived than normal input single-digit regions and were larger than regions defined by neuronal spiking activity. This study captures the dynamic reorganization of digit representations after dorsal column lesions and reveals differences between functional imaging and microelectrode recording maps. Our observations suggest that subthreshold activity plays an important role in the reactivation of deafferented cortex and could promote behavioral recovery.

Challenges in assessing regional distribution of inhaled drug in the human lungs

INTRODUCTION

Both the total amount of drug deposited in the lungs (whole lung deposition) and the amount deposited in different lung regions (regional lung deposition) are potentially important factors that determine the safety and efficacy of inhaled drugs. Radionuclide imaging is well established for quantifying the whole lung deposition of inhaled drugs, but the assessment of regional lung deposition is less straightforward, because of the complex nature of the lung anatomy.

AREAS COVERED

This review describes the challenges and problems associated with quantifying regional lung deposition by the two-dimensional (2D) radionuclide imaging method of gamma scintigraphy, and by the three-dimensional (3D) radionuclide imaging methods of single-photon-emission computed tomography (SPECT) and positron-emission tomography (PET). The advantages and disadvantages of each method for assessing regional lung deposition are discussed.

EXPERT OPINION

Owing to its 2D nature, gamma scintigraphy provides limited information about regional lung deposition. SPECT provides regional lung deposition data in three dimensions, but usually involves a (99m)Tc radiolabel. PET enables the regional lung deposition of radiolabeled drug molecules to be quantified in three dimensions, but poses the greatest logistical and technical difficulties. Despite their more challenging nature, 3D imaging methods should be considered as an alternative to gamma scintigraphy whenever the determination of regional lung deposition of pharmaceutical aerosols is a major study objective.

High-resolution functional magnetic resonance imaging mapping of noxious heat and tactile activations along the central sulcus in New World monkeys

This study mapped the fine-scale functional representation of tactile and noxious heat stimuli in cortical areas around the central sulcus of anesthetized squirrel monkeys by using high-resolution blood oxygen level-dependent (BOLD) fMRI at 9.4T. Noxious heat (47.5°C) stimulation of digits evoked multiple spatially distinct and focal BOLD activations. Consistent activations were observed in areas 3a, 3b, 1, and 2, whereas less frequent activation was present in M1. Compared with tactile activations, thermal nociceptive activations covered more area and formed multiple foci within each functional area. In general, noxious heat activations in area 3b did not colocalize with tactile responses. The spatial relationships of heat and tactile activations in areas 3a and 1/2 varied across animals. Subsequent electrophysiological mapping confirmed that the evoked heat and tactile BOLD signals were somatotopically appropriate. The magnitude and temporal profiles of the BOLD signals to noxious heat stimuli differed across cortical areas. Comparatively late-peaking but stronger signals were observed in areas 3b and 2, whereas earlier-peaking but weaker signals were observed in areas 3a, 1, and M1. In sum, this study not only confirmed the involvement of somatosensory areas of 3a, 3b, and 1, but also identified the engagements of area 2 and M1 in the processing of heat nociceptive inputs. Differential BOLD response profiles of the individual cortical areas along the central sulcus suggest that these areas play different roles in the encoding of nociceptive inputs. Thermal nociceptive and tactile inputs may be processed by different clusters of neurons in different areas. To critically bridge animal and human pain studies, human fMRI was related to primate fMRI and electrophysiology of nociceptive processing, examining the functional role of the primary somatosensory cortex in heat nociception and demonstrating that subregion areas 3a, 3b, 1, 2, and M1 are responsive to noxious heat stimuli.

Groupwise registration based on hierarchical image clustering and atlas synthesis

Groupwise registration has recently been proposed for simultaneous and consistent registration of all images in a group. Since many deformation parameters need to be optimized for each image under registration, the number of images that can be effectively handled by conventional groupwise registration methods is limited. Moreover, the robustness of registration is at stake due to significant intersubject variability. To overcome these problems, we present a groupwise registration framework, which is based on a hierarchical image clustering and atlas synthesis strategy. The basic idea is to decompose a large-scale groupwise registration problem into a series of small-scale problems, each of which is relatively easy to solve using a general computer. In particular, we employ a method called affinity propagation, which is designed for fast and robust clustering, to hierarchically cluster images into a pyramid of classes. Intraclass registration is then performed to register all images within individual classes, resulting in a representative center image for each class. These center images of different classes are further registered, from the bottom to the top in the pyramid. Once the registration reaches the summit of the pyramid, a single center image, or an atlas, is synthesized. Utilizing this strategy, we can efficiently and effectively register a large image group, construct their atlas, and, at the same time, establish shape correspondences between each image and the atlas. We have evaluated our framework using real and simulated data, and the results indicate that our framework achieves better robustness and registration accuracy compared to conventional methods.

Image-guided surgical planning using anatomical landmarks in the retrosigmoid approach

OBJECTIVE

The suboccipital lateral or retrosigmoid approach is the main neurosurgical approach to the cerebellopontine angle (CPA). It is mainly used in the treatment of CPA tumors and vascular decompression of cranial nerves. A prospective study using navigation registered with anatomical landmarks in order to identify the transverse and sigmoid sinuses junction (TSSJ) was carried out in a series of 30 retrosigmoid craniotomies. The goal of this study was to determine the accuracy of this navigation technique and to establish the relationship between the location of the asterion and the TSSJ.

METHODS

From March through November 2008, 30 patients underwent a retrosigmoid craniotomy for removal of CPA tumors or for surgical treatment of neurovascular syndromes. Magnetic resonance imaging (MRI) T1 sequences with gadolinium (FSPGR with FatSst, 1.5 T GE Signa) and frameless navigation (Vector vision, Brainlab) were used for surgical planning. Registration was performed using six anatomical landmarks. The position of the TSSJ indicated by navigation was the landmark to guide the craniotomy. The location of the asterion was compared with the position of the TSSJ. After craniotomy, the real TSSJ position was compared with the virtual position, as demonstrated by navigation.

RESULTS

There were 19 cases of vestibular schwannomas, 5 petroclival meningiomas, 3 trigeminal neuralgias, 1 angioblastoma, 1 epidermoid cyst and 1 hemifacial spasm. In all cases, navigation enabled the location of the TSSJ and the emissary vein, with an accuracy flaw below 2 mm. The asterion was located directly over the TSSJ in only seven cases. One patient had a laceration of the sigmoid sinus during the craniotomy.

CONCLUSIONS

Navigation using anatomical landmarks for registration is a reliable method in the localization of the TSSJ for retrosigmoid craniotomies and thereby avoiding unnecessary sinus exposure. In addition, the method proved to be fast and accurate. The asterion was found to be a less accurate landmark for the localization of the TSSJ using navigation.

Intra- and inter-subject variability of high field fMRI digit maps in somatosensory area 3b of new world monkeys

This study evaluates the intra- and inter-subject variability of digit maps in area 3b of anesthetized squirrel monkeys. Maps were collected using high field blood oxygenation level dependent (BOLD) functional magnetic resonance imaging (fMRI). BOLD responses to individual digit stimulations were mapped and their response properties (location, area of activation, % signal change, time to peak response) were compared within and across imaging sessions separated by up to 20 months. During single digit stimulation using a block design, the spatiotemporal response of the BOLD signal for individual runs within and across sessions and animals was well conserved, with a time to peak BOLD response of 20+/-4 s. The variability in the center of BOLD activation in area 3b was 0.41+/-0.24 mm (mean+/-SD) across individual 5-7 min runs within a scanning session and 0.55+/-0.15 mm across sessions. The average signal change across all animals, runs and sessions was 0.62+/-0.38%, and varied 32% within and 40% across sessions. In a comparison of the stability and reproducibility of the area of single digit activation obtained using three approaches, use of a fixed statistical threshold (P<10(-5)) yielded an average area of 4.8+/-3.5 mm(2) (mean+/-SD), adaptive statistical thresholding 1.32+/-1.259 mm(2) (mean+/-SD), and combined fixed statistical and adaptive BOLD signal amplitude 4.4+/-2.5 mm(2) (mean+/-SD) across image runs and sessions. The somatotopic organization was stable within animals across sessions, while across animals, there was some variation in overall activation pattern and inter-digit distances. These results confirm that BOLD activation maps of single digits in area 3b as characterized by activation center, signal amplitudes, and temporal profile are very stable. The activation sizes determined by various criteria are the most variable measure in this preparation, but adaptive statistical thresholding appears to yield the most stable and reproducible maps. This study serves as a baseline assessment of the limits imposed on the detection of plastic changes by experimental variations of the digit BOLD fMRI activation maps in normal animals, and as an indicator of the likely performance limits in human studies.

A comparative study of three CT and MRI registration algorithms in nasopharyngeal carcinoma

Objective: To evaluate the image registration accuracy and efficiency of CT and MRI fusion using three algorithms in nasopharyngeal carcinoma (NPC). Methods and materials: Twelve sets of CT and MRI scans of 12 NPC patients were fused using three image registration algorithms, respectively: Mark‐and‐link, Interactive, and Normalized Mutual Information (NMI). Registration accuracy was evaluated by performing statistical analysis of the coordinate differences between CT and MR anatomical landmarks along the x‐, y‐ and z‐axes. The time required to complete the registration process using three algorithms was also recorded. One‐way ANOVA was used to analyze the difference of the three registration methods. Results: The mean time required for CT/MRI registration using the three different registration algorithms, mark‐and‐link, interactive, and NMI, was 6.25 min, 5.25 min, and 5.15 min, respectively. The mark‐and‐link method was more time consuming (F=8.74,p=0.001); however no statistical difference was found between the time required using interactive and NMI methods (p=0.77). Mean registration errors of the three methods along the x‐axis were 0.66 mm, 0.70 mm, and 0.68 mm, respectively (F=0.09,p=0.91). Along the y‐axis, the mean registration errors were 1.03 mm, 1.04 mm, and 1.03 mm, respectively (F=0.02,p=0.98). Along the z‐axis, they were 0.58 mm, 0.64 mm, and 0.56 mm, respectively (F=0.21,p=0.81).

Conclusions: All three registration algorithms, mark‐and‐link, interactive, and NMI, can provide accurate CT/MRI registration. However the mark‐and‐link method was most time consuming.

PACS number: 87.57.nj

Evaluation of four volume-based image registration algorithms

We evaluated 4 volume-based automatic image registration algorithms from 2 commercially available treatment planning systems (Philips Syntegra and BrainScan). The algorithms based on cross correlation (CC), local correlation (LC), normalized mutual information (NMI), and BrainScan mutual information (BSMI) were evaluated with: (1) the synthetic computed tomography (CT) images, (2) the CT and magnetic resonance (MR) phantom images, and (3) the CT and MR head image pairs from 12 patients with brain tumors. For the synthetic images, the registration results were compared with known transformation parameters, and all algorithms achieved accuracy of submillimeter in translation and subdegree in rotation. For the phantom images, the registration results were compared with those provided by frame and marker-based manual registration. For the patient images, the results were compared with anatomical landmark-based manual registration to qualitatively determine how the results were close to a clinically acceptable registration. NMI and LC outperformed CC and BSMI, with the sense of being closer to a clinically acceptable result. As for the robustness, NMI and BSMI outperformed CC and LC. A guideline of image registration in our institution was given, and final visual assessment is necessary to guarantee reasonable results.

Automatic localization of anatomical point landmarks for brain image processing algorithms

Many brain image processing algorithms require one or more well-chosen seed points because they need to be initialized close to an optimal solution. Anatomical point landmarks are useful for constructing initial conditions for these algorithms because they tend to be highly-visible and predictably-located points in brain image scans. We introduce an empirical training procedure that locates user-selected anatomical point landmarks within well-defined precisions using image data with different resolutions and MRI weightings. Our approach makes no assumptions on the structural or intensity characteristics of the images and produces results that have no tunable run-time parameters. We demonstrate the procedure using a Java GUI application (LONI ICE) to determine the MRI weighting of brain scans and to locate features in T1-weighted and T2-weighted scans.

Field-warp registration for biomedical high-resolution thermal infrared images

Biomedical protocols based on thermal infrared images often require effective image registration. Algorithms specifically designed for registration of thermal images are hardly available and use of algorithms designed for other imaging techniques may result poorly reliable. In fact, registration algorithms developed for other biomedical images often require rigid-body assumption or limited range for intensity values. Such assumption may not be applicable for human body thermal images. Therefore, we present here an adaptation of a field-warp based method as a possible solution for thermal image registration. The method was applied for registering images taken from an experimental protocol aimed at comparing total body skin temperature distribution in natural or altered posture. The method appears to be effective into providing a reliable tool for objective intra and inter individual skin temperature distribution comparisons.

High-resolution maps of real and illusory tactile activation in primary somatosensory cortex in individual monkeys with functional magnetic resonance imaging and optical imaging

Although blood oxygenation level-dependent (BOLD) functional magnetic resonance imaging (fMRI) has been widely used to explore human brain function, questions remain regarding the ultimate spatial resolution of positive BOLD fMRI, and indeed the extent to which functional maps revealed by positive BOLD correlate spatially with maps obtained with other high-spatial-resolution mapping techniques commonly used in animals, such as optical imaging of intrinsic signal (OIS) and single-unit electrophysiology. Here, we demonstrate that the positive BOLD signal at 9.4T can reveal the fine topography of individual fingerpads in single-condition activation maps in nonhuman primates. These digit maps are similar to maps obtained from the same animal using intrinsic optical imaging. Furthermore, BOLD fMRI reliably resolved submillimeter spatial shifts in activation in area 3b previously identified with OIS (Chen et al., 2003) as neural correlates of the "funneling illusion." These data demonstrate that at high field, high-spatial-resolution topographic maps can be achieved using the positive BOLD signal, weakening previous notions regarding the spatial specificity of the positive BOLD signal.

Comparison of SPECT aerosol deposition data with a human respiratory tract model

Three-dimensional (3D) radionuclide imaging provides detailed information on the distribution of inhaled aerosol material within the body. Analysis of the data can provide estimates of the deposition per airway generation. In this study, two different nebulizers have been used to deliver radiolabeled aerosols of different particle size to 12 human subjects. Medical imaging has been used to assess the deposition in the body. The deposition pattern has also been estimated using the International Commission on Radiological Protection (ICRP) empirical model and compared to values obtained by experiment. The results showed generally good agreement between model and experiment for both aerosols for the deposition in the extrathoracic and conducting airways. However, there were significant differences in the fate of the remainder of the aerosol between the amount deposited in the alveolar region and that exhaled. The inter-subject variability of deposition predicted by the model was significantly less than that measured, for all regions of the body. The model predicted quite well the differences in deposition distribution pattern between the two aerosols. In conclusion, this study has shown that the ICPR model of inhaled aerosol deposition shows areas of good agreement with results from experiment. However, there are also areas of disagreement, which may be explained by hygroscopic particle growth and individual variation in airway anatomy.

Single-Photon Emission Computed Tomography/Computed Tomography in Brain Tumors

Anatomic imaging procedures (computed tomography [CT] and magnetic resonance imaging [MRI]) have become essential tools for brain tumor assessment. Functional images (positron emission tomography [PET] and single-photon emission computed tomography [SPECT]) can provide additional information useful during the diagnostic workup to determine the degree of malignancy and as a substitute or guide for biopsy. After surgery and/or radiotherapy, nuclear medicine examinations are essential to assess persistence of tumor, to differentiate recurrence from radiation necrosis and gliosis, and to monitor the disease. The combination of functional images with anatomic ones is of the utmost importance for a full evaluation of these patients, which can be obtained by means of imaging fusion. Despite the fast-growing diffusion of PET, in most cases of brain tumors, SPECT studies are adequate and provide results that parallel those obtained with PET. The main limitation of SPECT imaging with brain tumor-seeking radiopharmaceuticals is the lack of precise anatomic details; this drawback is overcome by the fusion with morphological studies that provide an anatomic map to scintigraphic data. In the past, software-based fusion of independently performed SPECT and CT or MRI demonstrated usefulness for brain tumor assessment, but this process is often time consuming and not practical for everyday nuclear medicine studies. The recent development of dual-modality integrated imaging systems, which allow the acquisition of SPECT and CT images in the same scanning session, and their co-registration by means of the hardware, has facilitated this process. In SPECT studies of brain tumors with various radiopharmaceuticals, fused images are helpful in providing the precise localization of neoplastic lesions, and in excluding the disease in sites of physiologic tracer uptake. This information is useful for optimizing diagnosis, therapy monitoring, and radiotherapy treatment planning, with a positive impact on patient management.

Image-guidance for surgical procedures

Contemporary imaging modalities can now provide the surgeon with high quality three- and four-dimensional images depicting not only normal anatomy and pathology, but also vascularity and function. A key component of image-guided surgery (IGS) is the ability to register multi-modal pre-operative images to each other and to the patient. The other important component of IGS is the ability to track instruments in real time during the procedure and to display them as part of a realistic model of the operative volume. Stereoscopic, virtual- and augmented-reality techniques have been implemented to enhance the visualization and guidance process. For the most part, IGS relies on the assumption that the pre-operatively acquired images used to guide the surgery accurately represent the morphology of the tissue during the procedure. This assumption may not necessarily be valid, and so intra-operative real-time imaging using interventional MRI, ultrasound, video and electrophysiological recordings are often employed to ameliorate this situation. Although IGS is now in extensive routine clinical use in neurosurgery and is gaining ground in other surgical disciplines, there remain many drawbacks that must be overcome before it can be employed in more general minimally-invasive procedures. This review overviews the roots of IGS in neurosurgery, provides examples of its use outside the brain, discusses the infrastructure required for successful implementation of IGS approaches and outlines the challenges that must be overcome for IGS to advance further.

Localization of anatomical point landmarks in 3D medical images by fitting 3D parametric intensity models

We introduce a new approach for the localization of 3D anatomical point landmarks. This approach is based on 3D parametric intensity models which are directly fitted to 3D images. To efficiently model tip-like, saddle-like, and sphere-like anatomical structures we introduce analytic intensity models based on the Gaussian error function in conjunction with 3D rigid transformations as well as deformations. To select a suitable size of the region-of-interest (ROI) where model fitting is performed, we also propose a new scheme for automatic selection of an optimal 3D ROI size based on the dominant gradient direction. In addition, to achieve a higher level of automation we present an algorithm for automatic initialization of the model parameters. Our approach has been successfully applied to accurately localize anatomical landmarks in 3D synthetic data as well as 3D MR and 3D CT image data. We have also compared the experimental results with the results of a previously proposed 3D differential approach. It turns out that the new approach significantly improves the localization accuracy.

Three-dimensional reconstruction of registered and fused Chinese Visible Human and patient MRI images

Radiological images are commonly used as important tools in medical diagnoses and treatment. Different modalities of medical images provide uniquely different content. Hence, it is natural and desirable to combine different image modalities to obtain additional new information to enhance clinical assessment. However, given the current technology, radiological images are not always sufficiently informative to permit diagnosis and treatment. In order to address this problem, we fused selected portions of the Chinese Visible Human (CVH) dataset with MRI images from a patient. Specifically, we segmented the caudate nucleus, the lentiform nucleus, and the thalamus in the CVH dataset and then registered and fused this dataset with corresponding MRI images using both rigid and nonrigid registration techniques. After rigid and nonrigid registration, the CVH and MRI images largely coincided with each other. The shape, relationship, and position of focal areas and neural structures were clearly displayed. Using volume and surface rendering, these images were three-dimensionally reconstructed to display the neural structures of interest within the brain. These structures can be rotated at will and observed from different angles. Our research indicates that the fusion of CVH and patients' MRI images can enhance the amount of neural information available to physicians and lay a foundation for the clinical use of the CVH dataset.

A three-dimensional registration method for automated fusion of micro PET-CT-SPECT whole-body images

Micro positron emission tomography (PET) and micro single-photon emission computed tomography (SPECT), used for imaging small animals, have become essential tools in developing new pharmaceuticals and can be used, among other things, to test new therapeutic approaches in animal models of human disease, as well as to image gene expression. These imaging techniques can be used noninvasively in both detection and quantification. However, functional images provide little information on the structure of tissues and organs, which makes the localization of lesions difficult. Image fusion techniques can be exploited to map the functional images to structural images, such as X-ray computed tomography (CT), to support target identification and to facilitate the interpretation of PET or SPECT studies. Furthermore, the mapping of two functional images of SPECT and PET on a structural CT image can be beneficial for those in vivo studies that require two biological processes to be monitored simultaneously. This paper proposes an automated method for registering PET, CT, and SPECT images for small animals. A calibration phantom and a holder were used to determine the relationship among three-dimensional fields of view of various modalities. The holder was arranged in fixed positions on the couches of the scanners, and the spatial transformation matrix between the modalities was held unchanged. As long as objects were scanned together with the holder, the predetermined matrix could register the acquired tomograms from different modalities, independently of the imaged objects. In this work, the PET scan was performed by Concorde's microPET R4 scanner, and the SPECT and CT data were obtained using the Gamma Medica's X-SPECT/CT system. Fusion studies on phantoms and animals have been successfully performed using this method. For microPET-CT fusion, the maximum registration errors were 0.21 mm +/- 0.14 mm, 0.26 mm +/- 0.14 mm, and 0.45 mm +/- 0.34 mm in the X (right-left), Y (upper lower), and Z (rostral-caudal) directions, respectively; for the microPET-SPECT fusion, they were 0.24 mm +/- 0.14 mm, 0.28 mm +/- 0.15 mm, and 0.54 mm +/- 0.35 mm in the X, Y, and Z directions, respectively. The results indicate that this simple method can be used in routine fusion studies.

Non-rigid image registration: theory and practice

Image registration is an important enabling technology in medical image analysis. The current emphasis is on development and validation of application-specific non-rigid techniques, but there is already a plethora of techniques and terminology in use. In this paper we discuss the current state of the art of non-rigid registration to put on-going research in context and to highlight current and future clinical applications that might benefit from this technology. The philosophy and motivation underlying non-rigid registration is discussed and a guide to common terminology is presented. The core components of registration systems are described and outstanding issues of validity and validation are confronted.