Kenneth J. Ruchala

The management of imaging dose during image-guided radiotherapy: Report of the AAPM Task Group 75

Radiographic image guidance has emerged as the new paradigm for patient positioning, target localization, and external beam alignment in radiotherapy. Although widely varied in modality and method, all radiographic guidance techniques have one thing in common--they can give a significant radiation dose to the patient. As with all medical uses of ionizing radiation, the general view is that this exposure should be carefully managed. The philosophy for dose management adopted by the diagnostic imaging community is summarized by the acronym ALARA, i.e., as low as reasonably achievable. But unlike the general situation with diagnostic imaging and image-guided surgery, image-guided radiotherapy (IGRT) adds the imaging dose to an already high level of therapeutic radiation. There is furthermore an interplay between increased imaging and improved therapeutic dose conformity that suggests the possibility of optimizing rather than simply minimizing the imaging dose. For this reason, the management of imaging dose during radiotherapy is a different problem than its management during routine diagnostic or image-guided surgical procedures. The imaging dose received as part of a radiotherapy treatment has long been regarded as negligible and thus has been quantified in a fairly loose manner. On the other hand, radiation oncologists examine the therapy dose distribution in minute detail. The introduction of more intensive imaging procedures for IGRT now obligates the clinician to evaluate therapeutic and imaging doses in a more balanced manner. This task group is charged with addressing the issue of radiation dose delivered via image guidance techniques during radiotherapy. The group has developed this charge into three objectives: (1) Compile an overview of image-guidance techniques and their associated radiation dose levels, to provide the clinician using a particular set of image guidance techniques with enough data to estimate the total diagnostic dose for a specific treatment scenario, (2) identify ways to reduce the total imaging dose without sacrificing essential imaging information, and (3) recommend optimization strategies to trade off imaging dose with improvements in therapeutic dose delivery. The end goal is to enable the design of image guidance regimens that are as effective and efficient as possible.

Fast free-form deformable registration via calculus of variations

In this paper, we present a fully automatic, fast and accurate deformable registration technique. This technique deals with free-form deformation. It minimizes an energy functional that combines both similarity and smoothness measures. By using calculus of variations, the minimization problem was represented as a set of nonlinear elliptic partial differential equations (PDEs). A Gauss-Seidel finite difference scheme is used to iteratively solve the PDE. The registration is refined by a multi-resolution approach. The whole process is fully automatic. It takes less than 3 min to register two three-dimensional (3D) image sets of size 256 x 256 x 61 using a single 933 MHz personal computer. Extensive experiments are presented. These experiments include simulations, phantom studies and clinical image studies. Experimental results show that our model and algorithm are suited for registration of temporal images of a deformable body. The registration of inspiration and expiration phases of the lung images shows that the method is able to deal with large deformations. When applied to the daily CT images of a prostate patient, the results show that registration based on iterative refinement of displacement field is appropriate to describe the local deformations in the prostate and the rectum. Similarity measures improved significantly after the registration. The target application of this paper is for radiotherapy treatment planning and evaluation that incorporates internal organ deformation throughout the course of radiation therapy. The registration method could also be equally applied in diagnostic radiology.

The use of megavoltage CT (MVCT) images for dose recomputations.

Megavoltage CT (MVCT) images of patients are acquired daily on a helical tomotherapy unit (TomoTherapy, Inc., Madison, WI). While these images are used primarily for patient alignment, they can also be used to recalculate the treatment plan for the patient anatomy of the day. The use of MVCT images for dose computations requires a reliable CT number to electron density calibration curve. In this work, we tested the stability of the MVCT numbers by determining the variation of this calibration with spatial arrangement of the phantom, time and MVCT acquisition parameters. The two calibration curves that represent the largest variations were applied to six clinical MVCT images for recalculations to test for dosimetric uncertainties. Among the six cases tested, the largest difference in any of the dosimetric endpoints was 3.1% but more typically the dosimetric endpoints varied by less than 2%. Using an average CT to electron density calibration and a thorax phantom, a series of end-to-end tests were run. Using a rigid phantom, recalculated dose volume histograms (DVHs) were compared with plan DVHs. Using a deformed phantom, recalculated point dose variations were compared with measurements. The MVCT field of view is limited and the image space outside this field of view can be filled in with information from the planning kVCT. This merging technique was tested for a rigid phantom. Finally, the influence of the MVCT slice thickness on the dose recalculation was investigated. The dosimetric differences observed in all phantom tests were within the range of dosimetric uncertainties observed due to variations in the calibration curve. The use of MVCT images allows the assessment of daily dose distributions with an accuracy that is similar to that of the initial kVCT dose calculation.

Assessment of parotid gland dose changes during head and neck cancer radiotherapy using daily megavoltage computed tomography and deformable image registration.

PURPOSE To analyze changes in parotid gland dose resulting from anatomic changes throughout a course of radiotherapy in a cohort of head-and-neck cancer patients. METHODS AND MATERIALS The study population consisted of 10 head-and-neck cancer patients treated definitively with intensity-modulated radiotherapy on a helical tomotherapy unit. A total of 330 daily megavoltage computed tomography images were retrospectively processed through a deformable image registration algorithm to be registered to the planning kilovoltage computed tomography images. The process resulted in deformed parotid contours and voxel mappings for both daily and accumulated dose-volume histogram calculations. The daily and cumulative dose deviations from the original treatment plan were analyzed. Correlations between dosimetric variations and anatomic changes were investigated. RESULTS The daily parotid mean dose of the 10 patients differed from the plan dose by an average of 15%. At the end of the treatment, 3 of the 10 patients were estimated to have received a greater than 10% higher mean parotid dose than in the original plan (range, 13-42%), whereas the remaining 7 patients received doses that differed by less than 10% (range, -6-8%). The dose difference was correlated with a migration of the parotids toward the high-dose region. CONCLUSIONS The use of deformable image registration techniques and daily megavoltage computed tomography imaging makes it possible to calculate daily and accumulated dose-volume histograms. Significant dose variations were observed as result of interfractional anatomic changes. These techniques enable the implementation of dose-adaptive radiotherapy.

Megavoltage CT image reconstruction during tomotherapy treatments.

An integrated tomotherapy system allows for improved radiotherapy verification by enabling the collection of megavoltage computed tomography (MVCT) images before or after treatment delivery. In this investigation, the possibility of collecting MV tomographic data and reconstructing images during a tomotherapy treatment is examined. By overcoming difficulties with the normalization of modulated treatment data and with the incompleteness of treatment data, it is possible to use data collected during tomotherapeutic treatments for MVCT reconstruction. The benefits of these techniques include potential increases in patient throughput, reductions in imaging dose, visualization of the patient in the treatment position and improvements in image contrast.

Automatic re-contouring in 4D radiotherapy

Delineating regions of interest (ROIs) on each phase of four-dimensional (4D) computed tomography (CT) images is an essential step for 4D radiotherapy. The requirement of manual phase-by-phase contouring prohibits the routine use of 4D radiotherapy. This paper develops an automatic re-contouring algorithm that combines techniques of deformable registration and surface construction. ROIs are manually contoured slice-by-slice in the reference phase image. A reference surface is constructed based on these reference contours using a triangulated surface construction technique. The deformable registration technique provides the voxel-to-voxel mapping between the reference phase and the test phase. The vertices of the reference surface are displaced in accordance with the deformation map, resulting in a deformed surface. The new contours are reconstructed by cutting the deformed surface slice-by-slice along the transversal, sagittal or coronal direction. Since both the inputs and outputs of our automatic re-contouring algorithm are contours, it is relatively easy to cope with any treatment planning system. We tested our automatic re-contouring algorithm using a deformable phantom and 4D CT images of six lung cancer patients. The proposed algorithm is validated by visual inspections and quantitative comparisons of the automatic re-contours with both the gold standard segmentations and the manual contours. Based on the automatic delineated ROIs, changes of tumour and sensitive structures during respiration are quantitatively analysed. This algorithm could also be used to re-contour daily images for treatment evaluation and adaptive radiotherapy.

A feasible method for clinical delivery verification and dose reconstruction in tomotherapy

Delivery verification is the process in which the energy fluence delivered during a treatment is verified. This verified energy fluence can be used in conjunction with an image in the treatment position to reconstruct the full three-dimensional dose deposited. A method for delivery verification that utilizes a measured database of detector signal is described in this work. This database is a function of two parameters, radiological path-length and detector-to-phantom distance, both of which are computed from a CT image taken at the time of delivery. Such a database was generated and used to perform delivery verification and dose reconstruction. Two experiments were conducted: a simulated prostate delivery on an inhomogeneous abdominal phantom, and a nasopharyngeal delivery on a dog cadaver. For both cases, it was found that the verified fluence and dose results using the database approach agreed very well with those using previously developed and proven techniques. Delivery verification with a measured database and CT image at the time of treatment is an accurate procedure for tomotherapy. The database eliminates the need for any patient-specific, pre- or post-treatment measurements. Moreover, such an approach creates an opportunity for accurate, real-time delivery verification and dose reconstruction given fast image reconstruction and dose computation tools.

Evaluation of geometric changes of parotid glands during head and neck cancer radiotherapy using daily MVCT and automatic deformable registration

BACKGROUND AND PURPOSE To assess and evaluate geometrical changes in parotid glands using deformable image registration and megavoltage CT (MVCT) images. METHODS A deformable registration algorithm was applied to 330 daily MVCT images (10 patients) to create deformed parotid contours. The accuracy and robustness of the algorithm was evaluated through visual review, comparison with manual contours, and precision analysis. Temporal changes in the parotid gland geometry were observed. RESULTS The deformed parotid contours were qualitatively judged to be acceptable. Compared with manual contours, the uncertainties of automatically deformed contours were similar with regard to geometry and dosimetric endpoint. The day-to-day variations (1 standard deviation of errors) in the center-of-mass distance and volume were 1.61mm and 4.36%, respectively. The volumes tended to decrease with a median total loss of 21.3% (6.7-31.5%) and a median change rate of 0.7%/day (0.4-1.3%/day). Parotids migrated toward the patient center with a median total distance change of -5.26mm (0.00 to -16.35mm) and a median change rate of -0.22mm/day (0.02 to -0.56mm/day). CONCLUSION The deformable image registration and daily MVCT images provide an efficient and reliable assessment of parotid changes over the course of a radiation therapy.

Patient Dose From Megavoltage Computed Tomography Imaging

PURPOSE Megavoltage computed tomography (MVCT) can be used daily for imaging with a helical tomotherapy unit for patient alignment before treatment delivery. The purpose of this investigation was to show that the MVCT dose can be computed in phantoms, and further, that the dose can be reported for actual patients from MVCT on a helical tomotherapy unit. METHODS AND MATERIALS An MVCT beam model was commissioned and verified through a series of absorbed dose measurements in phantoms. This model was then used to retrospectively calculate the imaging doses to the patients. The MVCT dose was computed for five clinical cases: prostate, breast, head/neck, lung, and craniospinal axis. RESULTS Validation measurements in phantoms verified that the computed dose can be reported to within 5% of the measured dose delivered at the helical tomotherapy unit. The imaging dose scaled inversely with changes to the CT pitch. Relative to a normal pitch of 2.0, the organ dose can be scaled by 0.67 and 2.0 for scans done with a pitch of 3.0 and 1.0, respectively. Typical doses were in the range of 1.0-2.0 cGy, if imaged with a normal pitch. The maximal organ dose calculated was 3.6 cGy in the neck region of the craniospinal patient, if imaged with a pitch of 1.0. CONCLUSION Calculation of the MVCT dose has shown that the typical imaging dose is approximately 1.5 cGy per image. The uniform MVCT dose delivered using helical tomotherapy is greatest when the anatomic thickness is the smallest and the pitch is set to the lowest value.

A simple fixed-point approach to invert a deformation fielda)

Inversion of deformation fields is applied frequently to map images, dose, and contours between the reference frame and the study frame. A prevailing approach that takes the negative of the forward deformation as the inverse deformation is oversimplified and can cause large errors for large deformations or deformations that are composites of several deformations. Other approaches, including Newtons method and scatter data interpolation, either require the first derivative or are very inefficient. Here we propose an iterative approach that is easy to implement, converges quickly to the inverse when it does, and works for a majority of cases in practice. Our approach is rooted in fixed-point theory. We build a sequence to approximate the inverse deformation through iterative evaluation of the forward deformation. A sufficient but not necessary convergence condition (Lipschitz condition) and its proof are also given. Though this condition guarantees the convergence, it may not be met for an arbitrary deformation field. One should always check whether the inverse exists for the given forward deformation field by calculating its Jacobian. If nonpositive values of the Jacobian occur only for few voxels, this method will usually converge to a pseudoinverse. In case the iteration fails to converge, one should switch to other means of finding the inverse. We tested the proposed method on simulated 2D data and real 3D computed tomography data of a lung patient and compared our method with two implementations in the Insight Segmentation and Registration Toolkit (ITK). Typically less than ten iterations are needed for our method to get an inverse deformation field with clinically relevant accuracy. Based on the test results, our method is about ten times faster and yet ten times more accurate than ITKs iterative method for the same number of iterations. Simulations and real data tests demonstrated the efficacy and the accuracy of the proposed algorithm.