A method for estimating radiation interaction coefficients for tissues from single energy CT.

A parametric model for the x-ray linear attenuation coefficient is used to describe the compositional dependence of Hounsfield numbers measured by medical CT scanners. Measurements with materials of known density and composition, that span and evenly sample the compositional range of tissues, are written as linear simultaneous equations and solved for model coefficients. An algorithm is identified for this purpose. Results are expressed as atomic cross-sections in units of barn per electron divided by the attenuation coefficient for water. With the CT scanner characterised, a virtual CT scan can be simulated to predict HN for tissues based upon their known density and composition. Similar calculations using the tabulations and mixture rule deliver attenuation coefficients and mass energy absorption coefficients for mono-energetic radiation 10 keV to 20 MeV. Results are presented for measurements with a radiotherapy CT simulator, the RMI-467 phantom with tissue substitute materials, plus common polymer materials and silicon. Published measurements with earlier generations of the phantom and tissue substitutes using different CT scanners are also considered. Measured atomic cross-sections differ from expectations for mono-energetic radiation due to the use of a filtered spectrum and energy integrating detection system. The cross-sections for different CT scanners are similar, without large variations with kVp. Results are presented showing the relationship between predicted HN for tissues, electron density and photon interaction coefficients for healthy tissues and mono-energetic radiation. A strategy is suggested for accommodating strongly attenuating materials such as calculi and metallic implants.

Capture and analysis of radiation dose reports for radiology.

Radiographic imaging systems can produce records of exposure and dose parameters for each patient. A variety of file formats are in use including plain text, bit map images showing pictures of written text and radiation dose structured reports as text or extended markup language files. Whilst some of this information is available with image data on the hospital picture archive and communication system, access is restricted to individual patient records, thereby making it difficult to locate multiple records for the same scan protocol. This study considers the exposure records and dose reports from four modalities. Exposure records for mammography and general radiography are utilized for repeat analysis. Dose reports for fluoroscopy and computed tomography (CT) are utilized to study the distribution of patient doses for each protocol. Results for dosimetric quantities measured by General Radiography, Fluoroscopy and CT equipment are summarised and presented in the Appendix. Projection imaging uses the dose (in air) area product and derived quantities including the dose to the reference point as a measure of the air kerma reaching the skin, ignoring movement of the beam for fluoroscopy. CT uses the dose indices CTDIvol and dose length product as a measure of the dose per axial slice, and to the scanned volume. Suitable conversion factors are identified and used to estimate the effective dose to an average size patient (for CT and fluoroscopy) and the entrance skin dose for fluoroscopy.

Feasibility study for DEXA using synchrotron CT at 20-35 keV.

A nonlinear model for the x-ray linear attenuation coefficient µ is employed for dual energy x-ray analysis (DEXA). Nonlinear simultaneous equations formed by µ and energy dependent model parameters are solved for the electron density N(e) and fourth compositional ratio R(4) which has the same 'units' as the atomic number. Computed tomography data was acquired at 20-35 keV using bending magnet synchrotron radiation, a double crystal monochromator, a rotation stage and an area detector. Test objects contained liquid samples as mixtures of ethanol, water and salt solutions with known density and composition. Various noise sources are identified and give µ uncertainties of 1-2%. A fan beam geometry allowed the detection of forward scattered radiation with measured µ being 6% lower than expectations for a narrow beam. Energy dependent model parameters were obtained by solving linear simultaneous equations formed by µ and material parameters based upon N(e) and R(4). DEXA accuracy was studied as a function of photon energy and sample composition. Propagation of errors analysis identifies the importance of the fractional compositional cross-products whose difference at the two beam energies should exceed 0.1, requiring 10 keV or more separation. For a reasonable approximation for the adjustable model parameters, the mean difference between the DEXA solution and true values (ΔN(e), ΔR(4)) are (1.0%, 0.5%) for soft tissue and (1.5%, 0.8%) for bone like samples.

Estimating photon interaction coefficients from single energy x-ray CT.

Single energy x-ray analysis is explored in the context of computed tomography (CT), whereby Hounsfield numbers (HN) are used to estimate electron density N(e) and parameters that describe composition. We examine measurements with tissue substitute materials and theoretical HN for a broad range of tissues. Results are combined with parametric models for the x-ray linear attenuation coefficient µ and energy absorption coefficient µ(en) to predict values at energies 10 keV to 20 MeV. At photon energies employed for CT, the fractional contribution to µ from composition is 0.1-0.4 for soft tissues to bone respectively, and is responsible for strong correlations between HN and N(e). The atomic density of tissues excluding lung is near constant allowing the models to be re-expressed as a function of N(e) alone. The transformed model is subjected to propagation of error analysis and results are presented as the ratio of uncertainties for µ or µ(en) to those for N(e). For soft tissues to bone the ratios are as follows: at photon energies 20-100 keV the ratio is 5.0-2.0, at intermediate energies it is unity and increases above 4 MeV to reach 1.5-2.0 at 20 MeV. Results are discussed in the context of attenuation correction and dosimetry calculations for the same range of photon energies.

Assessment of patient dose and image quality for cardiac CT with breast shields.

Breast shielding can reduce dose to the female breast, a radiosensitive organ receiving significant radiation during computed tomography (CT) chest examinations, particularly in cardiac CT, where Electrocardiogram dose modulation currently precludes the use of radial dose modulation to reduce breast dose. However, breast shields may produce artefacts affecting interpretation of coronary arteries. This study explores the dose savings and the effect of breast shields on image quality with torso and CT dose index body phantoms and an organ dose calculator. Change in dose calculated: 53-63 % (female breast), 82-85 % (lung), 79-84 % (oesophagus) and 76-80 % (effective dose) with larger dose reductions at lower kVp. Image quality is preserved when breast shields are placed after the scout no closer than 10 mm from the skin. Therefore, breast shields can be used in cardiac CT to reduce breast dose without compromising image quality. Revised conversion factors for dose length product to effective dose are suggested for cardiac CT without and with breast shields.

Feasibility study for a novel method of dual energy x-ray analysis.

Dual energy x-ray analysis (DEXA) is investigated using a nonlinear model for the x-ray linear attenuation coefficient µ that is expressed as a function of electron density N(e) and the fourth compositional ratio R4. Nonlinear simultaneous equations are solved using a least-squares algorithm based upon the method of Levenberg and Marquardt. Measurements of µ for low atomic number materials (containing elements hydrogen to calcium) at energies 32-66 keV are used to study DEXA accuracy as a function of sample composition, photon energy and their separation ΔE. Results are presented for ΔE = 5-30 keV, for 2% measurement precision, and the doses involved are quantified. The model is subject to propagation of error analysis and results are presented for the relationship between measurement uncertainties and those for N(e) and R4. The analysis shows how DEXA accuracy is controlled by the fractional compositional cross-product, which represents the contribution of composition to µ, and how this can be optimized by careful selection of beam energies according to the compositional range of interest. Accurate DEXA is achieved over restricted energy and compositional ranges: soft tissues only at approximately 15-25 keV, all tissues at approximately 30-80 keV and, for situations where a higher dose can be tolerated, all tissues at approximately 4-8 MeV.

A model for multi-energy x-ray analysis.

Multi-energy x-ray analysis (MEXA) uses measurements of the x-ray linear attenuation coefficient µ, obtained at different photon energies to determine parameters that characterize the density and composition of materials. The key to achieving this goal is an accurate parameterization for µ, allowing measurements to be written as simultaneous equations and then solved. This author has reported such a model where mixtures are characterized by four or more statistical moments that describe the distribution of atomic number. These can be re-expressed as the product of the electron density N(e) and four or more compositional ratios R(k) with the same 'units' as atomic number (i.e. dimensionless). The model was turned to MEXA where it delivered reliable estimates for N(e) and R(4) and not the intermediate compositional ratios. This report studies the relationships between compositional ratios for tissues and tissue substitute materials. Correlations are identified leading to a new parameterization that is expressed as a nonlinear function of N(e), R(4) and other coefficients. The properties of the transformed parameterizations for µ and the energy absorption coefficient µ(en) are considered for low atomic number materials at energies 15-100 keV, and for a broader range of materials at energies 5 keV to 20 MeV. The interpretation of the parameters N(e) and R(4) is explored in terms of basis materials. The general case of three basis materials cannot be solved for all contributions, but the special case of just two basis materials can be fully solved.

A method of dosimetry for synchrotron microbeam radiation therapy using radiochromic films of different sensitivity.

This paper describes a method of film dosimetry used to measure the peak-to-valley dose ratios for synchrotron microbeam radiation therapy (MRT). Two types of radiochromic film (manufactured by International Specialty Products, NJ, USA) were irradiated in a phantom and also flush against a microbeam collimator (beam width 25 microm, centre-to-centre spacing 200 microm) on beamline BL28 B2 at the SPring-8 synchrotron. Four experiments are reported: (1) the HD-810 and EBT varieties of radiochromic film were used to record 'peak' dose and 'valley' (regions in between peaks) dose, respectively; (2) a stack of HD-810 film sheets was microbeam-irradiated and analysed to investigate a possible dose build-up effect; (3) a very high MRT dose was delivered to HD-810 film to elicit a measurable valley dose to compare with the result obtained using broad beam radiation; (4) the half value layer of the beam with and without the microbeam collimator was measured to investigate the effect of the collimator on the beam quality. The valley dose obtained for films placed flush against the collimator was approximately 0.2% of the peak dose. Within the water phantom, the valley dose had increased to between 0.7 and 1.8% of the peak dose, depending on the depth in the phantom. We also demonstrated, experimentally and by Monte Carlo simulation, that the dose is not maximal on the surface and that there is a dose build-up effect. The microbeam collimator did not make an appreciable difference to the beam quality. The values of the peak-to-valley ratio reported in this paper are higher than those predicted by previously published Monte Carlo simulation papers.

Materials analysis using x-ray linear attenuation coefficient measurements at four photon energies.

The analytical properties of an accurate parameterization scheme for the x-ray linear attenuation coefficient are examined. The parameterization utilizes an additive combination of N compositional- and energy-dependent coefficients. The former were derived from a parameterization of elemental cross-sections using a polynomial in atomic number. The compositional-dependent coefficients are referred to as the mixture parameters, representing the electron density and higher order statistical moments describing elemental distribution. Additivity is an important property of the parameterization, allowing measured x-ray linear attenuation coefficients to be written as linear simultaneous equations, and then solved for the unknown coefficients. The energy-dependent coefficients can be determined by calibration from measurements with materials of known composition. The inverse problem may be utilized for materials analysis, whereby the simultaneous equations represent multi-energy linear attenuation coefficient measurements, and are solved for the mixture parameters. For in vivo studies, the choice of measurement energies is restricted to the diagnostic region (approximately 20 keV to 150 keV), where the parameterization requires N >or= 4 energies. We identify a mathematical pathology that must be overcome in order to solve the inverse problem in this energy regime. An iterative inversion strategy is presented for materials analysis using four or more measurements, and then tested against real data obtained at energies 32 keV to 66 keV. The results demonstrate that it is possible to recover the electron density to within +/-4% and fourth mixture parameter. It is also a key finding that the second and third mixture parameters cannot be recovered, as they are of minor importance in the parameterization at diagnostic x-ray energies.

A parameterization scheme for the x-ray linear attenuation coefficient and energy absorption coefficient.

A novel parameterization of x-ray interaction cross-sections is developed, and employed to describe the x-ray linear attenuation coefficient and mass energy absorption coefficient for both elements and mixtures. The new parameterization scheme addresses the Z-dependence of elemental cross-sections (per electron) using a simple function of atomic number, Z. This obviates the need for a complicated mathematical formalism. Energy dependent coefficients describe the Z-direction curvature of the cross-sections. The composition dependent quantities are the electron density and statistical moments describing the elemental distribution. We show that it is possible to describe elemental cross-sections for the entire periodic table and at energies above the K-edge (from 6 keV to 125 MeV), with an accuracy of better than 2% using a parameterization containing not more than five coefficients. For the biologically important elements 1 < or = Z < or = 20, and the energy range 30-150 keV, the parameterization utilizes four coefficients. At higher energies, the parameterization uses fewer coefficients with only two coefficients needed at megavoltage energies.