Daily patient setup error in prostate image guided radiation therapy with fiducial-based kilovoltage onboard imaging and conebeam computed tomography.

BACKGROUND

This study examined the interfraction setup error in patients undergoing prostate radiotherapy using fiducial markers and on-board imaging.

METHODS

Patients (n=53) were aligned to the treatment isocenter by laser followed by orthogonal kilovoltage (kV) radiographs to visualize bony anatomy and implanted fiducial markers. The magnitude and direction of couch shifts for isocenter correction required was determined by image registration for bony anatomy and fiducial markers. Twice weekly, 25 of the 53 patients also underwent conebeam computed tomography (CBCT) to measure any residual error in patient positioning. Based on individual coordinate shifts from CBCT, a net three-dimensional (3D) residual shift magnitude vector R was calculated.

RESULTS

The average couch shifts were 0.26 and 0.40 cm in inferior direction and 0.25 and 0.33 cm in superior direction for alignments made with bony anatomy and fiducial markers, respectively (P<0.0001). There were no significant differences noted in the vertical or lateral planes between the two image registration methods. In subset of 25 patients, no residual shift from fiducial plain film set up was required with CBCT matching in 66.5%, 52.4% and 57.9% of fractions for longitudinal, vertical and lateral planes, respectively, with majority (79%) of patients having a net residual 3D shifts of <0.3 cm. The use of CBCT increased average treatment time by approximately 6 min compared to kV radiographs alone.

CONCLUSIONS

The residual setup errors following daily kV image guided localization, as determined by CBCT, were small, which demonstrates high accuracy of kV localization when fiducial markers are present.