Comparison of imaging modalities for the accurate delineation of arteriovenous malformation, with reference to stereotactic radiosurgery

Purpose: To investigate the discrepancy between the arteriovenous malformations seen on magnetic resonance angiography (MRA) and on stereotactic digital subtracted angiography (DSA). Methods and Materials: The target volume on stereotactic DSA (VDSA ) and the target volume on MRA (VMRA ) were separately delineated in 28 intracranial arteriovenous malformations. The coordinates of the center and the outer edges of VDSA and VMRA were calculated and used for the analyses. Results: The standard deviations (mean value) of the displacement of centers of VMRA from VDSA were 2.67 mm (-1.82 mm) in the left-right direction, 3.23 mm (-0.08 mm) in the anterior-posterior direction, and 2.16 mm (0.91 mm) in the craniocaudal direction. VMRA covered less than 80% of VDSA in any dimensions in 9 cases (32%), although no significant difference was seen in the target volume between each method, with a mean value of 11.9 cc for VDSA and 12.3 cc for VMRA (p = 0.948). Conclusion: The shift of centers between each modality is not negligible. Considering no significant difference between VDSA and VMRA , but inadequate coverage of the VDSA by VMRA , it is reasonable to consider that the target on MRA might include the feeding artery and draining vein and possibly miss a portion of the nidus