Model-based noninvasive estimation of arterial input function from dynamic H215O PET images

N. Kudomi, H. Watabe, K. M. Kim, T. Hayashi, K. Hayashida, H. Iida

Research output: Contribution to conferencePaper

Abstract

For the application of a kinetic model to PET data, it is necessary to obtain the arterial input function. Since arterial blood sampling is invasive and labor intensive, a method to evaluate the input function without arterial blood sampling is important. In case of the [15O]water injection study, because of the noise in the time activity curve obtained from PET reconstructed image round carotid artery region, a model based evaluation method might be promised. In this study, a model function is composed of combination of blood time activity curve function and tissue time activity curve function which is generated from blood activity curve function. Present method was applied to 15 of dynamic PET scans of 15O water (10 normal subjects, 10 : Rest, and 5 : Diamox) for normal volunteers. The time activity curve obtained was fitted with the above model function. The shapes of input function obtained was in agreement with that by arterial blood sampling. The accuracy of height of the peak was 14 % and the accuracy of area under the curve was 10 %. This study supposed that present method could be used for direct extraction of carotid artery input function from dynamic PET image, and could be used in the quantitative CBF study using 15O water PET.

Original languageEnglish
Pages1559-1561
Number of pages3
Publication statusPublished - 2002 Dec 1
Externally publishedYes
Event2002 IEEE Nuclear Science Symposium Conference Record - Norfolk, VA, United States
Duration: 2002 Nov 102002 Nov 16

Other

Other2002 IEEE Nuclear Science Symposium Conference Record
CountryUnited States
CityNorfolk, VA
Period02/11/1002/11/16

ASJC Scopus subject areas

  • Computer Vision and Pattern Recognition
  • Industrial and Manufacturing Engineering

Fingerprint Dive into the research topics of 'Model-based noninvasive estimation of arterial input function from dynamic H<sub>2</sub><sup>15</sup>O PET images'. Together they form a unique fingerprint.

  • Cite this

    Kudomi, N., Watabe, H., Kim, K. M., Hayashi, T., Hayashida, K., & Iida, H. (2002). Model-based noninvasive estimation of arterial input function from dynamic H215O PET images. 1559-1561. Paper presented at 2002 IEEE Nuclear Science Symposium Conference Record, Norfolk, VA, United States.