| - PURPOSE. Radiotherapy patients are increasingly being treated with complex treatment techniques, like IMRT, often combined with high tumor doses. Reliable application in large patient groups requires a comprehensive dosimetric quality control program to ensure for each individual patient that the 3D dose distribution, as calculated by the treatment planning system (TPS) and approved by the physician, is indeed delivered to the patient during the entire fractionated treatment course. In contrast to existing techniques, the protocol should not only focus on the tumor, but on critical tissues as well. Important constraints are also a clinically acceptable workload and time required at the treatment unit. We propose that such a quality control protocol comprises for each patient: i) an independent calculation of the 3D dose distribution, ii) dosimetric verification of the fluence distributions realized by the treatment unit, prior to the start of the first treatment fraction, iii) daily 'in vivo' measurements at the treatment unit to derive the fluence maps delivered during treatment execution, iv) feed-back of the measured fluences and patient geometry in the TPS for daily assessment of the delivered 3D dose distribution. The purpose of this project is to develop the protocol, and to evaluate it clinically for all patient categories treated with IMRT. - PLAN OF INVESTIGATION. To verify the dose distribution calculation for each patient, a second TPS will be used to perform a fully independent forward 3D dose calculation (step i above). This procedure is the 3D extension of the widely used monitor unit check for individual patients, based on dose calculation in a single point in the tumor. The method aims at the fast detection of unacceptable errors related to the applied beam model in the TPS, (occasional) bugs in the system, and human failures related to its use. For these investigations a Monte Carlo based TPS is available. To compare the calculated 3D dose distributions, an independent environment will be developed in this project. For comparison, appropriate treatment plan parameters (e.g. dose-volume constraints) with corresponding action levels will be established, considering the constraints and objectives used for designing the IMRT plans. If clinically unacceptable differences are observed, additional investigations will be performed to identify their origin. In the end, this plan verification should run entirely automatically, requiring only human interference in case any action level is exceeded. Tools will be developed to derive the fluence distribution of IMRT beams, as realized by the linac, from portal dose images (PDIs) within about 1%. PDIs are measured with our electronic portal imaging device in absence of the patient (step ii), and daily during treatment delivery (step iii). To achieve the high accuracy level, our current algorithm to predict PDIs will be further developed for IMRT. A well-known problem in the use of 'in vivo' measured PDIs for assessment of delivered fluence maps are possible changes in the patient anatomy between acquisition of the planning CT scan and treatment delivery. We will investigate a new procedure that circumvents this problem, resulting in an expected major increase in the accuracy of in vivo dosimetry. For this purpose, each IMRT treatment field is split into a low dose, flat beam and a modulated beam, which are delivered immediately after each other. Using the measured PDI of both beams, the incident fluence distribution can be derived, irrespective of changes in the patient anatomy. In combination with detected set-up errors, the fluence maps are fed back into the TPS to assess the daily delivered 3D dose distribution (step iv). For the highest precision, exact knowledge of the patient's anatomy during treatment delivery is needed, which can be obtained using, e.g., a cone-beam CT. Prior to application and evaluation in clinical studies, the accuracy of the developed tools will be assessed and optimized in phantom studies. The investigations will include clinical evaluation of the tools and procedures in three radiotherapy departments in The Netherlands. - POSSIBLE RESULTS. A clinically evaluated, comprehensive system for individualized quality control and quality improvement will be available that allows for reliable application of high dose IMRT in large patient groups. The actual delivered 3D dose distribution in both the tumor and the sensitive structures is derived on a daily basis, taking into account variations in the performance of the linear accelerator and day-to-day variations in the patient's anatomy. Application of the protocol will enable determination of dose-volume-effect relationships for tumors and healthy tissues with high precision. |
