The purpose of this study was to evaluate how definition of multiple sclerosis lesions affects the measurement of quantitative magnetic susceptibility measurements in these lesions. Masks were drawn on FLAIR and QSM images at baseline and different follow-up time points. QSM was analyzed longitudinally across these lesion masks of the same lesion and compared to NAWM. A strong variability in longitudinal QSM was observed when masks were drawn on FLAIR images. The most consistnt behavior was seen on QSM based masks. The time point of lesion definition had a stronger influence when using FLAIR maks than with QSM based masks.