Background and purpose: We reported previously a model predicting the risk for radiation-induced complications following Gamma Knife radiosurgery for AVM. No factor other than the dose distribution was related to the risk. The aim of this study was to define if other parameters are of importance for the risk of complications.
Material and methods: The model above was used to calculate the risk for complications in all 1128 AVM patients Gamma Knife-treated at the Karolinska Hospital 1970-1993. The number of predicted complications was compared to the number of observed ones for a number of different parameters.
Results: The model underestimated the risk of complications for patients previously given radiation with multiple or single fractions. Neither age nor gender influenced the risk of complications. Centrally located AVM had a higher, and peripheral a lower incidence of complications as compared to the calculated risk, and a previous hemorrhage reduced the risk of complications. From the observed number of complications, parameters in the model were determined by a fitting procedure separately for three groups of AVM: central and peripheral with and without a previous hemorrhage. It is also shown that the assumption of a serial functional architecture is valid in the model. This was investigated by the use of a relative seriality model with a combined serial-parallel functional architecture.
Conclusions: The risk of complications following radiosurgical treatment of AVM is dependent on the clinical history, AVM location and whether the patient has received radiation earlier.