DIRART – A Software Suite for Deformable Image Registration and Adaptive Radiotherapy Research

Abstract

Purpose: Recent years have witnessed tremendous progress in image guide radiotherapy technology and a growing interest in the possibilities for adapting treatment planning and delivery over the course of treatment. One obstacle faced by the research community has been the lack of a comprehensive open-source software toolkit dedicated for adaptive radiotherapy ͑ART͒. To address this need, the authors have developed a software suite called the Deformable Image Registration and Adaptive Radiotherapy Toolkit ͑DIRART͒. Methods: DIRART is an open-source toolkit developed in MATLAB. It is designed in an objectoriented style with focus on user-friendliness, features, and flexibility. It contains four classes of DIR algorithms, including the newer inverse consistency algorithms to provide consistent displacement vector field in both directions. It also contains common ART functions, an integrated graphical user interface, a variety of visualization and image-processing features, dose metric analysis functions, and interface routines. These interface routines make DIRART a powerful complement to the Computational Environment for Radiotherapy Research ͑CERR͒ and popular image-processing toolkits such as ITK. Results: DIRART provides a set of image processing/registration algorithms and postprocessing functions to facilitate the development and testing of DIR algorithms. It also offers a good amount of options for DIR results visualization, evaluation, and validation. Conclusions: By exchanging data with treatment planning systems via DICOM-RT files and CERR, and by bringing image registration algorithms closer to radiotherapy applications, DIRART is potentially a convenient and flexible platform that may facilitate ART and DIR research.

Contextual Atlas Regression Forests: Multiple-Atlas-Based Automated Dose Prediction in Radiation Therapy

“…This is the first paper to test any of these approaches for direct dose prediction from CT images. In particular we compare two atlas-image matching algorithms: deformable atlas registration (DAR) via meansquared-error using the Horn-Schunck optical flow algorithm [34] as implemented in the open-source Deformable Image Registration and Adaptive Radiotherapy Matlab toolbox [35]; and our ARF. We defer discussion of our choice in deformable registration algorithm to Section IV.…”

“…The excellent upper-bound performance of the DAR on the breast-tangent and prostate plan classes indicates the algorithm is appropriate for this context. We also experimented with the Demons algorithm available in [35] on a small set of images and found the optical flow method to provide better results in less time. Comparing the upper-bounds of ARF with DARs demonstrates that canonical DARs are more specific, but ARFs are more generalizable.…”

“…Deformable registration is carried out between the CT images with the default parameters in [35]. Registration is performed in 4 multi-resolution phases at scales of: 1/32, 1/16, 1/8, and 1/4.…”

The correlation of tissue motion within the lung: implications on fiducial based treatments

“…Both inhalation and exhalation scans, along with the contours, were exported and loaded into custom software developed in MATLAB (2007, The MathWorks) in order to perform deformable registration. 17 Deformable registration was performed using the HornSchunck optical flow algorithm 18 in order to obtain motion vectors between exhalation and inhalation for each voxel within the thoracic cavity. After the vectors were obtained for each patient, they were manually inspected to ensure the vector magnitude increased near the diaphragm and the motion vectors matched with the motion of internal high contrast landmarks, such as bronchial branch points.…”

Automatic detection of patient identification and positioning errors in radiation therapy treatment using 3-dimensional setup images

“…Because of the inherent difference in CT number between MVCT and kVCT images for high atomic number tissues, an experimentally derived third-order polynomial was used to remap kVCT numbers to MVCT numbers. 19 Air voxels were filtered out using an absolute threshold of − 700 Hounsfield units.…”