Abstract
BANG-gel dosimetry offers the potential for measuring the dose delivered by a radiotherapy treatment technique, in three dimensions, with high spatial resolution and good accuracy. The ability to measure comprehensively a 3D dose distribution is a major advantage of the gel dosimeter over conventional planar and point-based dosimeter devices, particularly when applied to the verification of complex dose distributions characteristic of intensity-modulated radiotherapy (IMRT). In this paper an in-house manufactured BANG-gel dosimeter was applied to study the dose distributions of two irradiation experiments for which the distributions were known: (i) a dosimetrically simple parallel-opposed irradiation, and (ii) a more complex nine-field `static tomotherapy' intensity-modulated irradiation delivered with the Nomos MIMiC. The uniform distribution in (i) allowed a study of the magnetic resonance (MR) imaging parameters to achieve an optimal trade-off between noise and image resolution (optimum image resolution for the Siemens 1.5T Vision system was determined to be 
with a slice thickness of 2 mm). The spatial uniformity of gel sensitivity to radiation was found to depend strongly on the presence of oxygen, which must be eliminated for the gel dosimeter to be of use. The gel dosimeter was found to agree well with predicted dose distributions and accurately measured the steep penumbral fall-off of dose, even after many days, proving its potential for the verification of IMRT distributions. In the nine-field IMRT delivery (ii) the predicted dose was computed by both an in-house `component-delivery' dose algorithm and the Peacock planning-system dose algorithm. Good agreement was found between the two algorithms despite the latter's omission of the change in penumbral characteristics with aperture-size during delivery, lack of inhomogeneity correction and approximate modelling of leaf leakage. These effects were found to be small for the problem studied. The predicted distribution agreed well with the gel-measured distribution at medium and high doses (50-90% isodose lines) although differences of up to 10% were observed at lower doses (30% isodose line). The gel dosimeter was found to have the potential to verify IMRT distributions but required considerable care to achieve accurate results. Attention was required to achieve uniformity of gel sensitivity (to prevent oxygen contamination), and in the calibration process.