The main subject of this work is the rigid registration of surfaces dedicated to VirtualScope, a per-operative guiding system designed for oral implants surgery. It is based on a purely statistical approach. We first show how to compute and maximize a likelihood, based on a model of the data noise, for the landmarks registration problem. This approach justifies the use of the ICP algorithm, and a new multi-scale variant named ICP/EM, which improves accuracy, speed and robustness. We introduce new noise models specifically designed for the registration of sampled and noised surfaces. We discuss about the theoretical prediction of the registration accuracy, and use it for guiding the data acquisition. We analyze in detail the experimental performances of the algorithm, and provide methods for setting optimally the parameters and ensuring the correctness of registration results. The resulting algorithm is perfectly suited to the VirtualScope application. The second part of this work deals with the more general problem of the statistical modelisation of sampled and noised curves and surfaces. Based on previous works on Saliency and Tensor Voting notions, it defines a vote field that represent the probability of a curve or surface element, knowing another element. We provide basic yet easy to implement examples of such a field, which can handle the surface shape, the sampling strategy and the measurement errors. We apply them successfully to the registration problem, and suggest to use them to derive Bayesian methods to virtually all other computer vision problems involving sampled and noised curves and surfaces. This work could lead to the design of a common statistical and multi-scale framework for these various methods. |
With new irradiation techniques, the dose can be better matched to the contours of the tumour. The corollary is that greater precision is required. Recent intercomparison studies of treatment plans have emphasized the need to harmonise contouring practices. More of a consensus approach is based on using adaptive imaging modalities, expert group recommendations and automatic segmentation atlases, on harmonisation of dosimetric decisions through employing exhaustive nomograms for organs at risk, and on indexes for choosing optimal treatment plans. On another level, quality assurance and data pooling programmes have been set up, making use of DICOM-RT data transfer (image networks). The combination of several irradiation techniques (for example, intensity-modulated conformal radiation therapy plus CyberKnife((R)) boost and re-irradiation), making it possible to irradiate tumours better, requires the cumulative doses to be recorded by dose summation software. Real awareness has been achieved in recent years as regards improving the quality of treatment, pooling data and harmonising practices. |
PURPOSE: To propose an automatic atlas-based segmentation framework of the dental structures, called Dentalmaps, and to assess its accuracy and relevance to guide dental care in the context of intensity-modulated radiotherapy. METHODS AND MATERIALS: A multi-atlas-based segmentation, less sensitive to artifacts than previously published head-and-neck segmentation methods, was used. The manual segmentations of a 21-patient database were first deformed onto the query using nonlinear registrations with the training images and then fused to estimate the consensus segmentation of the query. RESULTS: The framework was evaluated with a leave-one-out protocol. The maximum doses estimated using manual contours were considered as ground truth and compared with the maximum doses estimated using automatic contours. The dose estimation error was within 2-Gy accuracy in 75\% of cases (with a median of 0.9 Gy), whereas it was within 2-Gy accuracy in 30\% of cases only with the visual estimation method without any contour, which is the routine practice procedure. CONCLUSIONS: Dose estimates using this framework were more accurate than visual estimates without dental contour. Dentalmaps represents a useful documentation and communication tool between radiation oncologists and dentists in routine practice. Prospective multicenter assessment is underway on patients extrinsic to the database. |
medInria is a free medical imaging software developed at Inria, which aims at providing clinicians with state-of-the-art algorithms dedicated to medical image processing and visualization. Efforts have been made to simplify the user interface, while keeping high-level algorithms. In this particular article, we will concentrate on its use in preoperative preparation for cardiac interventions, and how we handle the difficulties arising from the lack of standard format for data types such as meshes or fibers, the absence of a common programming interface for data processing algorithms, notably registration, and the issues of visualisation where display conventions would be beneficial. |
When performing statistics on elements of sets that possess a particular geometric structure, it is desirable to respect this structure. For instance in a Lie group, it would be judicious to have a notion of a mean which is stable by the group operations (composition and inversion). Such a property is ensured for Riemannian center of mass in Lie groups endowed with a bi-invariant Riemannian metric, like compact Lie groups (e.g. rotations). However, bi-invariant Riemannian metrics do not exist for most non compact and non-commutative Lie groups. This is the case in particular for rigid-body transformations in any dimension greater than one, which form the most simple Lie group involved in biomedical image registration. In this paper, we propose to replace the Riemannian metric by an affine connection structure on the group. We show that the canonical Cartan connections of a connected Lie group provides group geodesics which are completely consistent with the composition and inversion. With such a non-metric structure, the mean cannot be defined by minimizing the variance as in Riemannian Manifolds. However, the characterization of the mean as an exponential barycenter gives us an implicit definition of the mean using a general barycentric equation. Thanks to the properties of the canonical Cartan connection, this mean is naturally bi-invariant. We show the local existence and uniqueness of the invariant mean when the dispersion of the data is small enough. We also propose an iterative fixed point algorithm and demonstrate that the convergence to the invariant mean is at least linear. In the case of rigid-body transformations, we give a simple criterion for the global existence and uniqueness of the bi-invariant mean, which happens to be the same as for rotations. We also give closed forms for the bi-invariant mean in a number of simple but instructive cases, including 2D rigid transformations. For general linear transformations, we show that the bi-invariant mean is a generalization of the (scalar) geometric mean, since the determinant of the bi-invariant mean is the geometric mean of the determinants of the data. Finally, we extend the theory to higher order moments, in particular with the covariance which can be used to define a local bi-invariant Mahalanobis distance. |
To support the challenging task of early epithelial cancer diagnosis from in vivo endomicroscopy, we propose a content-based video retrieval method that uses an expert-annotated database. Motivated by the recent successes of non-medical content-based image retrieval, we first adjust the standard Bag-of-Visual-Words method to handle single endomicroscopic images. A local dense multi-scale description is proposed to keep the proper level of invariance, in our case to translations, in-plane rotations and affine transformations of the intensities. Since single images may have an insufficient field-of-view to make a robust diagnosis, we introduce a video-mosaicing technique that provides large field-of-view mosaic images. To remove outliers, retrieval is followed by a geometrical approach that captures a statistical description of the spatial relationships between the local features. Building on image retrieval, we then focus on efficient video retrieval. Our approach avoids the time-consuming parts of the video-mosaicing by relying on coarse registration results only to account for spatial overlap between images taken at different times. To evaluate the retrieval, we perform a simple nearest neighbors classification with leave-one-patient-out cross-validation. From the results of binary and multi-class classification, we show that our approach outperforms, with statistical significance, several state-of-the art methods. We obtain a binary classification accuracy of 94.2%, which is quite close to clinical expectations. |
Personalization is a key aspect of biophysical models in order to impact clinical practice. In this paper, we propose a personalization method of electromechanical models of the heart from cine-MR images based on the adjoint method. After estimation of electrophysiological parameters, the cardiac motion is estimated based on a proactive electromechanical model. Then cardiac contractilities on two or three regions are estimated by minimizing the discrepancy between measured and simulation motion. Evaluation of the method on three patients with infarcted or dilated myocardium is provided. |
EMPIRE10 (Evaluation of Methods for Pulmonary Image REgistration 2010) is a public platform for fair and meaningful comparison of registration algorithms which are applied to a database of intrapatient thoracic CT image pairs. Evaluation of nonrigid registration techniques is a nontrivial task. This is compounded by the fact that researchers typically test only on their own data, which varies widely. For this reason, reliable assessment and comparison of different registration algorithms has been virtually impossible in the past. In this work we present the results of the launch phase of EMPIRE10, which comprised the comprehensive evaluation and comparison of 20 individual algorithms from leading academic and industrial research groups. All algorithms are applied to the same set of 30 thoracic CT pairs. Algorithm settings and parameters are chosen by researchers expert in the configuration of their own method and the evaluation is independent, using the same criteria for all participants. All results are published on the EMPIRE10 website (http://empire10.isi.uu.nl). The challenge remains ongoing and open to new participants. Full results from 24 algorithms have been published at the time of writing. This paper details the organization of the challenge, the data and evaluation methods and the outcome of the initial launch with 20 algorithms. The gain in knowledge and future work are discussed. |
INTRODUCTION: Health related quality of life (HRQOL) is often affected in multiple sclerosis (MS). Nevertheless, to our knowledge, there is no longitudinal study in the literature about the correlation between MRI parameters and HRQOL in MS patients. METHODS: We included 28 patients with clinically definite relapsing remitting MS. All patients initiated subcutaneous interferon beta-1a therapy. To assess HRQOL, we used the SEP-59 scale, the French validated translation of MSQOL-54, and the MusiQoL scale. Conventional MRI was performed every year. Lesion load (LL) and brain atrophy were automatically measured using SepINRIA, a free software developed by INRIA in Sophia Antipolis. RESULTS: The mean EDSS score was 1.7 and disease duration was 2.5 years. Our results revealed that HRQOL was significantly correlated to T1 and T2-LL with both SEP-59 and MusiQoL scales. T1-LL was better correlated with physical dimensions and T2-LL was better correlated with mental components. At 1-year follow-up, patients whose MRI showed either an increase of T1 LL or at least one gadolinium enhancing lesion had a worse HRQOL at the end of the study. Initial brain parenchymal fraction (BPF) measure was also correlated with the long-term follow-up HRQOL. EDSS scored at the end of the study had not significantly changed (1.3; P>0.05). CONCLUSION: Our study revealed pertinent clinicoradiological correlations between HRQOL and MRI parameters in our cohort. |
Acquiring quantitative information on growing organs is an absolute requirement for a better understanding of morphogenesis in both plants and animals. However, detailed analyses of growth patterns at cellular resolution have remained elusive. To address this, we have designed a novel protocol where we image whole organs from multiple angles, and then computationally merge and segment these images to provide accurate cell identification within high-resolution 3-D reconstructions. We then extend this protocol to study growing organs in 4-D by using a temporal series of such reconstructions to automatically track cell lineages through multiple rounds of cell division during organ development. Using these methods, we carry out quantitative analyses of Arabidopsis flower development at cell resolution, and reveal differential growth patterns of key regions during the early stages of floral morphogenesis that were previously inaccessible. Lastly, using rice roots, we demonstrate that our approach is both generic and scalable. |
BACKGROUND AND PURPOSE: The physiopathologic bases underlying the signal intensity changes and reduced diffusibility observed in prion diseases (TSEs) are still poorly understood. We evaluated the interest of MRS combined with DWI both as a diagnostic tool and a way to understand the mechanism underlying signal intensity and ADC changes in this setting. MATERIALS AND METHODS: We designed a prospective study of multimodal MR imaging in patients with suspected TSEs. Forty-five patients with a suspicion of TSE and 11 age-matched healthy volunteers were included. The MR imaging protocol included T1, FLAIR, and DWI sequences. MRS was performed on the cerebellum, pulvinar, right lenticular nucleus, and frontal cortex. MR images were assessed visually, and ADC values were calculated. RESULTS: Among the 45 suspected cases, 31 fulfilled the criteria for probable or definite TSEs (19 sCJDs, 3 iCJDs, 2 vCJDs, and 7 genetic TSEs); and 14 were classified as AltDs. High signals in the cortex and/or basal ganglia were observed in 26/31 patients with TSEs on FLAIR and 29/31 patients on DWI. In the basal ganglia, high DWI signals corresponded to a decreased ADC. Metabolic alterations, increased mIns, and decreased NAA were observed in all patients with TSEs. ADC values and metabolic changes were not correlated; this finding suggests that neuronal stress (vacuolization), neuronal loss, and astrogliosis do not alone explain the decrease of ADC. CONCLUSIONS: MRS combined with other MR imaging is of interest in the diagnosis of TSE and provides useful information for understanding physiopathologic processes underlying prion diseases. |
Highly conformal irradiation techniques are associated with steep gradient doses. Accuracy and reproducibility of delineation are required to avoid geometric misses and to properly report dose-volume effects on organs at risk. Guidelines of the International Commission on Radiation Units have largely contributed to high quality treatments. The ICRU endeavors to collect and evaluate the latest data and information pertinent to the problems of radiation measurement and dosimetry. There remains a need for delineation guidelines and automatic segmentation tools in routine practice. Among these tools, atlas-based segmentation has been shown to provide promising results. The methodology used for head and neck cancer patients is illustrated. |
We present the Spherical Demons algorithm for registering two spherical images. By exploiting spherical vector spline interpolation theory, we show that a large class of regularizors for the modified Demons objective function can be efficiently approximated on the sphere using iterative smoothing. Based on one parameter subgroups of diffeomorphisms, the resulting registration is diffeomorphic and fast. The Spherical Demons algorithm can also be modified to register a given spherical image to a probabilistic atlas. We demonstrate two variants of the algorithm corresponding to warping the atlas or warping the subject. Registration of a cortical surface mesh to an atlas mesh, both with more than 160k nodes requires less than 5 minutes when warping the atlas and less than 3 minutes when warping the subject on a Xeon 3.2GHz single processor machine. This is comparable to the fastest non-diffeomorphic landmarkfree surface registration algorithms. Furthermore, the accuracy of our method compares favorably to the popular FreeSurfer registration algorithm. We validate the technique in two different applications that use registration to transfer segmentation labels onto a new image: (1) parcellation of in-vivo cortical surfaces and (2) Brodmann area localization in ex-vivo cortical surfaces. |
OBJECT: The localization of any given target in the brain has become a challenging issue because of the increased use of deep brain stimulation to treat Parkinson disease, dystonia, and nonmotor diseases (for example, Tourette syndrome, obsessive compulsive disorders, and depression). The aim of this study was to develop an automated method of adapting an atlas of the human basal ganglia to the brains of individual patients. METHODS: Magnetic resonance images of the brain specimen were obtained before extraction from the skull and histological processing. Adaptation of the atlas to individual patient anatomy was performed by reshaping the atlas MR images to the images obtained in the individual patient using a hierarchical registration applied to a region of interest centered on the basal ganglia, and then applying the reshaping matrix to the atlas surfaces. RESULTS: Results were evaluated by direct visual inspection of the structures visible on MR images and atlas anatomy, by comparison with electrophysiological intraoperative data, and with previous atlas studies in patients with Parkinson disease. The method was both robust and accurate, never failing to provide an anatomically reliable atlas to patient registration. The registration obtained did not exceed a 1-mm mismatch with the electrophysiological signatures in the region of the subthalamic nucleus. CONCLUSIONS: This registration method applied to the basal ganglia atlas forms a powerful and reliable method for determining deep brain stimulation targets within the basal ganglia of individual patients. |
BACKGROUND AND PURPOSE: Accurate conformal radiotherapy treatment requires manual delineation of target volumes and organs at risk (OAR) that is both time-consuming and subject to large inter-user variability. One solution is atlas-based automatic segmentation (ABAS) where a priori information is used to delineate various organs of interest. The aim of the present study is to establish the accuracy of one such tool for the head and neck (H&N) using two different evaluation methods. MATERIALS AND METHODS: Two radiotherapy centres were provided with an ABAS tool that was used to outline the brainstem, parotids and mandible on several patients. The results were compared to manual delineations for the first centre (EM1) and reviewed/edited for the second centre (EM2), both of which were deemed as equally valid gold standards. The contours were compared in terms of their volume, sensitivity and specificity with the results being interpreted using the Dice similarity coefficient and a receiver operator characteristic (ROC) curve. RESULTS: Automatic segmentation took typically approximately 7min for each patient on a standard PC. The results indicated that the atlas contour volume was generally within +/-1SD of each gold standard apart from the parotids for EM1 and brainstem for EM2 that were over- and under-estimated, respectively (within +/-2SD). The similarity of the atlas contours with their respective gold standard was satisfactory with an average Dice coefficient for all OAR of 0.68+/-0.25 for EM1 and 0.82+/-0.13 for EM2. All data had satisfactory sensitivity and specificity resulting in a favourable position in ROC space. CONCLUSIONS: These tests have shown that the ABAS tool exhibits satisfactory sensitivity and specificity for the OAR investigated. There is, however, a systematic over-segmentation of the parotids (EM1) and under-segmentation of the brainstem (EM2) that require careful review and editing in the majority of cases. Such issues have been discussed with the software manufacturer and a revised version is due for release. |
Intra-subject and inter-subject nonlinear registration based on dense transformations requires the setting of many parameters, mainly for regularization. This task is a major issue, as the global quality of the registration will depend on it. Setting these parameters is, however, very hard, and they may have to be tuned for each patient when processing data acquired by different centers or using different protocols. Thus, we present in this article a method to introduce more coherence in the registration by using fewer degrees of freedom than with a dense registration. This is done by registering the images only on user-defined areas, using a set of affine transformations, which are optimized together in a very efficient manner. Our framework also ensures a smooth and coherent transformation thanks to a new regularization of the affine components. Finally, we ensure an invertible transformation thanks to the Log-Euclidean polyaffine framework. This allows us to get a more robust and very efficient registration method, while obtaining good results as explained below. We performed a qualitative and quantitative evaluation of the obtained results on two applications: first on atlas-based brain segmentation, comparing our results with a dense registration algorithm. Then the second application for which our framework is particularly well suited concerns bone registration in the lower-abdomen area. We obtain in this case a better positioning of the femoral heads than with a dense registration. For both applications, we show a significant improvement in computation time, which is crucial for clinical applications. |
BACKGROUND: Increasing evidence supports the usefulness of brain magnetic resonance imaging (MRI) for the diagnosis of human prion diseases. From the neuroradiological point of view, fatal familial insomnia is probably the most challenging to diagnose because brain lesions are mostly confined to the thalamus. OBJECTIVE: To determine whether multisequence MRI of the brain can show thalamic alterations and establish pathoradiologic correlations in a patient with familial fatal insomnia. DESIGN: Radioclinical prospective study. We describe a patient with fatal familial insomnia and normal MRI images. Because the MRI study was performed only 4 days before the patient's death, we were able to compare radiological data with the lesions observed at the neuropathologic level. Patient A 55-year-old man with familial fatal insomnia. Main Outcome Measure Magnetic resonance spectroscopy combined with the measurement of apparent diffusion coefficient of water in different brain areas. RESULTS: The neuroradiological study showed, in the thalamus but not in the other brain regions studied, an increase of apparent diffusion coefficient of water and a metabolic pattern indicating gliosis. These alterations closely correlated with neuropathologic data showing an almost pure gliosis that was restricted to the thalami. CONCLUSION: Considering fatal familial insomnia as a model of thalamic-restricted gliosis, this case demonstrates that multisequences of magnetic resonance can detect prion-induced gliosis in vivo, as confirmed by a neuropathologic examination performed only a few days after radiological examination. |
PURPOSE: The aim of the present study was to quantitatively assess the performance of a block matching-based automatic registration algorithm integrated within the commercial treatment planning system designated ISOgray from Dosisoft. The accuracy of the process was evaluated by a phantom study on computed tomography (CT), magnetic resonance (MR) and positron emission tomography (PET) images. MATERIALS AND METHODS: Two phantoms were used to carry out this study: the cylindrical Jaszczak phantom and the anthropomorphic Liqui-Phil Head Phantom (the Phantom Laboratory), containing fillable spheres. External fiducial markers were used to quantify the accuracy of 41 CT/CT, MR/CT and PET/CT automatic registrations with images of the rotated and tilted phantoms. RESULTS: The study first showed that a cylindrical phantom was not adapted for the evaluation of the performance of a block matching-based registration software. Secondly, the Liqui-Phil Head Phantom study showed that the algorithm was able to perform automatic registrations of CT/CT and MR/CT images with differences of up to 40 degrees in phantom rotation and of up to 20-30 degrees for PET/CT with accuracy below the image voxel size. CONCLUSION: The study showed that the block matching-based automatic registration software under investigation was robust, reliable and yielded very satisfactory results. This phantom-based test can be integrated into a periodical quality assurance process and used for any commissioning of image registration software for radiation therapy. |
BACKGROUND AND PURPOSE: Conformal radiation therapy techniques require the delineation of volumes of interest, a time-consuming and operator-dependent task. In this work, we aimed to evaluate the potential interest of an atlas-based automatic segmentation software (ABAS) of brain organs at risk (OAR), when used under our clinical conditions. MATERIALS AND METHODS: Automatic and manual segmentations of the eyes, optic nerves, optic chiasm, pituitary gland, brain stem and cerebellum of 11 patients on T1-weighted magnetic resonance, 3-mm thick slice images were compared using the Dice similarity coefficient (DSC). The sensitivity and specificity of the ABAS were also computed and analysed from a radiotherapy point of view by splitting the ROC (Receiver Operating Characteristic) space into four sub-regions. RESULTS: Automatic segmentation of OAR was achieved in 7-8 min. Excellent agreement was obtained between automatic and manual delineations for organs exceeding 7 cm3: the DSC was greater than 0.8. For smaller structures, the DSC was lower than 0.41. CONCLUSIONS: These tests demonstrated that this ABAS is a robust and reliable tool for automatic delineation of large structures under clinical conditions in our daily practice, even though the small structures must continue to be delineated manually by an expert. |
BACKGROUND AND PURPOSE: The corpus callosum is an important predilection site for traumatic axonal injury but may be unevenly affected in head trauma. We hypothesized that there were local differences in axonal injury within the corpus callosum as investigated with diffusion tensor imaging (DTI), varying among patients with differing severity of traumatic brain injury (TBI). MATERIALS AND METHODS: Ethics committee approval and informed consent were obtained. Ten control subjects (7 men, 3 women; mean age, 37 +/- 9 years) and 39 patients with TBI (27 men, 12 women; 34 +/- 12 years) were investigated, of whom 24 had mild; 9, moderate; and 6, severe TBI. Regions of interest were selected in the callosal genu, body, and splenium to calculate fractional anisotropy (FA), apparent diffusion coefficient (ADC), and the number of fibers passing through. Statistical comparison was made through analysis of variance with the Scheffe post hoc analysis. RESULTS: Compared with controls, patients with mild TBI investigated <3 months posttrauma (n = 12) had reduced FA (P < .01) and increased ADC (P < .05) in the genu, whereas patients with mild TBI investigated > or =3 months posttrauma (n = 12) showed no significant differences. Patients with moderate and severe TBI, all investigated <3 months posttrauma, had reduced FA (P < .001) and increased ADC (P < .01) in the genu compared with controls and reduced FA in the splenium (P < .001) without significant ADC change. CONCLUSION: Mild TBI is associated with DTI abnormalities in the genu <3 months posttrauma. In more severe TBI, both the genu and splenium are affected. DTI suggests a larger contribution of vasogenic edema in the genu than in the splenium in TBI. |
BACKGROUND AND PURPOSE: Traumatic axonal injury is a primary brain abnormality in head trauma and is characterized by reduction of fractional anisotropy (FA) on diffusion tensor imaging (DTI). Our hypothesis was that patients with mild traumatic brain injury (TBI) have widespread brain white matter regions of reduced FA involving a variety of fiber bundles and show fiber disruption on fiber tracking in a minority of these regions. MATERIALS AND METHODS: Ethics committee approval and informed consent were obtained. Twenty-one patients with mild TBI were investigated (men:women, 12:9; mean age +/- SD, 32 +/- 9 years). In a voxel-based comparison with 11 control subjects (men:women, 8:3; mean age, 37 +/- 9 years) using z score analysis, patient regions with abnormally reduced FA were defined in brain white matter. MR imaging, DTI, and fiber tracking characteristics of these regions were described and analyzed using Pearson correlation, linear regression analysis, or the chi(2) test when appropriate. RESULTS: Patients had on average 9.1 regions with reduced FA, with a mean region volume of 525 mm(3), predominantly found in cerebral lobar white matter, cingulum, and corpus callosum. These regions mainly involved supratentorial projection fiber bundles, callosal fibers, and fronto-temporo-occipital association fiber bundles. Internal capsules and infratentorial white matter were relatively infrequently affected. Of all of the involved fiber bundles, 19.3% showed discontinuity on fiber tracking. CONCLUSION: Patients with mild TBI have multiple regions with reduced FA in various white matter locations and involving various fiber bundles. A minority of these fiber bundles show discontinuity on fiber tracking. |
In emission tomography imaging, respiratory motion causes artifacts in lungs and cardiac reconstructed images, which lead to misinterpretations, imprecise diagnosis, impairing of fusion with other modalities, etc. Solutions like respiratory gating, correlated dynamic PET techniques, list-mode data based techniques and others have been tested, which lead to improvements over the spatial activity distribution in lungs lesions, but which have the disadvantages of requiring additional instrumentation or the need of discarding part of the projection data used for reconstruction. The objective of this study is to incorporate respiratory motion compensation directly into the image reconstruction process, without any additional acquisition protocol consideration. To this end, we propose an extension to the maximum likelihood expectation maximization (MLEM) algorithm that includes a respiratory motion model, which takes into account the displacements and volume deformations produced by the respiratory motion during the data acquisition process. We present results from synthetic simulations incorporating real respiratory motion as well as from phantom and patient data. |
This paper describes the construction of an atlas of the human basal ganglia. The successive steps of the construction were as follows. First a postmortem specimen was subjected to a MRI acquisition prior to extraction of the brain from the skull. The brain was then cryosectioned (70 mum thickness). One section out of ten (80 sections) was Nissl-stained with cresyl violet, another series of 80 sections was immunostained for the calcium binding protein calbindin. Contours of basal ganglia nuclei including their calbindin-stained functional subdivisions, fiber bundles and ventricles (n=80 structures) were traced from histological sections and digitized. A novelty of this atlas is the MRI acquisition, which represents the core data element of the study. MRI was used for the coregistration of the atlas data and permitted, through multimodal (Nissl, calbindin, images of cryosectioning, T1 and T2 MRI) and 3D optimization, the production of anatomically and geometrically consistent 3D surfaces, which can be sliced through any desired orientation. The atlas MRI is also used for its deformation to provide accurate conformation to the MRI of living patients, thus adding information at the histological level to the patient's MRI volume. This latter aspect will be presented in a forthcoming paper. |
Introduction. Cognitive impairment is frequent in relapsing remitting Multiple Sclerosis and is often diagnosed after disruption of occupational and social relations. METHODS: We studied at baseline a homogeneous population of 32 RRMS patients, diagnosed for less than 5 years, with spontaneous memory complaints, and 20 controls. Sixteen patients were followed for 2 years, combining physical examination, neuropsychological tests, and brain MRI. Neuropsychological tests used evaluated memory capacities, attentional capacities, executive functions, language, and visuo-constructive praxis. Lesion load on brain MRI was measured with semi-automatic segmentation procedures and manual control. RESULTS: Eighty percent of patients presented cognitive impairment, and this proportion was higher than that found in the literature. These disorders were more marked for verbal episodic memory, attention, and executive functions. Patients with brain MRI that initially fulfilled the Barkhof criteria and those with callous lesions had more memory disorders. No link between global T1 and T2 lesion loads and neuropsychological scores was found. A statistical link between posterior fossa lesions and attentional disorders was shown. In the longitudinal follow-up, patients had better performances in memory and attentional domains, and a lower number of cognitive domains with dysfunction for each patient. This improvement on neuropsychological tests, whereas EDSS levels were stable, underlined a possible test-retest effect. CONCLUSION: During the initial phase of the disease, most of the relapsing remitting patients present a mild cognitive impairment. Early detection, therapeutic propositions, and recognition of disorders are necessary. |
Cardiovascular diseases remain the primary cause of death in developed countries. In most cases, exploration of possibly underlying coronary artery pathologies is performed using X-ray coronary angiography. Current clinical routine in coronary angiography is directly conducted in 2-D projection images from several static viewing angles. However, for diagnosis and treatment purposes, coronary artery reconstruction is highly suitable. The purpose of this study is to provide physicians with a 3-D model of coronary arteries, e.g. for absolute three-dimensional measures for lesion assessment, instead of direct projective measures deduced from the images, which are highly dependent on the viewing angle. In this article, we propose a novel method to reconstruct coronary arteries from one single rotational X-ray projection sequence. As a side result, we also obtain an estimation of the coronary artery motion. Our method consists of 3 main consecutive steps: (1) 3-D reconstruction of coronary artery centerlines, including respiratory motion compensation, (2) coronary artery 4-D motion computation, and (3) 3-D tomographic reconstruction of coronary arteries, involving compensation for respiratory and cardiac motions. We present some experiments on clinical datasets, and the feasibility of a true 3-D Quantitative Coronary Analysis is demonstrated. |
OBJECTIVE: Detailed information on microvascular network anatomy is a requirement for understanding several aspects of microcirculation, including oxygen transport, distributions of pressure, and wall shear stress in microvessels, regulation of blood flow, and interpretation of hemodynamically based functional imaging methods, but very few quantitative data on the human brain microcirculation are available. The main objective of this study is to propose a new method to analyze this microcirculation. METHODS: From thick sections of india ink-injected human brain, using confocal laser microscopy, the authors developed algorithms adapted to very large data sets to automatically extract and analyze center lines together with diameters of thousands of brain microvessels within a large cortex area. RESULTS: Direct comparison between the original data and the processed vascular skeletons demonstrated the high reliability of this method and its capability to manage a large amount of data, from which morphometry and topology of the cerebral microcirculation could be derived. CONCLUSIONS: Among the many parameters that can be analyzed by this method, the capillary size, the frequency distributions of diameters and lengths, the fractal nature of these networks, and the depth-related density of vessels are all vital features for an adequate model of cerebral microcirculation. |
Focal cortical dysplasia (FCD) is the most frequent malformation of cortical development in patients with medically intractable epilepsy. On MRI, FCD lesions are not easily differentiable from the normal cortex and defining their spatial extent is challenging. In this paper, we introduce a method to segment FCD lesions on T1-weighted MRI. It relies on two successive three-dimensional deformable models, whose evolutions are based on the level set framework. The first deformable model is driven by probability maps obtained from three MRI features: cortical thickness, relative intensity and gradient. These features correspond to the visual characteristics of FCD and allow discriminating lesions and normal tissues. In a second stage, the previous result is expanded towards the underlying and overlying cortical boundaries, throughout the whole cortical section. The method was quantitatively evaluated by comparison with manually traced labels in 18 patients with FCD. The automated segmentations achieved a strong agreement with the manuals labels, demonstrating the applicability of the method to assist the delineation of FCD lesions on MRI. This new approach may become a useful tool for the presurgical evaluation of patients with intractable epilepsy related to cortical dysplasia. |
Functional MRI is a technique of imaging which is developing fast as it allows non-aggressive evaluation of brain functions. Diffusion, perfusion and activation are each used to study brain responsiveness to a given task. As a pretherapeutic routine investigation, in brain tumours, it can be helpful as an additional tool to morphological MRI in evaluating the prognosis of patients. |
Spinal cord astrocytomas are rare neoplasms that can result in alteration of the spinal cord structural integrity, which can be assessed by using diffusion tensor imaging methods. Our objective was to visualize the deformation of the posterior spinal cord lemniscal and corticospinal tracts in 5 patients with low-grade astrocytomas compared with 10 healthy volunteers by using 3D fiber-tracking reconstructions. |
The study of cerebral microvascular networks requires high-resolution images. However, to obtain statistically relevant results, a large area of the brain (several square millimeters) must be analyzed. This leads us to consider huge images, too large to be loaded and processed at once in the memory of a standard computer. To consider a large area, a compact representation of the vessels is required. The medial axis is the preferred tool for this application. To extract it, a dedicated skeletonization algorithm is proposed. Numerous approaches already exist which focus on computational efficiency. However, they all implicitly assume that the image can be completely processed in the computer memory, which is not realistic with the large images considered here. We present in this paper a skeletonization algorithm that processes data locally (in subimages) while preserving global properties (i.e., homotopy). We then show some results obtained on a mosaic of three-dimensional images acquired by confocal microscopy. |
OBJECTIVES: To illustrate the value of diffusion tensor imaging and tractography in the diagnosis and follow-up of central pontine myelinolysis. CASE REPORT: We report a case of central pontine myelinolysis in a 29 year old woman, also anorexic, studied using MR Diffusion Tensor Imaging (DTI) and Fibre Tracking (FT) focused on the pons, and compared with the studies of 5 normal volunteers. Tractography showed a swollen aspect of the right corticospinal fiber tract correlating with mild left lower extremity deficit at clinical evaluation. The pontine fibers were posteriorly displaced but intact. The sensory tracts were also intact. Apparent Diffusion Coefficient values were increased and Fractional Anisotropy was decreased in the lesions. Follow up imaging showed persistent abnormal ADC and FA values in the pons although the left cortico-spinal tract returned to normal, consistent with the clinical outcome. CONCLUSION: Diffusion Tensor Imaging MR and Fiber tractography are a new method to analyse white matter tracts. It can be used to prospectively evaluate the location of white matter tract lesions at the acute phase of central pontine myelinolysis and follow up. |
Tensors are nowadays a common source of geometric information. In this paper, we propose to endow the tensor space with an affine-invariant Riemannian metric. We demonstrate that it leads to strong theoretical properties: the cone of positive definite symmetric matrices is replaced by a regular and complete manifold without boundaries (null eigenvalues are at the infinity), the geodesic between two tensors and the mean of a set of tensors are uniquely defined, etc. We have previously shown that the Riemannian metric provides a powerful framework for generalizing statistics to manifolds. In this paper, we show that it is also possible to generalize to tensor fields many important geometric data processing algorithms such as interpolation, filtering, diffusion and restoration of missing data. For instance, most interpolation and Gaussian filtering schemes can be tackled efficiently through a weighted mean computation. Linear and anisotropic diffusion schemes can be adapted to our Riemannian framework, through partial differential evolution equations, provided that the metric of the tensor space is taken into account. For that purpose, we provide intrinsic numerical schemes to compute the gradient and Laplace-Beltrami operators. Finally, to enforce the fidelity to the data (either sparsely distributed tensors or complete tensors fields) we propose least-squares criteria based on our invariant Riemannian distance which are particularly simple and efficient to solve. |
This manuscript tackles the reconstruction of 3D volumes via mono-modal registration of series of 2D biological images (histological sections, autoradiographs, cryosections, etc.). The process of acquiring these images typically induces composite transformations that we model as a number of rigid or affine local transformations embedded in an elastic one. We propose a registration approach closely derived from this model. Given a pair of input images, we first compute a dense similarity field between them with a block matching algorithm. We use as a similarity measure an extension of the classical correlation coefficient that improves the consistency of the field. A hierarchical clustering algorithm then automatically partitions the field into a number of classes from which we extract independent pairs of sub-images. Our clustering algorithm relies on the Earth mover's distribution metric and is additionally guided by robust least-square estimation of the transformations associated with each cluster. Finally, the pairs of sub-images are, independently, affinely registered and a hybrid affine/non-linear interpolation scheme is used to compose the output registered image. We investigate the behavior of our approach on several batches of histological data and discuss its sensitivity to parameters and noise. |
PURPOSE: Our aim was to study the fractional anisotropy (FA) variations and the fiber tracking (FT) patterns observed in patients with myelitis. MATERIAL AND METHODS: Fifteen patients with symptomatic myelitis and 11 healthy subjects were prospectively selected. We performed T2-weighted and diffusion tensor imaging on a 1.5T MR scanner. FA and apparent diffusion coefficient maps were computed in both healthy subjects and patients. In each patient, we performed FT to study pathologic aspects on this imaging method. FA data were analyzed by using z-scores. RESULTS: For the healthy subjects, averaged FA values ranged from 0.745 to 0.751. All abnormal areas seen on T2-weighted imaging had a significantly decreased FA value. In 9 patients (60%), FA maps showed decreased FA areas, whereas T2-weighted imaging findings were normal. These areas matched the neurologic deficit in 33%. Eighty percent of patients had multiple decreased FA areas. Five patients (33%) had increased FA values in normal T2-weighted areas. CONCLUSION: We observed specific FA and FT pattern variations in patients with myelitis. |
The ever-rising diffusion of cellular phones has brought about an increased concern for the possible consequences of electromagnetic radiation on human health. Possible thermal effects have been investigated, via experimentation or simulation, by several research projects in the last decade. Concerning numerical modeling, the power absorption in a user's head is generally computed using discretized models built from clinical MRI data. The vast majority of such numerical studies have been conducted using Finite Differences Time Domain methods, although strong limitations of their accuracy are due to heterogeneity, poor definition of the detailed structures of head tissues (staircasing effects), etc. In order to propose numerical modeling using Finite Element or Discontinuous Galerkin Time Domain methods, reliable automated tools for the unstructured discretization of human heads are also needed. Results presented in this article aim at filling the gap between human head MRI images and the accurate numerical modeling of wave propagation in biological tissues and its thermal effects. |
In studies on animal models of human brain pathologies, three-dimensional reconstruction from histological sections is particularly useful when assessing the morphologic, functional and biochemical changes induced by pathology. It allows assessing lesion heterogeneity in planes different from the cutting plane and allows correlating the histology with images obtained in vivo, such as by means of magnetic resonance imaging. To create a 3D volume from autoradiographic sections with minimal distortion, both cryosectioning as well as section registration need to be optimal. This paper describes a strategy whereby four external fiducial markers are positioned outside the rat brain with the use of a low temperature brain embedding procedure. The fiducial markers proposed here can be rapidly added to any frozen tissue block with no impact on the subsequent histological operations. Since embedding is performed at a low temperature, no tissue degradation occurs due to sample heating. The markers enable robust and almost error free registration, even in the presence of missing sections and poor image quality. Furthermore, the markers may be used to partially correct for global distortions. |
Saccular aneurisms illustrate usefulness and possible techniques of image-based modeling of flow in diseased vessels. Aneurism flow is investigated in order to estimate the rupture risk, assuming that the pressure is the major factor and that high-pressure zones are correlated to within-wall strong-stress concentrations. Computational flow is also aimed at providing additional arguments for the treatment strategy. Angiographies of aneurismal vessels of large and medium size are processed to provide three-dimensional reconstruction of the vessel region of interest. Different reconstruction techniques are used for a side and a terminal aneurisms. Reconstruction techniques may lead to different geometries especially with poor input data. The associated facetisation is improved to get a computation-adapted surface triangulation, after a treatment of vessel ends and mesh adaptation. Once the volumic mesh is obtained, the pulsatile flow of an incompressible Newtonian blood is computed using in vivo non-invasive flowmetry and the finite element method. High pressure zones are observed in the aneurism cavity. The pressure magnitude in the aneurism, the location and the size of high pressure zones depend mainly on the aneurism implantation on the vessel wall and its orientation with respect to the blood flux in the upstream vessel. The stronger the blood impacts on the aneurismal wall the higher the pressure. The state of the aneurism neck, where a high-pressure zone can occur, and the location of the aneurism, with an easy access or not, give arguments for the choice between coiling and surgical clipping. Mesh size and 3D reconstruction procedure affect the numerical results. Helpful qualitative data are provided rather than accurate quantitative results in the context of multimodeling. |
PURPOSE: Brain tumor radiotherapy requires the volume measurements and the localization of several individual brain structures. Any tool that can assist the physician to perform the delineation would then be of great help. Among segmentation methods, those that are atlas-based are appealing because they are able to segment several structures simultaneously, while preserving the anatomy topology. This study aims to evaluate such a method in a clinical context. METHODS AND MATERIALS: The brain atlas is made of two three-dimensional (3D) volumes: the first is an artificial 3D magnetic resonance imaging (MRI); the second consists of the segmented structures in this artificial MRI. The elastic registration of the artificial 3D MRI against a patient 3D MRI dataset yields an elastic transformation that can be applied to the labeled image. The elastic transformation is obtained by minimizing the sum of the square differences of the image intensities and derived from the optical flow principle. This automatic delineation (AD) enables the mapping of the segmented structures onto the patient MRI. Parameters of the AD have been optimized on a set of 20 patients. Results are obtained on a series of 6 patients' MRI. A comprehensive validation of the AD has been conducted on performance of atlas-based segmentation in a clinical context with volume, position, sensitivity, and specificity that are compared by a panel of seven experimented physicians for the brain tumor treatments. RESULTS: Expert interobserver volume variability ranged from 16.70 cm(3) to 41.26 cm(3). For patients, the ratio of minimal to maximal volume ranged from 48% to 70%. Median volume varied from 19.47 cm(3) to 27.66 cm(3) and volume of the brainstem calculated by AD varied from 17.75 cm(3) to 24.54 cm(3). Medians of experts ranged, respectively, for sensitivity and specificity, from 0.75 to 0.98 and from 0.85 to 0.99. Median of AD were, respectively, 0.77 and 0.97. Mean of experts ranged, respectively, from 0.78 to 0.97 and from 0.86 to 0.99. Mean of AD were, respectively, 0.76 and 0.97. CONCLUSIONS: Results demonstrate that the method is repeatable, provides a good trade-off between accuracy and robustness, and leads to reproducible segmentation and labeling. These results can be improved by enriching the atlas with the rough information of tumor or by using different laws of deformation for the different structures. Qualitative results also suggest that this method can be used for automatic segmentation of other organs such as neck, thorax, abdomen, pelvis, and limbs. |
We present a new algorithm to register 3D pre-operative Magnetic Resonance (MR) images to intra-operative MR images of the brain which have undergone brain shift. This algorithm relies on a robust estimation of the deformation from a sparse noisy set of measured displacements. We propose a new framework to co mpute the displacement field in an iterative process, allowing the solution to gradually move from an approximation formulation (minimizing the sum of a re gularization term and a data error term) to an interpolation formulation (least square minimization of the data error term). An outlier rejection step is i ntroduced in this gradual registration process using a weighted least trimmed squares approach, aiming at improving the robustness of the algorithm. We use a patient-specific model discretized with the finite element method (FEM) in order to ensure a realistic mechanical behavior of the brain tissue. To meet the clinical time constraint, we parallelized the slowest step of the algorithm so that we can perform a full 3D image registration in 35 seconds ( including the image update time) on a heterogeneous cluster of 15 PCs. The algorithm has been tested on six cases of brain tumor resection, presenting a brain shift of up to 14 mm. The results show a good ability to recover la rge displacements, and a limited decrease of accuracy near the tumor resection cavity. |
We propose a new model to simulate the 3D growth of glioblastomas multiforma (GBMs), the most aggressive glial tumors. The GBM speed of growth depends on the invaded tissue: faster in white than in gray matter, it is stopped by the dura or the ventricles. These different structures are introduced into the model using an atlas matching technique. The atlas includes both the segmentations of anatomical structures and diffusion information in white matter fibers. We use the finite element method (FEM) to simulate the invasion of the GBM in the brain parenchyma and its mechanical interaction with the invaded structures (mass effect). Depending on the considered tissue, the former effect is modeled with a reaction-diffusion or a Gompertz equation, while the latter is based on a linear elastic brain constitutive equation. In addition, we propose a new coupling equation taking into account the mechanical influence of the tumor cells on the invaded tissues. The tumor growth simulation is assessed by comparing the extit{in-silico} GBM growth with the real growth observed on two magnetic resonance images (MRIs) of a patient acquired with six months difference. Results show the feasibility of this new conceptual approach and justifies its further validation. |
This paper proposes an efficient method for removing tetrahedra from a tetrahedral mesh while keeping its manifold property. We first define precisely the notion of manifold tetrahedral mesh and stress its relevance in the context of real-time surgery simulation. We then provide a method for removing a tetrahedron that complies with the manifold definition. This removal may require in some cases the removal of neighboring tetrahedra. After providing an exhaustive description of the tetrahedron removal algorithm, its efficiency is evaluated for different mesh configurations. This algorithm is currently used in the context of real-time surgery simulation where the action of an ultrasonic lancet can be simulated by the removal of small set of tetrahedra from a tetrahedralisation. |
Chamfer distances are widely used in image analysis and many authors have investigated the computation of optimal chamfer mask coefficients. Unfortunately, these methods are not systematized: calculations have to be conducted manually for every mask size or image anisotropy. Since image acquisition (e.g. medical imaging) can lead to discrete anisotropic grids with unpredictable anisotropy value, automated calculation of chamfer mask coefficients becomes mandatory for e cient distance map computations. This article presents an automatic construction for chamfer masks of arbitrary sizes. This allows, first, to derive analytically the relative error with respect to the Euclidean distance, in any 3-D anisotropic lattice, and second, to compute optimal chamfer coefficients. In addition, the resulting chamfer map verifies discrete norm conditions. |
Pulmonary hypertensive disease is assessed by quantification of pulmonary vascular resistance. Pulmonary total arterial compliance is also an indicator of pulmonary hypertensive disease. However, because of difficulties in measuring compliance, it is rarely used. We describe a method of measuring pulmonary arterial compliance utilizing magnetic resonance (MR) flow data and invasive pressure measurements. Seventeen patients with suspected pulmonary hypertension or congenital heart disease requiring preoperative assessment underwent MR-guided cardiac catheterization. Invasive manometry was used to measure pulmonary arterial pressure, and phase-contrast MR was used to measure flow at baseline and at 20 ppm nitric oxide (NO). Total arterial compliance was calculated using the pulse pressure method (parameter optimization of the 2-element windkessel model) and the ratio of stroke volume to pulse pressure. There was good agreement between the two estimates of compliance (r = 0.98, P 10% in response to 20 ppm NO. As a population, the increase did not reach statistical significance. There was an inverse relation between compliance and resistance (r = 0.89, P < 0.001) and between compliance and mean pulmonary arterial pressure (r = 0.72, P < 0.001). We have demonstrated the feasibility of quantifying total arterial compliance using an MR method. |
We describe an automatic and reproducible method to analyze the histological design of the cerebral cortex as applied to brain sections stained to reveal myelinated fibers. The technique provides an evaluation of the distribution of myelination across the width of the cortical mantle in accordance with a model of its curvature and its intrinsic geometry. The profile lines along which the density of staining is measured are generated from the solution of a partial differential equation (PDE) that models the intermediate layers of the cortex. Cortical profiles are classified according to significant components that emerge from wavelet analysis. Intensity profiles belonging to each distinct class are normalized and averaged to produce area-specific templates of cortical myelo-architecture. |
La radioth{\'e}rapie est un domaine privil{\'e}gi{\'e} d'application des techniques de traitement des images de par l'utilisation importante de donn{\'e}es issues de l'imagerie. Celles-ci sont de plus en pleine expansion du fait de la progression des performances informatiques. Actuellement, les d{\'e}veloppements r{\'e}cents de la radioth{\'e}rapie (radioth{\'e}rapie de conformation, radioth{\'e}rapie conformationnelle avec modulation d'intensit{\'e}) procurent une place majeure {\`a} ces techniques. En effet, elles contribuent {\`a} r{\'e}pondre aux conditions de pr{\'e}cision exig{\'e}es par la radioth{\'e}rapie moderne et permettent d'envisager d'am{\'e}liorer les traitements. L'objectif de cet article est de pr{\'e}senter les diff{\'e}rentes techniques du traitement d'image utilis{\'e}es aujourd'hui en radioth{\'e}rapie (segmentation et recalage en particulier) au travers de la litt{\'e}rature. |
In this study, we used event-related functional magnetic resonance imaging to investigate whether visual mental images retinotopically activate early visual cortex. Six participants were instructed to visualize or view horizontally or vertically oriented flashing bow-tie shaped stimuli. When compared to baseline, imagery globally activated Area V1. When the activation evoked by the stimuli at the different orientations was directly compared, distinct spatial activation patterns were obtained for each orientation in most participants. Not only was the topography of the activation patterns from imagery similar to the topography obtained with a corresponding visual perception task, but it closely matched the individual cortical representation of either the horizontal or the vertical visual field meridians. These findings strongly support that visual imagery and perception share low-level anatomical substrate and functional processes. Binding of spatial features is suggested as one possible mechanism. |
INTRODUCTION: Magnetic resonance imaging (MRI) has transformed management of patients with multiple sclerosis. The exact contribution of brain MRI remains a subject of debate, but it is generally considered to provide a more specific and more sensitive outcome measure for monitoring purposes and for testing new therapies. The choice of MRI techniques, and measurement reproducibility for multiple sclerosis brain lesions are not defined with precision for routine practice. There are many sources of error when comparing successive images which can be overcome to some extent with repositioning and image processing techniques. METHODS: We evaluated the impact of image repositioning on treatment decision-making for twelve relapsing remitting patients. Brain MRIs were performed every three months for a one-year period. Two neurologists interpreted the non-repositioned and repositioned images giving their analysis of changes in the lesions visualized on the T2 sequences and their therapeutic decisions. RESULTS: For the first neurologist, analysis of the non-repositioned images yielded six patients whose lesions had worsened while for the repositioned images there were only three. For the second neurologist, four patients had more lesions with the non-repositioned images and only three with repositioning. The subjective interpretations were the same for the two neurologists when they used repositioned images. CONCLUSIONS: Comparison by two neurologists of non-repositioned and repositioned MRI, with no other image processing, affected the analysis and in certain cases propositions for treatment. |
Strains of mice, through breeding or the disruption of normal genetic pathways, are widely used to model human diseases. Atlases are an invaluable aid in understanding the impact of such manipulations by providing a standard for comparison. We have developed a digital atlas of the adult C57BL/6J mouse brain as a comprehensive framework for storing and accessing the myriad types of information about the mouse brain. Our implementation was constructed using several different imaging techniques: magnetic resonance microscopy, blockface imaging, classical histology and immunohistochemistry. Along with raw and annotated images, it contains database management systems and a set of tools for comparing information from different techniques. The framework allows facile correlation of results from different animals, investigators or laboratories by establishing a canonical representation of the mouse brain and providing the tools for the insertion of independent data into the same space as the atlas. This tool will aid in managing the increasingly complex and voluminous amounts of information about the mammalian brain. It provides a framework that encompasses genetic information in the context of anatomical imaging and holds tremendous promise for producing new insights into the relationship between genotype and phenotype. We describe a suite of tools that enables the independent entry of other types of data, facile retrieval of information and straightforward display of images. Thus, the atlas becomes a framework for managing complex genetic and epigenetic information about the mouse brain. The atlas and associated tools may be accessed at http://www.loni.ucla.edu/MAP. |
In the past years, the development of 3-D medical imaging has enabled the 3-D imaging of in vivo tissues, from an anatomical (MR, CT) or even functional (fMRI, PET, SPECT) point of view. However, despite immense technological progress, the resolution of these images is still short of the level of anatomical or functional details that in vitro imaging (e.g. histology, autoradiography) permits. The motivation of this work is to compare fMRI activations to activations observed in autoradiographic images from the same animals. We aim to fuse post-mortem autoradiographic data with a pre-mortem anatomical MR image. We first reconstruct a 3-D volume from the 2-D autoradiographic sections, coherent both in geometry and intensity. Then, this volume is fused with the MR image. This way, we ensure that the reconstructed 3-D volume can be superimposed onto the MR image that represents the reference anatomy. We demonstrate that this fusion can be achieved by using only simple global transformations (rigid and/or affine, 2-D and 3-D), while yielding very satisfactory results. |
The problem of increasing the slice resolution of functional MRI (fMRI) images without a loss in signal-to-noise ratio is considered. In standard fMRI experiments, increasing the slice resolution by a certain factor decreases the signal-to-noise ratio of the images with the same factor. For this purpose an adapted EPI MRI acquisition protocol is proposed, allowing one to acquire slice-shifted images from which one can generate interpolated super-resolution images, with an increased resolution in the slice direction. To solve the problem of correctness and robustness of the created super-resolution images from these slice-shifted datasets, the use of discontinuity preserving regularization methods is proposed. Tests on real morphological, synthetic functional, and real functional MR datasets have been performed, by comparing the obtained super-resolution datasets with high-resolution reference datasets. In the morphological experiments the image spatial resolution of the different types of images are compared. In the synthetic and real fMRI experiments, on the other hand, the quality of the different datasets is studied as function of their resulting activation maps. From the results obtained in this study, we conclude that the proposed super-resolution techniques can both improve the signal-to-noise ratio and augment the detectability of small activated areas in fMRI image sets acquired with thicker slices. |
Human functional MRI studies frequently reveal the joint activation of parietal and of lateral and mesial frontal areas during various cognitive tasks. To analyze the geometrical organization of those networks, we used an automatized clustering algorithm that g parcels out sets of areas based on their similar profile of task-related activations or deactivations. This algorithm allowed us to reanalyze published fMRI data (Simon, O., Mangin, J.F., Cohen, L., Le Bihan, D., Dehaene, S., 2002. Topographical layout of hand, eye, calculation, and language-related areas in the human parietal lobe. Neuron 33, 475-487) and to reproduce the previously observed geometrical organization of activations for saccades, attention, grasping, pointing, calculation, and language processing in the parietal lobe. Further, we show that this organization extends to lateral and mesial prefrontal regions. Relative to the parietal lobe, the prefrontal functional geometry is characterized by a partially symmetrical anteroposterior ordering of activations, a decreased representation of effector-specific tasks, and a greater emphasis on higher cognitive functions of attention, higher-order spatial representation, calculation, and language. Anatomically, our results in humans are closely homologous to the known connectivity of parietal and frontal regions in the macaque monkey. |
Recently, radiotherapy possibilities have been dramatically increased by software and hardware developments. Improvements in medical imaging devices have increased the importance of three-dimensional (3D) images as the complete examination of these data by a physician is not possible. Computer techniques are needed to present only the pertinent information for clinical applications. We describe a technique for an automatic 3D reconstruction of the eye and CT scan merging with fundus photographs (retinography). The final result is a "virtual eye" to guide ocular tumor protontherapy. First, we make specific software to automatically detect the position of the eyeball, the optical nerve, and the lens in the CT scan. We obtain a 3D eye reconstruction using this automatic method. Second, we describe the retinography and demonstrate the projection of this modality. Then we combine retinography with a reconstructed eye, using a CT scan to get a virtual eye. The result is a computer 3D scene rendering a virtual eye into a skull reconstruction. The virtual eye can be useful for the simulation, the planning, and the control of ocular tumor protontherapy. It can be adapted to treatment planning to automatically detect eye and organs at risk position. It should be highlighted that all the image processing is fully automatic to allow the reproduction of results, this is a useful property to conduct a consistent clinical validation. The automatic localization of the organ at risk in a CT scan or an MRI by automatic software could be of great interest for radiotherapy in the future for comparison of one patient at different times, the comparison of different treatments centers, the possibility of pooling results of different treatments centers, the automatic generation of doses-volumes histograms, the comparison between different treatment planning for the same patient and the comparison between different patients at the same time. It will also be less time consuming. |
Although numerous methods to register brains of different individuals have been proposed, no work has been done, as far as we know, to evaluate and objectively compare the performances of different nonrigid (or elastic) registration methods on the same database of subjects. In this paper, we propose an evaluation framework, based on global and local measures of the relevance of the registration. We have chosen to focus more particularly on the matching of cortical areas, since intersubject registration methods are dedicated to anatomical and functional normalization, and also because other groups have shown the relevance of such registration methods for deep brain structures. Experiments were conducted using 6 methods on a database of 18 subjects. The global measures used show that the quality of the registration is directly related to the transformation's degrees of freedom. More surprisingly, local measures based on the matching of cortical sulci did not show significant differences between rigid and non rigid methods. |
Standard group analyses of fMRI data rely on spatial and temporal averaging of individuals. This averaging operation is only sensible when the mean is a good representation of the group. This is not the case if subjects are not homogeneous, and it is therefore a major concern in fMRI studies to assess this group homogeneity. We present a method that provides relevant distances or similarity measures between temporal series of brain functional images belonging to different subjects. The method allows a multivariate comparison between data sets of several subjects in the time or in the space domain. These analyses assess the global intersubject variability before averaging subjects and drawing conclusions across subjects, at the population level. We adapt the RV coefficient to measure meaningful spatial or temporal similarities and use multidimensional scaling to give a visual representation of each subject's position with respect to other subjects in the group. We also provide a measure for detecting subjects that may be outliers. Results show that the method is a powerful tool to detect subjects with specific temporal or spatial patterns, and that, despite the apparent loss of information, restricting the analysis to a homogeneous subgroup of subjects does not reduce the statistical sensitivity of standard group fMRI analyses. |
OBJECT: The aim of this study was to correlate the clinical improvement in patients with Parkinson disease (PD) treated using deep brain stimulation (DBS) of the subthalamic nucleus (STN) with the precise anatomical localization of stimulating electrodes. METHODS: Localization was determined by superimposing figures from an anatomical atlas with postoperative magnetic resonance (MR) images obtained in each patient. This approach was validated by an analysis of experimental and clinical MR images of the electrode, and the development of a three-dimensional (3D) atlas-MR imaging coregistration method. The PD motor score was assessed through two contacts for each of two electrodes implanted in 10 patients: the "therapeutic contact" and the "distant contact" (that is, the next but one to the therapeutic contact). Seventeen therapeutic contacts were located within or on the border of the STN, most of which were associated with significant improvement of the four PD symptoms tested. Therapeutic contacts located in other structures (zona incerta, lenticular fasciculus, or midbrain reticular formation) were also linked to a significant positive effect. Stimulation applied through distant contacts located in the STN improved symptoms of PD, whereas that delivered through distant contacts in the remaining structures had variable effects ranging from worsening of symptoms to their improvement. CONCLUSIONS: The authors have demonstrated that 3D atlas-MR imaging coregistration is a reliable method for the precise localization of DBS electrodes on postoperative MR images. In addition, they have confirmed that although the STN is the main target during DBS treatment for PD, stimulation of surrounding regions, particularly the zona incerta or the lenticular fasciculus, can also improve symptoms of PD. |
This anatomic study presents an analysis of the distribution of calbindin immunohistochemistry in the human striatopallidal complex. Entire brains were sectioned perpendicularly to the mid-commissural line into 70-microm-thick sections. Every tenth section was immunostained for calbindin. Calbindin labeling exhibited a gradient on the basis of which three different regions were defined: poorly labeled, strongly labeled, and intermediate. Corresponding contours were traced in individual sections and reformatted as three-dimensional structures. The poorly labeled region corresponded to the dorsal part of the striatum and to the central part of the pallidum. The strongly labeled region included the ventral part of the striatum, the subcommissural part of the external pallidum but also the adjacent portion of its suscommissural part, and the anterior pole of the internal pallidum. The intermediate region was located between the poorly and strongly labeled regions. As axonal tracing and immunohistochemical studies in monkeys show a similar pattern, poorly, intermediate, and strongly labeled regions were considered as the sensorimotor, associative, and limbic territories of the human striatopallidal complex, respectively. However, the boundaries between these territories were not sharp but formed gradients of labeling, which suggests overlapping between adjacent territories. Similarly, the ventral boundary of the striatopallidal complex was blurred, suggesting a structural intermingling with the substantia innominata. This three-dimensional partitioning of the human striatopallidal complex could help to define functional targets for high-frequency stimulation with greater accuracy and help to identify new stimulation sites. |
Given a finite set of points ${\cal P}$ in ${\R}^d$, the diameter of $\cal P$ is defined as the maximum distance between two points of $\cal P$. We propose a very simple algorithm to compute the diameter of a finite set of points. Although the algorithm is not worst-case optimal, an extensive experimental study has shown that it is extremely fast for a large variety of point distributions. In addition, we propose a comparison with the recent approach of Har-Peled and derive hybrid algorithms to combine advantages of both approaches. |
The study of temporal series of medical images can be helpful for physicians to perform pertinent diagnoses and to help them in the follow-up of a patient: in some diseases, lesions, tumors or anatomical structures vary over time in size, position, composition, etc., either because of a natural pathological process or under the effect of a drug or a therapy. It is a laborious and subjective task to visually and manually analyze such images. Thus the objective of this work was to automatically detect regions with apparent local volume variation with a vector field operator applied to the local displacement field obtained after a non-rigid registration between two successive temporal images. On the other hand, quantitative measurements, such as the volume variation of lesions or segmentation of evolving lesions, are important. By studying the information of apparent shrinking areas in the direct and reverse displacement fields between images, we are able to segment evolving lesions. Then we propose a method to segment lesions in a whole temporal series of images. In this article we apply this approach to automatically detect and segment multiple sclerosis lesions that evolve in time series of MRI scans of the brain. At this stage, we have only applied the approach to a few experimental cases to demonstrate its potential. A clinical validation remains to be done, which will require important additional work. |
This paper presents an original method for three-dimensional elastic registration of multimodal images. We propose to make use of a scheme that iterates between correcting for intensity differences between images and performing standard monomodal registration. The core of our contribution resides in providing a method that finds the transformation that maps the intensities of one image to those of another. It makes the assumption that there are at most two functional dependencies between the intensities of structures present in the images to register, and relies on robust estimation techniques to evaluate these functions. We provide results showing successful registration between several imaging modalities involving segmentations, T1 magnetic resonance (MR), T2 MR, proton density (PD) MR and computed tomography (CT). We also argue that our intensity modeling may be more appropriate than mutual information (MI) in the context of evaluating high-dimensional deformations, as it puts more constraints on the parameters to be estimated and, thus, permits a better search of the parameter space. |
Apathy is the most frequent behavioral symptom in Alzheimer's disease and is also frequently reported in other brain organic disorders occurring in the elderly. Based on the literature, we hypothesized that apathy was related to an anterior cingulate hypofunction. Forty-one subjects were studied. According to ICD 10 diagnostic criteria, 28 patients had Alzheimer dementia (demented: diagnostic group 1), and 13 had organic personality disorders or mild cognitive impairment not attributable to dementia (nondemented: diagnostic group 2). Apathy was evaluated by the Neuro-Psychiatric Inventory. As a result each diagnostic group was divided into two symptomatic subgroups: apathetic or nonapathetic. Brain perfusion was measured by 99mTc-labeled bicisate (ECD) brain SPECT and the images were compared using Statistical Parametric Mapping (SPM96). We began by comparing apathetic vs nonapathetic patients, whatever their diagnostic group (whole population), then analyzed them within each group. Twenty-one subjects were apathetic (14 in group 1 and 7 in group 2) and 20 were not (14 in group 1 and 6 in group 2). For the whole population, the Z map showed a significant decrease in ECD uptake for the apathetic patients in the anterior cingulate (P < 0.002) bilaterally. This area was also identified as hypoactive by SPM analysis in the demented (P < 0.035) and in the nondemented (P < 0.02) apathetic patient groups. Finally, conjunction analysis indicated that the anterior cingulate was the common hypoactive structure of the two apathetic subgroups (Z = 4.35, P < 0.0009). These results point to a close relationship between apathy and the anterior cingulate region. |
We present a new image-based technique to rigidly register intraoperative three-dimensional ultrasound (US) with preoperative magnetic resonance (MR) images. Automatic registration is achieved by maximization of a similarity measure which generalizes the correlation ratio, and whose novelty is to incorporate multivariate information from the MR data (intensity and gradient). In addition, the similarity measure is built upon a robust intensity-based distance measure, which makes it possible to handle a variety of US artifacts. A cross-validation study has been carried out using a number of phantom and clinical data. This indicates that the method is quite robust and that the worst registration errors are of the order of the MR image resolution. |
OBJECTIVE: To improve the planning of hepatic surgery, we have developed a fully automatic anatomical, pathological, and functional segmentation of the liver derived from a spiral CT scan. MATERIALS AND METHODS: From a 2 mm-thick enhanced spiral CT scan, the first stage automatically delineates skin, bones, lungs, kidneys, and spleen by combining the use of thresholding, mathematical morphology, and distance maps. Next, a reference 3D model is immersed in the image and automatically deformed to the liver contours. Then an automatic Gaussian fitting on the imaging histogram estimates the intensities of parenchyma, vessels, and lesions. This first result is next improved through an original topological and geometrical analysis, providing an automatic delineation of lesions and veins. Finally, a topological and geometrical analysis based on medical knowledge provides hepatic functional information that is invisible in medical imaging: portal vein labeling and hepatic anatomical segmentation according to the Couinaud classification. RESULTS: Clinical validation performed on more than 30 patients shows that delineation of anatomical structures by this method is often more sensitive and more specific than manual delineation by a radiologist. CONCLUSION: This study describes the methodology used to create the automatic segmentation of the liver with delineation of important anatomical, pathological, and functional structures from a routine CT scan. Using the methods proposed in this study, we have confirmed the accuracy and utility of the creation of a 3D liver model compared with the conventional reading of the CT scan by a radiologist. This work may allow improved preoperative planning of hepatic surgery by more precisely delineating liver pathology and its relationship to normal hepatic structures. In the future, this data may be integrated with computer-assisted surgery and thus represents a first step towards the development of an augmented-reality surgical system. |
Detection of tubular structures in 3D images is an important issue for vascular detection in medical imaging. We present in this paper a new approach for centerline detection and reconstruction of 3D tubular structures. Several models of vessels are introduced for estimating the sensivity of the image second order derivatives according to elliptical cross-section, to curvature of the axis, or to partial volume effects. Our approach uses a multiscale analysis for extracting vessels of different sizes according to the scale. For a given model of vessel, we derive an analytic expression of the relationship between the radius of the structure and the scale at which it is detected. The algorithm gives both centerline extraction and radius estimation of the vessels allowing their reconstruction. The method has been tested on both synthetic and real images, with encouraging results. This work was done in collaboration with GEMS (General Electric Medical Systems, Buc, France). |
PROBLEM/BACKGROUND: In order to help hepatic surgical planning we perfected automatic 3D reconstruction of patients from conventional CT-scan, and interactive visualization and virtual resection tools. TOOLS AND METHODS: From a conventional abdominal CT-scan, we have developed several methods allowing the automatic 3D reconstruction of skin, bones, kidneys, lung, liver, hepatic lesions, and vessels. These methods are based on deformable modeling or thresholding algorithms followed by the application of mathematical morphological operators. From these anatomical and pathological models, we have developed a new framework for translating anatomical knowledge into geometrical and topological constraints. More precisely, our approach allows to automatically delineate the hepatic and portal veins but also to label the portal vein and finally to build an anatomical segmentation of the liver based on Couinaud definition which is currently used by surgeons all over the world. Finally, we have developed a user friendly interface for the 3D visualization of anatomical and pathological structures, the accurate evaluation of volumes and distances and for the virtual hepatic resection along a user-defined cutting plane. RESULTS: A validation study on a 30 patients database gives 2 mm of precision for liver delineation and less than 1 mm for all other anatomical and pathological structures delineation. An in vivo validation performed during surgery also showed that anatomical segmentation is more precise than the delineation performed by a surgeon based on external landmarks. This surgery planning system has been routinely used by our medical partner, and this has resulted in an improvement of the planning and performance of hepatic surgery procedures. CONCLUSION: We have developed new tools for hepatic surgical planning allowing a better surgery through an automatic delineation and visualization of anatomical and pathological structures. These tools represent a first step towards the development of an augmented reality system combined with computer assisted tele-robotical surgery. |
We present a general method to study the dissymmetry of anatomical structures such as those found in the human brain. Our method relies on the estimate of 3D dissymmetry fields, the use of 3D vector field operators, and T2 statistics to compute significance maps. We also present a fully automated implementation of this method which relies mainly on the intensive use of a 3D non-rigid inter-patient matching tool. Such a tool is applied successively between the images and their symmetric versions with respect to an arbitrary plane, both to realign the images with respect to the mid-plane of the subject and to compute a dense 3D dissymmetry map. Inter-patient matching is also used to fuse the data of a population of subjects. We then describe three main application fields: the study of the normal dissymmetry within a given population, the comparison of the dissymmetry between two populations, and the detection of the significant abnormal dissymmetries of a patient with respect to a reference population. Finally, we present preliminary results illustrating these three applications for the case of the human brain. |
OBJECTIVE: This article describes a preliminary work on virtual reality applied to liver surgery and discusses the repercussions of assisted surgical strategy and surgical simulation on tomorrow's surgery. SUMMARY BACKGROUND DATA: Liver surgery is considered difficult because of the complexity and variability of the organ. Common generic tools for presurgical medical image visualization do not fulfill the requirements for the liver, restricting comprehension of a patient's specific liver anatomy. METHODS: Using data from the National Library of Medicine, a realistic three-dimensional image was created, including the envelope and the four internal arborescences. A computer interface was developed to manipulate the organ and to define surgical resection planes according to internal anatomy. The first step of surgical simulation was implemented, providing the organ with real-time deformation computation. RESULTS: The three-dimensional anatomy of the liver could be clearly visualized. The virtual organ could be manipulated and a resection defined depending on the anatomic relations between the arborescences, the tumor, and the external envelope. The resulting parts could also be visualized and manipulated. The simulation allowed the deformation of a liver model in real time by means of a realistic laparoscopic tool. CONCLUSIONS: Three-dimensional visualization of the organ in relation to the pathology is of great help to appreciate the complex anatomy of the liver. Using virtual reality concepts (navigation, interaction, and immersion), surgical planning, training, and teaching for this complex surgical procedure may be possible. The ability to practice a given gesture repeatedly will revolutionize surgical training, and the combination of surgical planning and simulation will improve the efficiency of intervention, leading to optimal care delivery. |
Surgical simulation increasingly appears to be an essential aspect of tomorrow's surgery. The development of a hepatic surgery simulator is an advanced concept calling for a new writing system which will transform the medical world: virtual reality. Virtual reality extends the perception of our five senses by representing more than the real state of things by the means of computer sciences and robotics. It consists of three concepts: immersion, navigation and interaction. Three reasons have led us to develop this simulator: the first is to provide the surgeon with a comprehensive visualisation of the organ. The second reasons is to allow for planning and surgical simulation that could be compared with the detailed flight-plan for a commercial jet pilot. The third lies in the fact that virtual reality is an integrated part of the concept of computer assisted surgical procedure. The project consists of a sophisticated simulator which must include five requirements: a) visual fidelity, b) interactivity, c) physical properties, d) physiological properties, e) sensory input and output. In this report we describe how to obtain a realistic 3D model of the liver from bi-dimensional 2D medical images for anatomical and surgical training. The introduction of a tumor and the consequent planning and virtual resection is also described, as are force feedback and real-time interaction. |
We present a general scheme for automatically building a morphometric anatomical atlas. We detail each stage of the method, including the non-rigid registration algorithm, three-dimensional line averaging and statistical processes. We apply the method to obtain a quantitative atlas of skull crest lines. Finally, we use the resulting atlas to study a craniofacial disease; we show how we can obtain qualitative and quantitative results by contrasting a skull affected by a mandible deformation with the atlas. |
In this paper, we present the concept of diffusing models to perform image-to-image matching. Having two images to match, the main idea is to consider the objects boundaries in one image as semi-permeable membranes and to let the other image, considered as a deformable grid model, diffuse through these interfaces, by the action of effectors situated within the membranes. We illustrate this concept by an analogy with Maxwell's demons. We show that this concept relates to more traditional ones, based on attraction, with an intermediate step being optical flow techniques. We use the concept of diffusing models to derive three different non-rigid matching algorithms, one using all the intensity levels in the static image, one using only contour points, and a last one operating on already segmented images. Finally, we present results with synthesized deformations and real medical images, with applications to heart motion tracking and three-dimensional inter-patients matching. |
Surgical simulation increasingly appears to be an essential aspect of tomorrow's surgery. The development of a hepatic surgery simulator is an advanced concept calling for a new writing system which will transform the medical world: virtual reality. Virtual reality extends the perception of our five senses by representing more than the real state of things by the means of computer sciences and robotics. It consists of three concepts: immersion, navigation and interaction. Three reasons have led us to develop this simulator: the first is to provide the surgeon with a comprehensive visualisation of the organ. The second reason is to allow for planning and surgical simulation that could be compared with the detailed flight-plan for a commercial jet pilot. The third lies in the fact that virtual reality is an integrated part of the concept of computer assisted surgical procedure. The project consists of a sophisticated simulator which has to include five requirements: visual fidelity, interactivity, physical properties, physiological properties, sensory input and output. In this report we will describe how to get a realistic 3D model of the liver from bi-dimensional 2D medical images for anatomical and surgical training. The introduction of a tumor and the consequent planning and virtual resection is also described, as are force feedback and real-time interaction. |
This paper reports work in progress on X-ray angiography acquired under stereotactic conditions. The objective is to be able to match multimodality images (typically {MRI} and X-ray) without a stereotactic frame but with stereotactic precision. We have identified potential problems and have studied them in detail. We conclude that, although the overall application is feasible, much work remains to be done on the estimation of the X-ray system conic projection and on automatic matching based on vascular structures. |
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. |
Despite the large interest in simulators of minimally invasive surgery, it is still unclear to what extent simulators can achieve the task of training medical students in surgical procedures. The answer to that question is certainly linked to the realism of displays and force-feedback systems and to the level of interaction provided by the computer system. In this paper, we describe the virtual environment for anatomical and surgical training on the liver, currently under construction at INRIA. We specifically address the problems of geometric representation and physical modeling and their impact on the two aforementioned problems: realism and real-time interaction. |
Three-dimensional edge detection in voxel images is used to locate points corresponding to surfaces of 3D structures. The next stage is to characterize the local geometry of these surfaces in order to extract points or lines which may be used by registration and tracking procedures. Typically one must calculate second-order differential characteristics of the surfaces such as the maximum, mean, and Gaussian curvature. The classical approach is to use local surface fitting, thereby confronting the problem of establishing links between 3D edge detection and local surface approximation. To avoid this problem, we propose to compute the curvatures at locations designated as edge points using directly the partial derivatives of the image. By assuming that the surface is defined locally by a isointensity contour (i.e., the 3D gradient at an edge point corresponds to the normal to the surface), one can calculate directly the curvatures and characterize the local curvature extrema (ridge points) from the first, second, and third derivatives of the gray level function. These partial derivatives can be computed using the operators of the edge detection. In the more general case where the contours are not isocontours (i.e., the gradient at an edge point only approximates the normal to the surface), the only differential invariants of the image are in R^4. This leads us to treat the 3D image as a hypersurface (a three-dimensional manifold) in R^4. We give the relationships between the curvatures of the hypersurface and the curvatures of the surface defined by edge points. The maximum curvature at a point on the hypersurface depends on the second partial derivatives of the 3D image. We note that it may be more efficient to smooth the data in R^4. Moreover, this approach could also be used to detect corners of vertices. We present experimental results obtained using real data (X ray scanner data) and applying these two methods. As an example of the stability, we extract ridge lines in two 3D X ray scanner data of a skull taken in different positions. |
In order to detect small-scale deformations during disease propagation while allowing large-scale deformation needed for inter-subject registration, we wish to model deformation at multiple scales and represent the deformation at the relevant scales only. With the LDDMM registration framework, enforcing sparsity results in compact representations but with limited ability to represent deformation across scales. In contrast, the LDDKBM extension of LDDMM allows representations of deformation at multiple scales but it does not favour compactness and hence may represent deformation at more scales than necessary. In this paper, we combine a sparsity prior with the multi-scale framework resulting in an algorithm allowing compact representation of deformation across scales. We present a mathematical formulation of the algorithm and evaluate it on a dataset of annotated lung CT images. |
Non-linear image registration is one of the most challenging task in medical image analysis. In this work, we propose an extension of the well-established diffeomorphic Demons registration algorithm to take into account geometric constraints. Combining the deformation field induced by the image and the geometry, we define a mathematically sound framework to jointly register images and geometric descriptors such as fibers or sulcal lines. We demonstrate this framework by registering simultaneously T 1 images and 50 fiber bundles consistently extracted in 12 subjects. Results show the improvement of fibers alignment while maintaining, and sometimes improving image registration. Further comparisons with non-linear T 1 and tensor registration demonstrate the superiority of the Geometric Demons over their purely iconic counterparts. |
In the LDDMM framework, optimal warps for image registration are found as end-points of critical paths for an energy functional, and the EPDiff equations describe the evolution along such paths. The Large Deformation Diffeomorphic Kernel Bundle Mapping (LDDKBM) extension of LDDMM allows scale space information to be automatically incorporated in registrations and promises to improve the standard framework in several aspects. We present the mathematical foundations of LDDKBM and derive the KB-EPDiff evolution equations, which provide optimal warps in this new framework. To illustrate the resulting diffeomorphism paths, we give examples showing the decoupled evolution across scales and how the method automatically incorporates deformation at appropriate scales. |
{P}ersonalized simulation for therapy planning in the clinical routine requires fast and accurate computations. {F}inite-element ({FE}) simulations belong to the most commonly used approaches. {B}ased on medical images the geometry of the patient's anatomy must be faithfully represented and discretized in a way to find a reasonable compromise between accuracy and speed. {T}his can be achieved by adapting the mesh resolution, and by providing well-shaped elements to improve the convergence of iterative solvers. {W}e present a pipeline for generating high-quality, adaptive meshes, and show how the framework can be applied to specific cardiac simulations. {O}ur aim is to analyze the meshing requirements for applications in electrophysiological modeling of ventricular tachycardia and electromechanical modeling of {T}etralogy of {F}allot. |
Erratum: In the version available from LNCS, the equation on page 3 should be $z( ilde{D}) = \left[ \left( ilde{D}_{Log} - \bar{D}_{Log} \right)^{ op} C^{-1} \left( ilde{D}_{Log} - \bar{D}_{Log} \right) \right]^{1/2}$. This has been corrected in the pdf available here. |
Focal cortical dysplasia (FCD), a malformation of cortical development, is an important cause of intractable epilepsy. On magnetic resonance images (MRI), FCD lesions are difficult to distinguish from healthy cortex and defining their spatial extent is challenging. We previously introduced a method to segment FCD lesions on MRI, relying on a 3D deformable model driven by MR features of FCD. In the present paper, we propose to improve our approach by adding a second evolution step which expands the result towards the cortical boundaries. A quantitative evaluation was performed in 18 FCD patients by comparison with manually traced lesion labels. The proposed approach achieved a strong agreement with the manual labels and substantially improved the results obtained with our previous method. |
In inter-subject registration, one often lacks a good model of the transformation variability to choose the optimal regularization. Some works attempt to model the variability in a statistical way, but the re-introduction in a registration algorithm is not easy. In [1], we interpreted the elastic energy as the distance of the Green-St Venant strain tensor to the identity. By changing the Euclidean metric for a more suitable Riemannian one, we defined a consistent statistical framework to quantify the amount of deformation. In particular, the mean and the covariance matrix of the strain tensor could be efficiently computed from a population of non-linear transformations and introduced as parameters in a Mahalanobis distance to measure the statistical deviation from the observed variability. This statistical Riemannian elasticity was able to handle anisotropic deformations but its isotropic stationary version was locally inverse-consistent. In this paper, we investigate how to modify the Riemannian elasticity to make it globally inverse consistent. This allows to define a left-invariant "distance" between shape diffeomorphisms that we call the left-invariant Riemannian elasticity. Such a closed form energy on diffeomorphisms can optimize it directly without relying on a time and memory consuming numerical optimization of the geodesic path. |
Focal cortical dysplasia (FCD), a malformation of cortical development, is an important cause of medically intractable epilepsy. FCD lesions are difficult to distinguish from non-lesional cortex and their delineation on MRI is a challenging task. This paper presents a method to segment FCD lesions on T1-weighted MRI, based on a 3D deformable model, implemented using the level set framework. The deformable model is driven by three MRI features: cortical thickness, relative intensity and gradient. These features correspond to the visual characteristics of FCD and allow to differentiate lesions from normal tissues. The proposed method was tested on 18 patients with FCD and its performance was quantitatively evaluated by comparison with the manual tracings of two trained raters. The validation showed that the similarity between the level set segmentation and the manual labels is similar to the agreement between the two human raters. This new approach may become a useful tool for the presurgical evaluation of patients with intractable epilepsy. |
Erratum: The tensor $W_j$ should be defined as $W_j = J^T.J$ and not as $W_j = J.J^T$. |
We present a complete and fully automatic method to compute a tomographic 4D representation of coronary arteries from one single rotational monoplane X-ray sequence. The major steps of our method are the following: (1) images filtering, (2) arteries segmentation, (3) arteries matching and reconstruction, (4) parametric deformation field computation, and (5) deformation-compensated tomographic reconstruction. Steps (2) and (3) involve only a few frames, acquired at the same cardiac cycle phase, while the steps (1), (4), and (5) use the frames acquired at all cardiac cycle phases. This method has been successfully applied to 4 patient data sets. The 4D representation allows the visualization of coronary arteries anatomy from any point of view, and at any cardiac cycle time. |
This article addresses the problem of histogram matching in the context of medical image processing. Such a problem occurs while comparing two images of the same object, where intensity differences are due to different acquisition conditions. This can be compensated by histogram matching or equalization. To achieve this, we based our method on windowing techniques. This allows to match implicitly continuous probability density functions, yielding more robust results than the methods issued from discrete histograms. |
The skeleton and its associated medial axis give a very compact representation of objects, even in the case of complex shapes and topologies. They are powerful shape descriptors, bridging the gap between low-level and highlevel object representations. Surprisingly, skeletons have been used in a relatively small number of applications. This work deals with the question of using the potential strength of the skeleton and the medial axis. From the medial axis, we build adequate attributed relational graphs to organize in a structured way informations about object shape and topology contained in the medial axis. This representation then permits to compare in a meaningful way various objects using a graph matching algorithm. Synthetic results are presented. |
Traditional Content-Based Image Retrieval (CBIR) systems only deliver visual outputs that are not directly interpretable by the physicians. Our objective is to provide a system for endomicroscopy video retrieval which delivers both visual and semantic outputs that are consistent with each other. In a previous study, we developed an adapted bag-of-visual-words method for endomicroscopy retrieval that computes a visual signature for each video. In this study, we first leverage semantic ground-truth data to transform these visual signatures into semantic signatures that reflect how much the presence of each semantic concept is expressed by the visual words describing the videos. Using cross-validation, we demonstrate that our visual-word-based semantic signatures enable a recall performance which is significantly higher than those of several state-of-the-art methods in CBIR. In a second step, we propose to improve retrieval relevance by learning, from a perceived similarity ground truth, an adjusted similarity distance. Our distance learning method allows to improve, with statistical significance, the correlation with the perceived similarity. Our resulting retrieval system is efficient in providing both visual and semantic information that are correlated with each other and clinically interpretable by the endoscopists. |
The study of cerebral micro-vascular network requires high resolution images. However, to obtain statistically relevant results, a large area of the brain (about few square millimeters) has to be investigated. This leads us to consider huge images, too large to be loaded and processed at once in the memory of a standard computer. To consider a large area, a compact representation of the vessels is required. The medial axis seems to be the tools of choice for the aimed application. To extract it, a dedicated skeletonization algorithm is proposed. Indeed, a skeleton must be homotopic, thin and medial with respect to the object it represents. Numerous approaches already exist which focus on computational efficiency. However, they all implicitly assume that the image can be completely processed in the computer memory, which is not realistic with the size of the data considered here. We present in this paper a skeletonization algorithm that processes data locally (in sub-images) while preserving global properties (i.e. homotopy). We then show some results obtained on a mosaic of 3-D images acquired by confocal microscopy. |
This report describes the calculation of local errors in Chamfer masks both in two- and in three-dimensional anisotropic spaces. For these errors, closed forms are given that can be related to the Chamfer mask geometry. Thanks to these calculation, it can be obsrved that the usual Chamfer masks ({\em i.e.} 3x3x3 or 5x5x5) have an inhomogeneously distributed error. Moreover, it allows us to design dedicated Chamfer masks by controlling either the complexity of the computation of the distance map (or equivalently the number of vectors in the mask), or the error of the mask in $\mathbb{Z}^2$ or in $\mathbb{Z}^3$. Last, since Chamfer distances are usually computed with integer weights (and approximate the Euclidean distance up to a multiplicative factor), we demonstrate that the knowledge of the local errors allows a very efficient computation of these weights. |
Chamfer distances are widely used in image analysis, and many ways have been investigated to compute optimal chamfer mask coefficients. Unfortunately, these methods are not systematized: they have to be conducted manually for every mask size or image anisotropy. Since image acquisistion (e.g. medical imaging) can lead to anisotropic discrete grids with unpredictable anisotropy value, automated calculation of chamfer mask coefficients becomes mandatory for efficient distance map computation. This report presentes a systematized calculation of these coefficients based on the automatic construction of chamfer masks of any size associated with a triangulation that allows to derive analytically the relative error with respect to the Euclidean distance, in any 3-D anisotropic lattice and that also allows to compute norm constraints. |
This report tackles the registration of 2D biological images (histological sections or autoradiographs) to 2D images from the same or different modalities (e.g., histology or MRI). The process of acquiring these images typically induces composite transformations that we model as a number of rigid or affine local transformations embedded in an elastic one. We propose a registration approach closely derived from this model. Given a pair of input images, we first compute a dense similarity field between them with a block matching algorithm. A hierarchical clustering algorithm then automatically partitions this field into a number of classes from which we extract independent pairs of sub-images. Our clustering algorithm relies on the Earth mover's distribution metric and is additionally guided by robust least-square estimation of the transformations associated with each cluster. Finally, the pairs of sub-images are, independently, affinely registered and a hybrid affine/non-linear interpolation scheme is used to compose the output registered image. We investigate the behavior of our approach under a variety of conditions, and discuss examples using simulated and real medical images, including MRI, autoradiography, histology and cryosection data. We also detail the reconstruction of a 3-D volume from a series of 2-D histological sections and compare it against a reconstruction obtained with a global rigid approach. |
The delineation of anatomical structures based on images of the lower abdomen in the frame of dose calculation for conformational radiotherapy is very complex to automatize. We present here the first results of a semi-automatic delineation of the bladder in tomodensitometric (CT) images. The method we have used is based on deformable templates whose deformation is guided by the image and by the user as well, in case the latter desires to correct the automatic delineation. In order to validate our approach, we use a set of CT images that have been segmented by medical experts. These hand-made contours act in fact as "ground truth", allowing for an objective evaluation of the performance of our algorithm. |
OBJECTIVE: Detection and follow-up of cognitive disorders in relapsing-remitting multiple sclerosis (RRMS) patients with spontaneous memory complain. BACKGROUND: Sixty per cent of patients with RRMS present cognitive deficiency. Evaluation and characterization of these troubles are difficult even if different standardized batteries exist, which tend to harmonization.Early diagnosis of cognitive impairment in the first years of the disease could allow early taking care and follow-up. Clinical correlations with lesion load on brain MRI could be useful to distinguish specific patients profiles. DESIGN/METHODS: Prospective study with baseline and 2 annual follow-up, of thirty patients, combining clinical examination, neuropsychological tests, brain MRI and Tc99m-HMPAO-SPECT. Neuropsychological tests used were: Mnesic capacities: Mini mental status, Brief Repeatable Battery: Selective Reminding Test with delayed recall and 10/36 Spatial recall test with delayed recall, Grober&Buschke test. Attentional capacities: Paced Auditory Serial Addition Test (PASAT) and verbal fluency, Stroop test A, B, C, Trailmaking test, Mattis Scale. Linguistic capacities: Boston naming test. Visuo-spatial capacities: WAIS-R (cubes). Visuo-constructive praxis: BEC. MRI protocol : T2 SE 2 echos (2mm 256*256) + T1 3D acquisition GRE (1mm 256*256) + FLAIR (4mm, 256*256) + T1gado SE (2mm 256*256). SPECT acquisition was performed one hour after injection of 925 MBq of 99mTc-HMPAO,using a triple head camera equipped with UHR fan-beam collimators. RESULTS: 8 men and 22 women were recruited, mean age: 39.97 years (+/- 8.4); mean EDSS: 2.2 (+/- 1, 49). Mean evolution duration: 120.4 months (+/- 96, 71). We found 8 established mnesic deficiency (27%), 20 established attentional deficiency (66%), 9 executive dysfunctions (30%), 6 language disability (20%), 18 visuospatial perception alteration (60%) and 6 (20%) visuoconstructive praxis alteration. Five patients (17%) had no objective troubles revealed by serial tests. On 27 analyzed MRI, 22 patients were Barkhof + ; Executive functions alteration was positievly correlated to gadolinium enhancement and negatively to infratentorial structures (IS) lesion load (LL). T1 and IS LL also correlated to mnesic deficiency. Simultaneously, SPECT data were visually analyzed after registration with MRI in FSE-T2 and FLAIR-T2 modalities. Local hypoperfusions were related to T2 hypersignals and T1 hyposignal in the same area. No significant difference in cognitive function was found between patient groups with normal or abnormal SPECT. CONCLUSIONS: Our results indicate that most of patient had cognitive disorders at early stage of their RR disease. On the baseline analysis, patients without objective cognitive impairment had T2 lesions in CC and IS. T2 lesion load was correlated with EDSS and time from diagnosis. Annual follow-up will allow us to develop comments. |
Contribution to chap. 3 (Medical Imaging and Image Processing) and chap. 4 (Modelling the Human Body for Therapy Planning and Computer Assisted Interventions) |
This material is presented to ensure timely dissemination of scholarly and technical work. Copyright and all rights therein are retained by authors or by other copyright holders. All person copying this information are expected to adhere to the terms and constraints invoked by each author's copyright. In most cases, these works may not be reposted without the explicit permission of the copyright holder.
Les documents contenus dans ces répertoires sont rendus disponibles par les auteurs qui y ont contribué en vue d'assurer la diffusion à temps de travaux savants et techniques sur une base non-commerciale. Les droits de copie et autres droits sont gardés par les auteurs et par les détenteurs du copyright, en dépit du fait qu'ils présentent ici leurs travaux sous forme électronique. Les personnes copiant ces informations doivent adhérer aux termes et contraintes couverts par le copyright de chaque auteur. Ces travaux ne peuvent pas être rendus disponibles ailleurs sans la permission explicite du détenteur du copyright.
Last modified: Sun Dec 22 00:30:04 2024This document was translated from BibTEX by bibtex2html