MR evaluation of developmental venous anomalies: medullary venous anatomy of venous angiomas.

PURPOSE To present characteristic MR findings of developmental venous anomalies (DVAs) in terms of location of caput and draining veins, to correlate these findings with normal medullary venous anatomy, and to suggest an approach to the evaluation of DVAs by means of MR imaging. METHODS We reviewed the contrast-enhanced MR examinations of 61 patients with DVA, which were selected from 4624 consecutive cranial MR examinations. Site of the DVA and size and direction of draining veins were recorded. RESULTS Seventy-two DVAs with 78 draining veins were located: 18 were juxtacortical, 13 were subcortical, and 41 were periventricular or deep. Twenty-six of the DVA caputs were frontal, 16 were parietal, 13 were in the brachium pontis/dentate, seven were in the temporal lobe, three were in the cerebellar hemisphere, three were in the occipital lobe, three were in the basal ganglia, and one was in the pons. The draining veins were superficial in 29 cases and deep in 49. Of the 36 supratentorial deep draining veins, 16 were in the trigone/occipital horn, 11 were in the mid-body of the lateral ventricle, seven were in the frontal horn, and two were in the temporal horn. Among the 14 infratentorial deep draining veins, five were in the lateral recess of the fourth ventricle, four were anterior transpontine veins, three were lateral transpontine veins, and two were precentral cerebellar veins. CONCLUSION The DVA caputs and their draining veins occurred in typical locations that could be predicted from the normal medullary venous anatomy, with the frontal, parietal, and brachium pontis/dentate being the most common locations. Drainage can occur in superficial cortical veins or sinuses or in deep ventricular veins or in both, no matter where the caput is located. Whether drainage was superficial or deep could not be predicted on the basis of the site of the DVA caput. Contrast-enhanced T1-weighted MR images showed the DVAs best, but diagnosis could be made from T2-weighted MR images.