Studies of the dosimetric verification of intensity modulated radiotherapy

The ability to verify absorbed dose distributions produced by intensity modulated radiotherapy (IMRT) using dynamic multileaf collimation is of great concern in the clinical application of this technique. This thesis investigates two approaches using an Elekta SL/ linear accelerator operating in dynamic mode (Elekta Oncology Systems, Crawley, UK). A novel strip ionisation chamber array, located at the beam aperture, was designed and used in conjunction with a specialised electrometer. This also recorded cumulative accelerator monitor units (MU) via an isolated interface to the accelerator. The chamber signal, recorded as a function of MU, proved suitable for collimator position verification for the case of a dynamic wedge, but was found not to be suitable for more general cases in which the leaves moved independently. A tube camera-based electronic portal imaging device (EPID) (Theraview , Cablon Medical, Leusden, The Netherlands) was investigated in a further approach to verification. This EPID has not been previously studied for dosimetry and several unreported effects associated with the video system were identified. The phosphor Gd202S:Tb, which is used as the x-ray detector, was also studied by direct measurements of luminescence using a photomultiplier tube. It was confirmed that the optical signal was independent of accelerator pulse repetition frequency, and that there was no long- lived luminescence (afterglow) following prolonged irradiation, which is of concern in dosimetry of dynamic deliveries. The EPID was applied to the verification of collimator position using a specially constructed camera interface that triggered recording of the cumulative MU. The EPID was also assessed as a method of measuring the integrated dose distribution delivered during a dynamic sequence, and a method proposed to overcome unreliable triggering of image acquisition in such cases. Dark current and persistence of the camera target were found to complicate measurements. Images were also found to exhibit optical scattering, which is an inherent characteristic of camera- based EPIDs. Results of a physical means of reducing the effect using an optical rejection screen were compared to an ionisation chamber for static and dynamic cases, and it was shown that the optical rejection screen is limited in its effectiveness in removing optical scatter. Dose profiles obtained from the EPID agree with ionisation chamber measurements in-air within 6 % for plain fields, and within 15-25 % for static and dynamically produced wedged fields. It was concluded that both approaches studied can be applied to the verification of IMRT but with limitations, and that an ideal system has yet to be found.