EFFECTS OF SCATTER GENERATED BY BEAM-MODIFYING ABSORBERS IN MEGAVOLTAGE PHOTON BEAMS

Transmission through a beam-modifying absorber consists of attenuated primary beam and scattered radiation generated by the absorber. The pr imary component of the transmitted beam is characterized by the narrow beam attenuation coefficient which depends upon the energy of the bea m and type of the absorber. In addition to beam energy and absorber ma terial, the scatter component also depends on field size, thickness an d shape of the absorber, location of the absorber with respect to the source, and the point of calculation. Based upon Compton first-scatter , a method has been developed to calculate effective broad beam transm ission through any arbitrarily shaped absorber with variable thickness for any points on and off the central axis. The method requires prede termined narrow beam attenuation coefficients as a function of thickne ss. Transmission calculations for various absorbers such as wedges and attenuators were performed for cobalt-60 and 6-MV beams and were comp ared with measured data. For a cobalt-60 beam, the measured transmissi on fraction through a 1.33-cm-thick absorber (alloy, consisting of 55% bismuth and 45% lead) for a field size of 24X24 cm(2) is 17% higher t han the calculated value using a narrow beam attenuation coefficient. Also, for the same absorber, measured central axis transmission is as much as 3.6% higher compared to off-axis locations. The measured trans mission fraction through a 1.33-cm absorber was found to differ by as much as 13% and 14% for Cobalt-60 and 6 MV, respectively, as the chamb er-to-source distance was varied from 70 to 110 cm. The agreement betw een calculated and measured values is within 0.5% for both energies wh ereas conventional narrow beam calculations would have yielded errors of 18% and 19%, respectively. Similar agree ment was obtained when com paring calculated and measured wedge factors as a function of field si ze, with the maximum deviation being 0.7%. Measured scattered doses, d ue to an attenuator covering part of a beam, show a maximum for a thic kness of approximately one mean-free path. This is also predicted by c alculations with an agreement of 0.3%.