Impact of anatomical interventions on the localization of post-prostatectomy cancer patients.

PURPOSE: Anatomical deformations of prostate-bed, rectum, and bladder can compromise the targeting accuracy in post-prostatectomy cancer patients. In this work, the impact of anatomical interventions on the localization data from post-prostatectomy patients who received image-guided IMRT was analyzed.

METHODS: Patients were localized daily with online kilovoltage cone-beam computed tomography (kV-CBCT). The target and the organs at risk (OARs) positional and volumetric changes were evaluated and couch shifts were applied. For patients with large target or OAR volumetric changes, quantified by either a rectal or bladder wall displacement of >5 mm on the CBCT sagittal images compared to the planning CT, repeated localization CBCT scans were performed following an interventional procedure. The procedure involves insertion of a catheter to deflate the rectum, evacuation of stools, and/or adjustment of bladder filling. The required shifts were then evaluated, and the IMRT treatment was subsequently delivered after proper patient positioning. The pre- and post-intervention shifts were compared in the lateral [left-right (LR)], longitudinal [superior-inferior (SI)], and vertical [anterior-posterior (AP)] directions. The percentage of shifts larger than 5 mm in all directions was also compared. Clinical target volume to planning target volume (CTV-to-PTV) expansion margins were estimated based on the pre- and post-intervention localization data.

RESULTS: Intervention was performed on all patients (n=17) treated between October 2008 and March 2009. The number of interventions ranged from 2 to 12 with a median number of 5, resulting in a total of 96 pairs of pre- and post-intervention shifts. The mean value (sigma) and standard deviation (sigma) of the shifts from pre- versus post-intervention data were LR, 0.0 +/- 3.0 mm vs. 0.5 +/- 2.8 mm; SI, 0.2 +/- 3.1 mm vs. -1.0 +/- 2.1 mm; and AP, -2.6 +/- 5.8 mm vs. 1.7 +/- 3.9 mm. The mean 3D shift distance was 7.0 +/- 3.1 mm vs. 5.0 +/- 2.6 mm. The percentage of pre-intervention shifts larger than 5 mm were 7%, 7%, and 45% in the LR, SI, and AP directions, respectively, compared to 8%, 4%, and 21% for post-intervention. Localization data from pre- and post-intervention procedures suggest that treatments that do not include intervention to correct for rectum/bladder anatomical variations require an additional 3.3 mm CTV-to-PTV margin.

CONCLUSIONS: Anatomical interventions reduced the localization errors arising from large volume and shape changes in the rectum and/or bladder compared to treatments without interventions.