We read the article by Bugnicourt et al with great interest. Authors show that incomplete posterior circle of Willis (CW) is an independent risk factor of migraine. Our previous study disclosed that proportion of a fetal CW did not differ between migraine and control subjects. We also studied morphological changes of posterior CW in migraineurs, and we would like to compare between ours and the results of Bugnicourt et al. Among 141 migraine patients in our neurology department, 73 migraineurs (31 migraineurs with aura [MA] and 42 without aura [MO]) had brain MR imaging and angiography (MRA) using 3-dimensional time-of-flight sequence. A skilled neuroradiologist reviewed maximum intensity projection (MIP) imaging and source imaging of MRA. Complete posterior CW was defined as the presence of both posterior communicating arteries and both P1 segments of the posterior cerebral arteries. Other patterns of posterior CW were classified to incomplete type. CW morphology was compared between migraineurs and 100 age-matched control subjects (Table). Those participants who had hypertension, diabetes mellitus, dyslipidemia, or oral contraceptives were excluded. As compared with controls, migraineurs and MA sufferers significantly decreased the frequency of complete posterior CW. No significant changes of posterior CW existed between MO sufferers and controls. Posterior CW patterns were not correlated with other clinical aspects of migraineurs, including age, sex, onset age, and duration of migraine (Table). Incomplete posterior CW is associated with MA in our study whereas Bugnicourt et al showed no statistical differences of posterior CW patterns between MA and MO patients. The question arises whether incomplete posterior CW is a causative cofactor of migraine onset or whether these blood flow changes occur after migraine onset.We would like to know how onset age and duration of migraine influence posterior CW shapes in patients of Bugnicourt et al, as authors mention that their patients visit emergency department. Thirty of 47 migraineurs have atypical episodes, and 21 of 24 MA patients have prolonged aura more than 1 hour. Otherwise, our patients did not derive from emergency department. We excluded migraineurs with atypical episode or long visual aura. Posterior configuration of CW was assessed carefully by serial slices of source imaging in our subjects, besides