Volume 32, Issue 2, February 2005
- letters to the editor
- radiation therapy physics
- radiation imaging physics
- radiation measurement physics
- magnetic resonance physics
- nuclear medicine physics
- ultrasound physics
- thermotherapy physics
- anatomy and physiology
- radiation protection physics
- radiation protection
- ph. d. theses abstracts
- books and publications
Index of content:
- LETTERS TO THE EDITOR
- RADIATION THERAPY PHYSICS
32(2005); http://dx.doi.org/10.1118/1.1835611View Description Hide Description
A method to predict radiation-induced pneumonitis (RP) using an artificial neural network (ANN) was investigated. A retrospective study was applied to the clinical data from 142 patients who have been treated with three-dimensional conformal radiotherapy for tumors in the thoracic region. These data were classified, based on their treatment outcome, into two patient clusters: with RP and without RP . An ANN was designed as a classifier. To perform the classification, a patient-treatment outcome with RP was assigned a value of 1, and a patient treatment outcome without RP was assigned a value of −1. The input of the ANN was limited to the patient lungdose-volume data only. A volume vector that describes patient lung subvolumes receiving more than a set of threshold doses was used as the network input variable. A zero value was used as the threshold to set the output value into −1 or 1. Three ANNs (ANN_1, ANN_2, and ANN_3), each with three layers, were trained to perform this classification function and to show the effect of training data on the ANN performance. Radial basis function was applied as the hidden layer neuron activation function and a sigmoid function was selected as the output layer neuron function. Backpropagation with a conjugate gradient algorithm was used to train the network. ANN_1 was trained and tested by using the leave-one-out method. ANN_2 was trained by randomly selecting 2/3 of the patient data, and tested by the remaining 1/3 of the data. ANN_3 was trained by a user selecting 2/3 of the patient data, and tested by the remaining 1/3 of the data. The predictive accuracy was verified as the area under a receiver operator characteristic (ROC) curve. The correct classification rates of 73% for RP cases, and 99% for non-RP cases were obtained from ANN_1. The corresponding correct classification rates of 44% for RP cases, and 89% for non-RP cases were obtained from ANN_2. From the ANN_3 test phase, the corresponding correct classification rates of 55% for RP cases, and 95% non-RP cases were achieved. The area under ROC curve was , , and for ANN_1, ANN_2, and ANN_3, respectively, within its asymmetric 95% confidence interval. The sensitivity was 95%, 57%, and 71%, and the specificity was 94%, 88%, and 90% for ANN_1, ANN_2, and ANN_3, respectively. Preliminary results suggest that the ANN approach provides a useful tool for the prediction of radiation-induced lung pneumonitis, using the patient lungdose-volume information.
32(2005); http://dx.doi.org/10.1118/1.1843352View Description Hide Description
The orthogonal imaging method is commonly used for source localization in brachytherapy. In some cases, however, difficulty is encountered in determining the dummy sources because of the presence of either contrast materials or bony structures. We here offer a novel method for source localization utilizing a dual-energy, radiographic technique. In this approach, two sets of orthogonal radiographicimages (anterior–posterior and lateral views) are obtained using two different x-ray energies. Image processing (i.e., subtraction between two image sets) is carried out to enhance the source image. In a study performed using a laboratory developed pelvic phantom, it was demonstrated that the dual-energy method could significantly enhance the image quality of the dummy sources, and improve the achievable precision and relative accuracy in localization of source positions. When directly combined with digital imaging modalities, the dual-energy method can be a useful technique to improve the accuracy in brachytherapy source localization from planar radiographs.
32(2005); http://dx.doi.org/10.1118/1.1843354View Description Hide Description
We present a theoretical and empirical analysis of the errors associated with the spatial discretization of the dose grid employed in optimized intensity modulated radiation therapy(IMRT) treatment plans. An information theory based Fourier analysis of the accuracy of discrete representations of three-dimensional dose distributions is presented. When applied to beamlet-based IMRTdose distributions, the theory produces analytic integrals that can bound worst case aliasing errors that can occur regardless of the location and orientation of the dose grid. The predictions of this theory are compared to empirical results obtained by solving a linear-programming based fluence-map optimization model to global optimality. A reasonable agreement between worst case estimates and the empirical results is attributed to the fact that the optimization takes advantage of aliasing to produce an optimal plan. We predicted and empirically demonstrated that an isotropic dose grid with spacing is sufficient to prevent dose errors larger than a percent. However, we noted that in practice this resolution is mostly needed in high-dose target regions. Finally, a multiresolution 2–4–6 mm spacing model was developed and empirically tested where these spacings were applied to targets, structures, and tissue, respectively.
Dosimetric study using different leaf-width MLCs for treatment planning of dynamic conformal arcs and intensity-modulated radiosurgery32(2005); http://dx.doi.org/10.1118/1.1842911View Description Hide Description
This paper systematically studied the dosimetric difference between a micro multileaf collimator(MLC), a MLC, and a MLC for stereotactic radiosurgery using the Brainscan treatment planning system. Thirty-four cases treated with the dynamic conformal arcs technique and 20 cases treated with the intensity modulated radiosurgery∕fractionated radiotherapy (IMRS∕IMRT) technique were retrospectively studied. The conformity index, the percentage target coverage, and the dose–volume histogram (DVH) for organs-at-risk (OARs) were used for dosimetric analysis and comparison for different treatment techniques, target volumes, and treatment sites. For the dynamic conformal arcs technique, there were statistically significant differences in the conformity indices between different leaf-width MLCs. The ratio of the conformity indices between different MLCs depended on the target volume. The average conformity index ratios between the MLC and the MLC were , , and , respectively, for patients with the target volume in groups: (1) , (2) , (3) , and (4) . The average conformity index ratios between the 10 and MLCs were , , , and for patients in these four volume groups, respectively. No statistically significant difference was found for the target coverage among different MLCs. For the IMRS∕IMRT technique, the average conformity index and target coverage ratios were and , respectively, between the 5 and MLCs, and were and , respectively, between the 10 and MLCs. The MLC showed slightly better overall OAR DVHs than the 5 and MLCs, especially for the cranial site with small-volume OARs defined. The results suggest that for the dynamic conformal arcs technique, the narrower leaf-width MLC provides better dose conformity than the wider leaf-width MLCs. This advantage decreases when the target volume increases. For the IMRS∕IMRT technique, the narrower leaf-width MLC could have better sparing of small OARs than the wider leaf-width MLC.
A comparative study of rectal dose histograms in prostate brachytherapy: Some analytic and numerical results32(2005); http://dx.doi.org/10.1118/1.1852811View Description Hide Description
A cumulative dose histogram is the graph of an integral function integrated over a domain and is dubbed the dose-volume histogram (DVH), the dose-surface histogram (DSH) or the dose-wall histogram (DWH), depending on the dimension and structure of the region . This paper presents a comparative study of the three rectal dose histograms for sixty patients as well as for a cylindrical model of the rectum; in particular, the DSH and DVH for the cylindrical model with one point source are computed analytically in terms of elliptic integrals. The difference among the three relative dose histograms, averaged over the sixty patients, is less than 5%, whereas that between DVH and DWH for various wall-thickness can be as large as in the range . The paper also contains an error analysis using two simple models of the rectum, for which the true DSH and DWH can be computed via numerical integration, to evaluate the effect of digitization. The digitized computation agrees quite well with the pre-digitization numerical integration, within 1% or , because of the low dose-gradient effect near the rectum in prostate brachytherapy.
32(2005); http://dx.doi.org/10.1118/1.1851925View Description Hide Description
Today, inverse treatment planning for intensity modulated proton therapy (IMPT) usually employs a constant relative biological effectiveness (RBE). In this paper, the potential clinical relevance of RBE variations for scanning techniques in IMPT is investigated, and a new strategy to include the RBE into the inverse planning process is presented. Three-dimensional RBE distributions are calculated based on a phenomenological model that describes the RBE as a function of dose, linear energy transfer (LET) and tissue type in the framework of the linear-quadratic model. This RBE model is integrated into the optimization loop of inverse planning by using a modified version of the standard quadratic objective function, where the physical dose is replaced by the biological effect. This system for “biological optimization” was implemented into a research version of the inverse planning software KonRad and allows the direct optimization of the product of RBE and physical dose. Several treatment plans for a prostate case are presented, which compare the biological with the conventional physical doseoptimization for IMPT scanning techniques, in particular distal edge tracking (DET) and the full three-dimensional (3D) modulation of beam spots. Mainly due to their different LET distributions, the RBE effects for these two techniques are quite different: while the RBE distribution was more or less homogeneous in the planning target volume (PTV) for 3D modulation, considerable RBE variations within the PTV were observed for DET. These unfavorable effects could be compensated for by employing the new biological objective function, which led to a more homogeneous distribution of the product of RBE and physical dose in the PTV. The computation time increased by a factor of 2 compared to the optimization of the physical dose. In conclusion, the proposed method allows the simultaneous multifield optimization of the biological effect in a reasonable time, and is therefore well suited for studying the influence of a variable RBE in IMPT as well as for minimizing potentially adverse effects.
Use and uncertainties of mutual information for computed tomography/magnetic resonance (CT/MR) registration post permanent implant of the prostate32(2005); http://dx.doi.org/10.1118/1.1851920View Description Hide Description
Post-implant dosimetric analysis for permanent implant of the prostate benefits from the use of a computed tomography(CT) dataset for optimal identification of the radioactive source (seed) positions and a magnetic resonance(MR) dataset for optimal description of the target and normal tissue volumes. The CT/MR registration process should be fast and sufficiently accurate to yield a reliable dosimetric analysis. Since critical normal tissues typically reside in dose gradient regions, small shifts in the dose distribution could impact the prediction of complication or complication severity. Standard procedures include the use of the seed distribution as fiducial markers (seed match), a time consuming process that relies on the proper identification of signals due to the same seed on both datasets. Mutual information (MI) is more efficient because it uses image data requiring minimal preparation effort. A comparison of MI registration and seed-match registration was performed for twelve patients. MI was applied to a volume limited to the prostate and surrounding structures, excluding most of the pelvic bone structures (margins around the prostate gland were right–left, anterior–posterior, and superior–inferior). Seeds were identified on a slice CT dataset using an automatic seed identification procedure on reconstructed three-dimensional data. Seed positions on the slice thickness T2 MR data set were identified using a point-and-click method on each image. Seed images were identified on more than one MR slice, and the results used to determine average seed coordinates for MRimages and matched seed pairs between CT and MRimages. On average, 42% (19%–64%) of the seeds (19–54 seeds) were identified and matched to their CT counterparts. A least-squares method applied to the CT and MR seed coordinates was used to produce the optimum seed-match registration. MI registration and seed match registration angle differences averaged 0.5 degrees, which was not significantly different from zero. Translation differences averaged 0.6 (1.2 standard deviation) mm right–left, posterior-anterior, and inferior–superior. Registration error estimates were approximately for both the MI and seed-match methods. The observed standard deviations in the offset values were consistent with propagation of error. Registration methods as applied here using mutual information and seed matching are consistent, except for a small systematic difference in the inferior–superior axis for a minority of cases . Cases registered with mutual information and with bony anatomy misregistration of greater than should be evaluated for rescan or seed-match registration. The improvement in efficiency of use for the MI registration method is substantial, compared to several hours using seed match registration.
32(2005); http://dx.doi.org/10.1118/1.1852792View Description Hide Description
The use of rigid carbon fiber couch inserts in radiotherapy treatment couches is a well-established method of reducing patient set-up errors associated with couch sag. Several published studies have described such inserts as radiotranslucent with negligible attenuation of the radiation field. Most of these studies were conducted with the radiation field normally incident on the couch and there appears to be no evidence in the literature of the effect of the gantry angle on the extent of beam attenuation by the carbon fiber insert alone during external beam radiotherapy. In this study we examined the magnitude of this effect over a range of posterior oblique gantry angles using a cylindrical solid water phantom containing an ionization chamber placed isocentrically. It was found that a photon beam, field size, was attenuated significantly as the gantry angle approached the plane of the couch, from 2% at normal incidence and reaching 9% attenuation at angle of incidence . This could have serious implications regarding dose to the treatment volume for treatments requiring posterior oblique angles of incidence with a possible correction factor necessary in monitor unit calculations.
32(2005); http://dx.doi.org/10.1118/1.1844111View Description Hide Description
The performance of a variational regularization technique to improve robustness of inverse treatment planning for intensity modulated radiotherapy is analyzed and tested. Inverse treatment planning is based on the numerical solutions to the Fredholm integral equation of the first kind which is ill-posed. Therefore, a fundamental problem with inverse treatment planning is that it may exhibit instabilities manifested in nonphysical oscillations in the beam intensity functions. To control the instabilities, we consider a variational regularization technique which can be applied for the methods which minimize a quadratic objective function. In this technique, the quadratic objective function is modified by adding of a stabilizing functional that allows for arbitrary order regularization. An optimal form of stabilizing functional is selected which allows for both regularization and good approximation of beam intensity functions. The regularized optimization algorithm is shown, by comparison for a typical case of a head-and-neck cancer treatment, to be significantly more accurate and robust than the standard approach, particularly for the smaller beamlet sizes.
32(2005); http://dx.doi.org/10.1118/1.1852791View Description Hide Description
Patient-specific intensity-modulated radiotherapy (IMRT) verifications require an accurate two-dimensional dosimeter that is not labor-intensive. We assessed the precision and reproducibility of film calibrations over time, measured the elemental composition of the film, measured the intermittency effect, and measured the dosimetric accuracy and reproducibility of calibrated Kodak EDR2 film for single-beam verifications in a solid water phantom and for full-plan verifications in a Rexolite® phantom. Repeated measurements of the film sensitometric curve in a single experiment yielded overall uncertainties in dose of 2.1% local and 0.8% relative to . 547 film calibrations over an 18-month period, exposed to a range of doses from 0 to a maximum of or and using or energies, had optical density (OD) standard deviations that were 7%–15% of their average values. This indicates that daily film calibrations are essential when EDR2 film is used to obtain absolute dose results. An elemental analysis of EDR2 film revealed that it contains 60% as much silver and 20% as much bromine as Kodak XV2 film. EDR2 film also has an unusual 1.69:1 silver:halide molar ratio, compared with the XV2 film’s 1.02:1 ratio, which may affect its chemical reactions. To test EDR2’s intermittency effect, the OD generated by a single exposure was compared to the ODs generated by exposing the film , , and at a time to a total of . An ion chamber recorded the relative dose of all intermittencymeasurements to account for machine output variations. Using small MU bursts to expose the film resulted in delivery times of 4 to 14 minutes and lowered the film’s OD by approximately 2% for both 6 and beams. This effect may result in EDR2 film underestimating absolute doses for patient verifications that require long delivery times. After using a calibration to convert EDR2 film’s OD to dose values, film measurements agreed within 2% relative difference and criteria to ion chambermeasurements for both sliding window and step-and-shoot fluence map verifications. Calibrated film results agreed with ion chambermeasurements to within criteria for transverse-plane full-plan verifications, but were consistently low. When properly calibrated, EDR2 film can be an adequate two-dimensional dosimeter for IMRT verifications, although it may underestimate doses in regions with long exposure times.
32(2005); http://dx.doi.org/10.1118/1.1854776View Description Hide Description
Accurate knowledge of gantry angle is essential in megavoltage cone beam imaging (MVCBI) with an electronic portal imager. We present a method for determining the gantry angle by detectingmultileaf collimator(MLC) leaf positions in projection images. During image acquisition the gantry moves continuously and the MLC operates in dynamic arc mode. Our algorithm detects the leaf positions in the images and compares them with a stationary reference leaf. Comparison of the algorithm against angles determined from the locations of fiducial markers shows the accuracy ( error) to be sufficient for MVCBI.
32(2005); http://dx.doi.org/10.1118/1.1855015View Description Hide Description
Current dosimetric protocols based on the absorbed dose (AAPM TG-51 and IAEA TRS-398 protocols) require calibration measurements under reference conditions. For some radiotherapy systems, this requirement cannot be met, and calibration has to be performed under nonreference experimental conditions. In order to solve this problem, both protocols can be extended by inclusion of the measured-to-reference conversion factor, . In order to determine this factor, basic dosimetric quantities, like stopping power ratios, mass attenuation coefficients and chamber correction factors have to be calculated. If measurements are not feasible, accurate Monte Carlo modeling is required. The extension of the protocols is illustrated using the case of the helical tomotherapy radiation unit, where the typical calibration measurement conditions are the field size and the 85 cm surface source distance, limited by the system design. It was calculated that the factor for this conditions is close to unity . In addition, the deviation of the measurement conditions from the reference conditions results in the change of the quality conversion factor (approximately 0.995–0.998, depending on the ionization chamber used). This change is the same regardless of the used calibration protocol. For smaller field sizes the corrections become more significant, resulting in the total correction factor compared to the reference conditions of up to 1.5% for the smallest considered field size of .
32(2005); http://dx.doi.org/10.1118/1.1843471View Description Hide Description
A three-dimensional (3D) intensity-modulated radiotherapy (IMRT) pretreatment verification procedure has been developed based on the measurement of two-dimensional (2D) primary fluence profiles using an amorphous silicon flat-panel electronic portal imaging device(EPID). As described in our previous work, fluence profiles are extracted from EPIDimages by deconvolution with kernels that represent signal spread in the EPID due to radiation and optical scattering. The deconvolution kernels are derived using Monte Carlo simulations of dose deposition in the EPID and empirical fitting methods, for both 6 and 15 MV photon energies. In our new 3D verification technique, 2D fluence modulation profiles for each IMRT field in a treatment are used as input to a treatment planning system (TPS), which then generates 3D doses. Verification is accomplished by comparing this new EPID-based 3D dose distribution to the planned dose distribution calculated by the TPS. Thermoluminescent dosimeter(TLD) point dose measurements for an IMRT treatment of an anthropomorphic phantom were in good agreement with the EPID-based 3D doses; in contrast, the planned dose under-predicts the TLD measurement in a high-gradient region by approximately 16%. Similarly, large discrepancies between EPID-based and TPS doses were also evident in dose profiles of small fields incident on a water phantom. These results suggest that our 3D EPID-based method is effective in quantifying relevant uncertainties in the dose calculations of our TPS for IMRT treatments. For three clinical head and neck cancer IMRT treatment plans, our TPS was found to underestimate the mean EPID-based doses in the critical structures of the spinal cord and the parotids by (11%–14%). According to radiobiological modeling calculations that were performed, such underestimates can potentially lead to clinically significant underpredictions of normal tissue complication rates.
- RADIATION IMAGING PHYSICS
Performance comparison of an active matrix flat panel imager, computed radiography system, and a screen-film system at four standard radiation qualities32(2005); http://dx.doi.org/10.1118/1.1843451View Description Hide Description
Four standard radiation qualities (from RQA 3 to RQA 9) were used to compare the imaging performance of a computed radiography(CR)system (general purpose and high resolution phosphor plates of a Kodak CR 9000 system), a selenium-based direct flat panel detector (Kodak Direct View DR 9000), and a conventional screen-film system (Kodak T-MAT L/RA film with a 3M Trimax Regular screen of speed 400) in conventional radiography. Reference exposure levels were chosen according to the manufacturer’s recommendations to be representative of clinical practice (exposure index of 1700 for digital systems and a film optical density of 1.4). With the exception of the RQA 3 beam quality, the exposure levels needed to produce a mean digital signal of 1700 were higher than those needed to obtain a mean film optical density of 1.4. In spite of intense developments in the field of digital detectors, screen-film systems are still very efficient detectors for most of the beam qualities used in radiology. An important outcome of this study is the behavior of the detective quantum efficiency of the digital radiography (DR) system as a function of beam energy. The practice of users to increase beam energy when switching from a screen-film system to a CRsystem, in order to improve the compromise between patient dose and image quality, might not be appropriate when switching from screen-film to selenium-based DR systems.
32(2005); http://dx.doi.org/10.1118/1.1844433View Description Hide Description
We present a novel method for assessing the performance of computer-aided detection systems on unseen cases at a given sensitivity level. The sampling error introduced when training the system on a limited data set is captured as the uncertainty in determining the system threshold that would yield a certain predetermined sensitivity on unseen data sets. By estimating the distribution of system thresholds, we construct a confidence interval for the expected number of false positive markings per image at a given sensitivity. We present two alternative procedures for estimating the probability density functions needed for the construction of the confidence interval. The first is based on the common assumption of Poisson distributed number of false positive markings per image. This procedure also relies on the assumption of independence between false positives and sensitivity, an assumption that can be relaxed with the second procedure, which is nonparametric. The second procedure uses the bootstrap applied to the data generated in the leave-one-out construction of the FROC curve, and is a fast and robust way of obtaining the desired confidence interval. Standard FROC curve analysis does not account for the uncertainty in setting the system threshold, so this method should allow for a more fair comparison of different systems. The resulting confidence intervals are surprisingly wide. For our system a conventional FROC curve analysis yields 0.47 false positive markings per image at 90% sensitivity. The 90% confidence interval for the number of false positive markings per image is (0.28, 1.02) with the parametric procedure and (0.27, 1.04) with the nonparametric bootstrap. Due to its computational simplicity and its allowing more fair comparisons between systems, we propose this method as a complement to the traditionally presented FROC curves.
32(2005); http://dx.doi.org/10.1118/1.1854779View Description Hide Description
A concept of a photon counting cone beam CT is proposed. The system uses a new Multi Slit Multi Slice (MSMS) cone beam acquisition geometry utilizing a linear array photon countingdetectors. The MSMS cone beam acquisition is a direct analogy of the scanning multislit acquisition used in projection x-rayimaging. This geometry provides a CTimaging with dose efficient scatter rejection and allows for using available photon countingdetectors. The microchannel plate (MCP) detector is proposed as a linear array photon countingdetector for MSMS cone beam CT system. Initial testing of the MCP detector for CT application was performed. The field of view of the prototype MCP detector is . A delay line position encoding electronics was used. The electronics has a single channel input for evaluation of events from the entire detector field of view. This limits the system count rate at . The spatial resolution of this detector is FWHM at and FWHM at tube voltages. The detectornoise in CT projections is less than for the size. The CT projections contain quantum-limited and scatter free signal. Images of a contrast phantom and a small animal were acquired at and tube voltages. The CT numbers for different contrast elements were calculated for a given x-ray spectrum and compared with experimental values. The quantum efficiency of the current detector is 56% at , which is suboptimal because of the large channel diameter of these MCPs. The MCPs with smaller channels and higher efficiencies are being tested. The quantum efficiency was measured to be 70% for a new MCP with channel diameter. Design parameters of a clinically applicable photon counting MSMS cone beam CT for breast imaging was evaluated. System uses field of view MCP detectors based on channel MCPs and high count rate ASIC electronics. It was concluded that the MSMS cone beam CT with a photon counting MCP detector is feasible for volume breast imaging.
32(2005); http://dx.doi.org/10.1118/1.1851892View Description Hide Description
We conducted a study to evaluate the effectiveness of twelve different similarity measures in matching the corresponding masses on temporal pairs of current and prior mammograms. To perform this comparison we implemented each of the twelve similarity measures in the final stage of our multistage registration technique for automated registration of breast lesions in serial mammograms. The multistage technique consists of three stages. In the first stage an initial fan-shape search region was estimated on the prior mammogram based on the geometrical position of the mass on the currentmammogram. In the second stage, the location of the fan-shape region was refined by warping, based on an affine transformation and simplex optimization. A new refined search region was defined on the prior mammogram. In the third stage, a search for the best match between the lesion template from the currentmammogram and a structure on the prior mammogram was carried out within the search region. Our data set consisted of 318 temporal pairs. We performed three experiments, using a different subset of the 318 temporal pairs for each experiment. In each experiment we further tested how the performance of the similarity measures varied as the size of the search region increased or decreased. We evaluated the twelve similarity measures based on four criteria. The first criterion was the mean Euclidean distance, which was the average distance of the true location of the mass to the location detected by the similarity measure. The second criterion was the percentage of temporal pairs that were aligned so that 50% or more of the lesion area overlapped. The third criterion was the percentage of pairs that were aligned so that 75% or more of the lesion area overlapped. The fourth and final criterion was the robustness of the similarity measure. Our results showed that three of the similarity measures, Pearson’s correlation, the cosine coefficient, and Goodman and Kruskal’s Gamma coefficient, provide significantly higher accuracy in the task of matching the corresponding masses on serial mammograms than the other nine similarity measures.