Fig. 8

a Image intensity of normal and infarcted myocardium as a function of TI for magnitude reconstruction LGE. Simulations assumed a T1 of normal myocardium is 550 ms and infarcted myocardium is 350 ms, and blood is 300 ms. b Difference in image intensities between infarcted and normal myocardium as function of inversion time for magnitude reconstruction LGE. The time at which the magnetization of normal myocardium crosses zero is defined as the TI to null normal myocardium. This TI is considered ‘optimal’ because the image intensity difference between infarcted and normal myocardium is maximized (green box and line). Selection of a TI that is shorter than the optimal time (left of the green line) is problematic because the image intensity difference between infarct and normal myocardium can be significantly reduced (or possibly eliminated), and the relationship with the selected TI time is not linear. An indication that the TI is set ~ 20–50 ms too short is the presence of ‘etching’ (red box) which occurs at the interface between the blood pool and subendocardium due to the partial volume effect. The voxels situated at this interface include both myocardium (which is below the zero crossing) and blood pool (which is above the zero crossing), leading to a net magnetization of zero; these voxels appear black. The adjacent mid-myocardial voxels (below the zero crossing) and blood pool only voxels (above the zero crossing) are bright. Because of this artifact, we recommend that operators systematically select a TI that is slightly longer than optimal (~ 30 ms). The image intensity difference between infarct and normal myocardium with these slightly longer TIs (right of the green line) is predictable and stable. TI: inversion time