Infarct size is a major surrogate marker for prognosis in the context of myocardial infarction. There is a growing interest in validating a quantitative assessment approach in order to: (1) standardize these analyses; (2) to precise the individual prognosis of our patients. Several methods are available and were tested across their capacity to predict left ventricular (LV) remodeling at three months.

Late gadolinium enhancement-MRI was performed on day 5 and after a period of three months in 92 patients with STEMI. LV volumes and scar parameters were assessed visually (by using a four scale score) and quantitatively on day 5 and at three months. Dichotomous thresholds were defined first visually (VISUAL), then by 2, 5 and 6 standard deviations above remote myocardium, and by the full-width at half-maximum (FWHM) method.

All infarct sizing methods showed great relation to LV remodeling at three months (ROC analysis). Univariate predictors of an LV end-systolic volume index (LVESVi) superior to 70mL/m² were: heart failure, creatin kinase peak and infarct size at day 5. FWHM was shown to be the best of all quantitative methods. An infarct size superior to 44 grams predicted a LVESVi>70mL/m² with a sensitivity of 90% and a specificity of 92.5%. FWHM reproducibility was good (r=0.895, P<0.0001, Bland Altman bias of 0.8g).

In the context of STEMI, FWHM is a tough and reproducible algorithm to quantitatively assess late gadolinium hyperenhancement, greatly related to functional prognosis at three months follow-up.