Background and purpose. The objective of this study was to assess set-up reproducibility of pelvic treatments for obese patients in the prone position, with and without the use of an immobilisation shell.

Method and materials. The patients were treated using a posterior open field and opposed lateral wedged fields for 20 fractions over four weeks. Ten fractions were set-up using a pelvic shell and ten fractions without. The patients were randomised into two groups, whose initial set-up was either with or without a shell. Portal images were obtained on 14 treatment fractions (7 for each set-up) and compared against initial simulator images to assess movement. For each fraction when portal imaging was scheduled the treatment radiographers also assessed the ease at which the patient could be positioned. The aim was to determine the correlation between good visual positioning to surface marks and accurate location of bony landmarks in portal imaging.

Results. Early results from this study highlighted large random positioning errors in the superior/inferior plane when patients were positioned in the pelvic shell. The study was therefore abandoned at patient number four.

Conclusion. The use of the pelvic shell was found to increase positioning errors in the longitudinal (superior-inferior) direction. Reasons for this observation are discussed and modifications to the shell design suggested. The study also highlighted that ease of patient positioning using surface marks is not a good guide to positional accuracy and emphasises the importance of regular on-set verification.