MR tagging

Last revised by Joachim Feger on 11 Dec 2021

Cardiac MR tagging or myocardial tagging refers to a MRI based acquisition method designed for myocardial deformation analysis.

The method exploits tissue magnetization as a tissue property. A local magnetic saturation grid of dark lined tissue markers known as tags are induced onto the myocardium with radiofrequency pulses applied in orthogonal planes. The tag lines deform with the myocardial tissue during cardiac contraction and consequently the ‘tagging’ grid can be used to track and assess the displacement of the myocardial tissue as a basis for calculating cardiac strain 1-3.

Deformation assessment can be conducted in the following ways 2,3:

  • image preparation
  • detection of endocardial and epicardial borders
  • definition of the segments to be tracked
  • tag tracking
    • tracking of dark lines or intersections
    • harmonic phase analysis (HARP)
    • other optical flow techniques
  • motion reconstruction

MR tagging is considered and widely accepted as a reference standard in the assessment of regional myocardial strain imaging and has been used in large clinical trials e.g. the MESA trial 1,2.

However due to the requirement of a dedicated acquisition sequence, additional acquisition time and time-consuming post-processing algorithms MR tagging has been less widely used than other recent methods.

Advantages of MR tagging are 1:

  • direct measurement by physical tissue properties
  • extensive in vitro and in vivo validation

Disadvantages of MR tagging include 1,2:

  • low temporal resolution (20-30 frames / heart beat)
  • delay of tag deposition at the beginning of systole potentially leading to an underestimation
  • low accuracy at the endocardial border and in thin walled regions
  • tags fade through cycle

Cardiac MR tagging was developed in the late 1980s 5,6.

ADVERTISEMENT: Supporters see fewer/no ads

Updating… Please wait.

 Unable to process the form. Check for errors and try again.

 Thank you for updating your details.