Intrafraction motion of the prostate during an IMRT session: a fiducial-based 3D measurement with Cone-beam CT

Background: Image-guidance systems allow accurate interfractional repositioning of IMRT treatments, however, these may require up to 15 minutes. Therefore intrafraction motion might have an impact on treatment precision. 3D geometric data regarding intrafraction prostate motion are rare; we therefore assessed its magnitude with pre- and post-treatment fiducial-based imaging with cone-beam-CT (CBCT). Methods: 39 IMRT fractions in 5 prostate cancer patients after (125)I-seed implantation were evaluated. Patient position was corrected based on the (125)I-seeds after pre-treatment CBCT. Immediately after treatment delivery, a second CBCT was performed. Differences in bone- and fiducial position were measured by seed-based grey-value matching. Results: Fraction time was 13.6 +/- 1.6 minutes. Median overall displacement vector length of (125)Iseeds was 3 mm (M = 3 mm, Sigma = 0.9 mm, sigma = 1.7 mm; M: group systematic error, Sigma: SD of systematic error, sigma: SD of random error). Median displacement vector of bony structures was 1.84 mm (M = 2.9 mm, Sigma = 1 mm, sigma = 3.2 mm). Median displacement vector length of the prostate relative to bony structures was 1.9 mm (M = 3 mm, Sigma = 1.3 mm, sigma = 2.6 mm). Conclusion: a) Overall displacement vector length during an IMRT session is < 3 mm. b) Positioning devices reducing intrafraction bony displacements can further reduce overall intrafraction motion. c) Intrafraction prostate motion relative to bony structures is < 2 mm and may be further reduced by institutional protocols and reduction of IMRT duration.