Abstract
Between March 1993 and January 1997, stereotactic radiation techniques were used to irradiate 66 intraocular tumour patients with the Gamma Knife (Leksell Gamma Knife
, model B unit) at the University of Vienna, Austria. This study investigates the dosimetry for stereotactic irradiation of ocular structures.
For the dosimetry program KULA 4.4, Gamma Knife stereotactic irradiation of the eye represents an extreme frontal skull position. In addition, irradiation of the eye may be performed in the usual supine position in exceptional cases only. With the patient in the prone position, the dose planning program has to calculate with a significantly large number of single-beam extrapolations. In our first experiment we measured the isocentre dose for eight different 
-angle positions, both in prone and supine positions, using TLD measurements in an Alderson head phantom. We found a maximum deviation of 
using these individually calibrated TLDs.
In the second experiment we examined the dose cross profiles for the two most frequently used treatment positions (supine position, 
, and prone position, 
). For this purpose we implanted a specially designed TLD array into the orbit of a human cadaver head. We found excellent agreement of the dose values measured for the isocentre as well as the posterior part of the eye with orbit with deviations of less than -2.7%. However, for the anterior part of the eye, deviations between computer-generated calculations and the TLD measurements were found to range up to -30%. These differences were noticed both for supine and prone positions. For the Gamma Knife stereotactic irradiation of ocular tumours or pathologies, precautions should be taken to avoid significant underdosage in the anterior part of the radiation field.