Sources of variability in quantification of CMR infarct size and their impact on sample size calculations

Mean infarct size varied between 16.8% and 27.2% of LV mass depending on the method.
Even AUTO (no user interaction for infarct borders) resulted in significant within-patient
variability given the need to delineate endocardial/epicardial contours (CV=10.6%).
Adding user input to correct computer generated infarct borders resulted in a mild
improvement in reproducibility (AUTO-UC: CV=8.3%; p=0.045 for comparison with AUTO).
For manual and visual categories, explicitly adjusting for intermediate signal-intensities
led to improved reproducibility (MANUAL-ISI vs MANUAL: CV=8.3% vs 14.4%; p=0.03; VISUAL-ISI
vs VISUAL: CV=8.4% vs 10.9%; p=0.01). When the best techniques in each category were
compared, reproducibility was similar (AUTO-UC, MANUAL-ISI, and VISUAL-ISI: CV=8.3%,
8.3%, 8.4%, respectively). For these 3 techniques the within-patient variability due
to the quantification method was less than 10% of the total variability. Hence, there
were minimal differences between these methods in the calculated sample sizes needed
to detect a 3%, 5%, and 7% absolute reduction in acute infarct size.