The direct use of CT numbers in radiotherapy dosage calculations for inhomogeneous media

Estimation of size-specific dose estimate (SSDE) of CT scans using an effective diameter electron density

OBJECTIVE

An assessment of the effective diameter of a patient's body using electron densities of tissues inside the scan area (Deffρe) was proposed to overcome challenges associated with the estimation of water-equivalent diameter (Dw), which is used for size-specific dose estimate (SSDE). The aims of this study were to (1) investigate the Deffρe method in two different forms using a wide range of patient sizes and scanning protocols, and (2) compare between four methods used to estimate the patient size for SSDE.

MATERIALS AND METHODS

Under IRB approval, a total of 350 patients of varying sizes have been collected retrospectively from the Hospital. The Dw values were assessed over six different CT body protocols: (1) chest with contrast media, (2) chest High-Resolution Computed Tomography (HRCT) without contrast media, (3) abdomen-pelvis with contrast media, (4) abdomen-pelvis without contrast media, (5) chest-abdomen-pelvis with contrast media, and (6) pelvis without contrast media. A MATLAB-based code was developed in-house to assess the size of each patient using the conventional effective diameter method (Deff), Deffρe by correcting either both the lateral (LAT) and anterior-posterior (AP) dimensions (Deff,LAT+APρe) or LAT only (Deff,LATρe), and Dw at the mid-CT slice of the patient images.

RESULTS

The results of Deff,LAT+APρe and Deff,LATρe provided a better estimation for the chest protocols with the averages of absolute percentage difference (PD) values in the range of 3 - 7 % for all patient sizes as compared to the Dw method, whereas the averages of PD values for the Deff method were 9 - 15 %. However, Deff gave a better estimation for Dw values for the other body protocols, with differences of 2 - 4 %, which were lower than those obtained with the Deff,LAT+APρe and Deff,LATρe methods. For the chest protocols, statistically significant differences were found between Deff and the other methods, but there were no significant differences between all the methods for the other scanning protocols. The results show that the correction of both dimensions, LAT and AP, did not improve the accuracy of the Deffρe method, and, for most protocols, Deff,LAT+APρe gave larger range differences compared to those based on correction of the LAT dimension only.

CONCLUSION

If the Dw cannot be assessed, the Deff,LATρe method may only be considered for the chest protocols as an alternative approach. The Deff method may also be used for all regions taking into account the application of a correction factor for the chest protocols to avoid a significant under or overestimation of the patient dose.

Generation of synthetic megavoltage CT for MRI-only radiotherapy treatment planning using a 3D deep convolutional neural network

BACKGROUND

Megavoltage computed tomography (MVCT) has been implemented on many radiotherapy treatment machines for on-board anatomical visualization, localization, and adaptive dose calculation. Implementing an MR-only workflow by synthesizing MVCT from magnetic resonance imaging (MRI) would offer numerous advantages for treatment planning and online adaptation.

PURPOSE

In this work, we sought to synthesize MVCT (sMVCT) datasets from MRI using deep learning to demonstrate the feasibility of MRI-MVCT only treatment planning.

METHODS

MVCTs and T1-weighted MRIs for 120 patients treated for head-and-neck cancer were retrospectively acquired and co-registered. A deep neural network based on a fully-convolutional 3D U-Net architecture was implemented to map MRI intensity to MVCT HU. Input to the model were volumetric patches generated from paired MRI and MVCT datasets. The U-Net was initialized with random parameters and trained on a mean absolute error (MAE) objective function. Model accuracy was evaluated on 18 withheld test exams. sMVCTs were compared to respective MVCTs. Intensity-modulated volumetric radiotherapy (IMRT) plans were generated on MVCTs of four different disease sites and compared to plans calculated onto corresponding sMVCTs using the gamma metric and dose-volume-histograms (DVHs).

RESULTS

MAE values between sMVCT and MVCT datasets were 93.3 ± 27.5, 78.2 ± 27.5, and 138.0 ± 43.4 HU for whole body, soft tissue, and bone volumes, respectively. Overall, there was good agreement between sMVCT and MVCT, with bone and air posing the greatest challenges. The retrospective dataset introduced additional deviations due to sinus filling or tumor growth/shrinkage between scans, differences in external contours due to variability in patient positioning, or when immobilization devices were absent from diagnostic MRIs. Dose distributions of IMRT plans evaluated for four test cases showed close agreement between sMVCT and MVCT images when evaluated using DVHs and gamma dose metrics, which averaged to 98.9 ± 1.0% and 96.8 ± 2.6% analyzed at 3%/3 mm and 2%/2 mm, respectively.

CONCLUSIONS

MVCT datasets can be generated from T1-weighted MRI using a 3D deep convolutional neural network with dose calculation on a sample sMVCT in close agreement with the MVCT. These results demonstrate the feasibility of using MRI-derived sMVCT in an MR-only treatment planning workflow.

Dual-energy CT: minimal essentials for radiologists

Dual-energy CT, the object is scanned at two different energies, makes it possible to identify the characteristics of materials that cannot be evaluated on conventional single-energy CT images. This imaging method can be used to perform material decomposition based on differences in the material-attenuation coefficients at different energies. Dual-energy analyses can be classified as image data-based- and raw data-based analysis. The beam-hardening effect is lower with raw data-based analysis, resulting in more accurate dual-energy analysis. On virtual monochromatic images, the iodine contrast increases as the energy level decreases; this improves visualization of contrast-enhanced lesions. Also, the application of material decomposition, such as iodine- and edema images, increases the detectability of lesions due to diseases encountered in daily clinical practice. In this review, the minimal essentials of dual-energy CT scanning are presented and its usefulness in daily clinical practice is discussed.

A simple manual method to estimate water-equivalent diameter for calculating size-specific dose estimate in chest computed tomography

OBJECTIVES

The American Association of Physicists in Medicine (AAPM) Task Groups (TG) 204 and 220 introduced a method to estimate patient dose by introducing the Size-Specific Dose Estimate (SSDE). They provided patient size-specific conversion factors that could be applied to volumetric CT Dose Index CTDIvol to estimate patient dose in terms of SSDE based on either effective diameter (Deff) or water equivalent diameter (Dw). Our study presented an alternative method to manually estimate SSDE for the everyday clinical routine chest CT that can be readily used and does not require sophisticated computer programming.

METHODS

For 16 adult patients undergoing chest CT, the method employed an average relative electron density (ρelung = 0.3) for the lung tissue and a ρetissue of 1.0 for the other tissues to scale the lateral thickness and compute the effective lateral thickness on the patient's axial image. The proposed method estimated a "corrected" Deff (Deffcorr) to replace Dw and compared results with TG220 and a second method proposed by Huda et al, for the same set of CT studies.

RESULTS

The results showed comparable behavior for all methods. There is overall agreement especially between this study and TG220. Largest differences were +13.3% and+15.9% from TG220 and Huda values, respectively. Patient size correlation showed strong correlation with the TG220 and Huda et al methods.

CONCLUSIONS

A simple, quick manual method to estimate CT patient radiation dose in terms of SSDE was proposed as an alternative where sophisticated computer programming is not available. It can be readily used during any clinical chest CT scanning.

ADVANCES IN KNOWLEDGE

The paper is novel as it presents simple, quick manual method to estimate CT patient radiation dose in chest imaging. The process can be used as alternative in cases no sophisticated computer programming is available.

On the molecular relationship between Hounsfield Unit (HU), mass density, and electron density in computed tomography (CT)

Accurate determination of physical/mass and electron densities are critical to accurate spatial and dosimetric delivery of radiotherapy for photon and charged particles. In this manuscript, the biology, chemistry, and physics that underly the relationship between computed tomography (CT) Hounsfield Unit (HU), mass density, and electron density was explored. In standard radiation physics practice, quantities such as mass and electron density are typically calculated based off a single kilovoltage CT (kVCT) scan assuming a one-to-one relationship between HU and density. It is shown that, in absence of mass density assumptions on tissues, the relationship between HU and density is not one-to-one with uncertainties as large as 7%. To mitigate this uncertainty, a novel multi-dimensional theoretical approach is defined between molecular (water, lipid, protein, and mineral) composition, HU, mass density, and electron density. Empirical parameters defining this relationship are x-ray beam energy/spectrum dependent and, in this study, two methods are proposed to solve for them including through a tissue mimicking phantom calibration process. As a proof of concept, this methodology was implemented in a separate in-house created tissue mimicking phantom and it is shown that sub 1% accuracy is possible for both mass and electron density. As molecular composition is not always known, the sensitivity of this model to uncertainties in molecular composition was investigated and it was found that, for soft tissue, sub 1% accuracy is achievable assuming nominal organ/tissue compositions. For boney tissues, the uncertainty in mineral content may lead to larger errors in mass and electron density compared with soft tissue. In this manuscript, a novel methodology to directly determine mass and electron density based off CT HU and knowledge of molecular compositions is presented. If used in conjunction with a methodology to determine molecular compositions, mass and electron density can be accurately calculated from CT HU.

Multi-energy computed tomography and material quantification: Current barriers and opportunities for advancement

Computed tomography (CT) technology has rapidly evolved since its introduction in the 1970s. It is a highly important diagnostic tool for clinicians as demonstrated by the significant increase in utilization over several decades. However, much of the effort to develop and advance CT applications has been focused on improving visual sensitivity and reducing radiation dose. In comparison to these areas, improvements in quantitative CT have lagged behind. While this could be a consequence of the technological limitations of conventional CT, advanced dual-energy CT (DECT) and photon-counting detector CT (PCD-CT) offer new opportunities for quantitation. Routine use of DECT is becoming more widely available and PCD-CT is rapidly developing. This review covers efforts to address an unmet need for improved quantitative imaging to better characterize disease, identify biomarkers, and evaluate therapeutic response, with an emphasis on multi-energy CT applications. The review will primarily discuss applications that have utilized quantitative metrics using both conventional and DECT, such as bone mineral density measurement, evaluation of renal lesions, and diagnosis of fatty liver disease. Other topics that will be discussed include efforts to improve quantitative CT volumetry and radiomics. Finally, we will address the use of quantitative CT to enhance image-guided techniques for surgery, radiotherapy and interventions and provide unique opportunities for development of new contrast agents.

Evaluation of raw-data-based and calculated electron density for contrast media with a dual-energy CT technique

Objectives

The aim of the current study is to evaluate the accuracy and the precision of raw-data-based relative electron density (RED_raw) and the calibration-based RED (RED_cal) at a range of low-RED to high-RED for tissue-equivalent phantom materials by comparing them with reference RED (RED_ref) and to present the difference of RED_raw and RED_cal for the contrast medium using dual-energy CT (DECT).

Methods

The RED_raw images were reconstructed by raw-data-based decomposition using DECT. For evaluation of the accuracy of the RED_raw, RED_ref was calculated for the tissue-equivalent phantom materials based on their specified density and elemental composition. The RED_cal images were calculated using three models: Lung-Bone model, Lung-Ti model and Lung-Ti (SEMAR) model which used single-energy metal artifact reduction (SEMAR). The difference between RED_raw and RED_cal was calculated.

Results

In the titanium rod core, the deviations of RED_raw and RED_cal (Lung-Bone model, Lung-Ti model and Lung-Ti model with SEMAR) from RED_ref were 0.45%, 50.8%, 15.4% and 15.0%, respectively. The largest differences between RED_raw and RED_cal (Lung-Bone model, Lung-Ti model and Lung-Ti model with SEMAR) in the contrast medium phantom were 8.2%, -23.7%, and 28.7%, respectively. However, the differences between RED_raw and RED_cal values were within 10% at 20 mg/ml. The standard deviation of the RED_raw was significantly smaller than the RED_cal with three models in the titanium and the materials that had low CT numbers.

Conclusion

The RED_cal values could be affected by beam hardening artifacts and the RED_cal was less accurate than RED_raw for high-Z materials as titanium.

Advances in knowledge

The raw-data-based reconstruction method could reduce the beam hardening artifact compared with image-based reconstruction and increase the accuracy for the RED estimation in high-Z materials, such as titanium and iodinated contrast medium.

Simulation of photon-counting detectors for conversion of dual-energy-subtracted computed tomography number to electron density

For accurate tissue-inhomogeneity correction in radiotherapy treatment planning, the author previously proposed a conversion of the energy-subtracted computed tomography (CT) number to electron density (ΔHU-ρ_e conversion). The purpose of the present study was to provide a method for investigating the accuracy of a photon-counting detector (PCD) used in the ΔHU-ρ_e conversion by performing dual-energy CT image simulations of a PCD system with two energy bins. To optimize the tube voltage and threshold energy, the image noise and errors in ρ_e calibration were evaluated using three types of virtual phantoms: a 35-cm-diameter pure water phantom, 33-cm-diameter solid water surrogate phantom equipped with 16 inserts, and another solid water surrogate phantom with a 25-cm diameter. The third phantom was used to investigate the effect of the object's size on the ρ_e-calibration accuracy of PCDs. Two different scenarios for the PCD energy response were considered, corresponding to the ideal and realistic cases. In addition, a simple correction method for improving the spectral separation of the dual energies in a realistic PCD was proposed to compensate for its performance loss. In the realistic PCD case, there exists a trade-off between the image noise and ρ_e-calibration errors. Furthermore, the weakest image noise was nearly twice that for the ideal case, and the ρ_e-calibration error did not reach practical levels for any threshold energy. Nevertheless, the proposed correction method is likely to decrease the ρ_e-calibration errors of a realistic PCD to the level of the ideal case, yielding more accurate ρ_e values that are less affected by object size variation.

Simultaneous characterization of electron density and effective atomic number for radiotherapy planning using stoichiometric calibration method and dual energy algorithms

Relative electron densities of body tissues (ρ_e) for radiotherapy treatment planning are normally obtained by CT scanning of tissue substitute materials (TSMs) and producing a Hounsfield Unit-ρ_e calibration curve. Aiming for more accurate, simultaneous characterization of ρ_e and effective atomic number (Z_eff) of real tissues, an in-house phantom (including 10 water solutions plus composite cork as TSMs) was constructed and scanned at 4 kVps. Dual-energy algorithms were applied to 80-140 and 100-140 kVp combination scans, for better differentiation of tissues with same attenuation coefficient at 120 kVp but different ρ_e and Z_eff. Stoichiometric calibration and closeness of the ρ_e of the 11 TSMs to real tissues (≤ 0.5%) resulted in smaller ρ_e calculation discrepancies, compared to studies with commercial phantoms (p < 0.024). Applying an energy subtraction algorithm further mitigated errors by spectral separation and reduction of beam hardening artifacts and noise, reducing the mean and standard deviation of the absolute difference of ρ_e at 80-140 kVp (p < 0.003) and 100-140 kVp (p < 0.0001) scans, compared to 120 kVp scan, respectively. Moreover, a parametrization algorithm decreased the Z_eff discrepancy from real tissues at 80-140 kVp scans; for thyroid, the residual error was ≤ 0.18 units of Z_eff (vs. 0.2 with the Gammex 467 phantom from a previous study). These results further suggest that a dual-energy algorithm in combination with stoichiometry can decrease errors in calculation of the ρ_e of real tissues to ameliorate inhomogeneity for dose calculation in radiotherapy treatment planning, especially when the energy spectrum of the X-ray tube of the CT machine is not available.

Technical Note: exploring the limit for the conversion of energy-subtracted CT number to electron density for high-atomic-number materials

PURPOSE

For accurate tissue inhomogeneity correction in radiotherapy treatment planning, the authors had previously proposed a novel conversion of the energy-subtracted CT number to an electron density (ΔHU-ρe conversion), which provides a single linear relationship between ΔHU and ρe over a wide ρe range. The purpose of this study is to address the limitations of the conversion method with respect to atomic number (Z) by elucidating the role of partial photon interactions in the ΔHU-ρe conversion process.

METHODS

The authors performed numerical analyses of the ΔHU-ρe conversion for 105 human body tissues, as listed in ICRU Report 46, and elementary substances with Z = 1-40. Total and partial attenuation coefficients for these materials were calculated using the XCOM photon cross section database. The effective x-ray energies used to calculate the attenuation were chosen to imitate a dual-source CT scanner operated at 80-140 kV/Sn under well-calibrated and poorly calibrated conditions.

RESULTS

The accuracy of the resultant calibrated electron density,[Formula: see text], for the ICRU-46 body tissues fully satisfied the IPEM-81 tolerance levels in radiotherapy treatment planning. If a criterion of [Formula: see text]ρe - 1 is assumed to be within ± 2%, the predicted upper limit of Z applicable for the ΔHU-ρe conversion under the well-calibrated condition is Z = 27. In the case of the poorly calibrated condition, the upper limit of Z is approximately 16. The deviation from the ΔHU-ρe linearity for higher Z substances is mainly caused by the anomalous variation in the photoelectric-absorption component.

CONCLUSIONS

Compensation among the three partial components of the photon interactions provides for sufficient linearity of the ΔHU-ρe conversion to be applicable for most human tissues even for poorly conditioned scans in which there exists a large variation of effective x-ray energies owing to beam-hardening effects arising from the mismatch between the sizes of the object and the calibration phantom.

Dosimetry of a (90)Y-hydroxide liquid brachytherapy treatment approach to canine osteosarcoma using PET/CT

A new treatment strategy based on direct injections of (90)Y-hydroxide into the tumor bed in dogs with osteosarcoma was studied. Direct injections of the radiopharmaceutical into the tumor bed were made according to a pretreatment plan established using (18)F-FDG images. Using a special drill, cannulas were inserted going through tissue, tumor and bone. Using these cannulas, direct injections of the radiopharmaceutical were made. The in vivo biodistribution of (90)Y-hydroxide and the anatomical tumor bed were imaged using a time-of-flight (TOF) PET/CT scanner. The material properties of the tissues were estimated from corresponding CT numbers using an electron-density calibration. Radiation absorbed dose estimates were calculated using Monte Carlo methods where the biodistribution of the pharmaceutical from PET images was sampled using a collapsing 3-D rejection technique. Dose distributions in the tumor bed and surrounding tissues were calculated, showing significant heterogeneity with multiple hot spots at injection sites. Dose volume histograms showed that approximately 33.9% of bone and tumor and 70.2% of bone marrow and trabecular bone received an absorbed dose over 200Gy; approximately 3.2% of bone and tumor and 31.0% of bone marrow and trabecular bone received a total dose of over 1000Gy.

Measurement of electron density and effective atomic number by dual-energy scan using a 320-detector computed tomography scanner with raw data-based analysis: a phantom study

We evaluated the accuracy of the electron densities and effective atomic numbers determined by raw data-based dual-energy analysis on a 320-detector computed tomography scanner. The mean (SD) errors between the measured and true electron densities and between the measured and true effective atomic numbers were 1.3% (1.5%) and 3.1% (3.2%), respectively. Electron densities and effective atomic numbers can be determined with high accuracy, which may help to improve accuracy in radiotherapy treatment planning.

Computed tomography as a source of electron density information for radiation treatment planning

PURPOSE

To evaluate the performance of computed tomography (CT) systems of various designs as a source of electron density (rho(el)) data for treatment planning of radiation therapy.

MATERIAL AND METHODS

Dependence of CT numbers on relative electron density of tissue-equivalent materials (HU-rho(el) relationship) was measured for several general-purpose CT systems (single-slice, multislice, wide-bore multislice), for radiotherapy simulators with a single-slice CT and kV CBCT (cone-beam CT) options, as well as for linear accelerators with kV and MV CBCT systems. Electron density phantoms of four sizes were used. Measurement data were compared with the standard HU-rhoel relationships predefined in two commercial treatment-planning systems (TPS).

RESULTS

The HU-rho(el) relationships obtained with all of the general-purpose CT scanners operating at voltages close to 120 kV were very similar to each other and close to those predefined in TPS. Some dependency of HU values on tube voltage was observed for bone- equivalent materials. For a given tube voltage, differences in results obtained for different phantoms were larger than those obtained for different CT scanners. For radiotherapy simulators and for kV CBCT systems, the information on rhoel was much less precise because of poor uniformity of images. For MV CBCT, the results were significantly different than for kV systems due to the differing energy spectrum of the beam.

CONCLUSION

The HU-rho(el) relationships predefined in TPS can be used for general-purpose CT systems operating at voltages close to 120 kV. For nontypical imaging systems (e.g., CBCT), the relationship can be significantly different and, therefore, it should always be measured and carefully analyzed before using CT data for treatment planning.

Investigation of the usability of conebeam CT data sets for dose calculation

BACKGROUND

To investigate the feasibility and accuracy of dose calculation in cone beam CT (CBCT) data sets.

METHODS

Kilovoltage CBCT images were acquired with the Elekta XVI system, CT studies generated with a conventional multi-slice CT scanner (Siemens Somatom Sensation Open) served as reference images. Material specific volumes of interest (VOI) were defined for commercial CT Phantoms (CATPhan and Gammex RMI) and CT values were evaluated in CT and CBCT images. For CBCT imaging, the influence of image acquisition parameters such as tube voltage, with or without filter (F1 or F0) and collimation on the CT values was investigated. CBCT images of 33 patients (pelvis n = 11, thorax n = 11, head n = 11) were compared with corresponding planning CT studies. Dose distributions for three different treatment plans were calculated in CT and CBCT images and differences were evaluated. Four different correction strategies to match CT values (HU) and density (D) in CBCT images were analysed: standard CT HU-D table without adjustment for CBCT; phantom based HU-D tables; patient group based HU-D tables (pelvis, thorax, head); and patient specific HU-D tables.

RESULTS

CT values in the CBCT images of the CATPhan were highly variable depending on the image acquisition parameters: a mean difference of 564 HU +/- 377 HU was calculated between CT values determined from the planning CT and CBCT images. Hence, two protocols were selected for CBCT imaging in the further part of the study and HU-D tables were always specific for these protocols (pelvis and thorax with M20F1 filter, 120 kV; head S10F0 no filter, 100 kV). For dose calculation in real patient CBCT images, the largest differences between CT and CBCT were observed for the standard CT HU-D table: differences were 8.0% +/- 5.7%, 10.9% +/- 6.8% and 14.5% +/- 10.4% respectively for pelvis, thorax and head patients using clinical treatment plans. The use of patient and group based HU-D tables resulted in small dose differences between planning CT and CBCT: 0.9% +/- 0.9%, 1.8% +/- 1.6%, 1.5% +/- 2.5% for pelvis, thorax and head patients, respectively. The application of the phantom based HU-D table was acceptable for the head patients but larger deviations were determined for the pelvis and thorax patient populations.

CONCLUSION

The generation of three HU-D tables specific for the anatomical regions pelvis, thorax and head and specific for the corresponding CBCT image acquisition parameters resulted in accurate dose calculation in CBCT images. Once these HU-D tables are created, direct dose calculation on CBCT datasets is possible without the need of a reference CT images for pixel value calibration.

Study of the scan uniformity from an i-CAT cone beam computed tomography dental imaging system

OBJECTIVES

As part of an ongoing programme to improve diagnosis and treatment planning relevant to implant placement, orthodontic treatment and dentomaxillofacial surgery, a study has been made of the spatial accuracy and density response of an i-CAT, a cone beam CT (CBCT) dental imaging system supplied by Imaging Sciences International Inc.

METHODS

Custom-made phantoms using acrylic sheet and water were used for measurements on spatial accuracy, density response and noise. The measurements were made over a period of several months on a clinical machine rather than on a machine dedicated to research.

RESULTS

Measurements on a precision grid showed the spatial accuracy to be universally within the tolerance of +/-1 pixel. The density response and the noise in the data were found to depend strongly on the mass in the slice being scanned.

CONCLUSIONS

The density response was subject to two effects. The first effect changes the whole slice uniformly and linearly depends on the total mass in the slice. The second effect exists when there is mass outside the field of view, dubbed the "exo-mass" effect. This effect lowers the measured CT number rapidly at the scan edge furthest from the exo-mass and raises it on the adjacent edge. The noise also depended quasi-linearly on the mass in the slice. Some general performance rules were drafted to describe these effects and a preliminary correction algorithm was constructed.

Anatomical imaging for radiotherapy

The goal of radiation therapy is to achieve maximal therapeutic benefit expressed in terms of a high probability of local control of disease with minimal side effects. Physically this often equates to the delivery of a high dose of radiation to the tumour or target region whilst maintaining an acceptably low dose to other tissues, particularly those adjacent to the target. Techniques such as intensity modulated radiotherapy (IMRT), stereotactic radiosurgery and computer planned brachytherapy provide the means to calculate the radiation dose delivery to achieve the desired dose distribution. Imaging is an essential tool in all state of the art planning and delivery techniques: (i) to enable planning of the desired treatment, (ii) to verify the treatment is delivered as planned and (iii) to follow-up treatment outcome to monitor that the treatment has had the desired effect. Clinical imaging techniques can be loosely classified into anatomic methods which measure the basic physical characteristics of tissue such as their density and biological imaging techniques which measure functional characteristics such as metabolism. In this review we consider anatomical imaging techniques. Biological imaging is considered in another article. Anatomical imaging is generally used for goals (i) and (ii) above. Computed tomography (CT) has been the mainstay of anatomical treatment planning for many years, enabling some delineation of soft tissue as well as radiation attenuation estimation for dose prediction. Magnetic resonance imaging is fast becoming widespread alongside CT, enabling superior soft-tissue visualization. Traditionally scanning for treatment planning has relied on the use of a single snapshot scan. Recent years have seen the development of techniques such as 4D CT and adaptive radiotherapy (ART). In 4D CT raw data are encoded with phase information and reconstructed to yield a set of scans detailing motion through the breathing, or cardiac, cycle. In ART a set of scans is taken on different days. Both allow planning to account for variability intrinsic to the patient. Treatment verification has been carried out using a variety of technologies including: MV portal imaging, kV portal/fluoroscopy, MVCT, conebeam kVCT, ultrasound and optical surface imaging. The various methods have their pros and cons. The four x-ray methods involve an extra radiation dose to normal tissue. The portal methods may not generally be used to visualize soft tissue, consequently they are often used in conjunction with implanted fiducial markers. The two CT-based methods allow measurement of inter-fraction variation only. Ultrasound allows soft-tissue measurement with zero dose but requires skilled interpretation, and there is evidence of systematic differences between ultrasound and other data sources, perhaps due to the effects of the probe pressure. Optical imaging also involves zero dose but requires good correlation between the target and the external measurement and thus is often used in conjunction with an x-ray method. The use of anatomical imaging in radiotherapy allows treatment uncertainties to be determined. These include errors between the mean position at treatment and that at planning (the systematic error) and the day-to-day variation in treatment set-up (the random error). Positional variations may also be categorized in terms of inter- and intra-fraction errors. Various empirical treatment margin formulae and intervention approaches exist to determine the optimum strategies for treatment in the presence of these known errors. Other methods exist to try to minimize error margins drastically including the currently available breath-hold techniques and the tracking methods which are largely in development. This paper will review anatomical imaging techniques in radiotherapy and how they are used to boost the therapeutic benefit of the treatment.

The impact of peak-kilovoltage settings on heterogeneity-corrected photon-beam treatment plans

Differences were evaluated in external-beam treatment plan dose calculations that result from the use of different Hounsfield-unit to electron-density conversion curves with CT images acquired with various tube potentials. These differences were found to be clinically insignificant and it was concluded that the impact of CT tube potential on treatment planning is negligible.

Dosimetric impact of a CT metal artefact suppression algorithm for proton, electron and photon therapies

Accurate dose calculation is essential to precision radiation treatment planning and this accuracy depends upon anatomic and tissue electron density information. Modern treatment planning inhomogeneity corrections use x-ray CT images and calibrated scales of tissue CT number to electron density to provide this information. The presence of metal in the volume scanned by an x-ray CT scanner causes metal induced image artefacts that influence CT numbers and thereby introduce errors in the radiation dose distribution calculated. This paper investigates the dosimetric improvement achieved by a previously proposed x-ray CT metal artefact suppression technique when the suppressed images of a patient with bilateral hip prostheses are used in commercial treatment planning systems for proton, electron or photon therapies. For all these beam types, this clinical image and treatment planning study reveals that the target may be severely underdosed if a metal artefact-contaminated image is used for dose calculations instead of the artefact suppressed one. Of the three beam types studied, the metal artefact suppression is most important for proton therapy dose calculations, intermediate for electron therapy and least important for x-ray therapy but still significant. The study of a water phantom having a metal rod simulating a hip prosthesis indicates that CT numbers generated after image processing for metal artefact suppression are accurate and thus dose calculations based on the metal artefact suppressed images will be of high fidelity.

Tolerances for the accuracy of photon beam dose calculations of treatment planning systems

BACKGROUND AND PURPOSE

To design a consistent set of criteria for acceptability of photon beam dose calculations of treatment planning systems. The set should be applicable in combination with a test package used for evaluation of a treatment planning system, such as the ones proposed by the AAPM Task Group 23 or by the Netherlands Commission on Radiation Dosimetry.

RESULTS

Tolerances have been defined for the accuracy with which a treatment planning system should be able to calculate the dose in different parts of a photon beam: the central beam axis and regions with large and small dose gradients. For increasing complexity of the geometry, wider tolerances are allowed, varying between 2 and 5%. For the evaluation of a large number of data points an additional quantity, the confidence limit, has been introduced, which combines the influence of systematic and random deviations. The proposed tolerances have been compared with other recommendations for a number of clinically relevant examples, showing considerable differences, which are partly due to the way the complexity of the geometry is taken into account. Furthermore differences occur if criteria for acceptability of dose calculations are related either to the local dose value or to a normalized dose value.

CONCLUSIONS

Although it is acknowledged that the general aim must be to have good agreement between dose calculation and the actual dose value, e.g. within 2% or 2 mm, current day algorithms and their implementation into commercial treatment planning systems result often in larger deviations. A high accuracy can at present only be achieved in relatively simple cases. The new set of tolerances and the quantity confidence limit have proven to be useful tools for the acceptance of photon beam dose calculation algorithms of treatment planning systems.

A prototype 3D CT extension for radiotherapy simulators

In this paper we present a report on our custom 3D CT extension capable of producing fully 3D tomographic studies from conventional radiotherapy simulator cone-beam fluoro output. The extension consists of a common PC system on which a proprietary software toolkit provides the appropriate environment for reconstruction and acquisition of cone-beam data provided by commercial simulator fluoro video chains. The extension may compare favorably with CT options offered by simulator manufacturers: in particular, multi-slice single-scan reconstruction seems to be achievable while the current commercial solutions are based on single-slice single-scan. Radiotherapy applications include on-line treatment planning, patient treatment verification and registration.

Use of Monte Carlo computation in benchmarking radiotherapy treatment planning system algorithms

Radiotherapy treatments are becoming more complex, often requiring the dose to be calculated in three dimensions and sometimes involving the application of non-coplanar beams. The ability of treatment planning systems to accurately calculate dose under a range of these and other irradiation conditions requires evaluation. Practical assessment of such arrangements can be problematical, especially when a heterogeneous medium is used. This work describes the use of Monte Carlo computation as a benchmarking tool to assess the dose distribution of external photon beam plans obtained in a simple heterogeneous phantom by several commercially available 3D and 2D treatment planning system algorithms. For comparison, practical measurements were undertaken using film dosimetry. The dose distributions were calculated for a variety of irradiation conditions designed to show the effects of surface obliquity, inhomogeneities and missing tissue above tangential beams. The results show maximum dose differences of 47% between some planning algorithms and film at a point 1 mm below a tangentially irradiated surface. Overall, the dose distribution obtained from film was most faithfully reproduced by the Monte Carlo N-Particle results illustrating the potential of Monte Carlo computation in evaluating treatment planning system algorithms.

Correlation between CT numbers and tissue parameters needed for Monte Carlo simulations of clinical dose distributions

We describe a new method to convert CT numbers into mass density and elemental weights of tissues required as input for dose calculations with Monte Carlo codes such as EGS4. As a first step, we calculate the CT numbers for 71 human tissues. To reduce the effort for the necessary fits of the CT numbers to mass density and elemental weights, we establish four sections on the CT number scale, each confined by selected tissues. Within each section, the mass density and elemental weights of the selected tissues are interpolated. For this purpose, functional relationships between the CT number and each of the tissue parameters, valid for media which are composed of only two components in varying proportions, are derived. Compared with conventional data fits, no loss of accuracy is accepted when using the interpolation functions. Assuming plausible values for the deviations of calculated and measured CT numbers, the mass density can be determined with an accuracy better than 0.04 g cm(-3). The weights of phosphorus and calcium can be determined with maximum uncertainties of 1 or 2.3 percentage points (pp) respectively. Similar values can be achieved for hydrogen (0.8 pp) and nitrogen (3 pp). For carbon and oxygen weights, errors up to 14 pp can occur. The influence of the elemental weights on the results of Monte Carlo dose calculations is investigated and discussed.

An investigation into the use of polymer gel dosimetry in low dose rate brachytherapy

An investigation has been carried out into the properties of the BANG polymer gel and its use in the dosimetry of low dose rate brachytherapy. It was discovered that the response of the gel was reproducible and linear to 10 Gy. The gel was found to be tissue equivalent with a response independent of energy to within experimental accuracy (standard error of measurement +/- 5%). The slope of the calibration curve was found to increase from 0.28 +/- 0.01 s-1 Gy-1 to 0.50 +/- 0.02 s-1 Gy-1 for an increase in monomer concentration from 6 to 9%. Absorbed dose distributions for a straight applicator containing 36 137Cs sources were measured using the gel and the results compared with measurements made with thermoluminescent dosemeters (TLDs) and calculated values. Good agreement was found for the relative measurements. The root mean square residual percentage errors were 3%, 1% and 4% for the gel and the two groups of TLDs, respectively. There were some significant differences in absolute values of absorbed dose in the gel, possibly owing to the effects of oxygen. Measurements of a complex gynaecological insert were also made and compared with isodose curves from a planning system (Helax TMS), and in areas unaffected by oxygen diffusion the isodose levels from 100 to 50% agreed to within less than 0.5 mm.

Relative electron density calibration of CT scanners for radiotherapy treatment planning

Several authors have reported data on the variation of Hounsfield numbers with electron density in CT scanners. The data can be fitted with a double straight line approach. For non-bone tissues (or phantom materials with similar atomic numbers) the data from all authors can be fitted to a single straight line. For bone-like materials the line varies between authors. The method used to measure electron density has a greater effect than the differences between scanners, or the kilovoltage used on a given scanner. The effect of variation of these slopes on the accuracy of radiotherapy treatment planning is analysed. For typical radiotherapy beams, to produce a 1% error in dosimetry would require errors of over 8% in bone electron density. Using a single pair of calibration lines for all the scanners reported would give dosimetric errors of under 0.8%. A formula is recommended as a default for use in planning systems in circumstances where no data are available for a particular scanner.

A cone-beam megavoltage CT scanner for treatment verification in conformal radiotherapy

PURPOSE

A prototype scanner for large-volume megavoltage computed tomography (MVCT) in a clinical set-up is described. The ultimate aim is to improve treatment accuracy in conformal radiotherapy through patient set-up error reduction and transit dosimetry.

MATERIALS AND METHODS

The scanner consists of a custom-built 2D CsI(Tl) crystal array viewed by a lens and a CCD camera. Image acquisition is synchronized with radiation pulses. The 2D projections resulting from a single continuous 360 degrees gantry rotation are reconstructed using a cone-beam tomography algorithm. Prior to reconstruction, the raw projections are calibrated and corrected for centre of rotation movement and accelerator output fluctuation. The performance of the system has been evaluated by reconstructing projections of open fields, test objects and a humanoid phantom.

RESULTS

Hundreds of 2D projections can be acquired with a clinically-acceptable data collection time (about 2 min) and dose (approximately 40 cGy, with a possible four-fold reduction). A maximum density resolution of about 2% is achieved offering some soft tissue discrimination without using image enhancement tools. A spatial resolution of 2.5 mm is obtained. The reconstructed image intensity is linear with electron density over the range of interest. Coronal or sagittal slices through the 3D reconstruction of the humanoid phantom show a better delineation of structures than the corresponding portal images taken at the same orientation.

CONCLUSIONS

A similar image quality to our current single-slice MVCT scanner is achieved with the advantage of providing tens of tomographic slices for a single gantry rotation. This work demonstrates the feasibility of clinical cone-beam MVCT and indicates how this prototype can be improved.

Study of lung density corrections in a clinical trial (RTOG 88-08). Radiation Therapy Oncology Group

PURPOSE

To investigate the effect of lung density corrections on the dose delivered to lung cancer radiotherapy patients in a multi-institutional clinical trial, and to determine whether commonly available density-correction algorithms are sufficient to improve the accuracy and precision of dose calculation in the clinical trials setting.

METHODS AND MATERIALS

A benchmark problem was designed (and a corresponding phantom fabricated) to test density-correction algorithms under standard conditions for photon beams ranging from 60Co to 24 MV. Point doses and isodose distributions submitted for a Phase III trial in regionally advanced, unresectable non-small-cell lung cancer (Radiation Therapy Oncology Group 88-08) were calculated with and without density correction. Tumor doses were analyzed for 322 patients and 1236 separate fields.

RESULTS

For the benchmark problem studied here, the overall correction factor for a four-field treatment varied significantly with energy, ranging from 1.14 (60Co) to 1.05 (24 MV) for measured doses, or 1.17 (60Co) to 1.05 (24 MV) for doses calculated by conventional density-correction algorithms. For the patient data, overall correction factors (calculated) ranged from 0.95 to 1.28, with a mean of 1.05 and distributional standard deviation of 0.05. The largest corrections were for lateral fields, with a mean correction factor of 1.11 and standard deviation of 0.08.

CONCLUSIONS

Lung inhomogeneities can lead to significant variations in delivered dose between patients treated in a clinical trial. Existing density-correction algorithms are accurate enough to significantly reduce these variations.

Dosimetry of high energy electron therapy to the parotid region

Well-known inadequacies in currently available electron planning systems, and two cases of temporal lobe necrosis following electron therapy of the parotid stimulated a comprehensive head and neck phantom dosimetric study of the use of high energy electrons for parotid treatments. A typical electron field employed for the treatment of parotid malignancy was examined in an anthropomorphic head phantom from which air cavities had been excavated. Thermoluminescent dosimeter measurements were compared with predicted point doses obtained from a Theraplan Treatment planning system (V05). Data was examined for three different electron energies: 12, 16 and 20 MeV and with the addition of contoured bolus for 20 MeV. A number of significant discrepancies between the measured and predicted dose were observed. Measured doses were seen to exceed predicted doses by up to 23% in the temporal lobe. Further under-predictions of dose were found behind the mandible and in the nasal cavity. Over-predictions of dose by the planning algorithm of up to 22% were observed beside the oropharynx. Some of these discrepancies were found to relate to Theraplan under-estimation of the dose in the fall-off region. Other errors are attributable to the difficulties in predicting dose at density interfaces. Localised over- and under-predictions of this magnitude must be accounted for by the clinician prescribing treatment in terms of possible late effects on the temporal lobe and, in particular, the nominated dose specification point.

A comparison of three electron planning algorithms for a 16 MeV electron beam

PURPOSE

We report results of a comparison of three electron planning algorithms, an Age-Diffusion Pencil beam algorithm and two (2-D) and three dimensional (3-D) Hogstrom pencil beam algorithms, using simple 2 x 2 cm air and hard bone inhomogeneities and a complex anthropomorphic head and neck phantom.

METHODS AND MATERIALS

The simple inhomogeneities have variable dimensions outside the plane of calculation to test the effects of out of plane scattering on 2-D algorithms, compared with dose measured by film below the inhomogeneity in the dose fall-off range. Comparisons are also made of a parotid treatment field for 16 MeV electrons, and the dose measured by high sensitivity thermoluminescent dosimeters in the head and neck phantom.

RESULTS

Behind the simple inhomogeneities, the electron algorithms are found to underestimate the dose behind the air cavity by up to 40% and overestimated the dose behind bone by up to 30%. In the head phantom, the presence of inhomogeneities also presents problems for the algorithms, with overestimations of dose of up to 20% found behind bone-tissue interfaces, apparently due to shielding by high density bone. Overestimations of up to 17% are also found beside interfaces parallel to the beam. Underestimations of dose of up to 10% are found on the beam-side of interfaces, due to under-prediction of backscattered electrons. All three investigated algorithms underestimate the dose by up to 20% behind extreme surface curvature. One algorithm is found to underestimate the dose in the falloff region while another overestimates the dose around the 90% isodose.

CONCLUSION

Clinicians should be aware of the limitations of their planning systems.

Commissioning and quality assurance of treatment planning computers

The process of radiation therapy is complex and involves many steps. At each step, comprehensive quality assurance procedures are required to ensure the safe and accurate delivery of a prescribed radiation dose. This report deals with a comprehensive commissioning and ongoing quality assurance program specifically for treatment planning computers. Detailed guidelines are provided under the following topics: (a) computer program and system documentation and user training, (b) sources of uncertainties and suggested tolerances, (c) initial system checks, (d) repeated system checks, (e) quality assurance through manual procedures, and in vivo dosimetry, and (f) some additional considerations including administration and manpower requirements. In the context of commercial computerized treatment planning systems, uncertainty estimates and achievable criteria of acceptability are presented for: (a) external photon beams, (b) electron beams, (c) brachytherapy, and (d) treatment machine setting calculations. Although these criteria of acceptability appear large, they approach the limit achievable with most of today's treatment planning systems. However, developers of new or improved dose calculation algorithms should strive for the goal recommended by the International Commission of Radiation Units and Measurements of 2% in relative dose accuracy in low dose gradients or 2 mm spatial accuracy in regions with high dose gradients. For brachytherapy, the aim should be 3% accuracy in dose at distances of 0.5 cm or more at any point for any radiation source. Details are provided for initial commissioning tests and follow-up reproducibility tests. The final quality assurance for each patient is to perform an independent manual check of at least one point in the dose distributions, as well as the machine setting calculation. As a check of the overall treatment planning process, in vivo dosimetry should be performed on a select number of patients.

In vivo tissue characterization using quantitative computed tomography: a review

Over the past 15 years many attempts have been made to make use of the quantitative information contained within computed tomography (CT) scanner images. A survey of the various approaches to quantitative computed tomography (QCT) is reported. The technical limitations of QCT are discussed. It is concluded that the measurement of bone mineral content is currently the only clinically well-established QCT technique.

The effects of single dose oral hydralazine on blood flow through human lung tumours

Hydralazine has been shown to reduce tumour blood flow and to potentiate the cytotoxicity of melphalan and bioreductive agents in mice. In order to determine whether such a strategy might have clinical potential, a study was undertaken to investigate the effects of hydralazine on blood flow through human tumours. Twenty-two patients with carcinoma of the bronchus received a single oral dose of hydralazine in the range 25 to 150 mg (0.37-2.86 mg/kg) according to age and acetylator status. Tumour blood flow was assessed by single photon emission computed tomography (SPECT) performed 10 min following intravenous 99Tcm-HMPAO on two occasions 2-8 days apart, the second being performed 60 min after hydralazine administration. In 20 evaluable patients, hydralazine caused a 38% increase in blood flow through the whole tumour (p = 0.007) and a 28% increase in flow through the tumour centre (p = 0.03) with greater increases occurring in patients sustaining greater falls in peripheral resistance. Tumour vascular resistance fell indicating active vasodilation in arterioles supplying tumours. Side-effects due to hydralazine were reported by eight patients.

Evaluation of lung dose correction methods for photon irradiations of thorax phantoms

Radiation absorbed dose in lung is measured and calculated using several algorithms available on commercial treatment planning systems. Phantoms resembling the human thorax are used and irradiated with small and large photon beams of 60Co, 4, 6, and 10 MV X ray energies. The applicability and usefulness of the different calculation methods in clinical situations is discussed.

The performance characteristics of a simulator-based CT scanner

Computed tomographic (CT) images have been obtained from a modified radiotherapy simulator (the prototype Royal Marsden Hospital ;CT simulator'). The images are suitable for planning radiotherapy treatment in which the treatment volume is a whole organ, such as in the postoperative conservative management of breast cancer. The performance characteristics of the prototype equipment are reported for scanning conditions typical of those encountered in clinical practice.

Rapid three-dimensional treatment planning: I. Ray-tracing approach to primary component dose calculations

Algorithms for fully three-dimensional divergent-beam radiotherapy treatment planning have been developed to achieve very high sampling of dose in heterogeneous (inhomogeneous density) tissue throughout an arbitrarily oriented patient volume, in clinically acceptable times of calculation. Dose is calculated at points along numerous rays which sample each beam. To display the dose distribution, the calculated dose values for each beam are interpolated onto rectilinear grids of (arbitrary) parallel planes, scaled for beam weight and finally merged with the weighted dose contributions of other beams. In this paper we describe and demonstrate the algorithm for the primary component of the three-dimensional photon dose distribution delivered to a patient.

A 3-D beam subtraction method for inhomogeneity correction in high energy X-ray radiotherapy

Most methods of inhomogeneity correction in high energy X-ray beams, assume an infinite lateral extent of the heterogeneous volumes ("slab models") or require sophisticated time-consuming computer algorithms. We present here a method, developed for parallelepiped inhomogeneities based on a beam subtraction concept combined with a conventional "slab model". Provided that the conventional model is appropriately chosen, this method gives agreement with experimental results better than 1% in most cases. It accounts properly for the scatter modification according to the size and position of the inhomogeneous volume. One example of a computer application is also presented.