2007
A method for extracting the relevant MRI information from normoxic polymer gels exposed to low doses
R Martin, J BenComo, M Martin, J Bankson, M Heard, K Kaluarachchi, D Webb, N Wells, G Ibbott
Medical Physics, Vol. 34, No. 6, SU-FF-T-19, June 2007
Purpose: To extract the contribution due to the formation of polymer at low dose from the MRI multi-echo signal using a numerical method for Laplace transform inversion and to analyze the transverse relaxation rate spectrum Materials and Methods: Several samples of MAGIC gel (9% by weight of methacrylic acid) were prepared and irradiated with 60Co and 6 MV photon beams to doses in the range of 0.1 to 0.5 Gy. The samples were scanned in a 7 Tesla MRI Bruker BioSpec using a Carr-Purcell-Meiboom-Gill sequence. The signals were analyzed with the inversion algorithm. Samples of pure gelatin and a combination of gelatin and methacrylic acid were scanned in an MR spectrometer from Bruker. The samples were also scanned in an optical CT. Conclusions: Data analysis showed that polymer formation can be used as a better parameter than the average transverse relaxation rate for polymer gel dose calibration at low dose. The chemical shift spectrum showed that gelatin was not affected by radiation. The optical CT scan correlated well with the other results.
Fractal properties and critical exponents for tumor staging and classification
A. Quintana, M. Martín-Landrove, D. Pereira
III INTERNATIONAL CONGRESS ON COMPUTATIONAL BIOENGINEERING M. Cerrolaza, H. Rodrigues, M. Doublaré, J. Ambrosio, M. Viceconti (Eds.) Isla de Margarita, Venezuela, September 17 to 19, 2007
In general, tumors exhibit irregular borders with geometrical properties which are expected to depend upon their degree of malignancy. To appropriately evaluate these irregularities, it is necessary to apply segmentation procedures on the image to clearly define the active region of the tumor and its border. In the present work, T1 and T2 weighted magnetic resonance images of brain tumors were used to construct a nosologic map using in vivo magnetic resonance spectroscopy and biopsy for tissue classification and reference tumor staging. Different segmentation procedures were performed on the images, including gray level threshold and deformable contours (snakes). Several fractal properties were determined on the contour: the capacity or fractal dimension and the correlation dimension for different time series constructed from the original contour. Also a critical exponent coming from the contour roughness was calculated. The results obtained showed a good correlation between the roughness critical exponent and the degree of malignancy of the tumor.
Segmentation of brain tumor images using neural networks
Rodney Jaramillo, Marianela Lentini, Marco Paluszny
III INTERNATIONAL CONGRESS ON COMPUTATIONAL BIOENGINEERING M. Cerrolaza, H. Rodrigues, M. Doublaré, J. Ambrosio, M. Viceconti (Eds.) Isla de Margarita, Venezuela, September 17 to 19, 2007
Tumor image segmentation represents a crucial step not only in the diagnosis of the disease but also in its evaluation and monitoring of treatment. In the present work, neural networks are used to recognize the presence of tumoral tissue in brain on T2 and diffusion weighted magnetic resonance images. Relaxation and diffusion data were validated and categorized by means of in vivo spectroscopic data, used as a sort of virtual biopsy. For tumor classification and staging, all cases were correlated with histopathological results. Neural networks were trained with the set of validated data in a supervised mode, assuming at least three categories for the tissue: tumoral, normal or unaffected and liquid or necrosis. Segmentation performed in this way correlates closely with other methodologies previously developed, shortening drastically the processing time, what makes it very useful for its clinical application.
A quasi-analytical method for relaxation rate distribution determination of T2-weighted MRI in brain
Miguel Martín-Landrove, Giovanni Figueroa, Marco Paluszny, Wuilian Torres
Proceedings of the 29th Annual International Conference of the IEEE EMBS Cité Internationale, Lyon, France August 23-26, 2007
A quasi-analytical method for the determination of relaxation rate distribution functions in T2-weighted MRI in brain is proposed. The method solves analytically the set of non linear polynomial equations on the assumption that the transversal magnetization decay in Carr-Purcell-Meiboom-Gill (CPMG) T2-weighted MR brain images can be decomposed in a finite number of exponential decays, each one corresponding to a particular tissue class. The proposed method was validated by numerical simulations and applied to the calculation of relaxation rate distribution functions of tumoral lesions in brain.
Fractal analysis of tumoral lesions in brain
Miguel Martín-Landrove, Demian Pereira, María E. Caldeira, Salvador Itriago, María Juliac
Proceedings of the 29th Annual International Conference of the IEEE EMBS Cité Internationale, Lyon, France August 23-26, 2007
In this work, it is proposed a method for supervised characterization and classification of tumoral lesions in brain, based on the analysis of irregularities at the lesion contour on T2-weighted MR images. After the choice of a specific image, a segmentation procedure with a threshold selected from the histogram of intensity levels is applied to isolate the lesion, the contour is detected through the application of a gradient operator followed by a conversion to a “time series” using a chain code procedure. The correlation dimension is calculated and analyzed to discriminate between normal or malignant structures. The results found showed that it is possible to detect a differentiation between benign (cysts) and malignant (gliomas) lesions suggesting the potential of this method as a diagnostic tool.
Deformable models (snakes) for fractal analysis of brain tumors on T2-weighted images
A. Quintana, D. Pereira, M. Martín-Landrove
Proc. Intl. Soc. Mag. Reson. Med. 15 (2007)
Premature detection and classification for a disease process, is one of the most important objectives for a physician. In reference to the brain, MRI can detect abnormal structures from the early stages of development, showing the morphological characteristics of neoplasic tissue, specifically the presence of rugosity or irregularities on its boundaries. The main goal in this work is to employ the rugosity on the boundary of the lesion to classify a mass as belonging to any of the two kinds of lesions (benign or malign) present in T2-weighted MRI images of the brain. To reach this goal, tumor image is segmented and a deformable model (snake) is adapted to the boundary of the lesion. The energy density functional obtained from the deformable model is analyzed as an artificial time series to determine fractal correlation dimension. Also the fractal capacity dimension was calculated on the same model. To validate the proposed methodology, it was applied to a significant number of images with different types of lesions to establish a reliable classification method.
Brain tumor nosologic maps obtained from T2-weighted images
M. Martín-Landrove
Proc. Intl. Soc. Mag. Reson. Med. 15 (2007)
Tissue classification is a necessary step to obtain the spatial distribution of pathology, i.e., a nosologic map, and typically it is performed by the combination of different medical image modalities, sometimes including invasive histopathological studies. Different MRI techniques usually exhibit different spatial resolutions and as a consequence a partial volume problem is always present affecting considerably the precise determination of nosologic maps, i.e., in the case of T2-weighted or diffusion-weighted images, voxel intensity decays multiexponentially. Present work discusses different approaches to overcome the partial volume problem on T2-weighted images, by analysis of transversal relaxation rate distributions.
Application of a response scheme for collision handling among deformable objects
Bricelis J. Urbina, Omaira Rodríguez, Ernesto Coto
III INTERNATIONAL CONGRESS ON COMPUTATIONAL BIOENGINEERING, M. Cerrolaza, H. Rodrigues, M. Doblaré, J. Ambrosio, M. Viceconti, (Eds.), Isla de Margarita, Venezuela, September 17 to 19, 2007
Collision response schemes and deformable models have been recently an important subject of research, since many medical and entertainment applications require the simulation of real environments with deformable objects. Early research in this area comes from the field of Engineering, which propose very exact models but at a very high computational cost. New models propose plausible techniques which study the discrete nature of object representations in order to reduce the time required by these simulations. This paper presents the implementation of a very efficient collision response scheme which calculates the depth and direction of the penetration in the contact region, as well as the deformable model applied to the surface of the object. Furthermore, we present the application of the new collision detection scheme for the simulation of the interaction between a human abdominal organ and a surgery instrument.
Determinism, noise, and spurious estimations in a generalised
model of population growth
Harold P. de Vladar, Ido Pen
Physica A 373 (2007) 477–485
We study a generalised model of population growth in which the state variable is population growth rate instead of population size. Stochastic parametric perturbations, modelling phenotypic variability, lead to a Langevin system with two sources of multiplicative noise. The stationary probability distributions have two characteristic power-law scales. Numerical simulations show that noise suppresses the explosion of the growth rate which occurs in the deterministic counterpart. Instead, in different parameter regimes populations will grow with ‘‘anomalous’’ stochastic rates and (i) stabilise at ‘‘random carrying capacities’’, or (ii) go extinct in random times. Using logistic fits to reconstruct the simulated data, we find that even highly significant estimations do not recover or reflect information about the deterministic part of the process. Therefore, the logistic interpretation is not biologically meaningful. These results have implications for distinct model-aided calculations in biological situations because these kinds of estimations could lead to spurious conclusions.
Design of a Methodology for Assessing an Electrocardiographic Telemonitoring System
A. Alfonzo, M. K. Huerta, S. Wong, G. Passariello, M. Diaz, A. La Cruz, J. Cruz
Proceedings of the 29th Annual International Conference of the IEEE EMBS Cité Internationale, Lyon, France August 23-26, 2007
Recent studies in Bioengineering show a great interest in telemedicine projects, it is motivated mainly for the fast communication technologies reached during the last decade. Since then many telemedicine projects in different areas have been pursued, among them the electrocardiographic monitoring, as well as methodological reports for the evaluation of these projects. In this work a methodology to evaluate an electrocardiographic telemonitoring system is presented. A procedure to verify the operation of Data Acquisition Module (DAM) of an electrocardiographic telemonitoring system is given, taking as reference defined standards, and procedures for the measurement of the Quality of Service (QoS) parameters required by the system in a Local Area Network (LAN). Finally a graphical model and protocols of evaluation are proposed.
