Progress in the evaluation of electronic portal imaging-taking one step at a time

Abstract

There is increasing interest in the quality assurance of radiation therapy, particularly with regard to the quantification of field displacements due to inaccuracies in patient set-up. Although most studies are still based on the analysis of portal films, a number of recent reports have used electronic portal imagers (3, 4, 7, 10, 13, 21). In some studies the systematic set-up errors are dominant, while in others random day-to-day errors are also significant. Their relative magnitudes appear to depend on the treatment site, patient obesity, immobilization technique, and other conditions and procedures specific to each center. Electronic portal imaging (EPI) provides not only a more effective way of determining field displacements than film, but the immediate availability of the images opens up new possibilities for the reduction of set-up errors during the course of treatment. It is of interest to ask what is the current status of EPI, and what are the prospects for its routine application in the clinical setting? The study by Herman et al. (15) in this issue examines the clinical efficacy of EPI by asking whether the digital images are as good as portal films for identifying set-up errors. They found that in 63% they were as good, and in 7% were better than film. Improvements in image quality and larger detectors to fully accommodate pelvic fields will answer some of the concerns raised in this study. Of greater interest was the finding that visual comparison of the portal image with the simulator film was not a reliable indicator of set-up error, and 14% of the accepted fields still had set-up errors of 510 mm as determined by retrospective analysis. Indeed, 4% were misaligned by greater than 10 mm, even though the operators had judged them to be acceptable. While there are some difficulties with the study protocol (on-line set-up was judged by a simulator-portal comparison, and off-line analysis used the first portal for comparison), this report points to the problems encountered when image alignment is operator dependent, and indicates the need for automatic image registration and analysis, perhaps along the lines proposed for pelvic fields by van Herk et al. ( 14). The range of potential applications of EPI technology is far greater than simply replacing visualinspection of weekly portal films by the visual inspection of daily digital images. Automatic analysis of field size and shape can eliminate gross errors due to incorrect, misplaced or missing blocks, or improper collimator settings (19). The placement of lung attenuators in total-body irradiation can be verified (13), and the location of dosimeters has been determined using EPI (16). However, most current interest is concentrated on the determination and reduction of set-up errors, and the procedures required for this task. Systematic set-up errors can be reduced by daily portal imaging and intertreatment set-up corrections prior to the following fraction, in much the same way that portal films have been used for many years, except that the sampling frequency is much higher. The determination of limits of acceptability and decision criteria for implementing corrections is still the subject of debate (2,822). On the other hand, the reduction of random set-up errors implies intratreatment corrections, as reported by Herman et al. and previously by groups in Belgium (6), Canada (1 l), and Great Britain (12). This interventional technique, in which each treatment is verified in the first few seconds and, if necessary, corrected, is expensive in time and personnel, and can be justified by only the most stringent requirements for accuracy. Certainly one would want to ensure first that the best immobilization procedures are being used, and that systematic errors have been identified and minimized as far as possible. Only then would intratreatment correction be worthwhile for selected treatment sites and patient populations. In the short term we can expect most centers using EPI in routine clinical applications to view the digital display during each treatment to check against any gross errors in field size, shape, or placement, and to make subsequent off-line comparisons with the digitized simulator film to obtain a quantitative measure of field displacements, using either in-house software or programs which will be provided by the equipment manufacturers. Data accumulated