[Radiosurgery of cerebral arteriovenous malformations: methodology of the dose prescription and its implications for dosimetric results. Retrospective study of 408 cases]

To study dosimetric implications of our group dose prescription methodology on a series of 408 cerebral arteriovenous malformations.and method. Between January 1990 and July 1998, 408 patients with cerebral arteriovenous malformations that had never been irradiated before, were treated radiosurgically in a single fraction at Tenon Hospital. 223 patients were treated with a single isocentre and 185 with multiple isocenters. Dosimetric characteristics have been studied as a function of the global quality index of planning, defined our group as the standard deviation of the differential dose volume histogram calculated in the lesion.One percent correlation obtained between some dosimetric data and the global quality index of planning allowed to modelize by linear equations the connection between this quality index and minimum isodose in the lesion, mean doses and isodoses in the underdosed part of the lesion and in the complete lesion. This was done for mono isocentric cases and multi isocentric cases.Before the choice of the therapeutic dose and prescription isodose, it is therefore possible from the global quality index of planning to foresee isodoses modelized by the defined equations. So, the radiotherapist can compensate using the prescription since conformity is not perfect and thus attenuate healthy tissues overdosage.