演講者:Feng-Nan Hwang (黃楓南)(Department of Mathematics, National Central University)
演講者: Feng-Nan Hwang (黃楓南) (Department of Mathematics, National Central University)
標題:A dynamic contrast-enhanced MRI-based data-driven computational
technique for early detection of chronic liver diseases
時間:11月17日下午01點20分
地點:校本部第二綜合大樓8樓 B808 教室
摘要: Liver diseases are always on the list of the top ten causes of death
in Asian countries. Generally speaking, liver disease progression can
be classified into three stages: liver fibrosis, liver cirrhosis, and
liver cancer. One of the research focus for clinical practice is to
develop some noninvasive technique used for determining the status of
the liver disease. Early diagnosis of the liver's fibrosis with some
proper treatments can decrease the hepatocellular carcinoma chance. In
achieving the goal, we propose a data-driven computational technique
in conjunction with the dynamic contrast-enhanced MRI (DCI-MRI), which
has been shown promising for the early detection of chronic liver
diseases. The proposed technique's kernel is a Darcy solver weakly
coupled with an unsteady convection-diffusion solver used to simulate
the blood flows through the liver, assumedly as a kind of porous
medium, and the relative signal enhancement scanned by MRI varied in
time. Our approach consists of two phases: the online and online
stages. We correlate the values of the porosity in the mathematical
model to the degree of liver fibrosis during the online phase,
determined by the liver biopsy result using the clinical data. During
the online stage, to help the doctors' diagnosis, we perform the
numerical simulation by using the patient-specific data to determine
the liver's fibrosis stage. Our data-driven DCE-MRI computational
technique achieves a 93% success rate for diagnosing moderate liver
fibrosis status from the mild one correctly based on the real clinic
data.