Study of texture analysis on amyloid imaging in positron emission tomography

• Main Authors: Ming Je Tsai, 蔡佲哲
• Other Authors: I. T. Hsiao
• Format: Others
• Published: 2014
• Online Access: http://ndltd.ncl.edu.tw/handle/00802699727989343120

碩士 === 長庚大學 === 醫學影像暨放射科學系 === 102 === 18F-AV-45 is novel PET tracer for imaging the deposition of amyloid in Alzheimer's disease (AD). The amyloid PET SUVR images are widely applied for AD diagnosis. However, it is difficult to differentiate normal control and mild cognitive impairment significantly and has less ability of early diagnosis. This study utilized texture analysis which had been applied for different modality images in previous literature for the purpose of improving the shortage of SUVR images. We used phantom study to validate characteristics of feature calculation and evaluated the best texture properties of clinical amyloid imaging study. Aim of this study is to figure out the textural characteristic which can provide local information of the 18F-AV-45 labeled images between different severity trials and may be possibly provide more classification power of disease stages. The results demonstrated that the features obtained from texture analysis can provide local information of amyloid images, and enhanced the diagnosis differentiation ability in AD. As the results, texture analysis is an effective method for increasing extra information from image and aid the disease classification ability for medical images.