BACKGROUND AND PURPOSE: This work reports a detailed study carried out in two UK radiotherapy centres of the dosimetric accuracy of the collapsed cone algorithm of a commercial treatment planning system (Helax-TMS) in simulated clinical situations. MATERIALS AND METHODS: Initially the accuracy of the collapsed cone algorithm in homogeneous media is evaluated for an extensive set of simple and complex fields. Water, lung and bone substitute epoxy resin material were then used to assess the algorithm in inhomogeneous media and compare its accuracy with the pencil beam algorithm currently in clinical use. Finally a semi-anatomic phantom and an anthropomorphic phantom were employed to assess the dosimetric accuracy using simulated clinical set ups. Thermoluminescence dosimeter (TLD) measurements were made with the anthropomorphic phantom and ionisation chambers otherwise. Nominal 4, 6 and 15 MV photon beams were studied. RESULTS: For most homogeneous cases agreement between measured and calculated dose is within +/-2% or +/-2 mm. In cases with heterogeneities and simulated clinical situations it is observed that the accuracy is also generally within +/-2% or +/-2 mm. Specific instances where the difference between measured and calculated values exceed this are highlighted. CONCLUSIONS: It can be concluded that in clinical treatment planning situations where lung is present the collapsed cone algorithm should be considered in preference to pencil beam algorithms normally used but that there may still be some discrepancy between calculations and measurement.
BACKGROUND AND PURPOSE: This work reports a detailed study carried out in two UK radiotherapy centres of the dosimetric accuracy of the collapsed cone algorithm of a commercial treatment planning system (Helax-TMS) in simulated clinical situations. MATERIALS AND METHODS: Initially the accuracy of the collapsed cone algorithm in homogeneous media is evaluated for an extensive set of simple and complex fields. Water, lung and bone substitute epoxy resin material were then used to assess the algorithm in inhomogeneous media and compare its accuracy with the pencil beam algorithm currently in clinical use. Finally a semi-anatomic phantom and an anthropomorphic phantom were employed to assess the dosimetric accuracy using simulated clinical set ups. Thermoluminescence dosimeter (TLD) measurements were made with the anthropomorphic phantom and ionisation chambers otherwise. Nominal 4, 6 and 15 MV photon beams were studied. RESULTS: For most homogeneous cases agreement between measured and calculated dose is within +/-2% or +/-2 mm. In cases with heterogeneities and simulated clinical situations it is observed that the accuracy is also generally within +/-2% or +/-2 mm. Specific instances where the difference between measured and calculated values exceed this are highlighted. CONCLUSIONS: It can be concluded that in clinical treatment planning situations where lung is present the collapsed cone algorithm should be considered in preference to pencil beam algorithms normally used but that there may still be some discrepancy between calculations and measurement.