| Purpose: To report the inter-fractional prostate bed motion (PBM), and its correlation to the mean anterior rectal wall and mean posterior bladder wall shifts during post prostatectomy radiotherapy using cone-beam computed tomography (CBCT).;Methods and Materials: A total of 70 CBCT and 8 planning CT scans from 8 patients treated with prostate bed radiotherapy were retrospectively analyzed. For each CBCT-CT pair, two rigid image registrations were performed: one based on surgical clips and the other based on pelvic bony anatomy. Each registration gave a displacement vector. The difference between the two registration displacements represented the PBM. In addition, rectum/bladder contours on CT and CBCT scans were compared to identify the organ wall motion. Shifts of the anterior rectal wall and posterior bladder wall were assessed by averaging the slice-by-slice distances between contours on two image sets, measured along an anterior-posterior line passing through the symphysis pubis and spine. The correlation between the organ wall motions and PBM were calculated by the Pearson's correlation coefficient. The difference between the cranial and caudal half of the anterior rectal wall shifts was evaluated using Kruskal-Wallis test. The derived PTV-CTV margin was calculated using Marcel van Herk's margin recipe.;Results: The mean prostate bed motion (PBM) in the left-right (LR), anterior-posterior (AP) and superior-inferior (SI) directions were (0.1+/-0.8)mm, (0.9+/-1.5)mm, and (-0.4+/-1.8)mm, respectively. Pearson's correlation coefficients between PBM and anterior rectal wall (whole length) shifts, between PBM and anterior rectal wall (cranial half) shifts, and between PBM and posterior bladder wall shifts, were 0.43, 0.47, and 0.67, respectively. Significant difference was found between cranial and caudal half of the anterior rectal wall shifts (p<0.01). The derived PTV-CTV margin for LR, AP and SI motion was 3mm, 5mm, and 6mm, respectively, assuming daily radiographic alignment to bone.;Conclusion: The magnitude of PBM relative to pelvic bony anatomy in all three directions was small. The correlation between average anterior rectal wall shifts and PBM was weak-to-moderate, which may be due to rectum contouring inconsistency on the CBCT images. Possible sources for this inconsistency include: non-uniform rectal wall motion through its length, low soft tissue contrast on CBCT image, and artifacts caused by rectal filling and surgical clips. Significant correlation between average posterior bladder wall shifts and PBM suggests bladder wall motion may also be a suitable surrogate for PBM in the AP direction. |
