Catálogo de publicaciones - libros

A novel method for vertebral fracture quantification from X-ray images is presented. Using pairwise conditional shape models trained on a set of healthy spines, the most likely normal vertebra shapes are estimated conditional on all other vertebrae in the image. The differences between the true shape and the reconstructed normal shape is subsequently used as a measure of abnormality. In contrast with the current (semi-)quantitative grading strategies this method takes the full shape into account, it uses a patient-specific reference by combining population-based information on biological variation in vertebra shape and vertebra interrelations, and it provides a continuous measure of deformity.

The method is demonstrated on 212 lateral spine radiographs with in total 78 fractures. The distance between prediction and true shape is 1.0 mm for unfractured vertebrae and 3.7 mm for fractures, which makes it possible to diagnose and assess the severity of a fracture.

We present a new approach for cranial implant design which uses anatomical constrained deformation based on reference models. The methodological framework contains three steps: patient-specific generation of the reference model containing the anatomical constraints about the skull shape; determination of the spatial correspondence between the patient skull and the reference model by 3D matching; adaptive deformation of the fragment on the reference model corresponding to the defect area on the patient skull for implant design. The proposed method was validated by simulating the reconstruction of artificially generated defects on healthy skulls. The validation results show that this approach can generate implant geometry very fast and with satisfactory quality. This approach also outperforms the surface interpolation method in reconstructing cranial defects.

Shape correspondence is the foundation for accurate statistical shape analysis; this is usually accomplished by identifying a set of sparsely sampled and well-corresponded landmark points across a population of shape instances. However, most available shape correspondence methods can only effectively deal with complete-shape correspondence, where a one-to-one mapping is assumed between any two shape instances. In this paper, we present a novel algorithm to correspond 2D open-curve partial-shape instances where one shape instance may only be mapped to part of the other, i.e., the endpoints of these open-curve shape instances are not presumably corresponded. In this algorithm, some initially identified landmarks, including the ones at or near the endpoints of the shape instances, are refined by allowing them to slide freely along the shape contour to minimize the shape-correspondence error. To avoid being trapped into local optima, we develop a simple method to construct a better initialization of the landmarks and introduce some additional constraints to the landmark sliding. We evaluate the proposed algorithm on 32 femur shape instances in comparison to some current methods.

Reconstruction of patient-specific 3D bone surface from 2D calibrated fluoroscopic images and a point distribution model is discussed. We present a 2D/3D reconstruction scheme combining statistical extrapolation and regularized shape deformation with an iterative image-to-model correspondence establishing algorithm, and show its application to reconstruct the surface of proximal femur. The image-to-model correspondence is established using a non-rigid 2D point matching process, which iteratively uses a symmetric injective nearest-neighbor mapping operator and 2D thin-plate splines based deformation to find a fraction of best matched 2D point pairs between features detected from the fluoroscopic images and those extracted from the 3D model. The obtained 2D point pairs are then used to set up a set of 3D point pairs such that we turn a 2D/3D reconstruction problem to a 3D/3D one. We designed and conducted experiments on 11 cadaveric femurs to validate the present reconstruction scheme. An average mean reconstruction error of 1.2 mm was found when two fluoroscopic images were used for each bone. It decreased to 1.0 mm when three fluoroscopic images were used.

This paper presents results of a pilot study evaluating the efficacy of robotic assistance using novel steerable electrode arrays for cochlear implant surgery. The current surgical setup of cochlear implant surgery is briefly reviewed and its limitations are highlighted. In an effort to reduce trauma to the structure of the cochlea, the kinematics and path planning for novel cochlear steerable electrodes are developed to minimize the interaction forces between the electrode and the cochlea. An experimental robotic system is used to compare the electrode insertion forces of steerable implants with those of non-steerable electrodes. The results of these experiments show about 70% reduction in the insertion forces when steerable electrodes are used with our proposed path planning and control. A distance metric explaining this reduction in the insertion force is defined and experimentally validated. Although this is only a preliminary study, we believe that these results provide a strong indication to the potential of robot-assisted cochlear implant surgery to provide a significant reduction in trauma rates during cochlear implant surgery.

In contemporary brachytherapy procedures, needle placement at the desired target is challenging due to a variety of reasons. A robot-assisted brachytherapy system can improve the needle placement and seed delivery resulting in enhanced patient care. In this paper we present a 16 DOF (degrees-of-freedom) robotic system (9DOF positioning module and 7DOF surgery module) developed and fabricated for prostate brachytherapy. Techniques to reduce needle deflection and target movement have been incorporated after verifying with extensive experiments. Provisions for needle motion and force feedback have been included into the system for improving the robot control and seed delivery. Preliminary experimental results reveal that the prototype system is quite accurate (sub-millimeter) in placing brachytherapy needles.

We developed an image-guided robot system to achieve highly accurate placement of thin needles and probes into in-vivo rodent tumor tissue in a predefined pattern that is specified on a preoperative image. This system can be used for many experimental procedures where the goal is to correlate a set of physical measurements with a corresponding set of image intensities or, more generally, to perform a physical action at a set of anatomic points identified on a preoperative image. This paper focuses on the design and validation of the robot system, where the first application is to insert oxygen measurement probes in a three-dimensional (3D) grid pattern defined with respect to a PET scan of a tumor. The design is compatible with CT and MRI, which we plan to use to identify targets for biopsy and for the injection of adenoviral sequences for gene therapy. The validation is performed using a phantom and includes a new method for estimating the Fiducial Localization Error (FLE) based on the measured Fiducial Distance Error (FDE).

Real-time 3D ultrasound can enable new image-guided surgical procedures, but high data rates prohibit the use of traditional tracking techniques. We present a new method based on the modified Radon transform that identifies the axis of instrument shafts as bright patterns in planar projections. Instrument rotation and tip location are then determined using fiducial markers. These techniques are amenable to rapid execution on the current generation of personal computer graphics processor units (GPU). Our GPU implementation detected a surgical instrument in 31 ms, sufficient for real-time tracking at the 26 volumes per second rate of the ultrasound machine. A water tank experiment found instrument tip position errors of less than 0.2 mm, and an study tracked an instrument inside a beating porcine heart. The tracking results showed good correspondence to the actual movements of the instrument.

This paper presents a novel active surface segmentation algorithm using a multiscale shape representation and prior. We define a parametric model of a surface using spherical wavelet functions and learn a prior probability distribution over the wavelet coefficients to model shape variations at different scales and spatial locations in a training set. Based on this representation, we derive a parametric active surface evolution using the multiscale prior coefficients as parameters for our optimization procedure to naturally include the prior in the segmentation framework. Additionally, the optimization method can be applied in a coarse-to-fine manner. We apply our algorithm to the segmentation of brain caudate nucleus, of interest in the study of schizophrenia. Our validation shows our algorithm is computationally efficient and outperforms the Active Shape Model algorithm by capturing finer shape details.

Due to small training sets, statistical shape models constrain often too much the deformation in medical image segmentation. Hence, an artificial enlargement of the training set has been proposed as a solution for the problem. In this paper, the error sources in the statistical shape model based segmentation were analyzed and the optimization processes were improved. The method was evaluated with 3D cardiac MR volume data. The enlargement method based on non-rigid movement produced good results – with 250 artificial modes, the average error for four-chamber model was 2.11 mm when evaluated using 25 subjects.