Volume 33, Issue 6, June 2006
Index of content:
- Joint Imaging/Therapy Symposium: Valencia A
Young Investigators Symposium
SU‐CC‐ValA‐01: Automatic Comparison Between Reference and On Board Digital Tomosynthesis for Target Localization33(2006); http://dx.doi.org/10.1118/1.2240119View Description Hide Description
Purpose: Digital tomosynthesis (DTS) is a method for reconstructing 3D images from cone‐beam projection data acquired with limited angulation (e.g., 40°) of an x‐ray source, and is much faster and lower dose than full cone‐beam CT(CBCT). We previously developed a method for generating reference DTS images from a planning CT for registration with actual on‐board DTS images. This study examines the accuracy of 3D‐3D registration of reference and on‐board DTS images to assess the potential of DTS for image‐guidedradiation therapy(IGRT).Method and Materials: We simulated the online positioning of an anthropomorphic chest phantom with 6 noncoplanar reference BBs attached. Planning CT data of the phantom were acquired with a G.E. Lightspeed RT scanner. On‐board CBCT projection data were acquired with a Varian 21EX Clinac, equipped with a kV on‐board imager. On‐board DTS images were reconstructed from a subset of the CBCT projection data (81 projections, 44°). True alignment of planning and on‐board image data was achieved according to a 3D point‐based registration of the 6 reference BBs in the CT and CBCTimages. Single‐axis rotations up to +/− 10° and 3‐axis translations up to +/− 10 mm were simulated in the planning CT, prior to the generation of reference DTS images. A 67.5mm × 162.5mm × 20.8mm region of interest surrounding the spinal cord was extracted for registration. Mutual information‐based 3D‐3D registration of reference and on‐board DTS images was performed, and residual registration error was recorded. Results: Registration errors are within 0.7mm and 0.1 degree in all cases. The average registration error was 30% less for translations along the dimension of tomographic motion than for the other two dimensions. Conclusions: 3D‐3D rigid‐body registration of reference and on‐board DTS images is highly accurate, suggesting that DTS may be an effective IGRT technique. Partially supported by a Varian research grant.
33(2006); http://dx.doi.org/10.1118/1.2240120View Description Hide Description
Purpose: To determine more accurately the air‐kerma strength of low‐energy photon‐emitting brachytherapy sources. Method and Materials: Air‐kerma strength is the agreed upon metric for most brachytherapy sources and is defined as the product of the kerma rate in vacuo at distance d on the transverse axis of a seed (polar angle of 90 degrees), multiplied by the square of this distance d 2. This work introduces new methods to evaluate how anisotropy affects air‐kerma strength. First, NaI detector measurements yield in‐air anisotropy data that were previously reported only with Monte Carlo transport methods. In addition, multiple aperture sizes are used to evaluate large‐angle free‐air ionization chamber measurements. Lastly, a new method of Monte Carlo transport, which incorporates the distance‐dependent geometry effects of certain brachytherapy seeds, is used to determine more accurately the air‐kerma rate on the transverse axis of the source. These methods may be used to evaluate air‐kerma strength to a higher degree of accuracy. Results: Well‐collimated NaI detector measurements yield precise in‐air anisotropy measurements that may be compared directly with Monte‐Carlo transport simulations. Measurements of two seed types show deviations near the transverse axis of the source of at least 5%. Monte‐Carlo determined point‐detector simulations yield more accurate estimations of air‐kerma strength and show that air‐kerma strength is not constant for all distances in vacuo, as would be predicted by its definition. Conclusion: This work shows that with the combination of these new measurement techniques, air‐kerma strength may be evaluated to a higher degree of accuracy for low‐energy photon‐emitting brachytherapy sources. New experimental data on two seeds' in‐air anisotropy is also presented.
SU‐CC‐ValA‐03: A Directional Algorithm for An Electronically‐Collimated Gamma‐Ray Detector for Intraoperative Localization of Radiation Sources33(2006); http://dx.doi.org/10.1118/1.2240121View Description Hide Description
Purpose: An electronically‐collimated gamma‐radiation detector for intraoperative localization of sentinel lymph nodes and metastases is under development. Analogous to Compton telescopes and Compton cameras, localization is achieved using the coincidence detection of Compton‐scattered gamma rays. Electronic collimation allows the device to operate without physical collimation, providing high sensitivity while also allowing directional information to be determined. We report on implementation of algorithms to calculate the direction to the source.Methods and Materials: Two approaches to direction reconstruction were evaluated. The first technique backprojects each event onto the surface of a sphere centered on the device's primary detector. To use Fourier filtering methods for deblurring, the sphere's surface is mapped by stereographic projection onto a plane, filtered in Fourier space, and then projected back onto the sphere. The second technique also backprojects events onto the sphere, then determines the rectangle that circumscribes the backprojected cone; localization is obtained by intersection of all circumscribed rectangles. Results: Performance of the algorithm has been evaluated using randomly generated ideal Compton‐scatter events from point sources for our detector geometry. Direction angles are calculated within 5% accuracy for source positions up to 45° off‐axis for the filtering approach and ∼30° for the circumscription approach. Error in calculated direction angles depends on the arbitrary diameter of the sphere; optimally, the sphere should intersect the source. The circumscription technique converges to an estimate of direction angles in ∼50 events; the filtering approach requires ∼1000 events. Conclusion: The two methods complement each other in speed and field‐of‐view. Monte Carlo simulations and experimental testing of a prototype system are ongoing as a separate part of the overall project; data from these will further supplement evaluation of the algorithms. Acknowledgment: Supported by Homeland Security Advanced Research Projects Agency, and Space and Naval Warfare Systems Center San Diego; Contract No. N66001‐05‐C‐6024.
SU‐CC‐ValA‐04: Using Flow Information to Support 3D Vessel Reconstruction From Rotational Angiography33(2006); http://dx.doi.org/10.1118/1.2240122View Description Hide Description
Purpose: For the assessment of cerebral vessel diseases, it is very beneficial to obtain three dimensional morphologic and haemodynamic information about the vessel system. Our goal is to determine both concurrently using one rotational angiographysequence. To enable the extraction of flow information, the rotational angiographyimages should show inflow and outflow of contrast agent. Images with this property however, are not well suited to standard volume reconstruction algorithms. This work shows how flow information can support the vessel reconstruction to overcome this conflict. Method and Materials: In our method flow information is used as follows to determine, for every voxel, the likelihood of being inside a vessel: First, the rotational time intensity curve (R‐TIC) is determined from the image intensities at the projection points of the current voxel. Next, the arrival time of the contrast agent bolus at the voxel is estimated from the R‐TIC. Finally, a measure of the intensity and duration of the contrast enhancement is determined. The likelihood is used to steer the Fast Marching algorithm, which determines the order in which voxels are analyzed. This enables the centreline of the vessels to be extracted by backtracking. The proposed method was tested on 80 computer simulated rotational angiographysequences with systematically varied blood flow and contrast agent injection parameters. Results: The mean error in the 3D centreline and radius estimation was 0.62 mm and 0.28 mm respectively. Pulsatile blood flow was found to increase the error only slightly (0.05 mm). Conclusion: Under pulsatile and non‐pulsatile conditions, flow information can be used to enable a 3D vessel reconstruction from rotational angiography with inflow and outflow of contrast agent. Future work will aim to extract more quantitative flow information. Conflict of Interest: Research sponsored by Philips Research Aachen.
33(2006); http://dx.doi.org/10.1118/1.2240123View Description Hide Description
Purpose: To create a fast and accurate computer algorithm for simulating the emission x‐ray spectra from diagnostic tubes as a function of tube voltages, target material, and take‐off angles. Method and Materials: The method uses an integral model to determine the radiative losses from an electron as it slows down in arbitrary media. The effect of self‐absorption and backscatter is accurately described by distribution functions for electron number, electron depth, and angular distribution that are functions of electron slowing down energy. The Monte Carlo program PENELOPE was used to determine these three distribution functions. An exact accounting of electron orientation was found necessary due to large variations in the Bremsstrahlung cross‐section as a function of emission angle. These are accounted for in the integral model by pre‐computing tables based on the Kissel Bremsstrahlung shape function. Characteristic x‐ray emissions as a function of over‐voltage are described using Monte Carlo results for both direct and indirect production. The computer algorithm is implemented as a part of a larger program for computationally simulating x‐ray production, transmission, scattering, and detection for imagingsystems (XSPECT, V4.0). Results: We compared our program to the measured x‐ray spectra of Mercier, and spectral computations of PENELOPE. We found good agreement and an improvement over prior semi‐empirical estimates (Birch & Marshall, Tucker, Storm). Conclusion: We have developed a program that can simulate x‐ray spectra from tubes of arbitrary anode materials (including alloys), and target angles for tube voltages of 1 to 400 kV. After generation of target specific tables, the x‐ray spectra can be computed in a few seconds. The results are equivalent to Monte Carlo estimates that require days to compute a single spectra.
33(2006); http://dx.doi.org/10.1118/1.2240124View Description Hide Description
Purpose: Women with high mammographic breast density are at a 4‐ to 7‐ fold increased risk of developing breast cancer compared to women with fatty breasts. The purpose of this work is to investigate the potential of assessing breast density via acoustic velocity measurements obtained with ultrasoundcomputed tomography.Method and Materials: A sample of approximately 50 patients was imaged with our computed ultrasoundtomography clinical prototype. Each data set was comprised of 45 tomograms ranging from near the chest wall through the nipple region. Whole breast acoustic velocity was determined by creating image stacks and evaluating the sound speed frequency distribution. The acoustic measures of breast density were evaluated by comparing these results to two mammographic density measures: (1) qualitative, as determined by a certified radiologist using the BIRADS Categorical Assessment based on a 1 (fatty) to 4 (dense) scale, and (2) quantitative, via digitization and computerized analysis of archival mammograms. The former involved a radiologist's visual assessment of each mammogram, while the latter required scanning cranio‐caudal films with a Vidar VXR‐16 DosimetryPro digitizer and implementing semi‐automatic segmentation routines. Results: Approximately 60 m/s difference in acoustic velocity was found between the fatty and dense BIRADS categories. This investigation indicated a positive correlation between BIRADS category and acoustic velocity of the breast. In addition, a strong correlation between the mean acoustic velocity and quantitative measures of percent breast density was demonstrated (Pearson correlation coefficient 0.651, p < 0.001). Conclusion: These results support the hypothesis that utilizing acoustical velocity as an analogue to mammographic breast density is feasible. Our approach to evaluating breast density has the potential to provide a safer, non‐ionizing, and more quantitative means of evaluating breast density, thus better elucidating the relationship that exists between breast density and breast cancer risk.
SU‐CC‐ValA‐07: Air Kerma Rate Measurements From a Miniature X‐Ray Source Using Free‐Air Ionization Chambers33(2006); http://dx.doi.org/10.1118/1.2240125View Description Hide Description
Purpose: To measure the air kerma rate from a miniature x‐ray source using free‐air ionization chambers(FACs), and to transfer the source air kerma rate to a well‐type ionization chamber.Method and Materials: Air kerma rates from several Xoft AXXENT™ miniature x‐ray sources were measured in air along their transverse axes using three different FACs, each with a source‐to‐aperture distance of 100 cm. The sources were operated at 50 kV and 100 μA beam current. The University of Wisconsin (UW) Attix FAC was used for the initial measurements, and follow‐up measurements were performed using the Attix and Ritz FACs at the National Institute of Standards and Technology (NIST). Two different FAC aperture sizes were used for each of the air kerma rate measurements at NIST. Additional measurements for each source were performed using a well‐type ionization chamber with a custom‐built aluminum source holder. The ratio of air kerma rate at 100 cm to well chamber current was used to determine a well chamber calibration coefficient for each source. Results: The air kerma rates for the smaller aperture were generally within 2% of the rates measured with the larger aperture, indicating that the sources were aligned properly. The well chamber calibration coefficients demonstrated some source to source variation, with an overall standard deviation of 5.3%. The results suggest that most of this variation can be attributed to azimuthal anisotropy around the long axes of the sources, and not differences in the photonspectrum emitted from each source. Conclusion: Both the Attix and Ritz FACs are appropriate for measuring air kerma rates from the miniature x‐ray sources, but further work will be necessary to develop methods suitable for traceability to national measurements standards. Conflict of Interest:Funding for this research was provided by Xoft, Inc.
SU‐CC‐ValA‐08: Air‐Kerma Strength Determination of a 169Ytterbium High Dose Rate Brachytherapy Source33(2006); http://dx.doi.org/10.1118/1.2240126View Description Hide Description
Purpose: To provide an accurate determination of the new 169Yb high dose rate (HDR) bracytherapy source in terms of air kerma strength, based on an adaptation of the current, NIST traceable, in air measurement standard in use for HDR sources.Methods and Materials: Several modifications to the seven distance technique, which is the current standard for HDR source strength measurement, were required to adapt it to the 169Yb spectrum. An Exradin A4 spherical chamber was employed, which has a relatively flat chamber response to the range of energies in the 169Yb spectrum, and has been verified to accurately measure the air kerma strength of to within the reported uncertainty of the current standard measurement technique. To convert the electrometer readings to source strength, a chamber coefficient, Nk, was determined by using the NIST calibrated chamber coefficients from several NIST H‐Beams, whose energy spectrums fall strategically within the 169Yb spectrum. Several correction factors must be applied to these electrometer readings, including corrections for temperature and pressure, air attenuation, air scatter, ion recombination, and corrections for the finite size of the chamber. Results: The decay corrected average of fourteen measurement iterations was 8.063 × 10−3 Gy‐m2/hr. Analysis of uncertainty was performed on these experimental 169Yb air kerma measurements using the standard NIST method for evaluating uncertainty. This analysis established an overall k = 2 expanded uncertainty of 2.10%. Conclusion: It is shown that, with a few modifications, the current standard for high dose rate brachytherapysourcecalibration could be employed to accurately calibrate the new HDR brachytherapysource in terms of air kerma strength. The uncertainties as analyzed fall within those currently used for calibration.Conflict of Interest: Research funding provided by Implant Sciences Corporation.
33(2006); http://dx.doi.org/10.1118/1.2240127View Description Hide Description
Purpose: To study the characteristics of orthogonal bremsstrahlung photons produced by megavoltage electron pencil beams and to evaluate the suitability of their use for improved radiation therapy imaging. Method and Materials: A 10 MeV electron beam emerging through the research port of a Varian Clinac‐18 linac was made to strike targets of carbon,aluminum and copper. The quality of resulting forward and orthogonal bremsstrahlung beams was evaluated using PDD and attenuation measurements, and the experimental findings were compared with Monte Carlo‐calculated results using the EGSnrcMP code. Images of contrast objects were acquired with Agfa 400 diagnostic films and their contrast levels were analyzed. Results:Photon yield and mean energy of the forward bremsstrahlung spectra were determined to be essentially independent of the target's atomic number Z. In comparison with forward bremsstrahlung, the yield and effective energy were lower in the orthogonal direction, and this decrease was more pronounced for targets of lower atomic number. The effective energy of a spectrum produced by carbon dropped by a factor of 10 from 1535 keV in the forward direction to 151 keV in the orthogonal direction, while for aluminum it dropped by 77% to 425 keV, and for copper by 37% to 1107 keV. The image contrast of films exposed with orthogonal beams was qualitatively determined to be superior to that obtained using the forward megavoltage beams.Conclusions: Orthogonal bremsstrahlung beams produced by megavoltage electrons have a significantly lower mean energy compared to forward beams. In the orthogonal direction, higher Z targets create higher intensity, while lower Z targets provide a more desirable low energy spectrum. Using their relatively low effective energy, orthogonal bremsstrahlung beams produced by megavoltage electrons striking low atomic number targets yield images with a higher contrast than do forward bremsstrahlung beams.
33(2006); http://dx.doi.org/10.1118/1.2240128View Description Hide Description
Purpose: To propose and study a novel scanning trajectory, named reverse helical trajectory, for a cone‐beam CT(CBCT)imager mounted on a linear accelerator(LINAC)treatment system by applying an exact 3D backprojection filtration (BPF) algorithm. Method and Materials: A numerical study using 3D Shepp‐Logan phantom was performed. We applied the PI‐line‐based BPF algorithm to reconstruct exact 3D image from data acquired numerically for a reverse helical trajectory. It was revealed that there is a middle gap in the reconstructed image which is due to lack of PI‐lines passing through the gap. Application of a chord‐based BPF algorithm showed a reduction in the middle gap. Two kinds of line plus reverse helical trajectories were proposed and tested to reduce the middle gap further. One of them is a reverse helical trajectory with a short line segment between two helices separated apart. The other one is a reverse helical trajectory with a long line segment connecting the end points of the revere helices. Results: The middle gap in the reconstructed image was reduced by employing a chord‐based BPF algorithm. The gap was reduced further when we modified the scanning trajectory by inserting a line segment between two helices. The gap was removed completely when we used a reverse helical trajectory with a long line segment connecting the end points of the revere helices. Conclusion: A novel scanning geometry for a LINAC‐mounted CBCTimager was proposed, and a preliminary numerical study was performed. The middle gap in the reconstructed image obtained by PI‐line‐based BPF algorithm was effectively eliminated by using a chord‐based BPF algorithm with a line plus reverse helical trajectory.
President's Symposium: Regulations, Regulations, Regulations!
33(2006); http://dx.doi.org/10.1118/1.2241401View Description Hide Description
Patient Motion Modeling & Adaptive Planning Optimization for Radiotherapy
33(2006); http://dx.doi.org/10.1118/1.2241424View Description Hide Description
A quantitative understanding of respiratory motion is critical to improving radiation therapy for lung and upper abdominal cancers. Breathing motion impacts the quality of diagnostic and treatment planning images, causes conformal therapy portals to be larger than the cross‐sectional projection of the tumor, and increases irradiated normal organ volumes. Methods intended to reduce or eliminate the impact of breathing motion have been proposed, including breath hold, linear accelerator gating, and tracking either using the linear accelerator or the patient support assembly. A quantitative model of the patient's breathing motion, both tumor and normal organs, is necessary to optimize the gating or tracking methods.
The form of the respiratory model will depend on the ultimate use of the model. In the case of radiation therapy, we are interested in understanding the positions of the tumor and normal organs as a function of time, because our radiation delivery systems operate as a function of time. However, breathing is not sufficiently reproducible to use time directly as the independent model variable. A different, time‐dependent metric needs to be selected as the quantity that will be characterized as a function of time and, during acquisition of the motion model data and radiation treatment, be monitored. The metric needs to be: easily measured, quantitative, reproducible, and correlated with breathing motion. Metrics that have been proposed include abdomen or thorax height, abdomen circumference, and spirometry‐measured tidal volume.
The motion model requires input data to provide the patient‐specific parameters. The input data is typically derived from CT scans that are acquired while the patient undergoes simultaneous monitoring of the metric. This process is labeled “4D CT” in that multiple CT scans are acquired at each location, each scan acquired at a different time. CT scans are typically reconstructed or resorted at a variety of breathing phases. The reconstructed CT scans are then used to determine tumor and normal organ positions as a function of the breathing metric.
In the use of respiratory motion modeling there are some confusing and overlapping uses for the word “phase” that are worth differentiating. Firstly, the use of the term “breathing phase” is used to describe a general part of the breathing cycle, such as mid‐inhalation. Secondly, “phase angle” is used to describe a hypothetical angle used when the breathing cycle is described as a periodic function of time, and finally, “phase” itself is used for any quantitatively defined breathing state.
Prior to the development of a biophysically based breathing motion model, there have been two competing methods for describing the behavior of the metric as a function of time; phase‐angle and amplitude. Phase‐angle descriptions divide the breathing cycles between selected breathing phases, for example, inhalation. The time between successive inhalations is recast linearly as an angle from 0 to 2π (alternatively, some investigators separate the inhalation and exhalation processes, placing 0 and π at inhalation and exhalation, respectively, with linear time interpolation between these breathing phases). In the phase‐angle approach, each inhalation and exhalation is treated equally, irrespective of the depth of breathing, but the model can accurately characterize variations in breathing frequency (at least retrospectively). In this model, the description of motion as a function of phase angle can either be the positions as a function of angle or be written as periodic functions with parameters that provide the positions. The phase‐based process is capable of describing the hysteresis‐like motion of lungtumors well, but is not capable of adequately describing variations in breathing depth. Patient breathing training is often employed to reduce variations in breathing depth.
Amplitude‐based methods describe the tumor and organ positions as a function of the metric's amplitude, or numerical value. The time‐dependence of the breathing cycle is taken from the time‐dependence of the metric amplitude. The amplitude‐based approaches are capable of describing variations in breathing depth, but because modeling of hysteresis requires degeneracy in tumor positions as a function of amplitude, hysteresis is not easily described using the amplitude models.
While both amplitude and phase‐based models have been utilized to define and describe breathing motion, neither can adequately model even the simplest breathing motion, namely the amplitude‐variable hysteresis motion of lungtumors and normal organs that is known to exist. Recently a breathing model has been proposed that describes tissue positions as a function of tidal volume, namely the amount of air inhaled and exhaled during the breathing process. The model assumes that lungtissue positions vary as a function of tidal volume, or in other words, the deeper the breath, the farther the tissues move in their trajectories. Hysteresis is hypothesized to be due to pressure imbalances within the lungtissues that create the variations in trajectory between inhalation and exhalation. The pressure imbalances are assumed to be linearly proportional to the airflow (time derivative of the tidal volume). The position of a piece of lungtissue is therefore a function of its location at a reference breathing phase (e.g. tidal exhalation), the tidal volume and airflow relative to the reference breathing phase. Incidentally, while tidal volume has been used as the metric, any metric that is proportional to tidal volume and its temporal derivative can be used as the metric. For example, published reports indicate that abdomen height is linearly related to tidal volume for quiet respiration.
Understanding the variables that govern respiratory motion is insufficient to describe the positions; a mathematical model is still required. The simplest, namely a linear relationship between position and tidal volume and position and airflow, where the two position components are treated as independent, has been used and appears to provide good descriptions of breathing motion, although supporting data is still limited.
The process of modeling breathing motion is still in its beginning stages, but there are promising approaches being studied. Assuming that the models can accurately describe breathing motion, they will be key components in the treatment planning process.
33(2006); http://dx.doi.org/10.1118/1.2241425View Description Hide Description
Patient anatomical variation during the radiotherapy course can be modeled using a stochastic process. In this process, spatial position of each subvolume in patient organs of interest is defined as a random vector described using a probability distribution function (pdf). Two main parameters, the mean and the standard deviation, of the pdf have been historically used to characterize patient anatomical variation during the radiation treatment. It has been demonstrated that treatmentdose distribution in an organ of interest can be evaluated approximately using these two parameters alone, without the full knowledge of organ motion distribution. The approximation is, however, dependent on the scale of the standard deviation as well as the number of treatment delivery fractions. It is straightforward to estimate these two parameters if patient anatomical variation process is stationary. In this case, the two parameters are constants or time‐invariance during the treatment course. However, the estimation will be relatively difficult if patient anatomical variation process is non‐stationary.
Patient anatomical variation in radiotherapy can be systematically managed using image feedback adaptive treatment technique. The fundamental difference between adaptive treatment technique and other image‐guided techniques is the use of patient‐specific treatment information. Adaptive technique intends to use all patient‐specific dose information ‐ including what has been delivered in the previous treatments, what can be delivered at the present treatment, and what would be delivered in future treatments ‐ in the design of treatment plan. Therefore, treatment plan designed in adaptive radiotherapy is called 4D adaptive plan, which is in principle a treatmentcontrol law to manage treatment process. Mathematically, treatmentcontrol law is a spatial mapping from the parameter space of patient variation to the parameter space of treatment delivery control, which can be determined including the patient variation in the planning optimization or inverse planning. The objectives in adaptive planning optimization are constructed based on a selection of control strategies that could be either the online or the offline with one control action, multiple actions or continue actions. Selection of control strategy and number of control actions is, of course, dependent on the nature of patient variation process as well as the estimation uncertainties, and has to be determined considering also the clinical practice.
The lecture will provide an overview of the models and characteristics of patient anatomical variation process during the radiotherapy, the 4D dose summation methodology and the strategies of adaptive treatment process.
1. Understand the characteristics and dynamic model for patient anatomical variation during the course of radiotherapy.
2. Understand the model and methodology of 4D dose summation.
3. Understand the options of control strategy for image guided adaptive radiation treatment.
33(2006); http://dx.doi.org/10.1118/1.2241426View Description Hide Description
Organ motion blurs dose distributions. The blurring can be described in a statistical way by use of a motion probability (density) function (PDF). The motion‐blurred dose distribution is obtained by a convolution of the “sharp” (static case) dose distribution with the motion PDF. This holds true for both inter‐ and intra‐fraction motions. In the case of intra‐fraction motions an “interplay” effect is superimposed on top of the blurring effect. It has been shown that the interplay effect averages out during the course of a fractionated treatment, and that it is usually negligible after a typical number of fractions. The convolution model relies on the linear superimposition principle, which holds true for dose values but not for the biological effect. This issue has recently been addressed and will be discussed.
Several investigations have now looked at the feasibility of un‐doing the motion blur through the use if intensity‐modulation. In principle it should indeed be possible to de‐convolve the motion PDF from the intensity maps, in order to compensate motion effects. This approach has been called 4D optimization or 4D inverse planning. Motion de‐convolution cannot, however, compensate motion effects exactly and it cannot be applied in a naïve straight‐forward way, because that would lead to undeliverable intensity maps with sharp spikes and negative values. The method of choice is rather to include the motion PDF in the IMRToptimization process. It has been shown that this can indeed yield a surprisingly high degree of motion compensation and it can even compete with other motion compensation methods such as gated delivery. However, this is only true if the motion characteristics (the PDF) are known with great precision. If the actually realized motion PDF deviates substantially from the planned PDF, the method becomes less useful and can, in principle, make things worse. More recently, uncertainties in the knowledge of the motion characteristics have been taken into account by use of robust optimization techniques. With these one can now compensate for motion effects in an approximate way for a large class of motion characteristics. In terms of the sparing of normal structures, the results are in between the use of conventional margins and the idealistic case of perfect motion compensation. The resulting intensity maps exhibit “horns”, which can shave off a few mm from the margins.
1. Understand the concepts of motion blur and PDF.
2. Understand the idea of de‐blurring a dose distribution through “4D” motion optimization.
3. Be able to discuss the relative potential and limitations of 4D motion optimization in comparison with margins and gating.
Joint Imaging‐Therapy Symposium in Memoriam of Edward Webster: In‐Room Non‐Tomographic Guidance of Radiotherapy
33(2006); http://dx.doi.org/10.1118/1.2241764View Description Hide Description
Edward (Ted) W. Webster, Ph.D., 5th AAPM President, 12th Coolidge Awardee, and 16th Taylor Lecturer passed away on 17 December 2005 after a brilliant career of over a half century in medical physics. Besides making many contributions to the physics of diagnostic radiology and to the studies of the biological effects of low doses of radiation, Ted was a superb scholar and renown lecturer who greatly influenced the careers of many hundreds of radiologists and physicists.
33(2006); http://dx.doi.org/10.1118/1.2241765View Description Hide Description
Respiratory motion is both intra‐ and inter‐fraction. For lung and livertumors, the magnitude of inter‐fraction motion is often comparable with that of intra‐fraction motion. The position of a moving organ during one fraction can be decomposed into the average position and the instant position relative to the average position. Here we introduce a concept called daily home position of a moving organ, which is often the mean position averaged over the treatment fraction. However, it may be a position corresponding to a particular breathing phase. For example, if the treatment is gated at the exhale phase, it makes sense to define the tumor home position as the exhale position.
For lung and livertumors,image guided setup means the detection of daily tumor home position and the alignment of this position to the reference home position. If tomographic guidance is used to setup lung or livercancer patients, either using CT on‐rail or on board cone beam CT, the accurate daily tumor home position may be generated from a 4D scan, but not a 3D scan without patient breath holding. Daily breath hold 3D scan is not practical because it is difficult for patients with poor pulmonary function. For daily 4D scan, among others, how to reduce the potentially large amount of radiation dose and how to efficiently manage large amount of image data might be two major problems for its application to patient setup. For livers with implanted fiducial markers, it is easy to detect the marker positions, thus the daily tumor home position, either radiographically or fluoroscopically, using a gantry‐mounted or a room‐mounted x‐ray imagingsystem. For lung patients, we usually do not have any fiducial markers implanted in the lung, due to the concern about the risk of pneumothorax. Radiographic or fluoroscopic detection of lungtumor mass is not a trivial task. For some lungtumors, their projections in the images may be identified using current imaging techniques. For others, more advanced imaging techniques are required and this is still an on‐going research topic.
As to image guided delivery, the requirement on the amount of tumor location information is different, depending on the delivery technique. For beam tracking, the instant tumor position should be tracked. For gating, we only need to make sure the tumor home position is maintained within the gating window during the treatment fraction. It is difficult, if not impossible, to have tomographic guided delivery. Fluoroscopic tracking of implanted fiducial markers is relatively easy. However, we still need to pay attention to various practical issues, such as the changes in marker shape in projection images due to maker rotation for none‐spherical markers, occlusion by and confusion with bony structure and air bubbles, tracking closely located multiple markers, poor image quality with MV beam interference, etc. Fluoroscopic tracking of lungtumor mass without implanted fiducial markers is challenging. Some promising results have been seen recently. However, a lot more research is needed before those techniques can be clinically useful.
33(2006); http://dx.doi.org/10.1118/1.2241766View Description Hide Description
Intra‐fraction and inter‐fraction motion have continued to be areas where much study is needed in order to accurately target areas for radiation treatment. Many modalities have been explored that allow the user to evaluate the patient location from day to day, as well as during a single treatment session. Most of these applications are image‐related, including, both planar X‐rays (either KV or MV), CT scanning (KV or MV), as well as ultrasound. Each imaging modality has its unique set of issues including, time involvement, radiation exposure, and subjectivity in registration).
Non‐tomographic methods to deal with inter and intra‐fraction motion have been explored. For many years optical tacking with the use of infrared cameras have been used for radiosurgery and radiotherapy. These systems will be discussed in brief, but they are limited to external information and offer little information on internal organ motion. They are important in many applications related to target motion in that they can be used to create a model between the patient surface and internal motion. They can also be useful in monitoring other devices within the room. Another non‐tomographic method of aligning a patient relies on radiofrequency signals in order to track an implanted transponder. The Calypso® 4D tracking system is a system that provides RF tracking of Beacon® transponders that can be implanted or placed an immobilization device. Some befits of RF tracking include nearly real‐time (10 HZ frequency) update in location, accurate and objective coordinate information, and non‐ionizing radiation signals.
This lecture will provide an overview of non‐tomographic patient alignment techniques that can allow real‐time positioning. It will also discuss the strengths and weaknesses of such technology compared to tomographic modalities and specific concerns in radiation therapy.
1. Understand some common concerns in inter and intra‐radiation fraction motion.
2. Understand the differences between tomographic and non‐tomographic means of patient motion evaluation.
3. Understand the strengths and weaknesses of an RF tracking system in radiation therapytreatments.
33(2006); http://dx.doi.org/10.1118/1.2241768View Description Hide Description
Fluoroscopic, ultrasonic and 4D CT studies of the organs in the thorax and abdomen have shown that some organs may move as much as 4 cm due to respiratory motion. If the motion is not compensated for during external beam radiation therapy, the dose coverage to target may be compromised. On the other hand, if the motion is compensated for with an increase of margin, a significant amount of normal tissue may be irradiated unnecessarily. The issue of respiratory compensation becomes more important for hypofractionated treatments and even more so for single‐fraction extracranial radiosurgery applications. CyberKnife is an image‐guidedradiosurgery system that consists of a 6‐MV LINAC mounted to a robotic arm coupled through a control loop to a digital diagnostic x‐ray imaging system. The robotic arm can point the beam anywhere in space with six degrees of freedom, without being constrained to a conventional isocenter. The CyberKnife has been recently upgraded with a real‐time respiratory tracking and compensation system called Synchrony. Using external markers in conjunction with diagnostic x‐ray images, Synchrony helps to guide the robotic arm to move the radiation beam in real time such that the beam always remains aligned with the target. With the aid of the Synchrony, the tumor motion can be tracked in three dimensional space, and the motion induced dosimetric change to target can be minimized without an increase in margin. The working principles, advantages, limitations and our clinical experience with this new technology will be discussed.
33(2006); http://dx.doi.org/10.1118/1.2241769View Description Hide Description
Image guided radiotherapy has primarily been implemented with technologies that utilize subsurface imaging.Radiography and ultrasonograpy are established methods to localize the target or critical organs on a daily basis. Imaging the patient surface can also have an important role in IGRT. Video‐based surfaceimaging can be an alternative to radiography in patient setup of selected targets, such as partial breast irradiation (PBI). Video also has a role in continuous monitoring of patient position during treatment. There are various approaches to surfaceimaging; some measure the Cartesian coordinates of multiple discrete optical markers. We have utilized a stereo photogrammetric approach, where a speckle pattern is projected onto the patient surface, and surfacetopology is measured by cameras mounted from the ceiling. A user defined surface ROI from the treatment du jour is matched with the reference surfacetopology, and the rigid body translations and rotations required to bring the two surfaces into optimal congruence are determined.
The intrinsic performance of a surfaceimaging system can be <1mm and <1°, which is more than adequate for most clinical situations. Patient studies have been conducted to quantify the setup accuracy of laser, chestwall, and surfaceimaging, in comparison to ground truth (defined by radiographic clips). In these studies, effects of respiration, breast deformation, and posture variations are analyzed to determine their contribution to the uncertainties in targetry. The target registration error for surfaceimaging is ∼ 1mm relative to clips, and is more accurate than laser or chest wall setup, based on a statistically rigorous analysis.
This lecture will provide an overview of video based surfaceimaging as a tool for patient alignment in image guided therapy. While the focus is on PBI setup, the evaluation and analysis of how to compare IGRT approaches has general validity.
1. Understand basic principles of stereo photogrammetric surface mapping.
2. Understand approaches to characterize and validate system performance.
3. Understand clinical factors that affect the Target Registration Error in IGRT.