CME 1 Session - EANM'17

Physics: Challenges and Solutions for MR-Based Attenuation Correction of PET

Educational Objectives
  1. Learn  about  the  basic  methodology  of the  generation  of  maps  of  linear  attenuation coefficients (LAC, i.e. µ-maps) using MR signals for attenuation corrections of the PET signal (i.e. 511keV gamma rays).
  2. Gain insight in new developments and advancement of methods to improve attenuation correction using structural MRI data
  3. Understand atlas based and neuronal network based generation of CT-like µ-maps for attenuation correction in brain PET/MRI
  4. Learn about new methods of hard and software implementation to generate whole µ-maps for attenuation correction in PET/MRI
Other than using CT, MR structural imaging reveals a proton density of the imaged object. Thisis by no means representative for the attenuation of gamma rays traveling through thematerial and, hence, cannot be used to directly derive a map of the electron density or linearattenuation coefficients (LACs, µ-maps). Beginning with very basic segmentation algorithmsand the assignment of static LACs to classes of tissue, continuing with atlas based methods toderive more CT-like µ-maps up to neuronal networks that are trained to derive continuousvalued µ-maps from structural MRI data and involving more sophisticated MR imagingsequences to better detect bone, this session gives an overview of the methodology eitherwidely implemented or available for use in clinical and research settings. Methodology of MRbased attenuation correction is – at least for the brain – no longer a challenge. The wideimplementation and availability seems to remain the challenge.

Key Words:
PET/MRI, MR-based attenuation correction, tissue class Segmentation, µ-map
User action required
If you have the permission to view this publication, please log in.