|
The liver is the largest organ in the abdomen, and is prone to a large number of lesions and
benign and malignant primary and metastatic tumors.
Angiogenesis, the formation of new blood vessels, is a very important marker of the
aggressiveness of a tumor and its response to treatment. Thus the
assessment of tumors in their early stages
requires a quantitative evaluation of the blood supply to the liver. To achieve this
goal, it is important to develop precise pharmacokinetic approaches--describing how drugs
are absorbed, distributed, metabolized and eliminated--to the analysis
of the blood supply to and from the liver.
Dynamic Contrast Enhanced Magnetic Resonance Imaging (DCE-MRI) is extensively
used for the detection of tumors in the liver. To evaluate the blood supply to the liver,
the nature of the different tissues in the abdomen are characterized by analyzing
a series of these MRI images. A contrast substance is injected into the blood, and the
evolution of the contrast, viewed in the MRI images over a certain time span,
is used to characterize the tissues.
Different types of tissue define different coefficients for the blood flow rate, and these
coefficients are used to create parametric images such as the one shown.
In this parametric image the colors and intensities represent the different
values of the coefficients. To create the image, 12 dynamic MRI images were obtained,
co-registration was performed to deform and align the sequential images, the values
for each pixel were extracted for each time step, and the theoretical model for each pixel
was fitted. The process currently takes about 150 computing hours per patient, and we hope
to speed up the computation using grid computing.
The final objective of our work is to create a tool
to optimally select the parameters that describe the pharmacokinetic model of the
liver. This tool will use the Grid as a source of computing power and will offer a
simple and user-friendly interface. We have developed an application for the EGEE grid
infrastructure and have carried out a clinical trial on 20 patients. The computational cost
was reduced from 2623 to less than 20 hours using the grid.
Ignacio Blanquer, Vicente Hernández, José Carbonell, David Moratal,
and Montse Robles
Universidad Politécnica de Valencia, Valencia, Spain
Daniel Monleón and Bernardo Celda
Departamento de Química Física, Universitat de València, Valencia, Spain
Luis Martí-Bonmatí
Servicio de Radiología, Hospital Universitario Dr. Peset, Valencia, Spain
e-mail this article
|
