Gamma knife outcome models as a reference standard in the embolisation of cerebral arteriovenous malformations

We sought to utilise outcome models from Gamma Knife radiosurgery (GKRS) to cerebral arteriovenous malformations (AVM) as a reference standard in assessing the clinical outcome of embolisation, thus comparing the outcomes of two different management alternatives, in the same patients.

87 consecutive patients with 88 AVM were admitted during 1997-1999 for initial embolisation of an AVM. The clinical outcomes were recorded prospectively. Angiography under stereotactic conditions with measurement of AVM volume was performed before and after embolisation. GKRS outcome models were used to predict obliteration rate, complication rate and risk of haemorrhage before and after embolisation. The clinical outcome of embolisation followed by predicted outcome of adjunct GKRS was then compared with the predicted outcome of GKRS as the only treatment.

Eight patients were subjected to microcatheterisation but not to embolisation. By the end of the study period, embolisation had been terminated in 55 patients out of 80 (69%). The predicted outcome of GKRS alone was 58 obliterations and 12 complications while that of the combined management was 58 obliterations and 15 complications. The difference was not significant on the p < 0.1 level.

Volume measurement from angiography and outcome models from Gamma Knife radiosurgery are useful as a reference standard in the management of AVM. Absolute volume reduction from embolisation is most prominent for AVM > 10 ml and thus facilitates subsequent radiosurgery. For AVM < or = 10 ml, GKRS as the only treatment can be an alternative to primary embolisation, particularly if no significant volume reduction or obviously beneficial effect of targeted embolisation is expected. Further prospective studies are needed to identify subgroups in which one treatment has advantages over the other.