PLoS One. 2014 Oct 31;9(10):e111245. doi: 10.1371/journal.pone.0111245. eCollection 2014.
- 1Computational Science Research Center, San Diego State University, San Diego, California, United States of America.
- 2Electrical and Computer Eng. Dep/University of California San Diego, San Diego, California, United States of America.
- 3Department of Biology, San Diego State University, San Diego, California, United States of America.
- 4Department of Mathematics and Statistics, San Diego State University, San Diego, California, United States of America.
Abstract
Magnetic Resonance Imaging (MRI) and Computed Tomography (CT) scan are the two ubiquitous imaging sources that physicians use to diagnose patients with Cystic Fibrosis (CF) or any other Chronic Obstructive Pulmonary Disease (COPD). Unfortunately the cost constraints limit the frequent usage of these medical imaging procedures. In addition, even though both CT scan and MRI provide mesoscopic details of a lung, in order to obtain microscopic information a very high resolution is required. Neither MRI nor CT scans provide micro level information about the location of infection in a binary tree structure the binary tree structure of the human lung. In this paper we present an algorithm that enhances the current imaging results by providing estimated micro level information concerning the location of the infection. The estimate is based on a calculation of the distribution of possible mucus blockages consistent with available information using an offline Metropolis-Hastings algorithm in combination with a real-time interpolation scheme. When supplemented with growth rates for the pockets of mucus, the algorithm can also be used to estimate how lung functionality as manifested in spirometric tests will change in patients with CF or COPD.
No comments:
Post a Comment