Organ motion, set-up variation and treatment margins in radical radiotherapy of urinary bladder cancer

Abstract

Background and purpose: A major challenge in conformal radiotherapy of bladder cancer is to determine adequate treatment margins. For this purpose, we therefore quantified the internal motion of the urinary bladder as well as the external patient set-up variation during a course of fractionated radiotherapy. In the light of the recently introduced ICRU-62 concept, the planning organ at risk volume, we also studied the internal motion of nearby organs at risk, the rectum and intestine. Material and methods: Weekly CT scans and electronic portal images (EPIs) were sampled from 20 patients during radical, conformal bladder irradiation (60–64 Gy/2 Gy in five fractions weekly). The planning scans were acquired with 70 ml of bladder contrast instilled, and patients were instructed to void before the treatment/repeat scanning sessions. Internal motion of the bladder, rectum and intestine was measured by 3-D image matching of the repeat scans to the patients’ planning scans. Internal margins (CTV-to-ITV) were determined using both a direct empirical approach and an analytically derived margin recipe. The external patient set-up variability was determined by 2-D matching of front and lateral EPIs to corresponding digitally reconstructed radiographs. Results: A total of 149 CT scans (20 for planning, 129 during the treatment course) and 133 sets of EPIs were analysed. Bladder volumes were smaller during treatment than in the planning situation in 85% of the repeat scans. Nevertheless, we found the repeat scan bladder volumes to extend outside the planning scan bladder contours in 89% of the scans, on average with 9% of the volume (range: 0–47%). Eight patients (40%) had at least one repeat scan (25 scans in total) where displacements >15 mm were observed at one or more sides of the bladder. CTV-to-ITV margins of 10 mm inferior, 20 mm superior, 11 mm left, 8 mm right, 20 mm anterior and 14 mm posterior were required to simultaneously encompass all bladder deflections except for the largest outward deflection in all directions in 84% of the patients. Including patient set-up variation (CTV-to-PTV), we found that an additional safety margin of 2–6 mm had to be added in the various directions. The rectum expanded outside the planning contours in all repeat scans, on average with 24% of the volume (range: 2–69%). The volume of intestine found close to the bladder were significantly and negatively correlated to the bladder volume in almost half of the patients. Conclusion: This study documented both a large internal motion of the bladder and a substantial patient set-up variation. Our current treatment margins have been adjusted according to the findings of this study. Considerable variation in position and volume of the rectum and intestine was also documented.