Automated 2D–3D registration of portal images and CT data using line‐segment enhancement

Abstract

In prostate radiotherapy, setup errors with respect to the patient's bony anatomy can be reduced by aligning 2D megavoltage ͑MV͒ portal images acquired during treatment to a reference 3D kilovoltage ͑kV͒ CT acquired for treatment planning purposes. The purpose of this study was to evaluate a fully automated 2D-3D registration algorithm to quantify setup errors in 3D through the alignment of line-enhanced portal images and digitally reconstructed radiographs computed from the CT. The line-enhanced images were obtained by correlating the images with a filter bank of short line segments, or "sticks" at different orientations. The proposed methods were validated on ͑1͒ accurately collected gold-standard data consisting of a 3D kV cone-beam CT scan of an anthropomorphic phantom of the pelvis and 2D MV portal images in the anterior-posterior ͑AP͒ view acquired at 15 different poses and ͑2͒ a conventional 3D kV CT scan and weekly 2D MV AP portal images of a patient over 8 weeks. The mean ͑and standard deviation͒ of the absolute registration error for rotations around the right-lateral ͑RL͒, inferior-superior ͑IS͒, and posterior-anterior ͑PA͒ axes were 0.212°͑0.214°͒, 0.055°͑0.033°͒ and 0.041°͑0.039°͒, respectively. The corresponding registration errors for translations along the RL, IS, and PA axes were 0.161͑0.131͒ mm, 0.096͑0.033͒ mm, and 0.612͑0.485͒ mm. The mean ͑and standard deviation͒ of the total registration error was 0.778͑0.543͒ mm. Registration on the patient images was successful in all eight cases as determined visually. The results indicate that it is feasible to automatically enhance features in MV portal images of the pelvis for use within a completely automated 2D-3D registration framework for the accurate determination of patient setup errors. They also indicate that it is feasible to estimate all six transformation parameters from a 3D CT of the pelvis and a single portal image in the AP view.

2D‐3D registration for prostate radiation therapy based on a statistical model of transmission images

“…Algorithms for 2D-3D registration of 2D radiographs ͑projection radiographs͒ to 3D CT images can be broadly categorized as feature-based methods, 1-6 methods that use intensity-derived features, [7][8][9][10][11] intensity-based methods, 12,8,[13][14][15][16][17]9,18 and hybrid methods. [19][20][21] Key to intensity-based registration is the choice of an appropriate similarity measure to quantify the spatial alignment between the images.…”

A new multi-modal similarity measure for fast gradient-based 2D-3D image registration

“…The proposed 2D-3D registration algorithm also includes a new technique to calculate the gradients of the registration error with respect to out-of-plane rotations R x and R y for use in (7). The required formulae for these gradients are not easily defined since these parameters describe 3D motion Fig.…”

“…The optimization procedures used in previously reported single-plane 2D-3D algorithms include many non-gradient approaches (see for example [5,6]) and gradient-based approaches using numerically calculated gradients (see for example [7,8]). …”

Registration of clinical volumes to beams‐eye‐view images for real‐time tracking

“…Methods to confront the inherent low contrast with image enhancement techniques are an attractive option to assist in 2D/3D registration. 26 No such enhancement was used in this work and accuracy and robustness could potentially improve with additional image processing. However, for positions near EoE, where the tumor remains in the field of view, accurate registration is feasible for real time applications.…”