Dose Reconstruction and Dose Optimization
The determination of an exact individual dose received by each patient included in the cohort is crucial for the study in order to address the cancer induction rate by certain dose quantities (dose response). The dose for every organ, for which the risk of developing cancer is going to be evaluated, has to be calculated from each CT procedure as a CT examination of the brain may not expose the pelvis and is unlikely to cause a primary testicular cancer. The calculated organ doses will include, among others, also red bone marrow dose for analysis of leukaemia risk and the colon doses for analysis of all cancers as used in other studies like the atomic bombs survivor study to allow direct comparisons of derived risk estimates with them.
Individual dosimetry is very complex in CT scanning: the CT technology uses rotating X-ray sources, thin X-ray beams and numerous rotations, resulting in hundreds of single exposures per examination. Anatomical phantoms which simulate the human body are used to estimate organ doses inside the body. Nevertheless, they are only models of the human body and due to the variety of possible examination protocols used in practice it is impossible to measure doses from every examination type by using phantoms. A dose calculation algorithm which is based on the phantom measurements and uses several input parameters, is therefore developed for organ dose calculation. Patients' body measures have to be considered, as well as the exposed body region. In addition, a broad variety of technical parameters used to adapt the scanner settings for each specific examination has to be taken into account. Furthermore, the technology of the scanner is important and the resulting dose also depends on different models. Software dedicated to CT scanning will be used to calculate the individual organ doses based on these parameters for every examination.
As already mentioned in the Epidemiology section, the data collection period is split into two parts - before and after the introduction of PACS. In the later period, there is more information available for dose reconstruction. As extracting the data from individual images is too time-consuming, special software will be used to access the Digital Imaging and Communications in Medicine (DICOM) headers and upload information needed for dose reconstruction without downloading images. For the more distant past, only sparse information about scanner settings and technical parameters used for specific procedures can be obtained. For this period, a patchwork approach will be used retrieving information from specially developed questionnaires, published surveys, scientific publications, expert interviews and interpolations. This involves a number of uncertainties but the multisource approach is robust against major errors.
In addition, the information collected for the dose reconstruction will be used to assess technical reasons for variation in dose and image quality. Technical parameters and exposure settings used to produce images for the same CT model will be compared to the information available in the radiology departments. This will be used for defining the best strategies for dose optimization, in collaboration with paediatric radiologists and manufacturers.