Application of video imaging for improvement of patient set-up
Introduction
Portal imaging plays an important role for measurement and improvement of the patient set-up in external beam radiotherapy [1]. Because verification of patient set-up is widely performed, the magnitude of set-up errors for various treatment sites are well documented. There is evidence that for patients with prostate cancer the set-up error along the left–right (LR) axis is larger than the set-up error along the other axes [9]. In our hospital, the random set-up error along the LR axis is larger (1.9 mm standard deviation, SD) than along the anterior–posterior (AP) axis (1.4 mm SD) and the cranio-caudal (CC) axis (1.1 mm SD) [8].
Conventionally, the patient is positioned by aligning markers on the skin to the projection of room lasers. The larger magnitude of the set-up errors in the LR direction may be explained by the fact that only a single marker on the patient's abdomen is available for alignment to the room lasers. For the AP direction, accurate localization of the isocentre of patients treated in the pelvic area is often achieved by using a fixed-height technique [6]. For the CC direction, the set-up is based on the alignment of two lateral markers on the skin and the marker on the abdomen, although a mismatch of this latter marker is allowed. The lateral markers are likely to be more stable than the single marker on the patient's abdomen because of their proximity to the pelvic bone. Therefore the marker on the abdomen is ignored when it is not possible to align all three markers simultaneously.
Imaging of the skin of the patient may improve the set-up of the patient prior to irradiation. Many institutions use correction protocols to reduce systematic set-up errors [2]. When the initial patient set-up is improved, fewer corrections are required because the set-up errors remain within the limits. Since corrections are labour intensive, a reduction of the workload of the radiation technologists and of the number of acquired portal images might be achieved. This is especially important for intensity-modulated radiotherapy, where treatment fields are typically too small for patient set-up verification purposes, making the use of enlarged localization images necessary.
Various authors proposed incorporation of body contour information to improve the set-up accuracy [7], [13], [16]. Some studies showed improvement of the set-up accuracy of prostate cancer patients in the lateral and AP directions [13], [16]. Registration of surface data to a model of the patient requires a dedicated contouring device to acquire a three-dimensional (3D) surface description. Currently, there are no solutions commercially available for use in the treatment room.
Other groups utilize video cameras to improve the accuracy of the patient set-up [11], [14]. The methods described by Milliken et al. and Johnson et al. are aimed at computer-assisted set-up of patients. They used an interactive technique that is based on real-time subtraction of images. In this technique, live video images are aligned with a reference video image, usually taken during the first radiation fraction. The advantage of this technique is that it is fast, sensitive and easy to implement. A disadvantage, however, is that any set-up error at the time that the reference video image is acquired results in a systematic error over the course of the treatment. Furthermore, using the subtraction method, the accuracy may be compromised when the outer contour changes during treatment, i.e. due to respiration or weight loss.
In the above-mentioned studies, the inherent inaccuracy of video imaging with respect to portal imaging was not quantified in detail. When bone is considered as a reference of the target volume, movement of the skin of the patient with respect to the bony anatomy limits the overall accuracy. An estimate for the magnitude of movement of the skin of the patient relative to bony anatomy is therefore required to evaluate the potential use of a video-based technique.
In this paper a straightforward solution for improvement of the set-up accuracy is presented. We propose to use a video camera and planning CT data to improve patient positioning in the LR direction only, i.e. to solve problems related to the use of a single marker on the patient's abdomen. In this study, a retrospective evaluation of video images is performed and the results are compared with patient set-up errors determined from portal images. The aim of this feasibility study is to test the potential improvement of the patient set-up along the LR axis using a room-mounted video camera compared to the conventional patient set-up using the marker on the patient's abdomen.
Section snippets
Material and methods
Video images were acquired with a room-mounted camera. The position of this camera with respect to the isocentre and the scale of the images was established by imaging a phantom. Video images and portal images of patients with prostate cancer were acquired. Because two different modalities were used for estimating the set-up error, we will refer to the set-up error estimated from video images as the video set-up error and the error estimated from portal images as the portal set-up error. The
Calibration
Calibration measurements of the position of the camera were started directly after installation, at day 0. Shortly after installation, a relatively large drift with respect to the initial camera pose was observed (Fig. 5). The main source of this drift was sagging around the LR axis of the camera, probably caused by the weight of the zoom lens. Because the camera was attached to the ceiling of the treatment room under an oblique angle of approximately 50°, this sagging appeared as a translation
Discussion
Video imaging is an accurate technique for estimating the portal set-up error of prostate patients in the LR direction, i.e. a good correspondence between the results based on portal imaging and video imaging was observed. Because the set-up of the patients was verified by our correction protocol that aims to remove systematic errors, almost no reduction of the systematic portal set-up error in the LR direction was observed. However, a significant improvement of the random component of the
Conclusions
Video imaging is an accurate method for determining the LR position of patients with prostate cancer. The outline of the patient is a more accurate estimate of the set-up of the bony anatomy than the marker on the patient's abdomen. Thus, improvement of set-up of patients with prostate cancer in the LR direction by means of video imaging is possible, depending on the accuracy of set-up corrections. It is anticipated that the additional work of analysing the video images will be compensated by
Acknowledgements
This work was financially supported by the Dutch Cancer Society, Grant No. NKI 98-1682. The authors would like to thank Dr. Jan-Jakob Sonke and Jochem Wolthaus, MSc, for their careful review of the manuscript. We gratefully acknowledge Peter Groote, Elmi Westra and Ton Vlasveld for designing and constructing the phantom and Peter van de Ven and Nico Jessurun for deploying the camera system.
References (18)
- et al.
High-precision prostate cancer irradiation by clinical application of an offline patient setup verification procedure, using portal imaging
Int J Radiat Oncol Biol Phys
(1996) - et al.
Maximizing setup accuracy using portal images as applied to a conformal boost technique for prostatic cancer
Radiother Oncol
(1992) - et al.
Comparison of two methods for anterior–posterior isocenter localization in pelvic radiotherapy using electronic portal imaging
Int J Radiat Oncol Biol Phys
(1998) - et al.
Set-up verification using portal imaging; review of current clinical practice
Radiother Oncol
(2001) - et al.
Flat-panel cone-beam computed tomography for image-guided radiation therapy
Int J Radiat Oncol Biol Phys
(2002) - et al.
Initial clinical experience with a video-based patient positioning system
Int J Radiat Oncol Biol Phys
(1999) - et al.
Daily prostate targeting using implanted radiopaque markers
Int J Radiat Oncol Biol Phys
(2002) - et al.
Patient positioning using detailed three-dimensional surface data for patients undergoing conformal radiation therapy for carcinoma of the prostate: a feasibility study
Int J Radiat Oncol Biol Phys
(2001) - et al.
Performance of a video-image-subtraction-based patient positioning system
Int J Radiat Oncol Biol Phys
(1997)
Cited by (26)
Upright patient positioning for pelvic radiotherapy treatments
2022, Technical Innovations and Patient Support in Radiation OncologyCitation Excerpt :Based on bony registration, a good correlation between both devices was found. Ploeger et al analysed setup error of 22 patients undergoing prostate radiotherapy by comparing the shifts detected by video images with those obtained with an EPID in right-left direction and also reported a good correlation between both devices [8]. Bartocini et al also assessed the correlation between SGRT and EPID for 19 patients treated by prostate radiotherapy.
3D Reconstruction and alignment by consumer RGB-D sensors and fiducial planar markers for patient positioning in radiation therapy
2019, Computer Methods and Programs in BiomedicineCitation Excerpt :In this case, the system automatically corrects the position of the head by a motorized mechanism that corrects its position by translating it so that the isocenter is focused on the correct position. Later, Ploeger et al. [12] used image matching between a video recorded in the treatment phase and reference CT scan using body contours. They concluded that the outline of the patient’s body is a more accurate reference that the markers put on their abdomen.
Investigation of the geometric accuracy of proton beam irradiation in the liver
2012, International Journal of Radiation Oncology Biology PhysicsCitation Excerpt :Many methods have been investigated and clinically used to improve the geometric accuracy of irradiation. Methods of managing the interfraction setup variation include the use of daily portal images (1), fiducial markers (2), computed tomography (CT) simulation (3), and video cameras (4). Methods of managing intrafraction breathing movements include the use of breath control (5), respiratory-gated radiotherapy (6), and a tumor-tracking system (7).
Intraprostatic fiducials for image guidance: Workflow implications in a single linac department
2008, RadiographyCitation Excerpt :If we assume that treatment staff have the same level of competence correcting to fiducials as to bony anatomy then the implications for increased accuracy are clear. It has been long considered14,15,16 that to introduce a complete online correction protocol would require a considerable increase in workload, i.e. daily pre treatment orthogonal images, with very little benefit.14,15,17 However with such an approach using gold seed fiducials it is quite clear that margins can be significantly reduced and the possibilities of dose escalation are presented.
Use of the BrainLAB ExacTrac X-Ray 6D System in Image-Guided Radiotherapy
2008, Medical DosimetryCitation Excerpt :Image guidance plays an important role in radiosurgery and intensity modulated radiotherapy (IMRT) because it supports accurate target localization and avoidance of adjacent organs-at-risks (OAR). A number of imaging modalities including: ultrasound,1,2 video imaging,3,4 2-dimensional radiographic imaging (kV and MV),5–7 computed tomography (CT) (conventional CT as well as kV and MV cone beam CT),8–12 and magnetic resonance imaging (MRI)13–15 are used in image-guided radiotherapy (IGRT). Several IGRT systems are now commercially available and have been successfully implemented for clinical applications.