Patient dosimetry framework
In order to estimate patient effective dose levels, absorbed organ doses need to be measured. Traditionally, measurements with thermoluminescent dosimeters (TLDs) distributed at several locations within anthropomorphic phantoms are often used for this purpose [4]. However, this method requires the use of many TLDs, making it very time-consuming. In addition, the necessary phantoms are not always easily accessible. However, an alternative approach to estimate organ doses, would be to perform Monte Carlo (MC) simulations [4]. In a previous study in the author’s group, a hybrid framework was developed to simulate the complete CBCT imaging chain for dento-maxillofacial applications [5]. Therefore, in this study, MC simulations were considered the preferred method to determine patient effective dose levels.
Patient models
As the goal for this study was to determine population average patient effective dose levels, the ICRP reference pediatric and adult computational voxel phantoms were chosen to serve as patient models. Models representing a reference 5, 10 and 15 year old female and male patient, as well as a female and male adult model were available. The technical description of each of these phantoms can be found in ICRP publications 143 [6] and 110 [7], for the pediatric and adult patient models, respectively. Fig 1 and 2 show an overview of the used patient models.
The patient models were prepared to estimate the dose levels for different clinical examinations. The applications evaluated in this work were an ankle, hand and knee scan for extremity imaging, and a head scan for dento-maxillofacial/ENT imaging. For the latter, dependent on the chosen Field Of View (FOV) size, the focus was more on the dental, maxillo, sinus, or inner ear region.
DE CBCT system model
The existing simulation platform, employing the PENELOPE/penEasy MC software [8], was adapted and calibrated to match the behaviour of the NewTom 7G DE CBCT scanner. The DE acquisition consists of two consecutive scans; a low energy (LE) and high energy (HE) acquisition, using a tube voltage of respectively 80 kV and 120 kV. The x-ray spectrum for each of the selected tube voltages was determined using the SpekPy tool [9], in combination with first half-value-layer measurements performed on the real system. Furthermore, vendor specific information, such as the focus-to-object-distance, the rotation trajectory and scan angle, the number of projections per rotation, as well as their pulse width for different scan modes (i.e. regular and best) and FOV sizes, were provided by the manufacturer and incorporated in the simulation framework. A fixed non-modulated tube current of 59 mA and 48 mA was set for the LE and HE acquisitions, respectively. Finally, system calibration factors for both the LE and HE x-ray spectrum were determined based on tube output measurements, to relate the number of simulated photons to the number of photons in a real acquisition, and were expressed in µGy/mAs.
Patient effective dose levels were obtained for all the specified clinical examinations using the regular and best scan modes, for both a 10x10 cm² and 17x17 cm² FOV size. Table 1 shows for each combination of FOV size and scan mode the total tube current-exposure time product (mAs) applied during the scan. The statistical uncertainty on the MC obtained organ doses were less than 1%.
Effective dose calculation
Patient effective dose levels were obtained by applying the ICRP 103 [10] tissue weighting factors to the obtained organ doses from the simulations. The dose to the red bone marrow and endosteum were determined according to the recommendations of the ICRP [11].