Dose assessment in Computed Tomography based on Monte Carlo simulation for a 320 detector-row cone-beam scanner

Resumen

Introduction Computed Tomography (CT) has become one of the imaging techniques most used in the field of diagnostic radiology. Since the introduction of the multi-slice CT scanners, a continuous process of technological evolution has made possible a new range of diagnostic applications. These scanners allowed to obtain more anatomical information using advanced dose reduction techniques and to acquire with high resolution in order to improve image quality and reduce artefacts. But the continuous increase of the number of patients undergoing CT explorations and the impact of clinical applications have implied an exponential increase of the doses due to CT examinations in medicine. This is particularly relevant in some newest explorations that involve high doses, such as CT perfusion (CTP), CT angiography (CTA), dynamic scans (4D CT) and others. Additionally, the introduction of new CT equipment has introduced several technical advances in recent years, such as a 320 detector-row CT scanner (Aquilion ONE) for cone-beam acquisitions with a volume coverage of 160 mm in a single 0.35 s full rotation. These features lead to a reduction of the scan time and a minimization of the movement artefacts. With the appearance of the 320 detector-row CT scanners, the existing CT dose metrics had to be revised and new approaches for Computed Tomography dose index (CTDI) dosimetry for cone-beam scanners were suggested. In parallel, some specific CT explorations and acquisitions have been adapted to this scanner since its coverage allowed for imaging of whole organs in a single axial rotation. According to ALARA principle, radiation doses to patients have to be As Low As Reasonably Achievable. Hence, considering the fast technical advances in x-ray imaging during last years and the increase in the amount of CT examinations, there is a need to be aware about radiation exposure in current acquisition protocols in order to optimize the clinical application of CT and to minimize possible radiation-induced health effects. There are different methods to estimate doses such as dosimetric studies in anthropomorphic phantoms, Monte Carlo (MC) simulation studies or k-factors that convert the dose-length product (DLP) to effective dose. Besides, several dosimetric software tools have been developed to facilitate the determination of the effective dose and the organ doses from MC data sets. MC code simulates the radiation transport through voxelized space taking into account the geometry and x-ray characteristics of the CT scanner and compute the energy deposition from interactions of the beam in each voxel of the irradiated phantom. Hence, this method provides a versatile and accurate tool to estimate organ doses and effective dose in CT examinations. Regarding the literature, several studies about dose assessment in CT using MC simulation-based methods have been published. However, most are based on CT scanners with narrow collimation geometries and mathematical phantoms, becoming outdated with the publication of the latest recommendation from the International Commission on Radiological Protection (ICRP) such as the adult computational phantoms and the commercialization of the new CT contemporary scanners. For this reason, there is a need to develop methods for patient doses assessment in CT examinations that are up to date with current scanner technology and design, current acquisition protocols and the latest reference in computational phantoms. Motivation and objectives The motivation of this thesis was the development of a MC simulation tool taking into account all relevant technical characteristics of the 320 detector-row cone-beam CT scanner and the latest recommendations of the ICRP with the aim of assessing doses in patients undergoing CT examinations. In this framework, the starting hypothesis of this project is that the use of methods based on MC simulation using the ICRP adult computational phantoms can be a useful tool to estimate patient doses undergoing examinations with a 320 detector-row cone-beam CT scanner. The main specific objectives of this thesis are: 1. To devise a MC simulation tool for dose assessment that accurately reproduces the 320 detector-row cone-beam CT scanner operation taking into account all parameters that influence dose. 2. To validate the MC program by comparing results from simulations with the actual dose measurements in the scanner. 3. To apply the MC simulation program to assess doses using the ICRP adult computational phantoms. In particular, to calculate organ doses and effective doses of four clinical acquisition CT protocols used in the scanner with and without automatic tube current modulation (TCM). 4. To apply the MC simulation program to assess doses for the cardiac CT protocol performed in a international multicenter study (CORE320) which includes CT calcium scoring, CT coronary angiography and CT myocardial perfusion using the 320 detector-row scanner and the ICRP adult computational phantoms. To study the influence of the positioning of the patient during the cardiac CT scan. 5. To apply the MC simulation program to estimate the patient dose from perfusion CT examinations of the brain, lung tumors and the liver on a 320 detector-row scanner with the ICRP adult computational phantoms. 6. To determine the conversion factors (k-factors) to estimate effective dose from DLP for each CT protocol. 7. To develop a software based on MC simulation to easily assess and report doses for standard patients undergoing CT examinations in a 320 detector-row cone-beam scanner. Results This thesis is compounded from four papers. They described the procedure followed to tailor the scanner model in a MC simulation program, the validation of the MC code, the use of the program for dose estimation in different CT examinations and the development of the dosimetric software tool for dose assessment and reporting. Paper I was focused on the modelling and validation of a MC simulation program for patient dose assessment for a 320 detector-row CT scanner. All the technical features of the scanner were successfully reproduced. The MC program was validated by comparing simulations results with actual dose measurements acquired under the same conditions. Once validated, patient dose assessment was performed for four clinical axial acquisitions using the ICRP adult reference phantoms. The results were nearly always lower than those obtained from other dose calculator tools or published in other studies, which were obtained using mathematical phantoms in different CT systems. The influence on dose of TCM in one of the acquisitions was also analysed, leading to dose decrease and greater uniformity of the dose distributions. The MC simulation program was used in Paper II and Paper III in order to estimate organ absorbed dose and effective dose in standard patients for specific CT acquisition protocols. The former evaluated doses from a cardiac CT protocol, including CT calcium scoring, CT coronary angiography and CT myocardial perfusion with a 320 detector-row volumetric CT scanner according to the internationally recognized recommendations of the ICRP. The effect on patient dose of off-centering the patient and the positioning of the arms were evaluated. The latter was focused on perfusion CT examinations. The MC simulation tool was used to calculate the organ doses and the effective dose in the adult reference computational phantoms undergoing CTP Brain, CTP Lung Tumour and CTP Liver studies. Additionally dose assessment was performed for the skin and the eye lens. These CTP protocols performed in a 320 detector-row CT scanner operate safely below threshold doses for deterministic effects. Conversion factors (k-factors) were obtained to estimate effective doses from DLP in both cardiac and perfusion CT protocols. Results showed that in many studies the dose assessment for cardiac and body perfusion acquisitions is performed with a k-factor that underestimates effective doses. Contrarily, dose is overestimated in head examinations. Finally, the development of a software dosimetric tool for accurate dose assessment in CT with the 320 detector-row cone-beam scanner was described in Paper IV. The software, called SimDoseCT, was based on look-up tables generated from MC simulation. By selecting the appropriate acquisition technique through a graphical user interface, the software reports organ absorbed doses and effective doses from all possible acquisitions within the CT scanner, like axial (volumetric), helical and scanogram acquisitions. The validation and testing of the software demonstrated the accurate methodology of SimDoseCT and its usefulness for dose assessment in CT. Conclusions This dissertation presents a framework for dose estimates in standard patients undergoing CT examinations with a 320 detector-row cone-beam scanner. A set of k-factors and also a software dosimetric tool for an accurate estimation of organ absorbed doses and effective doses were created with the aim of improving the easily dose evaluation for standard adult patients in CT contemporary scanners. The initial intended goals of the study were covered with the methodology and results described in the four papers that constituted this thesis. Future research could be focused on technical extensions in the dosimetric software such as the implementation of the TCM, size-specific dose estimates (SSDE) or paediatric patients. In addition, new studies combining dosimetry with image quality evaluation could be perform for the optimization in CT latest technologies.