Object: A large number of clinical studies have been made on treatment outcomes of radiosurgery for arteriovenous malformations (AVMs), but the reported obliteration rates following this treatment vary significantly, perhaps reflecting the different methods and timings of the imaging studies used.
Methods: The authors retrospectively analyzed their experience with gamma knife surgery in 400 patients with AVMs (follow-up period 1-135 months, median 65 months), with special reference to the imaging modality used in each case. The calculated obliteration rates varied from 68.2 to 92%, depending on imaging modality and timing of evaluation. When only unquestionable imaging data such as demonstrations of a residual nidus on computerized tomography (CT) or magnetic resonance (MR) images or findings on angiograms were used in the calculation, the obliteration rates were 72% at 3 years and 87.3% at 5 years. Factors leading to a better obliteration rate were previous hemorrhage (p = 0.0084), smaller nidus (p = 0.0023), and higher radiation dose to the lesion's margin (p = 0.0495), as determined in a multivariate analysis. Factors leading to an earlier obliteration of the nidus were male sex (p = 0.0001), previous hemorrhage (p = 0.0039), smaller nidus diameter (p = 0.0006), and dose planning using angiography alone (p = 0.0201).
Conclusions: After the introduction of CT and MR images into dose planning, the conformity and selectivity of dosimetry improved remarkably, although the latency intervals until obliteration were prolonged. Imaging outcomes for AVMs should be evaluated using data provided by longer follow-up periods. The timing of additional treatments for residual AVMs should be decided cautiously, considering the size of the AVM, the patient age and sex, and the history of hemorrhage before radiosurgery.