Search

Predictive modeling of blood flow and pressure have numerous applications ranging from non-invasive assessment of functional significance of disease to planning invasive procedures. While several such predictive modeling techniques have been proposed, their use in the clinic has been limited due in part to the significant time required to perform virtual interventions and compute the resultant changes in hemodynamic conditions. We propose a fast hemodynamic assessment method to aid in interventional planning based on first constructing an exploration space of geometries, tailored to each patient, and subsequently building a physics driven reduced order model in this space. We demonstrate that this method can predict fractional flow reserve derived from coronary computed tomography angiography in response to changes to a patient-specific lumen geometry in real time while achieving high accuracy when compared to computational fluid dynamics simulations. We validated this method on over 1300 patients that received a coronary CT scan and demonstrated a correlation coefficient of 0.98 with an error of 0 . 005 ± 0 . 015 (95% CI: (-0.020, 0.031)) as compared to three-dimensional blood flow calculations. This technology is implemented in a product that has received clearance by the U.S. Food and Drug Administration and is being used clinically to enable physicians to predict changes in blood flow resulting from removal of coronary stenoses as might occur with percutaneous coronary interventions. This technology is also cleared for use in Japan and pending regulatory approval in Europe.

This study evaluated the predictive role of 1D, 2D and 3D quantitative, enhancement-based MRI regarding overall survival (OS) in patients with colorectal liver metastases (CLM) following intra-arterial therapies (IAT).This retrospective analysis included 29 patients who underwent transarterial chemoembolization (TACE) or radioembolization and received MRI within 6 weeks after therapy. Tumour response was assessed using 1D and 2D criteria (such as European Association for the Study of the Liver guidelines [EASL] and modified Response Evaluation Criteria in Solid Tumors [mRECIST]). In addition, a segmentation-based 3D quantification of overall (volumetric [v] RECIST) and enhancing lesion volume (quantitative [q] EASL) was performed on portal venous phase MRI. Accordingly, patients were classified as responders (R) and non-responders (NR). Survival was evaluated using Kaplan-Meier analysis and compared using Cox proportional hazard ratios (HR).Only enhancement-based criteria identified patients as responders. EASL and mRECIST did not predict patient survival (P = 0.27 and P = 0.44, respectively). Using uni- and multivariate analysis, qEASL was identified as the sole predictor of patient survival (9.9 months for R, 6.9 months for NR; P = 0.038; HR 0.4).The ability of qEASL to predict survival early after IAT provides evidence for potential advantages of 3D quantitative tumour analysis.• Volumetric assessment of colorectal liver metastases after intra-arterial therapy is feasible. • Early 3D quantitative tumour analysis after intra-arterial therapy may predict patient survival. • Volumetric tumour response assessment shows advantages over 1D and 2D techniques. • Enhancement-based MR response assessment is preferable to size-based measurements.

Design of a geometric vascular model.Non-parametric Bayesian model, learned by kernel density estimation from manually segmented datasets.Design of a new sampling scheme, Adaptive Auxiliary Particle Filtering (AAPF).Mean-Shift clustering for bifurcation detection and coronary tree extraction, and high computational efficiency.Experiments demonstrate the robustness of the proposed approach for complete vessel tree segmentation. Considering vessel segmentation as an iterative tracking process, we propose a new Bayesian tracking algorithm based on particle filters for the delineation of coronary arteries from 3D computed tomography angiograms. It relies on a medial-based geometric model, learned by kernel density estimation, and on a simple, fast and discriminative flux-based image feature. Combining a new sampling scheme and a mean-shift clustering for bifurcation detection and result extraction leads to an efficient and robust method. Results on a database of 61 volumes demonstrate the effectiveness of the proposed approach, with an overall Dice coefficient of 86.2% (and 92.5% on clinically relevant vessels), and a good accuracy of centerline position and radius estimation (errors below the image resolution).

We propose a fast, automatic and versatile framework for the segmentation of multiple anatomical structures from 2D and 3D images. We extend the work of [1] on implicit template deformation to multiple targets. Our variational formulation optimizes the non-rigid transformation of a set of templates according to image-driven forces. It embeds non-overlapping constraints ensuring a consistent segmentation result. We demonstrate the potential of our approach on the segmentation of abdominal organs (liver, kidneys, spleen and gallbladder) with an evaluation on CT volumes (50 for training and 50 for testing). Our method reaches state-of-the-art accuracy, ranging from 2mm (liver and kidneys) to 8mm (gallbladder).

To evaluate the diagnostic performance of three-dimensional ( 3D three-dimensional ) quantitative enhancement-based and diffusion-weighted volumetric magnetic resonance (MR) imaging assessment of hepatocellular carcinoma ( HCC hepatocellular carcinoma ) lesions in determining the extent of pathologic tumor necrosis after transarterial chemoembolization ( TACE transarterial chemoembolization ).This institutional review board-approved retrospective study included 17 patients with HCC hepatocellular carcinoma who underwent TACE transarterial chemoembolization before surgery. Semiautomatic 3D three-dimensional volumetric segmentation of target lesions was performed at the last MR examination before orthotopic liver transplantation or surgical resection. The amount of necrotic tumor tissue on contrast material-enhanced arterial phase MR images and the amount of diffusion-restricted tumor tissue on apparent diffusion coefficient ( ADC apparent diffusion coefficient ) maps were expressed as a percentage of the total tumor volume. Visual assessment of the extent of tumor necrosis and tumor response according to European Association for the Study of the Liver ( EASL European Association for the Study of the Liver ) criteria was performed. Pathologic tumor necrosis was quantified by using slide-by-slide segmentation. Correlation analysis was performed to evaluate the predictive values of the radiologic techniques.At histopathologic examination, the mean percentage of tumor necrosis was 70% (range, 10%-100%). Both 3D three-dimensional quantitative techniques demonstrated a strong correlation with tumor necrosis at pathologic examination (R(2) = 0.9657 and R(2) = 0.9662 for quantitative EASL European Association for the Study of the Liver and quantitative ADC apparent diffusion coefficient , respectively) and a strong intermethod agreement (R(2) = 0.9585). Both methods showed a significantly lower discrepancy with pathologically measured necrosis (residual standard error [ RSE residual standard error ] = 6.38 and 6.33 for quantitative EASL European Association for the Study of the Liver and quantitative ADC apparent diffusion coefficient , respectively), when compared with non- 3D three-dimensional techniques ( RSE residual standard error = 12.18 for visual assessment).This radiologic-pathologic correlation study demonstrates the diagnostic accuracy of 3D three-dimensional quantitative MR imaging techniques in identifying pathologically measured tumor necrosis in HCC hepatocellular carcinoma lesions treated with TACE transarterial chemoembolization .

Arterial spin labeling (ASL) perfusion measurements allow the follow-up of muscle perfusion with high temporal resolution during a stress test. Automated image processing is proposed to estimate perfusion maps from ASL images. It is based on two successive analyses: at first, automated rejection of the image pairs between which a large displacement is detected is performed, followed by factor analysis of the dynamic data and cluster analysis to classify pixels with large signal variation characteristic of vessels. Then, after masking these "vascular" pixels, factor analysis and cluster analysis are further applied to separate the different muscles between low or high perfusion increase, yielding a functional map of the leg. Data from 10 subjects (five normal volunteers and five elite sportsmen) had been analyzed. Resulting time perfusion curves from a region of interest (ROI) in active muscles show a good accordance whether extracted with automated processing or with manual processing. This method of functional segmentation allows automated suppression of vessels and fast visualization of muscles with high, medium or low perfusion, without any a priori knowledge.

The goal of this paper is to present the interest to define a geometric declarative modeler according to the needs of the associated solver. The modeler presented will seem, in a first view, of a complex architecture. A more detailed description of the modeler allows the presentation of the geometric constraints management. The interest of such approach is finally illustrated by the solver presentation. This solver, divided in three modules, allows a reliable and fast resolution of the geometric problems. Indeed, each module has been defined to optimize the resolution in terms of time and robustness.

were still experiencing menses and all reported significantly less menstrual bleeding. 12 women (60%) reported pre-procedural bulk symptoms and 10/12 (83%) described these as improved. 16/20 (60%) had post procedural MRI with an average decrease in uterine volume of 39%. 17/20 (85%) women were satisfied with the procedure and would undergo UAE again. Complications included one patient who developed tuboovarian abscesses and experienced large volume fibroid expulsion requiring drainage of abscesses and a vaginal myomectomy. Small volume fibroid expulsion not requiring further treatment occurred in 2 women (10%) and 3 women (15%) experienced post-procedure pain requiring treatment with IV narcotics in the emergency room. Conclusion: UAE outcomes in patients with megauterus are comparable to prior studies. Megauterus alone should not be a contraindication to UAE and may not be a key factor in predicting successful outcome of UAE.

International audience; Liver segmentation in 3D CT images is a fundamental step for surgery planning and follow-up. Robustness, automation and speed are required to fulll this task efficiently. We propose a fully-automatic workflow for liver segmentation built on state-of-the-art algorithmic components to meet these requirements. The liver is first localized using regression forests. A liver probability map is computed, followed by a global-to-local segmentation strategy using a template deformation framework. We evaluate our method on the SLIVER07 reference database and confirm its state-of-the-art results on a large, varied database of 268 CT volumes. This extensive validation demonstrates the robustness of our approach to variable fields of view, liver contrast, shape and pathologies. Our framework is an attractive tradeoff between robustness, accuracy (mean distance to ground truth of 1.7mm) and computational speed (46s). We also emphasize the genericity and relative simplicity of our framework, which requires very limited liver-specic tuning.

International audience; We propose a method for fast, accurate and robust localization of several organs in medical images. We generalize global-to-localcascades of regression forests [1] to multiple organs. A first regressor encodes global relationships between organs. Subsequent regressors refinethe localization of each organ locally and independently for improvedaccuracy. We introduce confidence maps, which incorporate informationabout both the regression vote distribution and the organ shape throughprobabilistic atlases. They are used within the cascade itself, to betterselect the test voxels for the second set of regressors, and to provide richerinformation than the classical bounding boxes thanks to the shape prior.We demonstrate the robustness and accuracy of our approach through aquantitative evaluation on a large database of 130 CT volumes.

With the ubiquity of imaging in medical applications (diagnostic, treatment follow-up, surgery planning. . . ), image processing algorithms have become of primary importance. Algorithms help clinicians extract critical information more quickly and more reliably from increasingly large and complex acquisitions. In this context, anatomy localization and segmentation is a crucial component in modern clinical workflows. Due to particularly high requirements in terms of robustness, accuracy and speed, designing such tools remains a challengingtask.In this work, we propose a complete pipeline for the segmentation of multiple organs in medical images. The method is generic, it can be applied to varying numbers of organs, on different imaging modalities. Our approach consists of three components: (i) an automatic localization algorithm, (ii) an automatic segmentation algorithm, (iii) a framework for interactive corrections. We present these components as a coherent processing chain, although each block could easily be used independently of the others. To fulfill clinical requirements, we focus on robust and efficient solutions. Our anatomy localization method is based on a cascade of Random Regression Forests (Cuingnet et al., 2012). One key originality of our work is the use of shape priors for each organ (thanks to probabilistic atlases). Combined with the evaluation of the trained regression forests, they result in shape-consistent confidence maps for each organ instead of simple bounding boxes. Our segmentation method extends the implicit template deformation framework of Mory et al. (2012) to multiple organs. The proposed formulation builds on the versatility of the original approach and introduces new non-overlapping constraintsand contrast-invariant forces. This makes our approach a fully automatic, robust and efficient method for the coherent segmentation of multiple structures. In the case of imperfect segmentation results, it is crucial to enable clinicians to correct them easily. We show that our automatic segmentation framework can be extended with simple user-driven constraints to allow for intuitive interactive corrections. We believe that this final component is key towards the applicability of our pipeline in actual clinical routine.Each of our algorithmic components has been evaluated on large clinical databases. We illustrate their use on CT, MRI and US data and present a user study gathering the feedback of medical imaging experts. The results demonstrate the interest in our method and its potential for clinical use., Avec l’utilisation de plus en plus répandue de l’imagerie dans la pratique médicale (diagnostic, suivi, planification d’intervention, etc.), le développement d’algorithmes d’analyse d’images est devenu primordial. Ces algorithmes permettent aux cliniciens d’analyser et d’interpréter plus facilement et plus rapidement des données de plus en plus complexes. Dans ce contexte, la localisation et la segmentation de structures anatomiques sont devenues des composants critiques dans les processus cliniques modernes. La conception de tels outils pour répondre aux exigences de robustesse, précision et rapidité demeure cependant un réel défi technique.Ce travail propose une méthode complète pour la segmentation de plusieurs organes dans des images médicales. Cette méthode, générique et pouvant être appliquée à un nombre varié de structures et dans différentes modalités d’imagerie, est constituée de trois composants : (i) un algorithme de localisation automatique, (ii) un algorithme de segmentation, (iii) un outil de correction interactive. Ces différentes parties peuvent s’enchaîner aisément pour former un outil complet et cohérent, mais peuvent aussi bien être utilisées indépendemment. L’accent a été mis sur des méthodes robustes et efficaces afin de répondre aux exigences cliniques. Notre méthode de localisation s’appuie sur une cascade de régression par forêts aléatoires (Cuingnet et al., 2012). Elle introduit l’utilisation d’informations a priori de forme, spécifiques à chaque organe (grâce à des atlas probabilistes) pour des résultats plus cohérents avec la réalité anatomique. Notre méthode de segmentation étend la méthode de segmentation par modèle implicite (Mory et al., 2012) à plusieurs modèles. La formulation proposée permet d’obtenir des déformations cohérentes, notamment en introduisant des contraintes de non recouvrement entre les modèles déformés. En s’appuyant sur des forces images polyvalentes, l’approche proposée se montre robuste et performante pour la segmentation de multiples structures. Toute méthode automatique n’est cependant jamais parfaite. Afin que le clinicien garde la main sur le résultat final, nous proposons d’enrichir la formulation précédente avec des contraintes fournies par l’utilisateur. Une optimisation localisée permet d’obtenir un outil facile à utiliser et au comportement intuitif. Ce dernier composant est crucial pour que notre outil soit réellement utilisable en pratique. Chacun de ces trois composants a été évalué sur plusieurs grandes bases de données cliniques (en tomodensitométrie, imagerie par résonance magnétique et ultrasons). Une étude avec des utilisateurs nous a aussi permis de recueillir des retours positifs de plusieurs experts en imagerie médicale. Les différents résultats présentés dans ce manuscrit montrent l’intérêt de notre méthode et son potentiel pour une utilisation clinique.