Abstract
Summary: Lung cancer is a disease which has poor prognosis and is very life-threatening, as the survival rate five years after being diagnosed is merely 20%.This is the reason why the scientific community continues research within the field of radiotherapy physics in order to increase this survival rate, as many patients who are diagnosed with this disease are subjected to radiotherapy. 18FDG PET based imaging is an excellent diagnostic tool for tracking down tumors as they are highly metabolic active and tend to absorb the FDG radiotracer and appear as bright spots in the images. Studies showed a great dependency between the amount of FDG uptake in the tumor and its aggressiveness. The higher the uptake, the more aggressive the tumor. For this reason, giving higher dose to the tumor could probably be the key to better treatment. However it’s not straightforward to induce a high dose to the tumor without harming the normal tissue with today’s treatment planning systems (TPS).In order to surpass this obstacle dose painting by numbers (DPBN) is proposed. This method employs direct use of the digital PET images in the radiotherapy planning, where higher radiation doses are given only to the parts of the tumor showing high image intensity, where the aggressive disease is located. In this project we used the CT images from a patient who had lung cancer in order to create an anthropomorphic phantom. This phantom simulated the thorax of the patient and contained 101 cavities, to be filled with alanine dosimeters. DPBN depends on the biological medical images employed. Based on the FDGuptake in PET images, a linear relationship was applied between PET-intensity and radiotherapy dose to create a DPBN prescription map. Based on this map, an inverse DPBN prescription map was also constructed to be used as a pretreated plan or mock plan during optimization in the TPS. The desired outcome is to achieve better local control of the tumor compared to conventional radiotherapy (RT). A series of VMAT treatment plans were constructed in the Varian Eclipse treatment planning system which were then relocated to the phantom. The phantom was irradiated at a Varian Trilogy linear accelerator at the Oslo University Hospital, and subsequently the alanine dosimeters were removed and they were transferred for read-out by an EleXsyS 560 Super X spectrometer. The dose of each dosimeter was estimated from accompanying calibration curves. The series of calibration dosimeters was irradiated immediately prior to the irradiation of the phantom. Finally, the planned doses in the TPS were compared with the prescribed doses and the measured doses. It was found that for all plans the correspondence, between each pair of observations was quiet good. The quality factors QF between planned-delivered dose was 2.3% and 1.90% for plan1 and plan2 pooled together, the extreme plan respectively. Thus we conclude that DPBN can be delivered quiet accurately, at least for the TPS and phantom used in this study.