Image registration and fusion algorithms exist in almost every software system that creates or uses images in radiotherapy. Most treatment planning systems support some form of image registration and fusion to allow the use of multimodality and time-series image data and even anatomical atlases to assist in target volume and normal tissue delineation. Treatment delivery systems perform registration and fusion between the planning images and the in-room images acquired during the treatment to assist patient positioning. Advanced applications are beginning to support daily dose assessment and enable adaptive radiotherapy using image registration and fusion to propagate contours and accumulate dose between image data taken over the course of therapy to provide up-to-date estimates of anatomical changes and delivered dose. This information aids in the detection of anatomical and functional changes that might elicit changes in the treatment plan or prescription.As the output of the image registration process is always used as the input of another process for planning or delivery, it is important to understand and communicate the uncertainty associated with the software in general and the result of a specific registration. Unfortunately, there is no standard mathematical formalism to perform this for real-world situations where noise, distortion, and complex anatomical variations can occur. Validation of the software systems performance is also complicated by the lack of documentation available from commercial systems leading to use of these systems in undesirable 'black-box' fashion.In view of this situation and the central role that image registration and fusion play in treatment planning and delivery, the Therapy Physics Committee of the American Association of Physicists in Medicine commissioned Task Group 132 to review current approaches and solutions for image registration (both rigid and deformable) in radiotherapy and to provide recommendations for quality assurance and quality control of these clinical processes.

The looming potential of deformable alignment tools to play an integral role in adaptive radiotherapy suggests a need for objective assessment of these complex algorithms. Previous studies in this area are based on the ability of alignment to reproduce analytically generated deformations applied to sample image data, or use of contours or bifurcations as ground truth for evaluation of alignment accuracy. In this study, a deformable phantom was embedded with 48 small plastic markers, placed in regions varying from high contrast to roughly uniform regional intensity, and small to large regional discontinuities in movement. CT volumes of this phantom were acquired at different deformation states. After manual localization of marker coordinates, images were edited to remove the markers. The resulting image volumes were sent to five collaborating institutions, each of which has developed previously published deformable alignment tools routinely in use. Alignments were done, and applied to the list of reference coordinates at the inhale state. The transformed coordinates were compared to the actual marker locations at exhale. A total of eight alignment techniques were tested from the six institutions. All algorithms performed generally well, as compared to previous publications. Average errors in predicted location ranged from 1.5 to 3.9 mm, depending on technique. No algorithm was uniformly accurate across all regions of the phantom, with maximum errors ranging from 5.1 to 15.4 mm. Larger errors were seen in regions near significant shape changes, as well as areas with uniform contrast but large local motion discontinuity. Although reasonable accuracy was achieved overall, the variation of error in different regions suggests caution in globally accepting the results from deformable alignment.

A B S T R A C T PurposeTo study whether changes of [ 18 F]fluorodeoxyglucose positron emission tomography (FDG-PET) during treatment correlate with post-treatment responses in tumor and normal lung in patients with non-small-cell lung cancer (NSCLC). Patients and MethodsPatients with stage I to III NSCLC requiring a definitive dose of fractionated radiation therapy (RT) were eligible. FDG-PET/computed tomography scans were acquired before, during, and after RT. Tumor and lung metabolic responses were assessed qualitatively by physicians and quantitatively by normalized peak FDG activity (the ratio of the maximum FDG activity divided by the mean of the aortic arch background). ResultsThe study reached the goal of recruiting 15 patients between February 2004 and August 2005. Of these, 11 patients had partial metabolic response, two patients had complete metabolic response, and two patients had stable disease at approximately 45 Gy during RT. The mean peak tumor FDG activity was 5.2 (95% CI, 4.0 to 6.4), 2.5 (95% CI, 2.0 to 3.0), and 1.7 (95% CI, 1.3 to 2.0) on pre-, during, and post-RT scans, respectively. None of the patients had appreciable changes in the lung during RT. The peak FDG activity of the lung was 0.47 (95% CI, 0.36 to 0.59), 0.52 (95% CI, 0.40 to 0.64), and 1.29 (95% CI, 0.82 to 1.76), on pre-, during-, and post-RT scans, respectively. The qualitative response during RT correlated with the overall response post-RT (P ϭ .03); the peak tumor FDG activity during RT correlated with those 3 months post-RT (R 2 ϭ 0.7; P Ͻ .001). ConclusionThis pilot study suggests a significant correlation in tumor metabolic response and no association in lung FDG activity between during RT scans and 3 months post-RT scans in patients with NSCLC. Additional study with a large number of patients is needed to validate these findings.

Contrast MRI reveals increases in Gd-DTPA uptake in the initially nonenhanced tumor region but not in the remaining brain during the course of RT, suggesting opening of the BTB. This finding suggests that the effect of conformal radiation is more selective on the BTB than the BBB, and there may be a window extending from 1 week after the initiation of radiotherapy to 1 month after the completion of treatment during which a pharmaceutical agent has maximum access to high-grade gliomas.

This paper provides an overview of image registration and data fusion techniques used in radiation therapy, and examples of their use. They are used at all stages of the patient management process; for initial diagnosis and staging, during treatment planning and delivery, and after therapy to help monitor the patients' response to treatment. Most treatment planning systems now support some form of interactive or automated image registration and provide tools for mapping information, such as tissue outlines and computed dose from one imaging study to another. To complement this, modern treatment delivery systems offer means for acquiring and registering 2D and 3D image data at the treatment unit to aid patient setup. Techniques for adapting and customizing treatments during the course of therapy using 3D and 4D anatomic and functional imaging data are currently being introduced into the clinic. These techniques require sophisticated image registration and data fusion technology to accumulate properly the delivered dose and to analyse possible physiological and anatomical changes during treatment. Finally, the correlation of radiological changes after therapy with delivered dose also requires the use of image registration and fusion techniques.

The advent of dynamic radiotherapy modeling and treatment techniques requires an infrastructure to weigh the merits of various interventions (breath holding, gating, tracking). The creation of treatment planning models that account for motion and deformation can allow the relative worth of such techniques to be evaluated. In order to develop a treatment planning model of a moving and deforming organ such as the lung, registration tools that account for deformation are required. We tested the accuracy of a mutual information based image registration tool using thin-plate splines driven by the selection of control points and iterative alignment according to a simplex algorithm. Eleven patients each had sequential CT scans at breath-held normal inhale and exhale states. The exhale right lung was segmented from CT and served as the reference model. For each patient, thirty control points were used to align the inhale CT right lung to the exhale CT right lung. Alignment accuracy (the standard deviation of the difference in the actual and predicted inhale position) was determined from locations of vascular and bronchial bifurcations, and found to be 1.7, 3.1, and 3.6 mm about the RL, AP, and IS directions. The alignment accuracy was significantly different from the amount of measured movement during breathing only in the AP and IS directions. The accuracy of alignment including thin-plate splines was more accurate than using affine transformations and the same iteration and scoring methodology. This technique shows promise for the future development of dynamic models of the lung for use in four-dimensional (4-D) treatment planning.

In this study we investigated the accumulation of dose to a deforming anatomy (such as lung) based on voxel tracking and by using time weighting factors derived from a breathing probability distribution function (p.d.f.). A mutual information registration scheme (using thin-plate spline warping) provided a transformation that allows the tracking of points between exhale and inhale treatment planning datasets (and/or intermediate state scans). The dose distributions were computed at the same resolution on each dataset using the Dose Planning Method (DPM) Monte Carlo code. Two accumulation/interpolation approaches were assessed. The first maps exhale dose grid points onto the inhale scan, estimates the doses at the "tracked" locations by trilinear interpolation and scores the accumulated doses (via the p.d.f.) on the original exhale data set. In the second approach, the "volume" associated with each exhale dose grid point (exhale dose voxel) is first subdivided into octants, the center of each octant is mapped to locations on the inhale dose grid and doses are estimated by trilinear interpolation. The octant doses are then averaged to form the inhale voxel dose and scored at the original exhale dose grid point location. Differences between the interpolation schemes are voxel size and tissue density dependent, but in general appear primarily only in regions with steep dose gradients (e.g., penumbra). Their magnitude (small regions of few percent differences) is less than the alterations in dose due to positional and shape changes from breathing in the first place. Thus, for sufficiently small dose grid point spacing, and relative to organ motion and deformation, differences due solely to the interpolation are unlikely to result in clinically significant differences to volume-based evaluation metrics such as mean lung dose (MLD) and tumor equivalent uniform dose (gEUD). The overall effects of deformation vary among patients. They depend on the tumor location, field size, volume expansion, tissue heterogeneity, and direction of tumor displacement with respect to the beam, and are more likely to have an impact on serial organs (such as esophagus), rather than on large parallel organs (such as lung).

No abstract