Gamma knife surgery of cerebral arteriovenous malformations

Aims of the study: to define and quantify parameters of importance for the treatment result in Gamma Knife (GK) surgery of arteriovenous malformations (AVM). To determine the natural course of AVM in order to calculate the impact of the GK treatment on the latency period between treatment and obliteration. Material and method: All AVM patients treated at the Karolinska Hospital 1970 - 1994 and at the University of Virginia 1989 - 1994, totally 2262 patients, were included in the material. All hospital records of these patients were reevaluated, and for all patients included in the studies of obliteration and risk for complications, all dose plans were reevaluated and recalculated in order to define dose/volume histograms and a number of treatment parameters. Result and conclusions: It could be shown that the lowest dose to the AVM periphery (=minimum dose) was the decisive factor for the treatment result. In addition, the average dose, reflecting the amount of energy delivered to the malformation, was linearly related to the length of the time period between treatment and obliteration. In the successfully treated AVM, the minimum dose decreased with increasing AVM volume. An index, K index, was defined as minimum dose * (AVM volume) 1/3. The obliteration rate was linearly increasing with the K index up to a value of 27, and thereafter a constant 80%. The age distribution at the initial AVM rupture was used to calculate the annual risk for hemorrhage in untreated AVM. The risk increases with increasing patient age, and is larger for larger AVM. The risk for rupture is 5% at 25, 7% at 50 and 10% at 70 years of age. It was shown that the risk for rupture is dose dependent and decreased in the latency period from treatment to verified obliteration. A model was presented that predicts the risk for complications accurately. Further studies revealed that not only the dose/volume distribution, but also AVM location and previous hemorrhage affected the risk for complications. Thus, different risk estimations were calculated for different groups of AVM