Use of portal images and BAT ultrasonography to measure setup error and organ motion for prostate IMRT: implications for treatment margins.

PURPOSE Traditionally, portal images have been used for verification of patient setup. More recently, direct prostate localization using ultrasound imaging has become available. The aim of this study was to use both modalities to measure daily setup error and prostate organ motion and their respective contributions to the overall uncertainty of prostate target localization. METHODS AND MATERIALS Thirty-five patients treated for prostate cancer with intensity-modulated radiotherapy (IMRT) between February 6 and July 2, 2001 underwent daily B-mode acquisition and targeting (BAT) ultrasound localization and weekly orthogonal portal imaging. RESULTS A total of 243 pairs of orthogonal portal films and the corresponding daily BAT images were reviewed. The mean shift +/- standard deviation in the right-left (RL), AP, and superinferior (SI) directions was 0.035 +/- 2.8 mm, -0.23 +/- 3.0 mm, and -0.013 +/- 2.0 mm, respectively, for portal films and -0.82 +/- 3.2 mm, -1.4 +/- 6.4 mm and -1.7 +/- 6.4 mm, respectively, for BAT images taken on the same day as the portal films. The mean prostate organ motion measurements were -0.89 +/- 3.3 mm (RL), -1.3 +/- 5.7 mm (AP), and -1.6 +/- 6.4 mm (SI). Without BAT localization, organ motion would have caused the clinical target volume to move outside the planning target volume margin in 23.3-41.8% of the treatments. Margins necessary to achieve complete coverage of the clinical target volume > 95% of the time without BAT would have been 5.3, 10.4 and 10.4 mm in the RL, AP, and SI dimensions, respectively. CONCLUSIONS Prostate organ motion appears to predominate over setup error as the major component of variation in target localization. Without the use of BAT ultrasound prostate imaging, misses of the prostate can occur in a high percentage of treatments, despite patient setup verification with portal images. Relatively large planning target volume margins in the AP and SI dimensions may be necessary to overcome this.