Gamma knife surgery of cerebral arteriovenous malformations: serial MR imaging studies after radiosurgery.

PURPOSE

To investigate the temporal sequence of post radiosurgery magnetic resonance imaging changes in cerebral arteriovenous malformations.

METHODS AND MATERIALS

Eighteen patients were regularly followed up after gamma knife surgery. The follow-up intervals ranged from one day to 44 months. High signal lesion in or around arteriovenous malformations on T2-weighted magnetic resonance images corresponding to the treatment volume and developing after radiosurgery were defined as the adverse reaction of the irradiation. This high signal and the regression of arteriovenous malformations nidus after radiosurgery were evaluated.

RESULTS

Adverse reaction of irradiation was observed in nine cases. Seven of them were symptomatic. The reactions presented as focal high signal in three cases and focal high signal with extension along the neural tracts in six cases. The reactions were seen either immediately after treatment (one case), between 3 and 14 months (seven cases), and 40 months after treatment (one case). The regression of the adverse reaction was observed to start 5 +/- 3 months after its appearance. Regression of the arteriovenous malformations' nidus was found in 16 cases. In two cases the AVMs became invisible on magnetic resonance images but the angiogram still demonstrated abnormal shunts. In another one case with angiogram showing total obliteration, the nidus was erroneously interpreted as incomplete obliteration on magnetic resonance images.

CONCLUSION

It is concluded that magnetic resonance imaging is a sensitive in vivo method for detecting cerebral radiation injury. Magnetic resonance imaging offers a method for evaluating the regression of arteriovenous malformations' nidus, but the diagnosis of complete obliteration of the nidus after radiosurgery still relies on the angiogram.