RESUMO
The impact of several physical quantities on the spatial resolution of an X-ray scintillating pixel detector for a micro cone beam CT (µCBCT) is investigated and discussed. The XtremeCT from SCANCO Medical AG was simulated using the EGSnrc/EGS++ Monte Carlo (MC) framework and extensively benchmarked in a previous work. The resolution of the detector was determined by simulating a titanium knife-edge to obtain the edge spread function (ESF) and the modulation transfer function (MTF). Propagation of the scintillation light through the scintillator and its coupling into the fiber optics system was taken into account. The contribution of particles scattered in the main scanner components to the detector signal is very low and does not affect the spatial resolution of the detector. The resolution obtained from the energy deposition in the scintillator without any blurring due to the propagation of the scintillation light into the fiber optics array was 31⯵m. By assuming isotropic light propagation in the scintillator, the resolution degraded to 360⯵m. A simple light propagation model taking into account the impact of the scintillator's columnar microstructures was developed and compared with the MANTIS Monte Carlo simulation package. By reducing the width of the model's light propagation kernel by a factor of 2 compared to the isotropic case, the detector resolution can be improved to 83⯵m, which corresponds well to the measured resolution of 86⯵m. The resolution of the detector is limited mainly by the propagation of the scintillation light through the scintillator layer. It offers the greatest potential to improve the resolution of the µCBCT imaging system.
Assuntos
Tomografia Computadorizada de Feixe Cônico/instrumentação , Método de Monte Carlo , Contagem de Cintilação/instrumentação , LuzRESUMO
High-resolution peripheral quantitative computed tomography (HR-pQCT) quantifies bone microstructure and density at the distal tibia where there is also a sizable amount of myotendinous (muscle and tendon) tissue (MT); however, there is no method for the quantification of MT. This study aimed (1) to assess the feasibility of using HR-pQCT distal tibia scans to estimate MT properties using a custom algorithm, and (2) to determine the relationship between MT properties at the distal tibia and mid-leg muscle density (MD) obtained from pQCT. Postmenopausal women from the Hamilton cohort of the Canadian Multicenter Osteoporosis Study had a single-slice (2.3 ± 0.5 mm) 66% site pQCT scan measuring muscle cross-sectional area (MCSA) and MD. A standard HR-pQCT scan was acquired at the distal tibia. HR-pQCT-derived MT cross-sectional area (MTCSA) and MT density (MTD) were calculated using a custom algorithm in which thresholds (34.22-194.32 mg HA/cm3) identified muscle seed volumes and were iteratively expanded. Pearson and Bland-Altman plots were used to assess correlations and systematic differences between pQCT- and HR-pQCT-derived muscle properties. Among 45 women (mean age: 74.6 ± 8.5 years, body mass index: 25.9 ± 4.3 kg/m2), MTD was moderately correlated with mid-leg MD across the 2 modalities (r = 0.69-0.70, p < 0.01). Bland-Altman analyses revealed no evidence of directional bias for MTD-MD. HR-pQCT and pQCT measures of MTCSA and MCSA were moderately correlated (r = 0.44, p < 0.01). Bland-Altman plots for MTCSA revealed that larger MCSAs related to larger discrepancy between the distal and the mid-leg locations. This is the first study to assess the ability of HR-pQCT to measure MT size, density, and morphometry. HR-pQCT-derived MTD was moderately correlated with mid-leg MD from pQCT. This relationship suggests that distal MT may share common properties with muscle throughout the length of the leg. Future studies will assess the value of HR-pQCT-derived MT properties in the context of falls, mobility, and balance.
Assuntos
Algoritmos , Músculo Esquelético/anatomia & histologia , Músculo Esquelético/diagnóstico por imagem , Tendões/anatomia & histologia , Tendões/diagnóstico por imagem , Tomografia Computadorizada por Raios X/métodos , Idoso , Idoso de 80 Anos ou mais , Tornozelo , Feminino , Humanos , Perna (Membro) , Pessoa de Meia-Idade , Tamanho do Órgão , Tíbia/diagnóstico por imagemRESUMO
PURPOSE: Micro CBCT scanners have a broad spectrum of applications in medicine and material science. However, CBCT suffers from scatter radiation and spectral effects such as beam hardening (BH). In this work an iterative BH- and scatter-correction algorithm was developed using Monte Carlo (MC) methods. METHODS: Two µCBCT scanner models (XtremeCT and µCT100 from SCANCO Medical AG) were simulated using EGSnrc/EGS++. The scanner measures the attenuation of X-rays passing through the phantom and interacting in a scintillating detector. The MC method is used to characterize the influence of the scattering and BH- effects. In addition, an analytical model is developed in order to correct for the scattering effect. For this purpose, by using MC methods, different scatter components were analyzed with respect to the number of scatter interactions within the given geometry. For the BH-effect, after assessing the detector response for an equivalent mono-energetic and scatter-free system using MC methods, an analytical correction model was developed. Both correction methods were implemented as an iterative reconstruction correction algorithm and were tested for various phantoms. RESULTS: MC simulations show that the amounts of correction for the scattering and BH- effects are in the same order of magnitude. The correction term for scattering effects is a function of the scatter to primary ratio only and is mainly due to single scattered particles. The contribution of multiple scattered particles to the total scatter signal is small and can be approximated by a constant. In all cases tested, the reconstructed linear attenuation coefficients converge to the mono-energetic reference values after 2-3 iteration steps with a deviation of about 1%. CONCLUSIONS: By using an iterative correction algorithm using single scatter approximation, BH- and scatter correction can be performed accurately for µCBCT scanners. This work was supported by CTI-10629.1 and SCANCO Medical AG.